Ecology and conservation of a new carnivorous

marsupial species: the silver-headed antechinus

(Antechinus argentus)

Eugene David Mason

B. App. Sci. (Hons)

School of Earth, Environmental and Biological Sciences Science and Engineering Faculty

Queensland University of Technology 2018

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

ii

Keywords

Antechinus, Australia, breeding biology, conservation, dasyurid, dietary strategy, ecology, fire ecology, habitat preference, Kroombit Tops, life-history,

mammal, mammal ecology, population ecology, semelparity, Queensland

iii

Abstract

The genus Antechinus comprises 15 species of small carnivorous marsupials endemic to Australia. Five of these have been described since 2012 on the basis of combined genetic, morphological and ecological data. One new species, the threatened silver-headed antechinus (Antechinus argentus) was described in 2013. When the present study commenced the species had only been found at one location: the south-eastern plateau of Kroombit Tops National Park, south-west of Gladstone in mid-east Queensland, Australia - an area encompassing less than 10 km2. However, the study that described the species was based on data obtained from only 15 individuals used for taxonomic description; thus, little was known about the biology of A. argentus. Therefore, the major component of the present study investigated three broad areas of the species’ ecology for which knowledge was sparse or non-existent, with the aim of providing foundational ecological knowledge to prioritise conservation management: 1. dietary strategy and composition; 2. life-history traits; 3. post-fire habitat use. Additional work was carried out to examine the potential influence of rainfall patterns on population dynamics of A. argentus, and widespread trapping surveys were undertaken with the aim of establishing a cogent distributional range of the species, which resulted in the discovery of a second population ~200 km distant from the type locality. During 2014, faecal pellets were collected each month (March-September) from a population at the type locality to gather baseline data on diet composition. A total of 38 faecal pellets were collected from 12 individuals (eight females, four males) and microscopic analysis of pellets identified seven invertebrate orders, with 70% combined mean composition of beetles (Coleoptera: 38%) and cockroaches (Blattodea: 32%). Other orders that featured as prey were ants, crickets/grasshoppers, butterflies/moths, spiders, and true bugs. Given that faecal pellets could only be collected from a single habitat type (Eucalyptus montivaga high-altitude open forest) and location, this was best described as a generalist insectivorous diet that is characteristic of other previously studied congeners.

iv

Antechinuses are well-known for their spectacular annual male die-off at the close of a one- to three-week mating period. The genus also displays sexual dimorphism for size—males are up to three times heavier than females. The A. argentus population studied at Kroombit Tops National Park over two years followed the trends of the genus, with strong evidence of both a synchronised male die-off (in June/July) and significantly larger males than females. Two proximate sites where A. argentus was previously known to occur were surveyed. Unexpectedly, there was a marked difference in A. argentus numbers between years and sites. It was hypothesised that the disparate capture rates between sites may be at least in part linked to the effects of fire on vegetation. Management of critical habitat for threatened species with small ranges requires location-specific, fine-scale survey data. Detailed vegetation surveys involving the collection of plant species diversity and structure data from three sites comprising the known habitat of A. argentus at Kroombit Tops were undertaken. This information was related to capture data obtained over two years. Differences in both vegetation and capture data between burnt and unburnt habitat were found. Leaf litter and grasstrees (Xanthorrhoea johnsonii) were the strongest vegetative predictors for A. argentus capture. The species declined considerably over the two years of the trapping study, and the present work raises concern for its continued survival at Kroombit Tops. Future work should focus on structural vegetative variables (specifically, the diameter and leaf density of grasstree crowns) and relate them to A. argentus occurrence. It was also recommended that a survey of invertebrate diversity in grasstrees and leaf litter be undertaken and compared to A. argentus prey. The data presented here illustrates how critical detailed monitoring is for planning habitat management and fire regimes, and highlights the utility of a high-resolution approach to habitat mapping. While a traditional approach to fire management contends that pyrodiversity encourages biodiversity, the present study demonstrates that some species prefer long-unburnt habitat. Additionally, in predicting the distribution of

v

rare species like A. argentus, data quality (i.e., spatial resolution) may prevail over data quantity (i.e., number of data). In Australia, rainfall is a major influence on temporal changes in small mammal populations. The present study postulated that population dynamics of Kroombit Tops A. argentus from 2014 to 2015 may have been influenced by both higher and lower than usual precipitation levels. Daily rainfall data dating back to 1993 was obtained from a pluviometer proximate to the study sites. Monthly rainfall amounts during the present study did not differ significantly to overall mean monthly values calculated from the 23 years of data available. Nevertheless, some trends were observed that indicated unusually high and low amounts of rainfall at potentially critical times for A. argentus breeding and growth. It was hypothesised that an overall drier spring in 2014 may have contributed to higher infant and juvenile A. argentus mortality, while an unusually high amount of rainfall in the following summer of 2014/2015 caused an increase in insect abundance, providing a surplus of food for the fewer surviving A. argentus and explaining larger mean weight data. Taken together, the present study afforded a study of the autecology of a species that was previously poorly understood. The results provide crucial foundational knowledge for ongoing and necessary conservation management of the species, and much needed clues on key habitat characteristics for locating other potential populations of this rare, cryptic species. A. argentus is almost certainly at risk of extinction. The two known populations have likely been isolated for millennia and today occur in relatively small areas of conservation land. The present research resulted in the species being listed as vulnerable in Queensland under the Nature Conservation Act 1992 and it is currently being considered for federal listing. Several processes are likely threatening both of these populations, namely inappropriate fire frequency, severity and patch size, and the impacts of introduced animals. Management actions must be undertaken if this threatened species is to persist in an increasingly impacted environment.

vi

Outcomes of the present study

Publications:

• Mason, E.D., Burwell, C.J., and A.M. Baker. 2015. Prey of the silver-headed antechinus (Antechinus argentus), a new species of Australian dasyurid marsupial. Australian Mammalogy 37(2): 164-169.

• Mason, E.D., Firn, J., Hines, H.B., and A.M. Baker. 2017. Breeding biology and growth in a new, threatened carnivorous marsupial. Mammal Research 62(2): 179-187.

• Mason, E.D., Firn, J., Hines, H.B., and A.M. Baker. 2017. Plant diversity and structure describes the presence of a new, threatened Australian marsupial within its highly restricted, post-fire habitat. Accepted for publication in PLOS ONE on 20/07/2017.

Threatened species listings:

• A. argentus listed as vulnerable under the Nature Conservation Act 1992 in Queensland

• A. argentus currently at public consultation stage for listing as endangered under the Environment Protection and Biodiversity Conservation Act 1999 in Australia

Conservation recommendations:

• Controlled fire regimes in areas containing A. argentus habitat should utilize cool fires that produce patchy burns, maximizing spatial variation in severity

• Designated unburnt patches should be maintained in A. argentus habitat for 10+ years

• Feral animal control measures should be intensified at Kroombit Tops, particularly for cats, cattle, horses and pigs

• Queensland threatened species listing for A. argentus should be raised to endangered

• A. argentus should be placed on the IUCN Red List

vii

Table of contents

Ecology and conservation of a new carnivorous marsupial species: the silver-headed antechinus (Antechinus argentus) ...................... iKeywords ............................................................................................................. iiAbstract .............................................................................................................. iiiOutcomes of the present study .......................................................................... viTable of contents ............................................................................................... viiList of figures ...................................................................................................... ixList of tables ....................................................................................................... xi

Statement of original authorship ................................................. xiiAcknowledgements .......................................................................................... xiii

Chapter 1: General introduction .................................................. 11.1 Mammal conservation ............................................................................. 11.2 Taxonomy of Australian mammals and discovery of A. argentus ......... 31.3 Ecology of the genus Antechinus ............................................................ 51.4 Outline of the present study .................................................................. 10

Chapter 2: Prey of the silver-headed antechinus (Antechinus argentus), a new species of Australian dasyurid marsupial ......... 15

Abstract: ......................................................................................................... 152.1 Introduction .......................................................................................... 162.2 Materials and methods .......................................................................... 172.3 Results ................................................................................................... 202.4 Discussion ............................................................................................. 222.5 Acknowledgements ............................................................................... 27

Chapter 3. Breeding biology and growth in a new, threatened carnivorous marsupial ................................................................. 28

Abstract: ......................................................................................................... 283.1 Introduction .......................................................................................... 293.2 Materials and methods ......................................................................... 313.3 Results ................................................................................................... 363.4 Discussion ............................................................................................. 413.5 Acknowledgments ................................................................................. 45

viii

Chapter 4. Plant diversity and structure describe the presence of a new, threatened Australian marsupial within its highly restricted, post-fire habitat. ......................................................... 46

Abstract: ......................................................................................................... 464.1 Introduction .......................................................................................... 474.2 Materials and methods ......................................................................... 494.3 Results ................................................................................................... 564.4 Discussion ............................................................................................. 604.4 Acknowledgements ............................................................................... 66

Chapter 5: General discussion ................................................... 675.1 Rainfall and photoperiod ...................................................................... 675.2 Detectability and distribution .............................................................. 725.3 Conservation recommendations .......................................................... 80Conclusion ..................................................................................................... 84

References ................................................................................... 85

Appendix 1. Methods and results of rainfall data collection and analysis ...................................................................................... 108

ix

List of figures

Fig 2.1 Mean percentage volume of invertebrate orders represented in faecal pellets (n = 34) of A. argentus collected over three seasons in 2014 - autumn (March, April, May [n=16]), winter (June, July [n=15]), and spring (September [n= 3]) and in faecal pellets of female (n= 29) and male (n= 9) individuals (n= 38). Data from pellets (n= 4) collected in December 2013 were omitted from the seasonal dataset due to the prolonged period until the next trapping month (March 2014). ......................................................................... 26

Fig 3.1 The Lookout site at Kroombit Tops National Park, with the trunk of a Eucalyptus montivaga in the foreground showing fire scarring, and a grass tree Xanthorrhoea johnsonii in front centre, which are prevalent at this site. Photograph by Eugene Mason. ......................................................................... 33

Fig 3.2 Location of the two main study sites within Kroombit Tops National Park. ................................................................................................................... 33

Fig 3.3 Number of individual A. argentus captured at Kroombit Tops in 2014 and 2015 and key life-history events. ............................................................... 39

Fig 3.4 Boxplot of female and male weights of A. argentus individuals recorded in March–September of 2014 and 2015. ........................................... 39

Fig 3.5 Timeline of captures for a regularly recaptured individual female A. argentus in 2014 and 2015 ............................................................................... 43

Fig 4.1 (a) The Lookout Unburnt site, (b) Lookout Burnt site, and (c) Northern site at Kroombit Tops National Park. Photographic credit to Eugene Mason and Harry Hines. ................................................................................... 52

Fig 4.2 Non-metric multidimensional scaling (nMDS) bubble plot of vegetation species diversity at each 10 m2 vegetation plot. The size of the bubbles indicates the number of A. argentus captures during 2014. N = Northern, LU = Lookout Unburnt, LB = Lookout Burnt. 0.1 represents vegetation plots with nil captures. .................................................................... 57

Fig 4.3 Principal component analysis (PCA) bubble plot of ground cover (%) of litter, grasses, ferns, bare soil, charcoal, bryophytes, herbs, shrubs and dead wood and and tree basal area (m2) at each 10 m2 vegetation plot. The size of

x

the bubbles indicates the number of A. argentus captures during 2014. N = Northern, LU = Lookout Unburnt, LB = Lookout Burnt. 0.1 represents vegetation plots with nil captures. .................................................................... 58

Fig 4.4 A. argentus captures at Kroombit Tops in 2014 and 2015. ................ 60

Fig 5.1 Number of individual A. argentus captured in 2014 and 2015, key life-history events and amount of rainfall each month at The Lookout. Capture data (histogram) applies to the y-axis scale shown at left (individuals captured) and rainfall data (dotted line) applies to the y-axis scale shown at right (rainfall). ................................................................................................... 68

Fig 5.2 Comparison of monthly rainfall at the Lookout from July 2013 to September 2015 (solid line), mean monthly rainfall at the Lookout from 1993 to 2015 (black dotted line) and Mean 3-year Foley’s index (actual rainfall for 3 years preceding each month minus expected 3-year rainfall divided by mean annual rainfall) from July 2013 to September 2015 (grey dotted line) (Foley 1957). Monthly rainfall data applies to the y-axis scale shown at left and Foley’s index data applies to the y-axis scale shown at right. .......................... 69

Fig 5.3 Trapping locations in central-eastern Queensland. u: Kroombit Tops Lookout site, :Kroombit Tops Northern site u: Collosseum Creek, u: Bulburin sites 1-4, u: Mt Robert sites 1-4, u: Blackdown Tableland site 1, u: Blackdown Tableland site 2 (A. argentus captured), u: Consuelo Tableland sites 1-2 (A. flavipes captured), u: Consuelo Tableland site 3. ........................ 77

xi

List of tables

Table 2.1 Number of pellets contributed by female (F) and male (M) A. argentus individuals by month. ........................................................................ 23

Table 2.2 Frequency of prey taxa in A. argentus faecal pellets (n = 38). ...... 23

Table 3.1 Summary of the six best generalized linear mixed effects models used to describe differences in weight among A. argentus, ordered from lowest AICC value (ie. the best-fitting model) to highest. AICC denotes the Akaike information criterion value corrected for small sample size. R2M represents the marginal coefficient explained by fixed factors, and R2C represents the conditional coefficient explained by both fixed and random factors. ............................................................................................................... 41

Table 4.1 Boosted regression tree model. Summary of the relative contributions (%) of predictor variables (tree and shrub species abundance) developed with cross validation on data using 1700 trees, tree complexity of 5 and learning rate of 0.003. ............................................................................... 59

Table 4.2 Boosted regression tree model. Summary of the relative contributions (%) of predictor variables (average Bray-Curtis values, tree basal area and ground cover [%]) developed with cross validation on data using 1300 trees, tree complexity of 5 and learning rate of 0.003. ........................... 59

Table 4.3 Average leaf litter cover. ................................................................. 60

Table 5.1 Minimum number of trap nights required to be 99% confident that A. argentus will be detected if it inhabits a specific locality, calculated based on 2014 data from two localities that A. argentus is known from. .................. 74

Table 5.2 Trapping undertaken at locations external to Kroombit Tops (colours indicate captures of Antechinus argentus and Antechinus flavipes).76

xii

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

Date: 03/06/2018

xiii

Acknowledgements I would like to respectfully acknowledge the Gangulu and Gureng Gureng people - the traditional owners of the land on which this study took place, and Elders both past and present. I would like to express special appreciation and thanks to my Principal Supervisor, Dr Andrew Baker, who is largely responsible for my decision to pursue a PhD in the first place. Andrew’s enthusiasm, guidance and friendship were unyielding throughout the project – on campus and in the field. I would also like to thank my Associate Supervisor Dr Jennifer Firn for her invaluable assistance and encouragement, organizing and shouting lab lunches and retreats, and for imparting a genuine passion for ecology. My thanks also goes to Harry Hines, who acted as an informal supervisor and general mentor on this study, and in particular provided indispensable advice, assistance and resources in the field. Harry’s knowledge of the Australian flora and fauna is encyclopedic, and his fascination with the biota of Kroombit Tops is infectious. I will always be grateful to Dr Thomas Mutton for inviting me to assist with trapping in the Samford Valley for his honours project when I was an undergraduate, which eventually led to my involvement in the present research. Thanks must also go to Dr Chris Burwell, who patiently assisted in the identification of invertebrate body parts on many long days at the Queensland Museum. This project was generously assisted by funding provided by several organizations: the Fitzroy Basin Association, the Burnett-Mary Regional Group and the Wildlife Preservation Society of Queensland. Special thanks go to Cassandra Tracey from FBA and Saranne Giudice from BMRG, who both also volunteered time and energy to helping carry out trapping surveys. Many Queensland Parks and Wildlife rangers assisted the project by providing accommodation, clearing roads, and allowing access to national parks. I’d particularly like to thank Peter Pickering for trusting myself and volunteers with using the barracks at Kroombit Tops – field work was never a chore with

xiv

the knowledge that there was a gas stove, fridge and a wood burner to go back to. I’d also like to thank Carlin Burns for providing access to the barracks at Blackdown Tableland, and for his interest in the research. I was fortunate to have a multitude of good people willing to volunteer and provide company on field trips. My deep thanks go to Coral Pearce, Emma Hawkes, Jarrah Wills, Karl Stone, Dave Warner, Thomas Mutton, Jordan Rochfort, Matt Turner, Laura Allen, William Mason, Jake Viel, Ed White, Kirsten Wallis, Paul O’Callaghan, Emily Coleman, Reece Newnham, Chai Glandfield, Temma Lee, Mie Geertsen, Rachael Collett, Neil Fordyce and Diana Hughes. Additionally, the extensive trapping survey that was organized by Harry Hines and undertaken at Kroombit Tops in June 2016 resulted in an unusual congregation of some of the most valuable individuals in Queensland conservation work - I thank Dr Ian Gynther, Luke Hogan, Geoff Smith, Angus McNab, Mark Sanders, Matthew Varghese, Melanie Venz, Symeon Marou, Ric Fennessy, Kiteesha Lawson, Stephen Mahony and Brendan Burmeister for their tireless work over a largely fruitless week. Heartfelt thanks go to my parents, Janet and Dave Mason, and my brother, Will Mason, for their enduring love and unquestioning support for whatever activities I choose to pursue. I thank the housemates I’ve shared living spaces with throughout this project. In particular I thank Hailey Atkins for continuing to live with me across many leases (and who I forgive for “tidying up” by moving important belongings of mine to various and increasingly more bizarre locations and then forgetting where said locations were [often the day before a field trip]), and Candice Hedge for the many late-night therapy sessions and who in the final weeks of this project miraculously survived a potentially fatal incident with sense of humour intact. My appreciation goes to Kellie Murphy and Jodan Longfield of SHED for facilitating a much-needed outlet for the strange ideas I would bring back from field trips. My deepest thanks go to Temma Lee for her unwavering support, affection and understanding through a period of my life when I was even more distracted than usual.

xv

Chapter 1: General introduction This chapter details the background (section 1.1, 1.2) and context (section 1.3) of the present study, before giving an outline of the remaining chapters of this thesis (section 1.4).

1.1 Mammal conservation

1.1.1 Global biodiversity status Global biodiversity is declining at a rate and magnitude that has not occurred since the last mass extinction event. It is now widely accepted that the global wave of extinction and local population declines of a range of species over the last 500 years is a result of human activities associated with natural resource consumption and large-scale modification of the environment (Barnosky et al., 2011). However, anthropogenic impacts on terrestrial animals remain under-recognized (Dirzo et al., 2014). Furthermore, animal biodiversity loss is not only a result of human activities on Earth, but is a driver of environmental change in and of itself. It has been shown that the mass extinction of vertebrates during the last ice age altered ecosystem processes and had cascading effects across a wide range of taxa (Dirzo et al., 2014, Doughty et al., 2013, Gill et al., 2009). Deforestation undeniably leads to the loss of biodiversity and is relatively easily quantified (Barlow et al., 2016, Panfil and Harvey, 2015). However, even though species and populations of terrestrial animals are often monitored through field trapping and camera techniques, faunal decline remains a largely cryptic phenomenon, occurring even in areas of protected habitat (Dirzo et al., 2014, Peres and Palacios, 2007). Population losses are a more sensitive indicator of biodiversity loss than species extinctions, partly because populations that decline are unlikely to go extinct as a species in the immediate future and therefore will not factor into extinction statistics (Ceballos and Ehrlich, 2002). In order to properly convey the depletion of biodiversity, analyses should ideally be narrowed down from the species to the population level (Ceballos and Ehrlich, 2002, Lomolino and

2

Channell, 1995). This is especially true for species with populations occupying isolated fragments of habitat (Turner, 1996). 1.1.2 Australian mammal status Despite having a low human population density by global standards, and most of the continent remaining sparsely settled (Sanderson et al., 2002), Australia has the worst mammal extinction rate of any country on Earth (Johnson, 2006). Of the 272 endemic terrestrial Australian mammals extant at the time of European invasion in 1788, 30 are now considered to be extinct (Woinarski et al., 2014). Moreover, many of the surviving native mammal species are rapidly declining (Woinarski et al., 2015). Yet the Australian public remains largely oblivious to the mammal extinction record (Flannery, 2012), perhaps because almost all extinct species (with the exception of the thylacine, Thylacinus cynocephalus) are cryptic, small rodents and marsupials (Woinarski et al., 2015). While the extinction and decline of the Australian mammal fauna are likely a result of the combined impacts of threatening processes, introduced predators have been implicated in the extinction of at least 16 species (Doherty et al., 2015, Johnson, 2006). 1.1.3 The predominant threat to Australian mammals – introduced fauna The domestic cat (Felis catus) is commonly kept by humans as a companion animal (Turner and Meister, 1988). In Australia, wild, feral populations of the species are now ubiquitous in all biomes across the continent (Denny and Dickman, 2010). Feral cats are opportunistic, efficient generalist carnivores, and are now considered to have a principal role in not only the extinction but also the recent decline of many native Australian mammal species (Doherty et al., 2015). Indeed, they are listed as a Key Threatening Process under the Environment Protection and Biodiversity Conservation Act 1999 (Department of the Environment, 2008). The red fox (Vulpes vulpes) was introduced to Australia slightly later than the domestic cat, and has a marginally less extensive spread (Abbott et al., 2014), but predation by it too is

3

considered to be a major contributing factor in the decline of the Australian terrestrial mammal fauna (Woinarski et al., 2015). In addition to feral carnivores, introduced herbivorous mammals are impacting native mammals in Australia. The now dominant form of land use across much of Australia is grazing for introduced cattle (Bos taurus) and sheep (Ovis aries) (and to a lesser extent other herbivores such as horses [Equus caballus], goats [Capra hircus], and rabbits [Oryctolagus cuniculus]). The impacts of pastoral land use are felt beyond the boundaries of private estates, as these introduced herbivores are collectively abundant across the entire Australian continent, including in conservation areas. Introduced herbivores have been implicated as a major cause of decline in small rodents and dasyurids (carnivorous marsupials), as shown by experimental removal of livestock (Legge et al., 2011a). Studies have also found differences in native mammal abundance between comparable areas with and without introduced herbivores (Kutt and Woinarski, 2007, Kutt and Gordon, 2012). Introduced fauna represent a major threat to Australian mammals, but they are not the only ecological disturbances manifesting in the 200+ years since European invasion. Other important factors include widespread habitat clearing and altered fire regimes (Short and Smith, 1994).

1.2 Taxonomy of Australian mammals and discovery of A. argentus

1.2.1 Mammal diversity Mammals include some 5,513 described living species worldwide, encompassing three major taxonomic groups: placentals, monotremes and marsupials (IUCN, 2016). Placentals are so named because a placenta provides nourishment for the foetus during gestation (Mess and Carter, 2007), and represent the vast majority of extant mammal species (>5000 species), including such diverse taxa as primates, rodents, and cetaceans. Monotremes are the surviving members of an ancient lineage (diverging from other modern mammals between 210 and 160 Myr ago), today represented by just five species: the platypus, Ornithorhynchus anatinus (Australia), and four echidna species (one in Australia, three in New Guinea) (Rowe et al., 2008,

4

Warren et al., 2008). Marsupials are traditionally characterised by the presence of a pouch, in which the early development of young occurs (Smith, 2001). The 334 described extant marsupial species inhabit parts of Australasia and the Americas; however, Australasia contains the most species-rich marsupial fauna (Ceballos and Brown, 1995). Included in this fauna are the diverse Dasyuridae (Australia and New Guinea), a family of carnivorous marsupials. 1.2.2 Taxonomy of Dasyuridae and Antechinus The Family Dasyuridae is currently recognised within the Order Dasyuromorphia, along with the families Thylacinidae and Myrmecobiidae (Wroe, 1997). With 74 described extant species, Dasyuridae exhibits the largest species-level diversity of the three dasyuromorph families (Thylacinidae – 12 extinct species; Myrmecobiidae – one extant species) and encompasses the dominant carnivorous and insectivorous terrestrial mammal fauna of Australia and New Guinea, collectively occurring in all major habitat types (Krajewski et al., 1994). Dasyuridae includes the subfamilies Dasyurinae and Sminthopsinae. Dasyurinae is comprised of the sister-tribes Phascogalini and Dasyurini (Krajewski et al., 2000a). The genus Antechinus is grouped (along with the genera Phascogale and Murexia) within the sister-tribe Phascogalini (Krajewski and Westerman, 2003). Fifteen species of Antechinus are currently formally recognised (Baker et al., 2015), five of which have been recently described, and among which a range of ecological diversity is displayed. 1.2.3 Discovery of A. argentus Between 2012-2015, a systematic revision of the dasyurid marsupial genus Antechinus was undertaken using combined genetic, morphological and ecological data (see Baker et al., 2012, Baker et al., 2013, Baker et al., 2014, Baker et al., 2015, Baker and Van Dyck, 2012, Baker and Van Dyck, 2013). This revision led to the description of five new Antechinus species. When the present study began, one of these, the silver-headed antechinus (A. argentus)

5

had only been found at one location: the south-eastern plateau of Kroombit Tops National Park, west of Gladstone in mid-east Queensland - an area encompassing less than 10 km2. At the time, this represented the smallest distribution of any Australian mammal. However, the Baker et al. (2013) study that described the species was based on data obtained from only 15 individuals used for taxonomic description; thus, almost nothing was known about the biology of A. argentus.

1.3 Ecology of the genus Antechinus

1.3.1 Dietary strategy in mammals A vast range of carnivorous (animal tissue-eating), herbivorous (plant-eating), granivorous (grain-eating) and omnivorous (consuming a variety of animal, plant, algae, fungi and bacteria) diets are found in mammals (see Dotta and Verdade, 2007, Ungar, 2010, Van Valkenburgh, 1988). The evolutionary importance of dietary strategy in mammals is apparent in the diverse mammalian dentary; many species can be identified by molar tooth structure alone (Carroll, 1988, Price et al., 2012). Indeed, the tribosphenic molar, which allows for more effective crushing and shearing when consuming food, is regarded as a key innovation in mammalian evolution, promoting more effective carnivory and omnivory (Luo, 2007). Carnivory, which broadly describes the primary consumption of vertebrate or invertebrate prey (Ewer, 1973), is found in all three modern mammal groups. Within the marsupials, primarily carnivorous forms are found in Australia, New Guinea and South America (Archer, 1982, Jones et al., 2003). The Australian and New Guinean dasyurids are a diverse carnivorous assemblage that, while predominantly insectivorous, will consume a range of vertebrates, invertebrates and carrion (Van Dyck and Strahan, 2008, p. 44). 1.3.2 Dietary strategy in Antechinus Among mammals, the diet of dasyurids is relatively well known. While larger dasyurids such as Tasmanian Devils (Sarcophilus harrisii) and quolls

6

(Dasyurus spp.) are known to primarily consume vertebrate prey (Jones and Barmuta, 1998), smaller species including Antechinus (e.g., A. agilis, A. minimus, A. stuartii, and A. swainsonii) feed on a variety of predominantly invertebrate prey (see Allison et al., 2006, Fox and Archer, 1984, Green, 1989, Hall, 1980, Lunney et al., 2001, Statham, 1982). These Antechinus species are regarded as dietary generalists, in that they consume prey opportunistically, differing from specialists that seek out a specific type of food (see Fisher and Dickman, 1993, Lunney et al., 2001, Read, 1987). Studies on A. minimus by Wainer (1989), Allison (2006) and Sale (2006) all concluded that the species exhibited a generalist diet, which included numerous invertebrate (and some vertebrate) orders (n=19 prey categories identified to order or above including reptiles and birds). A study by Lunney et al. (2001) found that A. agilis also exhibited a generalist diet, consuming a wide variety of invertebrate prey (n=13 orders), and also occasionally preying on vertebrates (including a small mammal, Acrobates pygmaeus). In contrast, Hall (1980) suggested that while A. stuartii and A. swainsonii were generalist predators (n=15 major taxa for A. stuartii, n=16 for A. swainsonii), both species actively seek out certain prey types, independent of their availability. Specifically, these species apparently preferred beetles and spiders, while largely avoiding more relatively abundant prey such as amphipods, isopods, and ants. Prey preference may also differ between sexes within a species. Lunney et al. (2001) found significant differences in the diets of male and female A. agilis and A. swainsonii, suggesting that they were foraging in different areas. This was also the case in a study by Lazenby-Cohen and Cockburn (1991), which suggested that A. stuartii (now A. agilis) females foraged in particular areas to sustain their greater need for resources during reproduction. Seasonal differences have also been found in the diets of A. swainsonii (now A. mimetes) and A. stuartii (now A. agilis), most notably between winter and summer (Green, 1989).

7

1.3.3 Life history in mammals The attributes of mammal life histories (such as metabolism and production rate) vary across genera (Bielby et al., 2007, Western, 1979). It has been proposed that there is a fundamental size-dependent constraint on mammalian life history attributes, such that metabolic and production rates in larger organisms are generally slower than in smaller ones (Sibly and Brown, 2007). In short, small mammals “live fast and die young”, with generally higher metabolic rates, shorter gestation time, larger litters, and higher mortality in comparison to larger-bodied species of mammals (Promislow and Harvey, 1990). Reproductive strategies are less variable than life histories among mammals. Almost all mammal species are iteroparous, reproducing multiple times during their life, yet a small number exhibit semelparity – the once-in-a-lifetime breeders (Braithwaite and Lee, 1979, Gotelli, 1995). Semelparity is found in both sexes of many modern fish, invertebrate and plant species, but the trait is highly unusual in mammals, at least in part due to the requirement for maternal rearing of young (Charnov and Schaffer, 1973). However, the evolution of divergent life history strategies between the sexes has occurred in a small number of dasyurid marsupials in which males are semelparous but females are not. One marsupial genus that exhibits this strategy is Antechinus. 1.3.4 Life history, breeding biology and growth in Antechinus All known Antechinus species exhibit a synchronised (and frenetic) breeding period of 1-3 weeks triggered by the onset of estrus in females, which is immediately followed by the death of all males in the population (Woolley, 1966). The timing of these reproductive events varies between the 15 known Antechinus species (Gray et al., 2017 provides a summary of the timing of Antechinus life history events). The death of males is the by-product of a testosterone-fueled increase in stress hormone (cortisol) level. Elevated testosterone causes a malfunction in the switch that turns off stress hormone production. Continuous production of cortisol is toxic and ultimately causes a

8

collapse of the immune system, resulting in annual die-off of all males in wild populations (Barker et al., 1978). Males living in the wild, therefore, have a maximum lifespan of 11.5 months (McAllan et al., 2006), although male Antechinus captured before the breeding period and raised in captive isolation have been known to live longer than this (Woolley, 1966). This is presumably because in captivity male testosterone levels are reduced in the absence of other males, with which they would otherwise vigorously compete for mating opportunities with females. Wild Antechinus females, without the problematic testosterone, in some cases survive through two, and rarely three, breeding seasons. This so-called “semelparous reproduction” has evolved in just four mammal genera (Antechinus, Dasykaluta, Phascogale, and Gracilinanus), all of which are marsupials (Fisher et al., 2013). There has been significant debate in the literature over why such an adaptation has been successful. Historically, probably the most popular hypothesis posits that because small mammals are particularly prone to environmental causes of death, surviving a full year past the breeding season is unlikely. Males are then likely driven by lethal competition during their first breeding season, as they are unlikely to survive to breed more than once (Braithwaite and Lee, 1979). This hypothesis reasons that a single, short annual estrus in females evolved to ensure that peak female energy demands coincide with maximum food abundance (the spring/summer flush of insects). Regardless of the adaptability of this trait, the timing of estrus in females has been shown to correlate with the rate of change of photoperiod (specifically that the breeding period coincides with the time of year of least sunlight) (McAllan and Dickman, 1986, McAllan et al., 2006). As Antechinus species are mainly nocturnal, this time of year may also allow for increased nightly foraging time, further assisting with the heightened energy demands of producing young. Braithwaite and Lee’s (1979) hypothesis has been challenged in recent years. A study by Fisher (2013) suggested that suicidal reproduction in males is driven primarily by sperm competition. The author’s hypothesis proposed that the short estrus and promiscuous mating in females escalates competition in

9

males and favours those that devote more of their energy resources to mating. By shortening the breeding season and mating with extreme promiscuity, females manipulate male behaviour and encourage sperm competition, which benefits females by increasing reproductive success (Fisher et al., 2013). 1.3.5 Biogeography and adaptation to fire in dasyurids and Antechinus The Australian marsupial fauna comprises four orders: Dasyuromorphia, Peramelemorphia, Notoryctemorphia and Diprotodontia. All but one of these appear to have diverged around the Cretaceous-Paleogene extinction event (Gallus et al., 2015, Mitchell et al., 2014). However, the diversity of the Dasyuromorphia as it currently exists has been considerably influenced by the changes in climate that occurred during the early Cenozoic (Byrne et al., 2008, Martin, 2006). Around this time, the dominant continental rainforest environments transitioned to open sclerophyll forest and grassland, and eventually widespread arid environments from the Pliocene onwards (Crowther and Blacket, 2003). The consequent drying that occurred caused the wet forests of Australia to contract to the coastal and montane regions in the east, likely generating biogeographical barriers and resulting in isolation and subsequent speciation of taxa (Byrne et al., 2011). The Dasyuromorphia as a whole are largely adapted to arid landscapes, likely coincident with the timing of the group’s radiation. While the majority of dasyurids occur in arid biomes, the genus Antechinus is collectively distributed across the Australian mesic zone (Van Dyck et al., 2013). Encompassing rainforests and open sclerophyll forests, the Australian mesic zone extends down the east coast and south west of the continent. While rare in the rainforest regions (Bowman, 2000), fire is pivotal to the functioning of Australian open sclerophyll ecosystems, and species are adapted to complex fire regimes (Bradstock et al., 2002). For example, there is recent evidence that some Antechinus species can endure fire events in situ, remaining within their burnt home range despite having unburnt areas in close proximity (Stawski et al., 2015). Furthermore, environmental cues resulting from fire such as charcoal, ash and smoke appear to induce torpor in

10

some Antechinus, allowing for energy conservation likely necessary due to food restriction (Stawski et al., 2017). Nevertheless, studies of numerous Antechinus species have indicated declines both in the short-term (Fox, 1982) and the long-term (Wilson et al., 2001) following fires. Due in part to extensive environmental modification and the rapidly accelerating effects of climate change, the important role of fire in Australian ecosystems has been significantly altered, and the frequency and severity of wildfires are increasing (Dale et al., 2001). This will have irreversible effects on ecosystems as a whole (Moritz et al., 2014), but its impact on various species is not yet well understood.

1.4 Outline of the present study

This study consists of three complementary data chapters that together provide a study of the autecology of the silver-headed antechinus, A. argentus. At the beginning of this research, the species had just been described, and had not been the focus of any ecological study. The present research therefore details novel information that covers a range of fundamental aspects of the species’ ecology, including diet, life history (growth, breeding and abundance), habitat preferences, distribution and how population dynamics relate to various environmental factors such as fire and rainfall. Taken together, it is hoped these data will form a crucial baseline for future conservation management of the species. Chapter Two aims to describe the dietary contents and strategy of A. argentus. Antechinus show considerable differences in dietary composition between species. As previously mentioned, A. stuartii (now A. agilis) and A. swainsonii (now A. mimetes) apparently selectively pursue types of prey at times, regardless of availability (Hall, 1980). However, A. agilis and A. minimus appear to exhibit a completely generalist diet (Allison et al., 2006, Lunney et al., 2001, Sale et al., 2006, Wainer, 1989). Furthermore, dietary variation occurs within congener species, with some exhibiting differences between sexes and seasons. Males and females of A. agilis have been shown to forage

11

in different areas (Lazenby‐Cohen and Cockburn, 1991, Lunney et al., 2001),

as have A. stuartii (now A. agilis) and A. swainsonii (now A. mimetes), which also exhibit notable differences in diet between winter and summer (Green, 1989). When this study commenced nothing was known of the diet of A. argentus. To determine how the diet of A. argentus compares to congeners, faecal pellets were collected from individual A. argentus over seven months and analyzed under microscope to determine percent composition of invertebrate orders. The faecal pellet data was formally tested in order to answer the following research questions:

1. What is the dietary composition of A. argentus? Is the species a dietary generalist like most of its congeners or is it selective of a few specific prey types?

2. Is there a difference in dietary composition between sexes? 3. Is there a difference in dietary composition between seasons?

The present author published this chapter with primary authorship, and it is presented here unedited from the final proof. In this paper, coauthors had the following responsibilities: the present author conceived the design, led the field work, sorted scat contents, assisted in identification of scat contents, analysed the data and wrote the manuscript; C.J. Burwell (Queensland Museum) confirmed identification of scat contents, assisted with analysis and helped revise the manuscript; A.M. Baker (QUT) assisted with study design and helped revise the manuscript. The paper was submitted on 1 December 2014, accepted 10 March 2015, and published online 15 May 2015. The citation is as follows: Mason, E.D., Burwell, C.J., and A.M. Baker. 2015. Prey of the silver-headed antechinus (Antechinus argentus), a new species of Australian dasyurid marsupial. Australian Mammalogy 37(2): 164-169. Chapter Three seeks to investigate the breeding biology, growth and population dynamics of the species. As previously mentioned, all known Antechinus species exhibit a male-specific semelparous breeding strategy that

12

results in a complete, annual male die-off (Braithwaite and Lee, 1979, Jones et al., 2003). Antechinus are also sexually dimorphic for size, with males considerably larger than females (Marlow, 1961, Sale et al., 2013). Yet when the present study began, the apparent life-history norms of the genus had not been observed in A. argentus. To address these knowledge gaps, the first long-term (two year) mark-recapture study was undertaken on the (at the time) only known population of the species. It was hypothesized that, like its congeners, A. argentus would exhibit a synchronised breeding period concluding with a male die-off, significantly larger males than females, and similar population abundance between years. Across two years of monthly trapping, the specific aims were to determine if A. argentus:

1. exhibited highly synchronised breeding with a male die- off, and if so compare its timing in relation to congeners;

2. males were larger than females and to track patterns in growth and development of both sexes over time;

3. abundance varied between years and sites. These aims were formally tested and presented in a publication with the present author as primary author, and as with Chapter Two it is presented here unedited from the final proof. In this paper, coauthors had the following responsibilities: the present author conceived the design, led the field work, analysed the data and wrote the manuscript; J. Firn (QUT) assisted with analysis and helped revise the manuscript; H.B. Hines (Queensland Parks and Wildlife Service) assisted with study design, field work and helped revise the manuscript; A.M. Baker (QUT) assisted with study design, field work and helped revise the manuscript. The paper was submitted on 6 March 2016, accepted on 1 December 2016, and published online 13 December 2016. The citation for this paper is as follows: Mason, E.D., Firn, J., Hines, H.B., and A.M. Baker. 2017. Breeding biology and growth in a new, threatened carnivorous marsupial. Mammal Research 62(2): 179-187.

13

The restricted range of A. argentus at Kroombit Tops likely renders the species highly sensitive to disturbance. Furthermore, A. argentus is absent from widespread, seemingly comparable connected habitat at Kroombit. Months after A. argentus was described and just prior to the commencement of this study (October 2013), the majority of the habitat at the type locality was burnt by a wildfire. While the effects of fire have been documented in some species of Antechinus, none have possessed a range as small as the type population of A. argentus. Furthermore, the present-day threat of habitat fragmentation likely means that the impact of fires are potentially greater than in the evolutionary history of species (Robinson et al., 2013). The general use of “patchy” fire mosaics for the conservation of animal species dominates the literature (e.g., Bradstock et al., 2005, Christensen and Kimber, 1975, Woinarski, 1999). But the importance of heterogeneity in fire-age, size of patches and severity of the fire likely depends on context, and in particular small mammals may depend on the retention of older vegetation (Kelly et al., 2012). Knowing a species’ habitat requirements may inform effective fire regimes (Kelly et al., 2017, Kelly and Brotons, 2017). Therefore, Chapter Four aimed to investigate if vegetation floristics and structure are significantly different between burnt and unburnt areas of A. argentus habitat, and to relate this data to A. argentus captures. Specifically, it aimed to investigate whether:

1. differences in plant community composition and structure following a fire correlate with differences in capture rates of A. argentus;

2. plant community composition and structure influence the occurrence of A. argentus;

3. A. argentus abundance changed as the post-fire habitat recovered.

It was hypothesized that there would be differences in vegetation between the burnt and unburnt areas, and that A. argentus would be more abundant in the unburnt area. To address these hypotheses, fine-scale vegetation surveys were undertaken at Kroombit Tops, and this data was formally tested. The strongest vegetative predictor variables for the occurrence of A. argentus were

14

also determined. This chapter was submitted on 5 April 2017, accepted on 16 July 2017, and published online on 10 August 2017. It is presented here unedited from the final proof. In this paper, coauthors had the following responsibilities: the present author conceived the design, led the field work, identified plants, analysed the data and wrote the manuscript; J. Firn (QUT) confirmed plant identification, assisted with study design and analysis and helped revise the manuscript; H.B. Hines (Queensland Parks and Wildlife Service) confirmed plant identification, assisted with study design, and helped revise the manuscript; A.M. Baker (QUT) assisted with study design and helped revise the manuscript. The citation is as follows: Mason, E.D., Firn, J., Hines, H.B., and A.M. Baker. 2017. Plant diversity and structure describes the presence of a new, threatened Australian marsupial within its highly restricted, post-fire habitat. PLOS ONE 12(8): e0182319 The research concludes with a general discussion concerning the potential influence of rainfall on A. argentus in the context of the preceding data chapters, and compares the breeding time of the species with congeners, giving consideration to the importance of photoperiod as a reproductive cue. This section also details the trapping effort conducted in areas external to Kroombit Tops and the implications capture rate may have on species detectability. The final section of the general discussion provides conservation management recommendations based on the preceding research.

15

Chapter 2: Prey of the silver-headed

antechinus (Antechinus argentus), a new

species of Australian dasyurid marsupial Eugene D. MasonA,D, Chris J. Burwell B,C and Andrew M. BakerA

AEarth, Environmental and Biological Sciences School, Queensland University of Technology,

2 George Street, Brisbane, Qld 4001, Australia. BNatural Environments Program, Queensland Museum, PO Box 3300, South Brisbane, Qld

4101, Australia. CEnvironmental Futures Research Institute and Griffith School of Environment, Griffith

University, Nathan,

Qld 4111, Australia. DCorresponding author. Email: ed.mason@qut.edu.au

Australian Mammalogy, 2015, 37, 164–169

DOI 10.1071/AM14036

Abstract:

The silver-headed antechinus (Antechinus argentus) is one of Australia’s most recently described mammals, and the single known population at Kroombit Tops in south-east Queensland is threatened. Nothing is known of the species, ecology, so during 2014 we collected faecal pellets each month (March-September) from a population at the type locality to gather baseline data on diet composition. A total of 38 faecal pellets were collected from 12 individuals (eight females, four males) and microscopic analysis of pellets identified seven invertebrate orders, with 70% combined mean composition of beetles (Coleoptera: 38%) and cockroaches (Blattodea: 32%). Other orders that featured as prey were ants, crickets/ grasshoppers, butterflies/moths, spiders, and true bugs. Given that faecal pellets could only be collected from a single habitat type (Eucalyptus montivaga high-altitude open forest) and location, this is best described as a generalist insectivorous diet that is characteristic of other previously studied congeners.

16

Additional keywords: Dasyuridae, insectivorous diet, Kroombit Tops.

Received 1 December 2014, accepted 10 March 2015, published online 15 May 2015

2.1 Introduction

Knowing what an animal eats is key to understanding its ecology. As a group, mammal diets have been studied intensively but least well known are those of small, secretive species, such as marsupials of the genus Antechinus. Antechinuses are primarily insectivorous dasyurid marsupials endemic to Australia (Krajewski et al., 2000b). Species discovery in the genus has a history dating back to 1803, when Geoffrey Saint-Hilaire, a French naturalist colleague of Lamarck, described A. minimus. There followed numerous descriptions of new species throughout the 1800s and 1900s, culminating in 10 recognised species when Van Dyck (2002) reviewed the genus. But discovery of aberrant museum specimens over the following decade prompted further taxonomic investigation, during which three new species were named in as many years: A. mysticus (Baker et al., 2012), A. argentus (Baker et al., 2013), and A. arktos (Baker et al., 2014). The silver-headed antechinus (A. argentus) apparently has a highly restricted distribution; despite numerous surveys in recent years, A. argentus is known only from the type locality, an area of ~10 km2 on the south-eastern plateau of Kroombit Tops National Park, 60 km south-west of Gladstone, in mid-east Queensland (Baker et al., 2013). This represents perhaps the smallest distribution of any known Australian mammal, and a threatened species listing is pending for Queensland. Our present research on this species focuses on the only known, small population at the type locality, with the aim of defining the species, status, general breeding ecology and diet. Many dasyurids are dietary generalists (see Lunney et al., 1986, Read, 1987, Fisher and Dickman, 1993), including species of Antechinus (e.g. A. agilis, A. minimus, A. stuartii, and A. swainsonii), which consume a wide variety of predominantly invertebrate prey (see Hall, 1980, Statham, 1982, Fox and

17

Archer, 1984, Green, 1989, Lunney et al., 2001, Allison et al., 2006) and, in some cases, nectar (Goldingay, 2000). Here, we aimed to describe the dietary constituents of A. argentus faecal pellets; we hypothesised that, like its congeners, A. argentus will be a generalist, consuming a diversity of arthropod orders limited by its highly restricted distribution. This study contributes to our ecological knowledge of a newly discovered, threatened Australian mammal and is the first investigation of the diet of A. argentus.

2.2 Materials and methods

2.2.1 Study sites Kroombit Tops National Park is a mesic, temperate island standing above the surrounding drier, hotter subtropical lowlands (McDonald and Sharpe, 1986, Hines, 2014). The area experiences warm to hot summers, and fine, cool winters; mean annual rainfall in the vicinity of the type locality is estimated at 1400-1800 mm (McDonald and Sharpe, 1986, Baker et al., 2013, Queensland Department of Natural Resources and Mines, Bundaberg Hydrographic Section. pers. comm. 2013). The area is biogeographically interesting: many temperate plant, fungi, and animal species occur here at their northern limit or as northern disjunct populations (McDonald and Sharpe, 1986, Monteith, 1986, Hines, 2014). The type locality of A. argentus is a gently undulating plateau at an elevation of ~850-900 m above sea level, which is bounded to the east by an escarpment with a sheer edge that falls away with cliffs up to ~50m in height (Schulz and De Oliveira, 1995). The habitat upon the plateau and adjacent to the escarpment comprises Eucalyptus montivaga shrubby tall open forest with occasional brown bloodwood (Corymbia trachyphloia) (Regional Ecosystem 12.9-10.20: see Baker et al. 2013; Queensland Herbarium 2013). The area is often cloaked in mist, and cloud stripping likely adds significantly to available soil moisture. Two study sites were used. One (Lookout) was located close to the Lookout on the eastern edge of the escarpment (24.396 S,

18

151.044 E). The habitat at this site is broadly described as a tall open Eucalyptus forest. The site’s trap grid was split by a dirt road down the middle, with half of the Elliott traps (75 per night of a total 150) laid out on each side. In October 2013, a wildfire spread across much of the escarpment, heavily burning the site on one side of the road, while the other side remained unburnt (H. Hines, pers. comm.). The road therefore divides the trap grid between recently burnt and unburnt habitat. The second site (Northern) was located ~6.5 km north-west of the Lookout site, and closer to the escarpment edge (24.352 S, 150.999 E). Floristically, it is broadly similar to the Lookout site, but is slightly more open. The entire Northern site was comprehensively burnt by the October 2013 wildfire. Additionally, unpublished data obtained from radio-tracking has shown that A. argentus individuals move readily between burnt and unburnt habitat. It was therefore assumed that locality would not affect prey composition. 2.2.2 Animal trapping Small mammals were trapped between December 2013 and September 2014 using standard Type A Elliott traps. The December 2013 trapping survey of 300 trap-nights was undertaken over four nights at the Northern site before establishing our regular trapping grids and regime. Following this, trapping surveys were undertaken for a week of each month from March to September of 2014. At each site, 150 traps were laid out in six straight transects, forming a grid. Each transect comprised 25 traps laid out at 10-m intervals along a consistent bearing. Each trap was baited with a mixture of peanut butter and oats. Traps were opened for three nights during each trapping period, totalling 450 trap-nights at each site per month. When A. argentus individuals were captured at a site, the traps at that site were closed for the following night to allow the animals to forage naturally. Faecal pellets were collected from captured A. argentus individuals. These pellets were stored in 90% ethanol for dietary analysis. Pellets that were obviously contaminated with bait were discarded. Traps were thoroughly cleaned with detergent after any small

19

mammal capture. Captured A. argentus individuals were also weighed, sexed and PIT-tagged (for mark-recapture data) before release at the location of capture (these data form a parallel study and are not shown here). 2.2.3 Dietary analyses Analyses of gut contents or faecal contents are two often-used methods of assessing the diet of insectivorous mammal species (Dickman and Huang, 1988). In comparison to gut content analysis, faecal analysis has the obvious advantage of not reducing the population size of the study animal. Additionally, fragments of invertebrate prey are more concentrated in faeces than in the gut, making them easier to work with (Black, 1972). However, studies of faecal contents potentially misrepresent diversity of taxa in the overall diet (for example, due to the differential digestibility of prey items). Specifically, the relative proportion of hard-bodied invertebrates in the diet is likely to be overestimated due to the low digestibility of hard parts (and therefore high detectability in faeces), whereas consumption of soft-bodied organisms such as worms that would presumably be almost (if not completely) digested potentially goes undetected (Dickman and Huang, 1988, Allison et al., 2006). Nevertheless, faecal analysis has been shown to be a relatively accurate method of assessing the diet of animals that are generalist insectivores, and that consume predominantly hard-bodied prey (Dickman and Huang, 1988). Methods of faecal pellet analysis adopted here follow Pavey et al. (2009). Individual pellets were placed in a Petri dish and treated with the direct application of four or five drops of 10% KOH. They were then teased apart using fine forceps and submerged in 70% ethanol. Each faecal pellet was systematically searched for identifiable prey material (we note that in adopting this method we will have overlooked any pollen that may have been consumed if A. argentus was feeding on nectar; see Goldingay, 2000). This was then separated from the unidentifiable material. Prey fragments were identified to the lowest possible taxonomic level. The percentage volume of each major prey group (typically order) was visually estimated to the nearest

20

5%. Unidentifiable material was not included in the percentage volume data. Taxa that contributed less than 5% by volume were not included in percentage volume estimates. Taxa such as ticks and mites were likely to have been incidentally ingested, and thus were not included in percentage volume estimates. 2.2.4 Statistical analyses The limited number of captures and faecal pellets precluded formal statistical analyses of dietary change with respect to individual A. argentus or month; however, we were able to test for differences in diet between sexes across the entire sample, and seasonal differences in diet between autumn and winter. Differences in the frequency of occurrence of orders in faecal pellets between sexes, and between autumn and winter, were assessed using a Chi-square test (Pearson, 1900). Where orders appeared in <5 pellets of a particular category, a Fisher’s exact test was used (Fisher, 1922). In these cases, only the P-value was listed. Statistical tests were carried out in SPSS (IBM, 2012).

2.3 Results

A total of 33 pellets from the Lookout site were examined; these came from 10 individuals, of which six were females (EDM005, EDM035 - 1 pellet; EDM026 - 2 pellets; EDM007 - 5 pellets; EDM013 - 6 pellets; EDM000 - 9 pellets) and four were males (EDM009 - 1 pellet; EDM040, EDM044 - 2 pellets; EDM003 - 4 pellets). A total of five pellets from the Northern site were examined; these were from two individuals, both of which were females (HH001 - 4 pellets; EDM001 - 1 pellet) (see Table 2.1). The small total sample size (n=38; 33 from Lookout, 5 from Northern site) represents all available faecal material from captures made at the only known A. argentus localities throughout the six-month 2014 trapping period and the preceding December 2013 trapping survey. A. argentus consumed a broad variety of prey, with the remains of six insect orders and one arachnid order identified from pellets (Table 2.2). Whole mites

21

(Arachnida: Acari) were detected on occasion, but it was presumed that these were consumed incidentally, and thus they were omitted from the results. Ants (Hymenoptera) may also have been incidental prey, being attracted to and consumed along with bait inside Elliott traps. However, several ant genera were sometimes detected in single pellets, raising doubt that ant presence could be entirely explained by bait consumption. Therefore, due to the frequency of ants in pellets, we included them as prey items. Coleoptera and Blattodea dominated the identifiable material in pellets, both in terms of the frequency of occurrence (Table 2.2) and percentage volume (Fig 2.1). Overall mean percentage volume for Coleoptera and Blattodea were 37.9% and 32.1%, respectively. Hymenoptera were the next most common as prey (12.2%), followed by Orthoptera (7.8%), Lepidoptera (4.9%), Araneae (4.3%), and Hemiptera (0.8%). The percentage volume contribution of Blattodea in pellets increased substantially from autumn (12.3%) through winter (28.6%) to spring (68%) (Fig 2.1), although there was only a small sample size in spring (n=3), precluding its inclusion in formal analyses. Despite this seasonal trend, there was no significant association between season (autumn and winter) and the frequency of occurrence of any invertebrate order in A. argentus pellets (autumn: n=16; winter: n=15; Blattodea: χ2=1.642, df=1, P=0.200; Hymenoptera: χ2=0.987, df=1, P=0.320; Orthoptera: P= 1.000; Lepidoptera: P=1.000; Araneae: P=0.654; Hemiptera: P=0.600), except for Coleoptera, which occurred significantly more frequently in faecal pellets from winter than in those from autumn (χ2= 4.288, df=1, P=0.038). Several trends were apparent in the relative proportion of prey orders consumed by males versus females (Fig 2.1). Mean percentage volume of cockroaches (Blattodea) was higher in females than in males (34.7% versus 23.9% respectively). The reverse was true of beetles (Coleoptera), which represented 49.4% of mean male prey volume, compared with 34.3% for females. Spiders (Araneae) were also more prominent in male pellets (11.7%) than those of females (6%). There was a slightly larger contribution of

22

Lepidoptera to female (6%) than male (1%) pellets. Representation of Hymenoptera was comparable between sexes (11.7% in females, 13.9% in males). Finally, female pellets contained two insect orders that were not identified in male pellets: Orthoptera (10%) and Hemiptera (1%) (Fig 2.1). However, notwithstanding these trends, there was no significant association between sex and the frequency of occurrence of any invertebrate order in A. argentus pellets (females n=29, males n=9; Coleoptera: χ2=0.186, df=1, P=0.666; Blattodea: χ2=2.418, df=1, P=0.120; Hymenoptera: χ2=0.004, df= 1, P=0.949; Orthoptera: P=0.159; Lepidoptera: P=1.000; Araneae: P= 1.000; Hemiptera: P=0.554).

2.4 Discussion

Our study demonstrates that A. argentus consumes a variety of arthropod prey, exhibiting a diet that is characteristic of a generalist insectivore. A total of eight invertebrate orders were identified in A. argentus pellets. Of these, seven were considered to be actual dietary components, while mites (Acarina) were presumed to have been consumed incidentally. Similar dietary components have been found in other species of Antechinus. However, other species are known to consume a larger number of major prey taxa. For example, studies on A. minimus by Wainer (1989), Allison et al. (2006) and Sale et al. (2006) all concluded that the species had a generalist diet, which included all the invertebrate orders identified in our study of A. argentus faecal pellets, plus numerous others (n=7 orders for A. argentus, n=19 prey categories identified to order or above including reptiles and birds for A. minimus). However, A. minimus is a larger species, and has a much broader distribution, being found in a variety of habitat types throughout Tasmania, the central highlands of Victoria and the Mount Gambier region of South Australia (Magnusdottir et al., 2008), and dietary studies of this species have incorporated animals from a range of habitat types (Allison et al., 2006, Sale et al., 2006). Our data are from the only known A. argentus population, which inhabits an area of only ~10 km2. Given that only a single population and habitat were investigated here, a relatively smaller level of prey diversity might be expected. Lunney et al. (2001) found that A. agilis had a generalist

23

diet, consuming a wide variety of invertebrate prey (n=13 orders), and occasionally preyed on vertebrates (including a small mammal, Acrobates pygmaeus). In contrast, Hall (1980) suggested that while A. stuartii (now A. agilis) and A. swainsonii were generalist predators (n=15 major taxa for A. stuartii, n=16 for A. swainsonii), both species actively seek out certain prey types, independent of their availability (see also Green, 1989). Specifically, these species apparently preferred beetles and spiders, while steering away from relatively more abundant prey such as amphipods, isopods, and ants. On the basis of our results, A. argentus fits the generalist diet description of Antechinus, and more broadly many dasyurids (e.g., Fox and Archer, 1984, Read, 1987, Fisher and Dickman, 1993). Table 2.1 Number of pellets contributed by female (F) and male (M) A. argentus individuals

by month.

Dec. Mar. Apr. May. Jun. Jul. Sep.

No. of trap-nights 300 900 900 900 900 900 900

No. of individuals 1F 2F, 1M 1F, 2M 4F, 1M 3F, 3M 1F 1F

No. of pellets 4 3 4 9 10 5 3

Table 2.2 Frequency of prey taxa in A. argentus faecal pellets (n=38).

Class Prey category No. of pellets

Insecta Coleoptera 25

Blattodea 24

Unidentified to family 22

Laxta 2

Hymenoptera 12

Unidentified to family 1

Formicidae 11

Unidentified to genus 4

Pheidole 5

Rhytidoponera 1

Aphaenogaster 1

Camponotus 1

Crematogaster 1

Nylanderia 1

Solenopsis 1

Carebara 1

24

Orthoptera 6

Lepidoptera 4

Hemiptera 4

Unidentified to suborder 1

Heteroptera 2

Auchenorrhyncha 1

Arachnida Araneae 7

Although it was sometimes possible to identify prey items at lower taxonomic levels (such as the genus Laxta within Blattodea, the suborders Heteroptera and Auchennoryncha within Hemiptera, and eight ant genera: see Table 2.2), in general the degree of fragmentation of prey in faecal matter meant that identification only to order was possible, which is typical of diet studies in small dasyurids (e.g., Lunney et al., 2001, Burwell et al., 2005). It is worth noting that prey diversity for A. argentus may have been underestimated due to the relatively small number of individuals that were trapped (38 faecal pellets were assessed, from 12 individuals [eight females and four males]). In particular, one of these individuals, a female, contributed a large number of pellets (n=9) across the course of the seven-month trapping period. Lunney et al. (2001) found significant differences in the diets of male and female A. agilis and A. swainsonii, suggesting that they were foraging in different areas. This was also the case in a study of A. stuartii (now A. agilis) by Lazenby-Cohen and Cockburn (1991) , who suggested that A. stuartii females foraged in particular areas to sustain their greater need for resources during reproduction. Our study found no significant differences between prey composition of pellets from males and females (Fig 2.1). Although the lack of a significant difference in diet of males and females may be partly attributable to the small number of available pellets, it may also be due to the restricted habitat of this species. The diversity of potential prey is unlikely to vary notably within the ~10 km2 known range of A. argentus. There was no significant association between season (autumn and winter) and the frequency of occurrence of any invertebrate order except for Coleoptera, which occurred significantly more frequently in faecal pellets in winter than in autumn. It was notable that Coleoptera and Blattodea increased in prey item

25

proportion from autumn to winter, and clearly represent the largest proportion of identified A. argentus prey in both sexes (Fig 2.1, Table 2.2). However, we need to quantify the diversity and abundance of invertebrates at the Kroombit Tops study sites to help determine whether these prey composition results reflect a degree of selection, or simply reflect prey availability. Hall (1980) found that the frequency of prey types in the faeces of A. swainsonii and A. stuartii (now A. agilis) did not correspond with their frequency in leaf litter, and suggested that these species may be actively searching for beetles and spiders, while eschewing more abundant invertebrates such as amphipods and ants. But Hall (1980) also noted that these species feed on most of the prey types available, and can therefore best be described as opportunistic. They also pointed out that this type of diet would be beneficial for animals that exist in environments subject to disturbance. Certainly, the habitat of A. argentus at Kroombit Tops is prone to disturbance from sporadic wildfires. One such event occurred in October 2013, when a fire spread and burnt through large parts of our trapping sites (see Methods). This may have further restricted the (already limited) habitat of the species, as most A. argentus captures in 2014 were made in the unburnt section of the study sites. Another wildfire, while not reaching the present study sites, affected nearby areas within Kroombit Tops in October 2014. A flexible dietary strategy would greatly benefit an insectivorous predator that inhabits such an area where disturbance could affect invertebrate species composition.

26

Fig 2.1 Mean percentage volume of invertebrate orders represented in faecal pellets (n=34) of

A. argentus collected over three seasons in 2014 - autumn (March, April, May [n=16]), winter

(June, July [n=15]), and spring (September [n=3]) and in faecal pellets of female (n=29) and

male (n=9) individuals (n=38). Data from pellets (n=4) collected in December 2013 were

omitted from the seasonal dataset due to the prolonged period until the next trapping month

(March 2014).

Coupling our data with previous work on congeners (e.g., Wainer, 1989, Lunney et al., 2001, Allison et al., 2006, Sale et al., 2006), we may best describe Antechinus species as generalist insectivores that vary in diet depending on species, habitat type, distributional range and/or opportunity. Further study of comparative insect diversity and availability within A. argentus habitat would enable us to assess more specific questions about dietary preferences in this species. The data presented here provide a fundamental baseline description of diet in a recently discovered marsupial. This study forms part of a broader, ongoing study concerning the ecology, distribution and status of A. argentus. Taken together, data obtained from

0% 20% 40% 60% 80% 100%

Autumn

Winter

Spring

F

M

Mean % Composition

Coleoptera

Blattodea

Hymenoptera

Orthoptera

Lepidoptera

Araneae

Hemiptera

27

this research will inform sustainable management practices to try to ensure this threatened species is secure into the future.

2.5 Acknowledgements

This study was generously funded by the Fitzroy Basin Association, the Burnett-Mary Regional Group, and the Wildlife Preservation Society of Queensland. E. D. Mason was assisted by an Australian Postgraduate Award scholarship for research. Legends of the highest calibre enthusiastically volunteered their help on field trips: David Warner, Jarrah Wills, Thomas Mutton, Jordan Rochfort, Matthew Turner, Laura Allen, and William Mason. The guru Harry Hines of NPRSR donated Elliott traps, assisted in planning and setting up trap grids, and regularly imparted a vast knowledge gained from his years of exploring Kroombit Tops. Peter Pickering of NPRSR and his team at the Moonford Rangers, Office generously provided the use of the Rangers, Barracks for accommodation at Kroombit Tops. Bench space at the Queensland Museum was kindly offered by Heather Janetzki and the Biodiversity Program.

28

Chapter 3. Breeding biology and growth in a

new, threatened carnivorous marsupial

Eugene D. Mason1,4, Jennifer Firn1, Harry B. Hines2 & Andrew M. Baker1,3 1Earth, Environmental and Biological Sciences School, Queensland University of Technology,

2 George St, Brisbane, QLD 4001, Australia 2Queensland Parks and Wildlife Service, Department of National

Parks, Sport and Racing, Level 19, 111 George St, Brisbane, QLD 4000, Australia 3Queensland Museum, PO Box 3300, South Brisbane, Qld 4101,

Australia 4Corresponding author. Email: ed.mason@qut.edu.au

Mammal Research, 2017, 62(2), 179–187

DOI 10.1007/s13364-016-0303-z

Abstract:

The silver-headed antechinus, A. argentus, is a recently-discovered, threatened carnivorous marsupial known from only two small, isolated montane populations within central-eastern Queensland, Australia. Here, we present the first study of the species’ life-history characteristics. Antechinuses are well-known for their spectacular annual male die-off at the close of a one- to three-week mating period. The genus also displays sexual dimorphism for size—with males up to three times heavier than females. The A. argentus population we studied at Kroombit Tops National Park over 2 years fitted the norm for the genus, with strong evidence of both a synchronised male die-off (in June/July) and significantly larger males than females. We surveyed two proximate sites at Kroombit Tops where A. argentus was previously known to occur. Unexpectedly, there was a marked difference in A. argentus numbers between years and sites. We hypothesise that the disparate capture rates between sites may be at least in part linked to the effects of fire on vegetation. The marked fluctuation in capture numbers between years, coupled with

29

annual male die-off raises concerns over the future of the Kroombit Tops population of this threatened mammal.

Keywords: Antechinus, Kroombit Tops, dasyurid, marsupial, life-history, autecology

Received 6 March 2016, accepted 1 December 2016, published online 16 December 2016

© Mammal Research Institute, Polish Academy of Sciences, Białowieża, Poland 2016

3.1 Introduction

Semelparity is a reproductive strategy in which organisms undergo a single breeding event before their death (Stearns, 1976). In mammals, this ‘suicidal breeding’ strategy is found in just two families, both marsupials: the Dasyuridae (21 species, Australia and New Guinea) and Didelphidae (five species, South America) (Cockburn, 1997, Lopes and Leiner, 2015). In the primarily insectivorous dasyurid genus Antechinus, this unusual breeding strategy is well-documented (Braithwaite and Lee, 1979, Bradley et al., 1980, Dickman, 1985, Jones et al., 2003). Female antechinuses may produce two or (rarely) three annual litters over their lifetime, but males live only a maximum of 11.5 months, until the conclusion of a frenetic, highly synchronised 1- to 3-week annual breeding period (Woolley, 1966, Dickman, 1985, Jones et al., 2003). Several weeks of testosterone-fuelled mating (each session up to 14 h long) and fighting off rival suitors leaves males with a highly compromised physiology. Soaring testosterone levels ensure the eventual malfunction of the stress hormone (cortisol) cut-off switch. Flooded with uncontrolled cortisol, males are systemically poisoned, ravaged by disease and internal bleeding resulting from immune system shutdown (Bradley et al., 1980). This inevitably results in the annual death of all males before a single young is born, effectively halving Antechinus populations annually and imposing the threat of local extinction (Woolley, 1966, Dickman, 1985, Jones et al., 2003). Because key life-history characteristics of antechinuses are predictably synchronised to within days (Woolley, 1966, Braithwaite and Lee, 1979), their timing in unstudied species or populations can be determined relatively

30

accurately by routinely trapping a population over successive years (Smith, 1984). All known Antechinus species are sexually dimorphic for size, with males growing up to three times heavier than females (Marlow, 1961, Sale et al., 2013). Males gain weight rapidly prior to mating, enabling them to eat little during this period and focus on the strenuous mating effort (Wood, 1970). In 2010, there were 10 named species of Antechinus. Since that time, the genus has been revised based on combined DNA and morphological analyses and now includes 15 species (see Baker et al., 2012, Baker et al., 2013, Baker et al., 2014, Baker et al., 2015). The buff-footed antechinus (A. mysticus), found in eastern Queensland, was named after it was discovered sheltering under the geographically and taxonomically expansive umbrella of yellow-footed antechinus, A. flavipes (Baker et al., 2012, Baker and Van Dyck, 2012). The search for further populations of A. mysticus led to the discovery of its purported sister species, the silver-headed antechinus (Antechinus argentus) (Baker et al., 2013). At present, only two small, isolated populations of A. argentus are known. Both occur in central-eastern Queensland, Australia: one at Kroombit Tops National Park (70 km SSW of the coastal city of Gladstone) and another at Blackdown Tableland National Park (220 km W of Gladstone). Our study focused on Kroombit (Baker et al., 2013), as the Blackdown population was not discovered until July 2015. Here, we present results of the first mark-recapture study of A. argentus, describing core life-history events. Across 2 years of monthly trapping, our aims were to determine if A. argentus:

1. exhibited highly synchronised breeding with a male die- off, and if so compare its timing in relation to congeners;

2. males were larger than females and to track patterns in growth and development of both sexes over time;

31

3. abundance varied between years and sites. With its habitat apparently highly restricted, threatening processes such as increased wildfire frequency and intensity and the presence of feral animals are likely impacting A. argentus (Baker et al., 2013, Mason et al., 2015). In 2015, the species was listed as vulnerable under the Queensland Nature Conservation Act 1992. The present work provides information on the ecology of A. argentus critical for its conservation management.

3.2 Materials and methods

3.2.1 Study sites Kroombit Tops National Park is a mesic, temperate plateau that stands above the surrounding drier, hotter subtropical lowlands (McDonald and Sharpe 1986; Hines 2014). Temperatures upon the plateau are notably cooler than those below: summers are generally warm to hot, winters are fine and cool. Mean annual rainfall in the vicinity of the A. argentus type locality is estimated at 1400–1800 mm (Baker et al., 2013). Biogeographically, the area is unusual: many temperate plant, fungi, and animal species occur here at their northern limit or as northern disjunct populations (McDonald and Sharpe, 1986, Monteith, 1986, Hines, 2014). Despite substantial ongoing survey effort by us and others within Kroombit Tops (~10,000 trap nights [one trap night = one trap open for one night] across more than 20 sites), A. argentus is only known from two sites, about 6.5 km apart, on the eastern edge of a gently undulating sandstone plateau (elevation of 850–900 m), bounded on the eastern side by an escarpment with cliffs up to 50 m in height (Baker et al., 2013) (Figs 3.1 and 3.2). The habitat at these two sites and much of the intervening area comprises a blackbutt, Eucalyptus montivaga, with a subdominant brown bloodwood, Corymbia trachyphloia, shrubby tall open forest (Regional Ecosystem 12.9–10.20 Eucalyptus montivaga woodland on sedimentary rocks, Queensland Herbarium 2015; Baker et al., 2013). The escarpment has a strong orographic

32

effect, so this area is often blanketed in clouds resulting in higher precipitation and lower temperatures (Hines, 2014). However, after prolonged dry spells, these forests are highly flammable. In October 2013, immediately prior to our study, a wildfire burnt a large proportion of the eastern plateau and escarpment. Intensity of this fire was mostly moderate, with areas immediately adjacent to drainage lines unburnt, but there were some areas of high intensity with complete crown scorch (Queensland Parks and Wildlife Service, unpublished data). In the present study, trapping was undertaken at these two sites. The southern or Lookout site (lat. 24.396 S, long. 151.044 E) comprised a grid of 150 traps on the plateau immediately west of the escarpment. We elected to split the grid at this site north-south across a dirt road (Kroombit Forest Drive). The habitat south of the road was long unburnt, but the area to the north was burnt in October 2013 under moderate intensity. The second or Northern site (lat. 24.355 S, long. 151.005 E) fell well within the perimeter of the October 2013 wildfire and was burnt with moderate to high intensity. The two sites are broadly similar floristically although sclerophyllous shrubs and grass trees Xanthorrhoea are more prevalent at the Lookout and grasses and ferns more prevalent at the Northern site. A detailed floristic study will be reported elsewhere. These two sites had been trapped at various times from 2011 to 2013. The Northern site yielded 10 captures from 700 trap nights and the Lookout site 12 captures from 2050 trap nights (Baker et al., 2013).

33

Fig 3.1 The Lookout site at Kroombit Tops National Park, with the trunk of a Eucalyptus

montivaga in the foreground showing fire scarring, and a grass tree Xanthorrhoea johnsonii

in front centre, which are prevalent at this site. Photograph by Eugene Mason.

Fig 3.2 Location of the two main study sites within Kroombit Tops National Park.

34

3.2.2 Animal trapping A trapping grid of 150 traps was established at each site using a measuring tape and compass. The location of each trap point was marked in the field with flagging tape with the trap number written on it, to ensure consistent placement of the traps across all trapping periods. The layout of the grids varied somewhat between the two sites due to the recent fire history and the presence of roads and tracks. However, each grid comprised six lines of 25 traps laid out at 10 m intervals. We used box traps (Type A Elliott traps, 23 × 9 × 8 cm, Elliott Scientific, Upwey, Victoria, Australia). Trapping was carried out for a week of each month from March to September of 2014 and 2015. Traps were baited with a mixture of peanut butter and oats as per standard (Pearson and Ruggiero, 2003). Traps were opened for three nights during each trapping period, totalling 450 trap nights at each site per month. When A. argentus individuals were captured at a site, the traps were closed for the following night to allow the animals to forage naturally. Open traps were rebaited daily. Traps were thoroughly cleaned with detergent, rinsed with water and then rebaited after any small mammal capture (Tasker and Dickman, 2001). Captured A. argentus individuals were weighed, sexed and PIT tagged (using Biomark GPT12 PIT tags) to ensure accurate mark-recapture data, before immediate release at the point of capture.

3.2.3 Age and reproductive condition To determine if captured A. argentus individuals were considered adult, we ensured the presence of a fully descended third upper premolar tooth (as per Archer, 1976). To assess reproductive condition of females, the pouch area was examined. We scored the condition of the pouch with grades assigned 1– 4, as per Woolley (1966) and Gray (2013). All known Antechinus species lack a true pouch, but in immature females the pouch area is visible as an area on the abdomen covered with white hairs, and nipples are almost indiscernible (grade 1). Immediately prior to mating, the pouch area typically be- comes

35

red, and the nipples become visible (grade 2). During pregnancy, the pouch area expands (grade 3) and finally develops a granular appearance (grade 4) (Woolley, 1966). Following lactation, the pouch decreases in size, but never returns to the immature state. Nipples remain enlarged and clearly visible. When females were carrying a litter, the crown-rump measurements of three randomly chosen young were measured in-situ using a Mitutoyo dial caliper to the nearest 0.1 mm.

3.2.4 Body weight data Captured A. argentus individuals were placed in calico bags and weighed to the nearest 0.1 g using a Pesola spring scale (Pesola AG, Schindellegi, Switzerland). After recording this value, the captured individuals were removed and the bags were weighed. The weight of the bag alone was then subtracted from the first weight to obtain the weight of the A. argentus individual.

3.2.5 Data analyses The A. argentus population at Kroombit Tops is only known from our two sites (Lookout and Northern). We therefore were required to consider the limitations of using just two sites for statistical analyses. In particular, we were aware of the potential for misleading results due to autocorrelation and pseudo-replication. Recognizing this, we analysed data using generalized linear mixed models (GLMMs), treating individuals as random effects, and weight as the response variable. We modelled sex, site and month as functions of weight. As antechinuses exhibit a synchronised annual life-history pattern, we treated months for each year as the same. To account for autocorrelation due to repeated captures of an individual within a single month, we implemented an AR1 correlation structure into the model. The models that best fit the data were selected based on the Akaike Information Criterion (AIC) (Johnson and Omland, 2004). It was determined that models that fell within four AIC values of each other could be considered equivalent (Burnham and Anderson, 2003, Grueber et al., 2011). R2 values were calculated as per

36

Nakagawa and Schielzeth (2013). Residuals from the six models in the best- model set were plotted to determine whether they exhibited spatial structure, thereby testing the assumptions of the analysis (in particular, independence and linearity) (Bolker et al., 2009). GLMM analyses were fit using the lme4 package (Bates et al., 2014), and all model comparisons were undertaken using the model selection function in the MuMIn package (Bartoń 2013) in R version 3.2.2 (R Core Team, 2015). We were unable to test for differences in weight between years with the GLMM analysis, because the variable comprised only one degree of freedom (two years), rendering it rank-deficient and incompatible with the method (Bolker et al., 2009). Therefore, a one-way analysis of variance (ANOVA) was carried out to analyse for differences in weight within sexes and between years. Because repeated captures of one female were made across the two years, an error structure for individuals was used in the ANOVA for females. These tests were undertaken using R version 3.2.2 (R Core Team, 2015).

3.3 Results

3.3.1 Aim 1 The total trap effort at the Lookout and Northern sites throughout 2014 and 2015 was 12,600 trap nights. From this trap effort, there was a total of 70 A. argentus captures that included 19 individuals (10 females and 9 males, Fig 3.3). Taken over the entire trapping period, this equates to an overall trap success rate of just 0.56% (one capture per 175 trap nights). All A. argentus individuals captured throughout the study (March–September) were adults, as indicated by a fully descended third upper premolar tooth (Archer, 1976). In 2014, the last A. argentus male capture occurred on 15 June. The individual appeared healthy, energetic and had turgid testes. The remaining trapping for 2014 (July 20–26; August 13–18; September 13–18) resulted in only female captures. Comparably, in 2015, the last male capture (also healthy) occurred on 12 June, with only female captures observed for the remainder of the year’s trapping (July 14–20; August 14–19; September 12–17). The last male

37

captured in June of each year was made on the last trap night of their respective monthly survey. In the following month of each year, only females were captured (July 20–26, 2014; July 14–20, 2015). Timing of birth and size of young are also useful in determining synchronicity of male die–off in Antechinus (Fisher et al., 2006). Female A. argentus with pouch young were first observed in the August trapping of both years (August 14, 2014; August 16, 2015). Mean size of these young across both years was 6.83 mm. Laboratory breeding studies on the closely related and similar-sized A. flavipes (Marlow, 1961) suggest young are about 4.9 mm at birth and reach 7.5 mm by about one week of age. A. flavipes have a gestation of about 31.5 days (Marlow, 1961). The young of A. mysticus, purported sister species of A. argentus (Baker et al., 2013), have a crown-rump length of about 3 mm at 1 day old, and grow to approximately 5.8 mm by one week of age (Gray, 2013). Applying these data to our observations of A. argentus suggests that young observed in August were about 1 week old, with ovulation predicted to commence at about July 6–8. Trapping was stopped after September of both years, as applying past research on Antechinus breeding biology to our measurements of young indicated that the young were likely to detach from the mother imminently (Marlow, 1961). Although at this point the young do not remain fused to the mother’s teats, they are still reliant on her breast milk for nutrition (Woolley, 1966). Capturing a female A. argentus at this time could therefore result in death of the young from malnutrition, as mothers often visit the nested young several times during the night between foraging bouts (Cockburn et al., 1985). Taken together, our results suggest that as for all congeners (Woolley, 1966,

Braithwaite and Lee, 1979, Bradley et al., 1980, Kraaijeveld‐Smit et al., 2003,

Sale et al., 2009, Gray, 2013), A. argentus undergoes a synchronised, annual male die-off. The mating period of A. argentus at Kroombit Tops ostensibly takes place from mid-late June to early July, with the male die-off plausibly concluding in early July and birth of young occurring in late July to early August.

38

3.3.2 Aim 2 A. argentus were found to be sexually dimorphic for size, with males larger than females. Mean weights of A. argentus individuals of both sexes were higher in 2015 than 2014 (Fig 3.4). However, while there was a significant difference in weight between years for females (F[1,32] = 31.35, p < 0.001), this was not the case for males: (F[1,20] = 1.121, p = 0.302). In 2014, female weights taken from 9 captured individuals ranged from 16.4 g in March - 28.5 g in September (mean 27.2 g), while in 2015 the female weight range recorded from two captured individuals was 24.8 g in April—30.8 g in September (mean 20.2 g). However, the lower and upper bounds of female weights in 2015 were taken from a single individual that had been previously captured in 2014 and was therefore determined be in her second year of life. Indeed, all female captures in 2015 (except one, an individual captured in August weighing 25.3 g) were of this individual. Male weights taken from six individuals captured in 2014 ranged from 28.8 g in March to 51.7 g in June (mean 38.5 g). In 2015, male weights of the three individuals captured ranged from 38.2 g in March to 50.1 g in June (mean 43.2 g). Notably, the lowest male weight recorded in March 2015 was 9.4 g higher than the highest male weight recorded in March 2014. The six models included in the best model set (Table 3.1) each explained large amounts of variation within the data, with marginal and conditional R2 values ranging between 0.8 and 0.81. Every model included month, sex and site as functions of weight, indicating that each variable explains a significant amount of variation in weight data. Specifically, weights differed between sexes, and also showed variation between months and sites. The variation in weight between months was consistent with studies on congeners, with males typically increasing in size up until the breeding period (Wood, 1970, Dickman, 1989).

39

Fig 3.3 Number of individual A. argentus captured at Kroombit Tops in 2014 and 2015 and

key life-history events.

Fig 3.4 Boxplot of female and male weights of A. argentus individuals recorded in March–

September of 2014 and 2015.

The typical expansion and granular appearance of the pouch exhibited in pregnant females of congeners was never observed in A. argentus, most likely because of our 1 week/month trapping program. Rather, we observed pouches in the immature (stage 1) state of Woolley’s (1966) pouch scoring system from

40

months March–July in 2014, and April–July in 2015, and then with young attached (from months August– September in both years). Intermediate stages 2–4 of sinking, expanding pouch and reddening of skin (respectively) were assumed to occur in females between the trapping periods in July and August. One A. argentus female was regularly recaptured in both years of the study (Fig 3.5). In 2014, this individual was first captured on June 12. At this time, the pouch area was covered in hair and nipples were not apparent, which combined with meristic data indicated that she was a first year animal (Woolley, 1966). In July 2014, the individual was captured twice (July 23 and 25), with hair loss around the pouch area barely noticeable (stage 1). The individual was then captured twice in August (14 and 16) 2014 bearing eight pouch young with mean crown-rump length increasing from 6.68 to 7.02 mm. Finally, the individual was captured once in September of 2014 (15), still bearing eight pouch young with mean crown-rump length 18.53 mm (age estimated at 5 weeks). This female was ‘trap-happy’ and continued to be captured regularly throughout 2015, in every month from April–September. Throughout 2015, eight nipples were clearly visible on the individual’s pouch area, typical of female Antechinus in their second year of life (Woolley, 1966). On August 16, 2015, this female was observed carrying eight young with mean crown-rump length of 6.97 mm, which was of comparable size to the young observed on the same date in 2014 (χ = 7.02 mm). However, mean crown-rump length of the eight young in September 2015 was markedly lower than in September 2014 suggesting an earlier conception during the synchronised breeding period and/or a variation in growth-rate (13.67 mm on September 15, 2015, compared to 18.5 mm on September 15, 2014) (Fig 3.5).

41

Table 3.1 Summary of the six best generalized linear mixed effects models used to describe

differences in weight among A. argentus, ordered from lowest AICC value (ie. the best-fitting

model) to highest. AICC denotes the Akaike information criterion value corrected for small

sample size. R2M represents the marginal coefficient explained by fixed factors, and R2C

represents the conditional coefficient explained by both fixed and random factors.

Terms df logLik AICc delta weight R2M R2C

123467 11 -150.96 329.00 0.00 0.42 0.8 0.8

123456 10 -153.19 330.53 1.52 0.20 0.8 0.8

12346 9 -154.67 330.67 1.67 0.18 0.8 0.8

12345 9 -155.62 332.58 3.58 0.07 0.8 0.8

1234 8 -156.99 332.59 3.59 0.07 0.81 0.81

124 7 -158.44 332.89 3.88 0.06 0.81 0.81

Term codes:

Month Sex Site Month:Sex Month:Site Sex:Site Month:Sex:Site

1 2 3 4 5 6 7

3.3.3 Aim 3 Annual captures (March–May) at the two primary study sites (Lookout and Northern) were markedly higher in 2014 than 2015, with 54 captures of 15 individuals (nine females and six males) recorded in 2014, and just 16 captures of 5 individuals (two females and three males) recorded in 2015. In 2014, the highest monthly capture rate was recorded in June (20 captures of 8 individuals), while in 2015, the highest monthly capture rate was recorded in May (5 captures of 3 individuals). The vast majority of captures were made at the Lookout site, with just one female and two males captured once each over the two-year trapping period at the Northern site.

3.4 Discussion

A. argentus exhibits semelparous life-history characteristics comparable to its congeners, being considerably sexually dimorphic (Table 3.1, Fig 3.4), exhibiting growth in the months before breeding (especially in males), with a male die-off (Fig 3.3) at the conclusion of a synchronous, annual breeding

42

period that occurs relatively early among south-eastern Australian Antechinus (Scott, 1986, McAllan et al., 2006, Gray, 2013, Pearce et al., in review). Despite an intensive trapping effort over 7 months in each of 2 years, captures in the present study overall were very low, decreased strikingly from 2014 to 2015 (Fig 3.3), and were almost entirely made at one of the two sites (Lookout). We hypothesise low abundance of the species at Kroombit Tops. A. argentus exhibits synchronised breeding followed by male die-off over a short period, typical of all Antechinus species studied to date. McAllan et al. (2006) suggested that the timing of reproduction in Antechinus may allow females to time the birth of young with an increase in food availability— specifically a peak in insect abundance characteristic of the warmer months that would be crucial to development of young in this group of generalist predators (Wainer, 1976, Braithwaite and Lee, 1979, Fisher et al., 2013). We therefore expected that A. argentus would ovulate at a similar time to other members of its phylogenetic clade (A. flavipes and A. mysticus), which occupy a comparative mid-latitudinal range (Baker et al., 2015). On the contrary, our data indicated that A. argentus at Kroombit (lat. -24.4, alt. 850–900 m) began ovulating between July 6 and 8. This is 4–6 weeks earlier than A. flavipes (September) from Brisbane (lat. −27.3, alt. 150 m) and A. mysticus (August) from both Brisbane (lat. −27.3, alt. 150 m) and Eungella (lat. −21.2, alt. 700 m). A. argentus does not appear to fit well with a prevailing pattern in Antechinus of ovulation timing being driven by rate of change of photoperiod (latitude and altitude) (McAllan et al., 2006).

43

Fig 3.5 Timeline of captures for a regularly recaptured individual female A. argentus in 2014

and 2015.

Intermittent trapping surveys in 2011–2013 yielded comparable A. argentus capture rates at both sites (to date, the only known A. argentus habitats at Kroombit Tops) (Baker et al., 2013), but the present study (commencing in March 2014) resulted in only one female and two males captured once each at

44

the Northern site, with the remaining captures (67 captures, nine females and seven males) occurring at the Lookout site. Plausibly, these uneven capture rates between sites may be partially explained by a reduction in habitat complexity caused by a wildfire that affected the entire Northern site and half of the Lookout site in October 2013. Fox (1982) suggested that changes in small mammal abundance after fire may be a response to changes in vegetation, as opposed to the fire itself (Fox, 1982). Furthermore, a study on A. flavipes found that higher complexity in forest structure and higher litter cover positively affected the presence of individuals in certain habitat (Kelly and Bennett, 2008). Other studies have found that abundance of antechinuses is negatively affected by fire both in the short-term (A. stuartii has been known to recolonise burnt areas, albeit at low levels, within 6 months of a fire) (Fox, 1982) and the long-term (populations of A. minimus have been driven to extinction in some areas after fires) (Wilson et al., 2001). A parallel study by EM aims to quantify vegetation diversity and regrowth structure in association with A. argentus density, which may shed more light on the putative relationship between habitat complexity and population dynamics. Ongoing widespread trapping effort throughout a range of habitat types suggests that Kroombit Tops A. argentus is limited to the ~10 km2 area on the eastern edge of the plateau that encompasses our study sites (data not shown). This suggests that this species is highly habitat-specific, and therefore likely particularly sensitive to disturbance. Although no formal analysis of feral animal impacts was undertaken here, cattle (Bos taurus), pigs (Sus scrofa), horses (Equus caballus), cats (Felis catus) and dogs (Canis familiaris) are present throughout Kroombit Tops (Queensland Parks and Wildlife Service un- published data). There is growing evidence that the removal of vegetation cover caused by intense fire facilitates increased predation on small mammals, particularly by feral cats (Leahy et al., 2016, McGregor et al., 2014). Future planned trapping surveys of A. argentus at Kroombit Tops will help us to better understand and quantify the long-term response of this threatened species to fire and ferals.

45

In summary, like its congeners, A. argentus exhibits sexual dimorphism for size, male-specific semelparity, and a synchronised breeding period, occurring between late June and early July. Marked yearly population fluctuations are evident. Further investigation into the impacts of fire and feral animals is needed. With A. argentus being listed as vulnerable (Nature Conservation Act 1992) in 2015, managing any threats will be imperative for the conservation of the species. Indeed, as one of only two known (restricted and fragmented) populations of the species, understanding the drivers of growth, reproduction and population dynamics is crucial to their survival. We aim to gather this information in the coming years with a view to obtaining a federal threatened species listing for A. argentus. An action plan will follow, in the hope of sustainably managing this species as it faces an uncertain future.

3.5 Acknowledgments

Our study was generously funded by the Fitzroy Basin Association and the Burnett Mary Regional Group. EDM was assisted by an Australian Postgraduate Award scholarship for research. QUT granted the use of 4WD vehicles, without which this study would not exist. Bulk gratitude goes out to field trip volunteers: Dave Warner, Jarrah “Steve” Wills, Thomas “Forgot to come” Mutton, Jordan Rochfort, Matt “Zig Zag” Turner, Laura Allen, William Mason, Karl “Dente” Stone, Jake “Jakey BOI” Viel, “B.I.G.” Ed White, Kirsten Wallis, Paul O’Callaghan, Emily Coleman, Reece Newnham, Chai “Animals are such dogs” Glandfield, Temma Lee and Mie “Mad Dog” Geertsen. Thanks to Peter Pickering Ranger-Charge Kroombit Tops National Park for supporting the work, especially for allowing use of the barracks for accommodation. Research was conducted under permit WITK14454914 granted by the Department of Environment and Heritage Protection, and was approved by the Australian Ethics Committee (project number 1400000003).

46

Chapter 4. Plant diversity and structure

describe the presence of a new, threatened

Australian marsupial within its highly restricted,

post-fire habitat.

Eugene D. Mason1,4, Jennifer Firn1, Harry B. Hines2,3, and Andrew M. Baker1,2

1Earth, Environmental and Biological Sciences School, Queensland University of Technology,

2 George St, Brisbane, Queensland 4001, Australia 2Queensland Museum, PO Box 3300, South Brisbane, Qld 4101, Australia 3Queensland Parks and Wildlife Service, Department of National Parks, Sport and Racing,

Level 19, 111 George St, Brisbane, Queensland 4000, Australia 4Corresponding author. Email: eugene.mason@connect.qut.edu.au

PLOS ONE, 2017, 12(8):e0182319

DOI 10.1371/journal.pone.0182319

Abstract:

Management of critical habitat for threatened species with small ranges requires location-specific, fine-scale survey data. The silver-headed antechinus (Antechinus argentus) is known from only two isolated, fire-prone locations. At least one of these populations, at Kroombit Tops National Park in central-eastern Queensland, Australia, possesses a very small range. Here, we present detailed vegetation species diversity and structure data from three sites comprising the known habitat of A. argentus at Kroombit Tops and relate it to capture data obtained over two years. We found differences in both vegetation and capture data between burnt and unburnt habitat. Leaf litter and grasstrees (Xanthorrhoea johnsonii) were the strongest vegetative predictors for A. argentus capture. The species declined considerably over the two years of the trapping study, and we raise concern for its survival at Kroombit Tops. We suggest that future work should focus on structural vegetative variables (specifically, the diameter and leaf density of grasstree

47

crowns) and relate them to A. argentus occurrence. We also recommend a survey of invertebrate diversity in grasstrees and leaf litter with a comparison to A. argentus prey. The data presented here illustrates how critical detailed monitoring is for planning habitat management and fire regimes, and highlights the utility of a high-resolution approach to habitat mapping. While a traditional approach to fire management contends that pyrodiversity encourages biodiversity, the present study demonstrates that some species prefer long-unburnt habitat. Additionally, in predicting the distribution of rare species like A. argentus, data quality (i.e., spatial resolution) may prevail over data quantity (i.e., number of data).

Received 5 April 2017, accepted 16 July 2017, published online 10 August 2017

4.1 Introduction

In Australia, loss of habitat means that natural processes such as fire have become a disproportionately larger threat to fauna (Burbidge and McKenzie, 1989). The Australian fauna have adapted to the destructive effects of fire in part by seeking refuge in surrounding habitat (Robinson et al., 2013). However, the destruction and fragmentation of habitat by humans has resulted in less available refugia, especially for species with small ranges (Andren, 1994, Robinson et al., 2014). Climate change is predicted to increase the frequency and intensity of wildfires, further exacerbating the risk (Lucas et al., 2007). With this comes an additional threat, especially for small and medium-sized mammals: increased predation by invasive carnivores (Dickman, 1996). Feral cats in particular can travel large distances specifically to hunt within a recently burned area (Leahy et al., 2016). With habitat loss accelerating and the frequency and intensity of fires increasing, it is critical that conservation measures for threatened species in fire-prone areas are improved (Ceballos et al., 2015, Dale et al., 2001). It is important that the specific needs of threatened species are understood, thereby allowing existing conservation areas to be managed in such a way as to reduce the likelihood of their extinction. Species with small ranges tend to

48

be rare within these ranges, increasing the risk of extinction (Brown, 1984). Given that these species are likely highly habitat-specific, investigating their fine-scale habitat use may be beneficial for conservation management. Additionally, while the effects of fire on both vegetation (e.g., Purdie and

Slatyer, 1976, Russell‐Smith et al., 1998) and small mammals (e.g., Letnic et

al., 2004, Lunney et al., 1987) in Australia have been extensively studied, relating the two may be useful for best conservation practice. Of particular concern are threatened wildlife species existing in restricted, fire-prone habitats. The silver-headed antechinus (Antechinus argentus) is one such species. When described in 2013, it was known from a small area in the eastern part of Kroombit Tops National Park in central eastern Queensland, Australia (Baker et al., 2013). Despite extensive trapping at Kroombit and some other montane areas of similar habitat in central Queensland since its discovery, only two populations of the species are known: one at the type location, an ~10 km2

area within Kroombit Tops National Park, central eastern Queensland, Australia; and one at Blackdown Tableland National Park, ~200 km to the west of Kroombit (Mason et al., 2017). As well as having a highly restricted range, concern for the species’ status is heightened by males being semelparous - all die at the end of a highly synchronised annual breeding period in late June-early July of each year, eliminating any chance of breeding again and effectively halving the population every year (Mason et al., 2017, Woolley, 1966). In addition, other threatening processes are known from its habitat including feral predators and grazing from cattle and horses. Due to these factors, A. argentus was listed as vulnerable under the Queensland Nature Conservation Act 1992 in 2015. The present study was undertaken on the type population of A. argentus at Kroombit Tops. Just months after A. argentus was described, the majority of its habitat at Kroombit Tops was burnt by a wildfire. Research on other Antechinus species indicates that a lengthy period of vegetation regrowth is needed for populations to recover after a fire (see Aberton, 1996, Swinburn et al., 2008, Wilson et al., 2001), and a recent study on A. argentus hypothesised

49

that this species may be sensitive to fire (Mason et al., 2017). Particularly concerning is that although Kroombit Tops comprises a 74.6 km2 area, and in spite of an ongoing, extensive trapping effort of ~10,000 trap nights by us and others, A. argentus is only known from a very small (<10 km2) section of the national park even though similar habitat is much more extensive. We therefore hypothesised that A. argentus may have specific plant-related habitat preferences. The present study related detailed post-fire plant community composition and structural data to the distribution and abundance of one of only two known populations of A. argentus, within its highly restricted range at Kroombit Tops National Park in Queensland, Australia. We aimed to investigate whether: 1. Differences in plant community composition and structure following a

fire correlate with differences in capture rates of A. argentus; 2. Plant community composition and structure influence the occurrence

of A. argentus; 3. A. argentus abundance changed as the post-fire habitat recovered. By relating detailed vegetation data to our small mammal trapping data, we aimed to uncover abundance and habitat preference patterns of A. argentus within its small range at Kroombit Tops. We compared our research to existing studies that took similar approaches, and discuss whether this approach may be useful for effective management of threatened wildlife with small, isolated populations in fire-prone habitats.

4.2 Materials and methods

4.2.1 Study sites Kroombit Tops National Park is a montane plateau 70 km SSW of the city of Gladstone in Queensland, Australia. Although it is situated in the subtropics, Kroombit Tops has been described as a “mesic temperate island” (Hines, 2014, McDonald and Sharpe, 1986). Mean annual precipitation at the A.

50

argentus type locality is estimated at 1400-1800 mm (Baker et al., 2013), and is mostly concentrated to the summer months, when temperatures are warm to hot. Winters are fine and cool, and frosts are relatively common (Baker et al., 2013). The present study was undertaken at two sites ~6.5 km apart that lie on the eastern edge of a gently undulating sandstone plateau at an elevation of 850-900 m which is bounded on the eastern side by an escarpment with cliffs up to 50 m in height (Baker et al., 2013). An ongoing, extensive trapping effort of ~10,000 trap nights has been undertaken by us and others, yet A. argentus at Kroombit Tops is only known from these two small sites. The habitat at the two sites and the intervening area is broadly described as Eucalyptus montivaga and Corymbia trachyphloia forest with a grassy, ferny or shrubby understory (Baker et al., 2013) (Regional Ecosystem 12.9-10.20, [Queensland Herbarium, 2016]). In October 2013, both sites were burnt by a wildfire. The entire Northern site (lat. 24.355 S, long. 151.005 E) was burnt with moderate to high intensity (as defined for forest ecosystems in southeast Queensland by Queensland Parks and Wildlife Service [QPWS] 2013, p16), while only half of the southern or Lookout site (lat. 24.396 S, long. 151.044 E) was burnt, with mostly moderate intensity. Because half of the Lookout site remained unburnt due to an intervening dirt road (Kroombit Forest Drive) that acted as a fire break, we elected to subdivide the site (into halves) along this road for the present study so that it comprised two smaller sites: the Lookout Burnt and Lookout Unburnt sites. Prior to the October 2013 wildfire, the Northern and Lookout Burnt sites were last burnt in a wildfire in September 2001, and a planned burn in March 2008 (QPWS fire mapping), although no information is available on intensity at our study sites. However the Lookout Burnt site was either unburnt during the 2008 planned burn or burnt at a very low severity, based on the size and presence of fire-sensitive trees and shrubs at that site in June 2013 (assessed from a series of HBH photographs). The Lookout Unburnt site falls within a fire management block attributed with being burnt in a planned burn in 2011 (QPWS fire mapping), but visual assessments (degree of skirting on grasstrees, persistence of fire-scarring on trunks of E. montivaga – see for example Fig 4.1.) strongly suggest that the

51

area falling within our trap grid was not burnt at that time. There are no other records of fires in that block as far back as 1991, indicating that the Lookout Unburnt site is long-unburnt (20+ years).

52

Fig 4.1 (a) The Lookout Unburnt site, (b) Lookout Burnt site, and (c) Northern site at

Kroombit Tops National Park. Photographic credit to Eugene Mason and Harry Hines.

53

4.2.2 Vegetation surveys Vegetation surveys were undertaken twice (November 2014 and January 2016) to capture the change in vegetation over time following the fire. The layout of the surveys followed already-established mammal trapping grids at the three sites, as per Mason et al. (2017). Six transects formed the Northern site, and three transects each formed the Lookout Burnt and Lookout Unburnt sites. On each of these transects, five 10 m2 vegetation plots were laid out at equally spaced intervals (see Trapping surveys section below for spatial arrangement) as per Peet et al. (1998), totalling 30 plots at the Northern site, 15 plots at the Lookout Burnt site, and 15 plots at the Lookout Unburnt site. Within each of these vegetation plots, a count of all tree and shrub species was undertaken and the diameter at breast height (DBH) of all trees and shrubs >10 cm was measured (DBHs <10 cm were recorded as 10 cm) from which tree basal area was derived. Ground cover of eight variables (shrubs, grasses, ferns, herbs, bryophytes, litter, charcoal, bare soil) was estimated as the percentage of each of three randomly placed 1 m2 quadrats inside each 10 m2 plot. The percentage of ground concealed by the aboveground projection of each variable was estimated (Peet et al., 1998). No single ground cover variable could exceed 100%, but total ground cover estimates of all variables within a quadrat could exceed 100% (Daubenmire, 1959). 4.2.3 Trapping surveys Small mammal trapping was undertaken at the three sites for a week of each month from March to September in 2014 and 2015. We used aluminium folding traps (Type A Elliott traps, 23 x 9 x 8 cm, Elliott Scientific, Upwey, Victoria, Australia), and baited them with a standard mixture of oats, peanut butter and vegetable oil (Pearson and Ruggiero, 2003). The trapping grid consisted of 150 traps laid out along the six transects at the Northern site, 75 along the three transects at the Lookout Burnt site, and 75 along the three transects at the Lookout Unburnt site. The layout of traps was such that every

54

fifth trap location formed the centre of one of the 10 m2 vegetation plots. Traps were opened for three nights of each monthly trapping survey period. This totalled 450 trap nights at the Northern site each month, and 225 trap nights at each of the Lookout Burnt and Lookout Unburnt sites for each month. If any A. argentus individuals were captured, traps were closed for the following night to reduce our disturbance of the population, and to allow the animals to forage naturally. 4.2.4 Data analyses Because A. argentus is only known at Kroombit Tops from two areas (and one of these with adjacent burnt and unburnt sites), the scope of our data analyses was unavoidably constrained – an issue for most threatened species, but particularly an issue for species like A. argentus, which have only recently been discovered and are already threatened. Furthermore, highly limited A. argentus captures in 2015 (16 captures of five individuals [two females and three males]) (Mason et al., 2017), precluded formal comparative vegetation and A. argentus capture analysis of the 2015 data. These limitations notwithstanding, comparative statistical analyses were undertaken using the November 2014 vegetation data and the March-September 2014 trapping data. For the plant species composition data, we first square root transformed the data to reduce the dominant contribution of abundant species. We then constructed a resemblance matrix using the Bray-Curtis similarity statistic (Bray and Curtis, 1957). Utilising the permutational multivariate analysis of variance (PERMANOVA) pair-wise test function in PRIMER 7 (version 7.0.1 with add-on PERMANOVA+1), we tested for differences in plant species composition between the three sites (Anderson, 2001, Clarke and Gorley, 2015, McArdle and Anderson, 2001). We treated the 10 m2 vegetation plots as nested within transects, and transects nested within sites. To visualise the differences between vegetation plots, sites and A. argentus captures, we generated a non-metric multidimensional scaling (nMDS) bubble plot, with each data point representing one of the 10 m2 vegetation plots, and the size of

55

the bubbles representing the number of A. argentus captures in 2014. As the vegetation plots were located at every fifth trap location, we counted the A. argentus captures from the trap in the centre of the plot as well as two traps either side of the plot in the transect, following our assumption that these individuals would be accessing proximate areas as habitat. For vegetation structure (ground cover [%] and tree basal area [m2]) data, we constructed a resemblance matrix using the rank transform and the Euclidean distance similarity measure to avoid strong skewness in the distribution over samples, and again tested for differences between the three sites using the PERMANOVA pair-wise test function (McArdle and Anderson, 2001, Anderson, 2001, Clarke and Gorley, 2015). To visualise differences in the structure data between vegetation plots, sites and A. argentus captures, we generated a principal component analysis (PCA) bubble plot, again representing each 10 m2 vegetation plot as a data point and number of A. argentus captures (as derived above) determining the size of bubbles. To assess the relative importance of specific predictor variables for the capture (and therefore occurrence) of A. argentus at our sites, we developed two boosted regression tree (BRT) models (Elith et al., 2008, Friedman, 2001). BRT models are simple classification or rule-based models that use a series of binary splits dependent on predictor variables to partition observations into groups based on similar values in the response variable (Ridgeway, 2006). The boosting algorithm then iteratively fits models to the data in a forward, stage-wise procedure. This method allows for different types of predictor variables to be included in a single model. For our BRT models, we represented the response variable of A. argentus captures as presence or absence data, and therefore used a Bernoulli distribution. The first model included the species abundance data, and the relative influence of each species is shown in Table 4.1. The second boosted regression model included the ground cover (%) and tree basal area (m2) data, as well as the average Bray-Curtis similarity for each 10 m2 vegetation plot (Bray and Curtis, 1957). To obtain this value, the Bray-Curtis similarity of ground cover

56

and tree basal area of each vegetation plot was compared to every other vegetation plot in the survey, then the average of these values was calculated for each vegetation plot. The relative influence of each of these variables is shown in Table 4.2. We also show the number of trees needed in each of the BRT models and the estimates of the correlation between observed and expected response variables for both of the final models. The BRT models were fitted in the R statistical computing program version 3.2.2 using the gbm package version 2.1.1 (Ridgeway, 2006, R Core Team, 2015).

4.3 Results

1. Do differences in plant community composition and structure following a fire correlate with differences in capture rates of A. argentus? While the three sites were not unambiguously separated based on nMDS of vegetation species diversity, two-dimensional structure was suggested by the grouping of the Lookout Unburnt (red) and Lookout Burnt (green) vegetation plots into diagonal lines (Fig 4.2). Additionally, there was structure in the number of A. argentus captures, as indicated by the size of the bubbles representing each vegetation plot. In particular, the majority of captures were made at the Lookout Unburnt site, followed by the Lookout Burnt site. Multivariate analysis of species abundance showed a significant difference between the Northern and Lookout Unburnt sites in 2014 (PERMANOVA, t=2.05, P=0.01), while evidence against the null hypothesis was weak for comparison of the Northern and Lookout Burnt sites (PERMANOVA, t=1.52, P=0.08), and weaker still for comparison of the Lookout Burnt and Lookout Unburnt sites (PERMANOVA, t=1.15, P=0.39).

57

Fig 4.2 Non-metric multidimensional scaling (nMDS) bubble plot of vegetation species

diversity at each 10 m2 vegetation plot. The size of the bubbles indicates the number of A.

argentus captures during 2014. N = Northern, LU = Lookout Unburnt, LB = Lookout Burnt.

0.1 represents vegetation plots with nil captures.

Similar results were found for the vegetation structure (ground cover and tree basal area). Again, explicit separation was not evident between the three sites based on PCA, although the Lookout Unburnt (red) vegetation plots were grouped in a diagonal line (Fig 4.3). Multivariate analysis of vegetation structure showed there was a significant difference between the Northern and Lookout Unburnt sites in 2014 (PERMANOVA, t=2.92, P=0.01), but evidence against the null hypothesis was weak for comparison of the Lookout Burnt and Lookout Unburnt sites (PERMANOVA, t=2.83, P=0.09) and weaker still for comparison of the Northern and Lookout Burnt sites (PERMANOVA, t=1.12, P=0.34).

58

Fig 4.3 Principal component analysis (PCA) bubble plot of ground cover (%) of litter, grasses,

ferns, bare soil, charcoal, bryophytes, herbs, shrubs and dead wood and tree basal area (m2) at

each 10 m2 vegetation plot. The size of the bubbles indicates the number of A. argentus

captures during 2014. N = Northern, LU = Lookout Unburnt, LB = Lookout Burnt. 0.1

represents vegetation plots with nil captures.

2. Does plant community composition and structure influence the occurrence of A. argentus? Nineteen tree and shrub species were recorded from the vegetation plots, but of these only five were found to exhibit a relative contribution to A. argentus captures in the tree and shrub diversity and abundance BRT model (Table 4.1). The model indicated that the most important tree or shrub species for the capture of A. argentus was the grasstree X. johnsonii (38.3%, Table 4.1) followed by the bloodwood C. trachyphloia (27.4%, Table 4.1). The estimate of the correlation between observed and predicted response variables was 74%.

59

Table 4.1 Boosted regression tree model. Summary of the relative contributions (%) of

predictor variables (tree and shrub species abundance) developed with cross validation on

data using 1400 trees, tree complexity of 2 and learning rate of 0.004.

Variable Relative influence (%)

Xanthorrhoea johnsonii 38.3

Corymbia trachyphloia 27.4

Allocasuarina torulosa 15.5

Elaeocarpus reticulatus 11.2

Eucalyptus montivaga 7.6

The partial responses of A. argentus captures for the six most influential variables in the ground cover (%), tree basal area and average Bray-Curtis model indicated a species that favours areas with ground cover dominated by high amounts of leaf litter and the presence of herbs and grasses, and that are relatively different to the overall habitat (as indicated by lower average Bray-Curtis similarity values). By far the most important variable was leaf litter (41.2%, Table 4.2). The estimate of the correlation between observed and predicted response variables was 87%. Table 4.2 Boosted regression tree model. Summary of the relative contributions (%) of

predictor variables (average Bray-Curtis values, tree basal area and ground cover [%])

developed with cross validation on data using 1200 trees, tree complexity of 2 and learning

rate of 0.004.

Variable Relative influence (%)

litter 41.2

Average Bray-Curtis 14.2

herbs 14.1

grasses 14.0

tree basal area 6.4

ferns 4.7

charcoal 3.6

bare soil 1.3

dead wood 0.6

3. Does A. argentus abundance change as the post-fire habitat recovers? Despite a longer period of post-fire vegetation recovery, A. argentus captures dropped dramatically overall from 2014 (54 captures of 15 individuals) to 2015

60

(16 captures of 5 individuals) (Fig 4.4). The highest capture rate per site for both years was at the Lookout Unburnt site (35 captures of 9 individuals in 2014; 11 captures of 3 individuals in 2015). Across both years, more captures were recorded at the Lookout Burnt site (2014: 17 captures of 6 individuals; 2015: 3 captures of 3 individuals) than the Northern site (2014: 2 captures of 2 individuals; 2015: 2 captures of 1 individual).

Fig 4.4 A. argentus captures at Kroombit Tops in 2014 and 2015.

Average ground cover (%) of the most influential structure variable for A. argentus captures in 2014 (leaf litter) increased at the Northern (55.9-87.3%) and Lookout Burnt (64.2-82.1%) sites between vegetation surveys, but decreased slightly at the Lookout Unburnt site (98.4-87.6%) (Table 4.3). Table 4.3 Average leaf litter cover.

Date Northern Lookout Burnt Lookout Unburnt

Nov 2014 56.0 64.2 98.4

Jan 2016 87.2 82.1 87.5

4.4 Discussion

Differences in plant community composition and structure between recently burnt and unburnt habitat were found to correlate with the occurrence of A. argentus. A. argentus appears to have a preference for habitat not recently

0 5

10 15 20 25 30 35 40

Northern 2014

Lookout Burnt 2014

Lookout Unburnt

2014

Northern 2015

Lookout Burnt 2015

Lookout Unburnt

2015

Cap

ture

s (#

)

Site and year

61

burnt. Within unburnt habitat, the fine-scale occurrence of the species was influenced by ground cover attributes and the presence and abundance of certain plant species, in particular the grasstree X. johnsonii. The data presented here may prove important for habitat management of a highly restricted, isolated population of this threatened species. Our results also highlight the potential utility of relating detailed vegetation data to the ecology of threatened species that have small ranges within fire-prone areas. These aspects of the data are addressed in more detail below. Plant community composition and structure differed between the three sites, with the strongest differences between the Lookout Unburnt site and the Northern site (Figs 4.2, 4.3). The Northern site was completely burnt at moderate to high intensity in October 2013. This pattern was paralleled by the 2014 A. argentus capture rate data, which was highest at the Lookout Unburnt site, and lowest at the Northern site (Figs 4.2, 4.3). The three sites are proximate and well connected with similar habitat; plausibly, these differences could largely be explained by the effects of the October 2013 wildfire (five months prior to the commencement of the present study) on habitat variables, in turn influencing A. argentus capture rate. The effects of fire on other Antechinus species are relatively well-documented (e.g., Aberton, 1996, Chia et al., 2016, Fox, 1982, Hindmarsh and Majer, 1977, Wilson et al., 2001, Penn et al., 2003, Recher et al., 2009). However, these studies focused on Antechinus species with comparatively larger geographic ranges than A. argentus, the latter only being known from two isolated montane plateaux. Moreover, the abundance of A. argentus at Kroombit Tops is apparently very low. A study on A. minimus found that patchily-burnt areas were recolonised at low abundance after a fire, but areas that were fully burnt resulted in local extinction of the species (Wilson et al., 2001). Likewise, Fox (1982) found that a population of A. stuartii recolonized an area six months after a fire, but at a lower abundance (57%) than the pre-fire population, and proposed that this might be a response to changes in the vegetation structure and composition due to fire.

62

In temperate and semi-arid areas of Australia, studies have found that plant community composition and structure influences small mammal distribution (Kelly et al., 2013, Holland and Bennett, 2007, Monamy and Fox, 2010). A number of studies have noted the importance of dense low-level foliage for Antechinus species (e.g., Bennett, 1993, Friend and Taylor, 1985, Knight and Fox, 2000, Newell and Wilson, 1993), but few have described the preferences of Antechinus species for specific vegetation types. Banks et al. (2005) found that abundant large eucalypts had a positive effect on A. agilis population size, with populations decreasing as the distance from eucalypt-dominated forest increased and transitioned into other forest types. Studies on A. stuartii have suggested this species also has a preference for habitat containing grasstree (Xanthorrhoea) species (Wakefield and Warneke, 1967, Wilson et al., 1986, Wilson et al., 1990). Likewise, Swinburn et al. (2008) found that A. flavipes use Xanthorrhoea as important foraging and nesting resources, and Whelan et al. (1996) found evidence of frequent visits from A. stuartii to flowering inflorescence stems of Xanthorrhoea. Furthermore, in post-fire environments these species showed a preference for habitats with numerous surviving Xanthorrhoea individuals. Like several of its congeners and the very closely related A. flavipes, our study found that the grasstree X. johnsonii was the most important floristic predictor for captures of A. argentus. Grasstrees provide habitat for at least 315 invertebrate species (Borsboom, 2005). While other Antechinus species have diverse insectivorous diets, sometimes supplemented with soft, ground-dwelling invertebrates or small vertebrates (Allison et al., 2006, Green, 1989, Lunney et al., 2001), A. argentus has a diet dominated by just two invertebrate orders: Coleoptera and Blattodea (Mason et al., 2015). Both of these are typically found in unburnt or regenerating Xanthorrhoea with dense skirts (Hindmarsh and Majer, 1977, Swinburn et al., 2008), indicating that X. johnsonii may provide an important foraging resource for A. argentus. While we were not able to directly observe any A. argentus foraging or nesting in an X. johnsonii, we did observe an A. argentus individual launching outward at human head height from within a large X. johnsonii skirt at the Lookout Unburnt site during a trapping survey (pers. obs.).

63

In contrast to other species of Antechinus, A. argentus capture rates dropped drastically from 2014 to 2015 (Fig 4.4) despite a longer time period since fire. Leaf litter markedly increased at both of the burnt sites by 2016 (Northern and Lookout Burnt) (Table 4.3). We found that leaf litter was the most important ground cover variable in terms of relative contribution for the occurrence of A. argentus at Kroombit Tops (Table 4.1). A thick leaf litter layer has been shown to influence occurrence of A. flavipes (Stokes et al., 2004). Arthropods found in leaf litter are a necessary food component for carnivorous small mammals including Antechinus, and therefore this might be partially explained by the role of leaf litter for the abundance of invertebrate prey (Andrew et al., 2000, York, 1999). The lower amount of leaf litter at the burnt sites in November 2014 is likely explained by its removal through combustion in the October 2013 wildfire. The volume of flammable leaf litter in the ground cover assemblage can be a determining factor in the initiation of wildfires, and leaf litter accumulation has been shown to be a function of the regeneration age of eucalypt forests since a fire (Fox et al., 1979, Van Loon, 1970). This is reflected in the increase in leaf litter ground cover at the two burnt sites from Nov 2014 to Jan 2016. However, the A. argentus population did not become more abundant in the burnt sites as the time since fire increased. In fact, it apparently declined across all three sites from 2014 to 2015 (Fig 4.3). Plausibly, this could indicate some broader effects of recent fire damage, such as increased predation by introduced fauna. Studies have found evidence that the loss of vegetation cover caused by higher intensity fire allows for a higher proportion of the habitat to be available to feral predators (Leahy et al., 2016). In particular, feral cats (Felis catus) have been shown to travel large distances to reach recently burnt habitat, amplifying predation on exposed native species in the absence of sufficient cover (Dickman, 1996, McGregor et al., 2014, McGregor et al., 2017). Recent evidence has linked the declines of small mammal species with these amplified feral predation effects due to fire, as well as the effects of grazing by introduced herbivores (Legge et al., 2011a). During the present study, we directly or indirectly observed cattle (Bos taurus), pigs (Sus scrofa), horses (Equus caballus), cats (Felis catus) and dogs/dingos (Canis sp.) at the study

64

sites. Grazing by cattle and horses appeared to considerably reduce or slow recovery of palatable species post-fire (mostly grasses), resulting in open areas with very cropped grass between patches of fern, or shrub or fallen timber cover (pers. obs.). Cats are widespread at Kroombit Tops (QPWS unpublished data) and limited camera trapping at the Lookout sites in late 2015 confirmed the presence of one large individual. The combined pressures of grazing and predation from introduced carnivores in a post-fire landscape may be explanatory factors in the decline of the Kroombit A. argentus population, but this is a preliminary hypothesis until future work investigates direct interactions. While the approach of the present study emphasised vegetation composition, we suggest that future research focus on structural habitat elements such as the size and shape of vegetation, for example the diameter and leaf density of grasstree crowns. We also suggest that relating invertebrate species diversity in grasstree crowns to A. argentus prey preference may reveal important patterns. Nevertheless, the apparent importance of specific habitat components such as X. johnsonii for the occurrence of A. argentus highlights the utility of relating fine-scale habitat data to wildlife occurrence. This data will allow future trapping surveys to be targeted in areas of similar habitat in the hope of uncovering additional populations of the species. It also adds to a growing body of literature concerning the response of wildlife to changing fire regimes. Recent work has suggested that fire is likely a prominent feature in the evolutionary history of Antechinus, finding that smoke and a substrate of ash and charcoal may act as a cue for the onset of torpor in A. flavipes (Stawski et al., 2017). However, in contrast to congeners (see Aberton, 1996, Fox, 1982, Wilson et al., 2001), our results indicate that long-unburnt vegetation is important for A. argentus. The present study therefore highlights the importance of a localised approach to fire management, especially in habitats that house threatened species. The future of A. argentus at Kroombit Tops is uncertain; a trapping survey in June 2016 at our three study sites resulted in just one A. argentus capture (data not shown). As one of only two known populations of the species, effective conservation management is vital

65

if it is to survive. As global habitat loss accelerates, unfortunately there are likely to be more species facing comparable threats to A. argentus. The interactions between fire and biodiversity are complex (Gill et al., 2013, Tingley et al., 2016). Many plants and animals require fire to survive, but even in fire-prone habitats some species are highly sensitive to fire (Kelly and Brotons, 2017). A landmark study by Martin and Sapsis (1992) hypothesised that temporal and spatial variation in fires can promote biodiversity by creating a wider variety of ecological niches available for species. A growing body of literature has supported this hypothesis (see Hutto et al., 2016, Ponisio et al., 2016, Tingley et al., 2016). However, pyrodiversity doesn’t always promote biodiversity. A recent study found that in Australian semiarid eucalypt woodland, increasing variation in fire regimes didn’t correlate with increasing bird diversity because long-unburnt vegetation provided disproportionately important habitat (Kelly et al., 2017). Nevertheless, the landscape of Australia has changed dramatically over the last 250 years as a result of human activity, and as a result the ability for fauna to sustain the effects of fire has likely been lowered (Hobbs, 2005). As an important driver of ecology in Australia, appropriate fire regimes will benefit from research into the critical limits of severity and patch size (Kelly and Brotons, 2017). This is a relatively new area of research, but studies suggest that fire management approaches are best tailored to local conditions (see Berry et al., 2015, Moritz et al., 2013), and will rely on specific studies (Kelly and Brotons, 2017) such as the present one. Understanding the effects of fire on threatened species is vital if we are to properly ensure their survival, as ongoing global climate change may ensure increased fire frequency and intensity (Lucas et al., 2007). This is likely to be especially important for species that are rare, highly restricted and that inhabit fire-prone areas. For these species, data quality (i.e., spatial resolution) appears to prevail over data quantity (i.e., number of data) (Engler et al., 2004) and therefore a more fine-scale approach to habitat mapping may inform conservation management strategies. Additionally, fine-scale habitat data can be utilized to predict unsurveyed sites of high potential of occurrence (Guisan and Thuiller, 2005). Effective management of conservation areas that house rare and endangered species is becoming increasingly important, and a

66

greater understanding of their distribution and localised habitat use will help to mitigate threats such as invasive predators and fire.

4.4 Acknowledgements

Our study was generously funded by the Fitzroy Basin Association and the Burnett Mary Regional Group. EDM was assisted by an Australian Postgraduate Award scholarship for research. QUT granted the use of 4WD vehicles, without which this study would not exist. Many thanks to field trip volunteers: Coral Pearce, Emma Hawkes, Jarrah Wills, Karl Stone, Dave Warner, Thomas Mutton, Jordan Rochfort, Matt Turner, Laura Allen, William Mason, Jake Viel, Ed White, Kirsten Wallis, Paul O’Callaghan, Emily Coleman, Reece Newnham, Chai Glandfield, Temma Lee and Mie Geertsen. Gratitude goes out to Peter Pickering Ranger-Charge Kroombit Tops National Park for supporting the work, especially for allowing use of the barracks for accommodation. Research was conducted under permit WITK14454914 granted by the Department of Environment and Heritage Protection, and was approved by the Australian Ethics Committee (project number 1400000003).

67

Chapter 5: General discussion The present research consisted of three complimentary data chapters that together provided a study of the autecology of A. argentus, which was virtually unknown beyond its description as a species when the study began. This document has thus far covered three broad areas: diet composition and strategy (Chapter 2), life-history characteristics (Chapter 3), and post-fire habitat use (Chapter 4). The present chapter proceeds with a discussion of the potential influence of rainfall and photoperiod in the context of A. argentus ecology (Section 5.1), then discusses the detectability of the species and the further surveys undertaken in areas external to Kroombit Tops to determine distributional range (Section 5.2), and finally provides management recommendations with a view towards conservation of the species (Section 5.3).

5.1 Rainfall and photoperiod

5.1.1 Rainfall In Australia, rainfall is often considered to be a major influence on temporal changes in small mammal populations (Dickman et al., 1999, Kelly et al., 2013, Predavec and Dickman, 1994). The present study found that the Kroombit Tops A. argentus population declined considerably from 2014 to 2015 (Chapter 3). The previous chapters have considered the influence of fire and feral animals on the population dynamics of A. argentus, but as yet the potential role of rainfall has not been addressed.

The present study found a decrease in abundance of A. argentus at Kroombit Tops from 2014 to 2015 (Fig 5.1, see Chapter 3).

68

Fig 5.1 Number of individual A. argentus captured in 2014 and 2015, key life-history events

and amount of rainfall each month at The Lookout. Capture data (histogram) applies to the y-

axis scale shown at left (individuals captured) and rainfall data (dotted line) applies to the y-

axis scale shown at right (rainfall). For methods detailing the generation of this figure see

Appendix 1.

In contrast, the mean weight of male and female A. argentus increased from 2014 to 2015, albeit based on low numbers of 2015 captures. Parrott et al. (2007) found no relationship between precipitation and abundance in A. swainsonii. However, they did find that rainfall had a strong effect on abundance and weight of sympatric A. agilis, and proposed that the amount of rainfall around the period of pregnancy and early lactation is important. Similarly, Magnusdottir et al. (2008) found that precipitation had a major effect on population dynamics of A. minimus, with unusually high rainfall followed by a dramatic increase in abundance of the species, and hypothesised that this was due to an increase in prey availability. Van Dyck (1981) argued that while at higher latitudes (e.g., in Tasmania) dasyurid food peaks are driven by a rise in temperature, at lower latitudes (such as in central-east Queensland), they are most affected by rainfall, which tends to increase later in the season (McAllan and Dickman, 1986). A recent long-term study found that small mammals (including A. agilis and A. swainsonii) in fire-prone, temperate ecosystems may display rainfall-driven ‘boom and bust’ phases,

69

with A. swainsonii rarely or not recorded during dry periods and several years after a fire (Hale et al., 2016). While the present study did not detect a significantly abnormal amount of rainfall at Kroombit Tops over the two-year study period, some clear trends were observed (Fig 5.2).

Fig 5.2 Comparison of monthly rainfall at the Lookout from July 2013 to September 2015

(solid line), mean monthly rainfall at the Lookout from 1993 to 2015 (black dotted line) and

Mean 3-year Foley’s index (actual rainfall for 3 years preceding each month minus expected 3-

year rainfall divided by mean annual rainfall) from July 2013 to September 2015 (grey dotted

line) (Foley 1957). Monthly rainfall data applies to the y-axis scale shown at left and Foley’s

index data applies to the y-axis scale shown at right. For methods detailing the generation of

this figure see Appendix 1.

Specifically, a higher amount of rainfall was recorded in August-October 2014 (221 mm) compared to August-October 2013 (47 mm) (Fig 5.2). However, overall moisture due to rainfall in the preceding months was likely considerably lower in August-October 2014 compared to August-October 2013, as indicated by Foley’s drought index (0.76-0.77 Aug-Oct 2014; 2.08-1.85 Aug-Oct 2013) (Fig 5.2). In contrast, rainfall was unusually high in Nov 2014-Jan 2015 (606 mm) compared to Nov 2013-Jan 2014 (345 mm) (Fig

70

5.2), with overall moisture exhibiting a slight increase in Nov 2014-Jan 2015 compared to a sharp decrease in Nov 2013-2014 (Foley’s index: 0.75-0.81 Nov 2014-Jan 2015; 1.76-1.03 Nov 2013-Jan 2014) (Fig 5.2). Among Antechinus species it is typical for young to be attached to the nipples in the pouch for 4-6 weeks, after which they detach but are left in a nest for a further six weeks, reliant on the mother’s milk for nourishment (Fisher et al., 2006, Marchesan and Carthew, 2004, Wood, 1970). Juvenile antechinuses therefore leave the nest at about 10-12 weeks of age and begin hunting and consuming prey independently (Braithwaite, 1979, Dickman, 1982, Cockburn et al., 1985). Assuming A. argentus complies with this strong general pattern for the genus, our data suggest that A. argentus young would leave the nest in late October/early November (Figs 5.1, 5.2), presumably a time when an abundant invertebrate food source is of great importance. Therefore, the amount of rainfall around this time may be important for survival in any given cohort. The data collected here is not powerful enough to show a direct link between rainfall and A. argentus population dynamics. However, timing and amount of rainfall is postulated to be important inter-annually in affecting recruitment and weight of A. agilis (Parrott et al., 2007) and A. minimus (Magnusdottir et al., 2008, Sale et al., 2008) due to its effect on prey availability. It might then be hypothesised that an overall drier spring in 2014 may have contributed to higher infant and juvenile A. argentus mortality, while an unusually high amount of rainfall in the following summer of 2014/2015 (Nov-Jan) (Fig 5.2) caused an increase in insect abundance, providing a surplus of food for the fewer surviving A. argentus and explaining the larger mean weight data (more rainfall and thus food supporting increased growth in fewer individuals). To formally test this hypothesis, data on insect abundance and diversity (pitfall/leaf litter) in the habitat and in A. argentus faecal samples (i.e., food preference) would need to be collected and compared with rainfall and survival rates in A. argentus.

5.1.2 Photoperiod An interesting finding of the present study was that the timing of ovulation (and hence, the breeding period) in A. argentus was comparatively earlier in

71

the year than the other members of its phylogenetic clade, A. flavipes and A. mysticus. It has been suggested that an increase in the rate of change of photoperiod acts as a cue for ovulation in most Antechinus, potentially allowing females to time the birth of young with an increase in food availability - specifically a peak in insect abundance characteristic of the spring/summer months that would be crucial to development of young in this group of generalist insectivorous predators (Fisher et al., 2013, McAllan et al., 2006, McAllan and Dickman, 1986, McAllan and Geiser, 2006). The timing of the beginning of ovulation in A. agilis, A. flavipes, A. stuartii, A. subtropicus, A. adustus, A. leo, A. mysticus and A. arktos ranges from early July to mid-October (Gray, 2013, Gray et al., 2017, McAllan et al., 2006, Mutton et al., 2017, Pearce, 2016). A. bellus and A. godmani begin ovulation slightly earlier in the year (between late June and late July), likely due to their lower latitudinal range (far northern Australia) (McAllan et al., 2006). Phylogenetically, A. swainsonii and A. minimus form a clade (along with the recently described A. arktos, A. mimetes, and A. vandycki) that is separate to all other Antechinus species (Baker et al., 2014, Baker et al., 2015, Westerman et al., 2016). Perhaps owing to this, these species apparently don’t cue their ovulation timing primarily with a rate of change of photoperiod, ovulating comparatively earlier in the year (beginning between late May and late September). McAllan et al. (2006) pointed out that A. swainsonii and A. minimus prey on a wider variety of species than other Antechinus species, and therefore timing the birth of young to the spring flush of insects may be less important to these two species. It is therefore plausible that the genetic trait of reproductive photoperiodism found in most Antechinus either isn’t present in (or is at least less important to) A. swainsonii and A. minimus. Based on the existing knowledge concerning photoperiodism in Antechinus, it was expected that A. argentus would respond to a comparable photoperiodic cue and ovulate at a similar time to the other members of its phylogenetic clade (A. flavipes and A. mysticus), which occupy a comparative mid-latitudinal range. On the contrary, the data presented and discussed in

72

Chapter 3 indicated that A. argentus began ovulating between July 6-8, when the rate of change of photoperiod was ~26.5 s. This is earlier than A. flavipes and A.mysticus, and comparable to A. swainsonii, A. minimus, and northern Australian A. bellus and A. godmani. Furthermore, unlike A. swainsonii and A. minimus, A. argentus consumes a relatively lower diversity of invertebrate prey (Mason et al., 2015), suggesting that timing the birth of young with the spring flush of insects should be more beneficial to the species. Consider that the sister taxon of A. argentus, A. mysticus (Baker et al., 2013) ovulates considerably later; a recent study of A. mysticus studied breeding in two populations, Brisbane (lat. -27.3) and Eungella (lat. -21.2), and both populations bred in August (Pearce, 2016). This is some 4-6 weeks later than A. argentus at Kroombit Tops, which is latitudinally (-24.4) situated midway between the two A. mysticus populations and altitudinally (850-900 m asl) similar to the Eungella population of A. mysticus (700 m asl). Based on the present study’s findings in relation to ovulation timing and photoperiodic cues, A. argentus does not appear to fit well with the strong pattern of photoperiodism exhibited by closely-related congeners (McAllan et al. 2006).

5.2 Detectability and distribution

5.2.1 Detectability The present study resulted in A. argentus being listed as vulnerable in Queensland under the Nature Conservation Act 1992. At time of writing, a threatened species listing is under review at the federal level. Several factors can contribute to labelling a species as threatened, not least of which is distributional range (Ceballos and Ehrlich, 2002). A larger range and the occupation of more diverse habitats is beneficial to species living in a time of great ecological destruction, as they are more likely to find refuge in areas spared from deforestation and industrialization (Sisk et al., 1994). On the other hand, species limited to a smaller area or number of habitat types are less likely to persist under encroaching disturbance (McKinney and Lockwood, 1999). The ability to detect species accurately to determine their distribution and range is therefore critical in determining the level of threat to their survival, and to effectively implement conservation measures (Gu and

73

Swihart, 2004). Determining whether a species is present or absent within an area of interest is a significant undertaking - detectability varies considerably between species (Boulinier et al., 1998). Furthermore, detectability within a species can vary depending on individual behaviour, habitat type, climatic conditions, temporal factors such as seasons, or indeed numerous ecological variables (Watkins et al., 2010). In order to determine the presence or absence of a species with sufficient confidence, it is therefore desirable to use a quantifiable, statistical method of estimation based on existing data. Recently, Garrard (2015) proposed a quantitative method of determining appropriate survey requirements when searching for rare species. This method gives an estimation of D, which is the probability that a species will be detected given a survey of prespecified effort, where effort can be represented by a number of discrete units (such as trap nights – i.e., the number of traps deployed per night). Assuming that each trap at a single locality is equally likely to capture an individual, then:

D = Pr(detected│present) = 1 – (1 – p)n, where p is the probability of capturing the species in a single trap night if it is present at the locality, and n is the number of trap nights. Under this method, the probability of detecting the species when it is present increases with trapping effort. Adopting this method in a small mammal species distribution study should reduce the likelihood of making Type II errors (not detecting the species when it is present, or in other words accepting a H0 when it is false) by estimating the minimum number of trap nights required to be 99% (or whatever desired %) confident that the species is absent from the specified locality if it is not captured. Adopting the method detailed above, Table 5.1 shows the minimum trap nights required to be 99% confident that A. argentus will be detected if it is at the locality for each trapping month, assuming the species has similar abundance and trappability between sites. These values were calculated based

74

on data collected from the two Kroombit Tops study sites (Lookout and Northern). Table 5.1 Minimum number of trap nights required to be 99% confident that A. argentus will

be detected if it inhabits a specific locality, calculated based on 2014 data from two localities

where A. argentus is known from. “N/a” indicates that no captures were made at that site in

that month and therefore minimum trap nights could not be calculated.

Lookout Northern

March 1044 2300

April 228 n/a

May 174 2300

June 102 n/a

July 416 n/a

August 1044 n/a

September 2300 n/a

Given the apparent population decline observed in 2015 (see Chapter 3), the capture rate data are taken from just 2014. As mentioned previously, one of these sites (Northern) was burnt heavily in a wildfire immediately prior to this research study. The fire significantly reduced the vegetation cover at the site (see Chapter 4), and likely also reduced the abundance of A. argentus. This may explain the relatively low capture rate in comparison to the other site (Lookout) (see Table 5.1). Captures during early surveys that took place before the commencement of this project yielded relatively equal capture rates between the two sites (H. Hines, pers. comm., data not shown). This is reflected in the minimum trap nights calculated from each site’s data, with the Northern site typically displaying inconsistently high minimum trap night values, while the Lookout’s data more closely reflects the pattern that would be expected based on all other Antechinus life histories (a steady increase in activity/trappability leading up to breeding, before a reduction in population numbers and trappability following the male die-off and while females are bearing pouch young). In an attempt to resolve the distribution of A. argentus, the present study used these minimum trap night calculations to inform trapping surveys of areas outside Kroombit Tops.

75

5.2.2 Distribution When the present study began, A. argentus was only known from the two Kroombit Tops study sites focused on in Chapters 2, 3, and 4 (Baker et al., 2014, Mason et al., 2015). At the time, this represented the smallest known range of any Australian mammal. A range of habitat types and microclimates are represented at Kroombit Tops. In addition to the Eucalyptus montivaga moist, tall open forest dominating the two A. argentus study sites, the Kroombit plateau includes warm temperate rainforest at its eastern end and dry open forest in the west (Baker et al., 2013). In June 2015, a large-scale trapping survey at Kroombit Tops was undertaken in collaboration with Queensland Parks and Wildlife Service (QPWS), Department of Environment and Heritage Protection (DEHP), Burnett-Mary Regional Group (BMRG) and Queensland Herbarium staff that included a total of 6,600 trap nights comprising 600 trap nights simultaneously at each of 11 sites encompassing all the major habitat types in the national park (QPWS unpublished data). This was carried out concurrently with the present study’s regular trapping regime at the Northern and Lookout sites. The survey resulted in nil A. argentus captures at any site except the Lookout, where a single female and male (both already PIT-tagged earlier in the year) were captured once each. This is in contrast to the results of the previous June’s trapping at the Lookout (2014), in which 20 total captures of four females and four males were recorded. As previously mentioned, annual captures at the two primary study sites (Lookout and Northern) were markedly lower in 2015 than they were in 2014. Based on 2014 captures, it was calculated that 102 trap nights in June would be required to be 99% sure that A. argentus were not present at a given site (based on Garrard et al., 2015) (Table 5.1). This determined the trapping strategy for the larger survey in June 2015 of 600 trap nights at each site, an amount that was estimated to be almost six times the required amount. However, based on June 2015 results at the Lookout of just two captures, 1,044 trap nights would have been the minimum effort required to rule out presence (with 99% confidence) of A. argentus at adjacent sites in 2015 (see

76

Table 5.1). With the employed effort of 600 trap nights for the large survey, based on June 2015 captures, there is 93% confidence that A. argentus were not present elsewhere at the other sites simultaneously surveyed on Kroombit Tops. Unfortunately, limitations in trap numbers and human resources necessitated the trapping effort to remain at 600 trap nights per site. Nevertheless, based on these efforts, the Kroombit Tops A. argentus population is most likely limited to the very small area (approximately 10 km2) adjacent to the escarpment edge that bounds the north east of the plateau. In addition to the regular trapping surveys being undertaken at Kroombit from March-September of 2014 and 2015, additional trap nights were carried out in areas of relatively undisturbed, vegetated habitat throughout central-eastern Queensland (Table 5.2, Fig 5.3). Table 5.2 Trapping undertaken at locations external to Kroombit Tops (colours indicate

captures of Antechinus argentus and Antechinus flavipes).

Date Site Trap nights Antechinus captures

24-25/07/2014 Collosseum Creek (-24.39, 151.47) 225 0

27-30/07/2014 Bulburin NP site 1 (-24.55, 151.48) 225 0

27-30/07/2015 Bulburin NP site 2 (-24.58, 151.50) 150 0

12-13/05/2015 Mt Robert site 1 (-24.56, 151.32) 100 0

12-13/05/2015 Mt Robert site 2 (-24.55, 151.31) 100 0

12-13/05/2015 Mt Robert site 3 (-24.55, 151.31) 100 0

12-13/05/2015 Mt Robert site 4 (-24.55, 151.31) 100 0

24-25/05/2015

Blackdown Tableland NP site 1 (-23.79,

149.07) 300 0

25-26/05/2015

Blackdown Tableland NP site 2 (-23.79,

149.12) 350 6 A. argentus

28-29/05/2015

Bulburin NP site 3 (-24.51, 151.52)

200 0

28-29/05/2015 Bulburin NP site 4 (-24.53, 151.53) 200 0

10-12/05/2016

Consuelo Tableland, Carnarvon NP, Mt

Moffatt section site 1 (-24.93, 148.06) 600 4 A. flavipes

13-14/05/2016

Consuelo Tableland, Carnarvon NP, Mt

Moffatt section site 2 (-24.91, 148.21) 375 2 A. flavipes

15/05/2016

Consuelo Tableland, Carnarvon NP, Mt

Moffatt section site 3 (-24.94, 148.13) 150 0

77

Fig 5.3 Trapping locations in central-eastern Queensland. u: Kroombit Tops Lookout site,

:Kroombit Tops Northern site u: Collosseum Creek, u: Bulburin sites 1-4, u: Mt Robert

sites 1-4, u: Blackdown Tableland site 1, u: Blackdown Tableland site 2 (A. argentus

captured), u: Consuelo Tableland sites 1-2 (A. flavipes captured), u: Consuelo Tableland site

3.

Historical records of Antechinus existed for all of these areas (Atlas of Living Australia, 2015), but individuals had not been genetically and morphologically assessed in detail due to their age. Moreover, many of the locations had been subject to disturbance (fire, exotic flora and fauna) since the records were taken. The unusual biota of Kroombit Tops also directed particular interest towards some specific locations. Kroombit is a mesic temperate outlier, and is known to support four endemic species including two frogs - Taudactylus pleione and Litoria kroombitensis, the former critically endangered federally (Czechura, 1986, Hoskin et al., 2013). Furthermore, 70+ species occur at their northern limit (Hines, 2014). Indeed, the broad biogeographical affinity of the Kroombit Tops biota is to temperate regions of the south (Hines, 2014). It has been suggested that some temperate taxa have historically become isolated at Kroombit Tops following the contraction of wet forests to montane and coastal regions of eastern Australia due to the climatic cooling and drying that occurred during the late Pliocene/early Pleistocene (Byrne, 2008, Czechura, 1986, Hines, 2014, Ponniah et al., 2004). This is further evident by the

78

significant genetic divergence of many taxa (apparent even in the Kroombit populations of some widespread species) at Kroombit Tops (Chapple et al., 2011, James and Moritz, 2000, Nicholls and Austin, 2005). However, the biogeography and geology of Kroombit Tops also exhibit similarities to montane regions found to the west and north-west. Approximately 215 Ma, the Kroombit Tops Caldera apparently erupted, depositing the Winterbourne Volcanic rocks that form much of the present surface geology (Hines, 2014, Willmott and Moore, 2006). It is postulated that broad rivers flowed across these rocks about 190 Ma, resulting in the deposition of horizontal beds of Precipice Sandstone (Willmott and Moore, 2006). The eastern plateau and escarpment that include the present study’s two sites are composed of the erosion-resistant Winterbourne Volcanic rocks and capped with Precipice Sandstone, outcropping at an elevation of 800-930 m from the comparably eroded lowlands (Hines, 2014). The sheer, eastern edges of the escarpment (up to 50 m high) are the result of erosion from the steep-flowing streams of the Boyne catchment. Blackdown Tableland to the north-west is another outcrop of Precipice Sandstone with comparable elevation to Kroombit, as are the Carnarvon Ranges to the northwest (Willmott and Moore, 2006). Furthermore, these three areas share important biological values. At least nine tree, shrub and fern species found at Kroombit are restricted to sandstone areas of central Queensland, and especially to Precipice Sandstone (Hines, 2014). The highest point of the residual sandstone plateaus in Queensland is at Consuelo Tableland in the Carnarvon Ranges, which reaches 1200 m elevation (Moore and Monteith, 2004). A Gondwanan Stag Beetle, Sphaeonognathus munchowae is confined to this highest, wettest part of Carnarvon, and also occurs at the highest, wettest part of Blackdown (Moore and Monteith, 2004), and the butterfly species Trapezites taori also occurs in unconnected populations within the two regions (Johnson and Valentine, 2004). Both Blackdown and Consuelo were therefore of particular interest in the search effort to secure further populations of A. argentus.

79

In May 2015 (predicted to be a time of year when A. argentus are relatively trappable, see Chapter 3, Table 5.1), six Antechinus individuals were captured at one site at Blackdown Tableland National Park, approximately ~200 km WNW of Kroombit Tops (Table 5.2, Fig 5.3). Based on initial external morphological assessment and subsequent genetic and morphological (skull) assessment, it was confirmed that these Antechinus indeed represented another apparently isolated population of A. argentus. The location of the capture site at Blackdown exhibits marked similarities to the Kroombit Tops study site. As previously noted, like Kroombit, Blackdown Tableland is an outcrop of Precipice Sandstone that rises above the surrounding lowlands, representing a mesic temperate outlier (Johnson and Valentine, 2004, Willmott and Moore, 2006). The capture site is in the wetter eastern edge of the park in close proximity to the sheer escarpment edge, at an elevation of approximately 850 m (Kroombit study sites ~850-900 m, on the wettest eastern side) (Baker et al., 2013). The discovery of the Blackdown Tableland A. argentus population directed interest to the third significant outcrop of Precipice Sandstone, the Carnarvon Ranges. Similar to Blackdown, scant few Antechinus capture records from the 1980s and 1990s existed for the area, but these were all recorded as A. flavipes (and dated well prior to the description of A. argentus) (Atlas of Living Australia, 2015). All these Queensland Museum specimens did not have tissue specimens taken at the time of capture and all attempts at retrieval of DNA on preserved tissue proved unsuccessful. Morphological analysis of both ethanol-faded pelts and skull material could also not clearly resolve a species identification. Thus in May 2016, a trapping effort of 1,125 trap nights across three sites on Consuelo Tableland in Carnarvon National Park was undertaken which resulted in six Antechinus captures (Table 5.1). However, subsequent morphological and genetic analysis of these individuals determined that they were indeed A. flavipes. This does not rule out the presence of A. argentus in the Carnarvon Ranges – the protected area is relatively large, and multiple species of Antechinus are known to occur sympatrically in other areas (e.g., Green, 1989, Mutton, 2017, Parrott et al., 2007). Indeed, the numerous surveys undertaken during this study cannot unequivocally rule out the presence of A. argentus in the broader areas of Collosseum Creek, Bulburin National Park, or Mt Robert (Table 5.2,

80

Fig 5.3), all of which were targeted because of their high elevation, protected habitat, proximity to Kroombit and general geomorphological affinities. However, based on the data obtained from 2014 capture rates at the Kroombit Tops Lookout site (Table 5.1), trap nights undertaken at these areas indicate with reasonable confidence that the species is likely absent at least from the specific sites that were trapped (Garrard et al., 2015). With the exception of Collosseum Creek and the initial Bulburin efforts, all trapping surveys undertaken external to Kroombit Tops were in May, when 2014 Lookout site capture data indicated a minimum of 174 trap nights are necessary to be 99% confident that the species will be detected if it is present at a specific site (Table 5.1). Of course, given the recent discovery and lack of knowledge of the long-term population dynamics of A. argentus, these capture rate data lack robustness. Continuing surveys at Kroombit Tops and potential discoveries of further populations will consolidate the minimum trap nights required to detect this apparently rare species, and further resolve its distribution. Importantly, the tissue samples sequenced from the A. argentus population at Blackdown Tableland exhibited notable differences to the Kroombit Tops population (1.4-1.7% at the mtDNA gene cytochrome B) (Mutton, 2017). This divergence suggests that the two populations have been isolated at these elevated, mesic habitats for a significant length of time (likely tens of thousands of generations).

5.3 Conservation recommendations

A. argentus apparently occurs in restricted montane habitats, which have been identified as one of the most vulnerable ecosystems in Australia (Laurance et al., 2011). Most likely, the species is climatically confined. As global temperatures rise, these habitats are expected to retract further upwards along elevational gradients (La Sorte and Jetz, 2010). This process likely places A. argentus at considerable risk of extinction. Fire (both natural and intentional) will play an important role in A. argentus’ uncertain future. Determining the role of fire to the ecology of this species was

81

not a central focus of the present work, although the partial burning of the Kroombit escarpment in 2013 immediately prior to this work provided an opportunity to assess how the population recovered in the short term to a burned landscape. Thus, some recommendations can be made here regarding fire regimes. Unlike populations of some other Antechinus species that can endure fire in situ (such as A. stuartii in Guy Fawkes River National Park, NSW) (Stawski et al., 2015, Stawski et al., 2017), the present study found that A. argentus apparently has a preference for relatively long-unburnt habitat. This conclusion is drawn from regular trapping data from the isolated Kroombit Tops population of the species (see Chapter 3) and ongoing monitoring work in fire-affected Blackdown Tableland National Park (data not shown). A prominent paradigm in conservation management is that fire mosaics (spatial variation in fire and creating a diversity of fire-ages) are an effective strategy for the persistence of animal species by mediating heterogeneity in habitats (e.g., Christensen and Kimber, 1975, Woinarski, 1999, Short and Smith, 1994), and it has been shown that the retention of older vegetation within a fire mosaic is important for Australian small mammals (Kelly et al., 2010, Kelly et al., 2011, Kelly et al., 2012). However, the effectiveness of this strategy is likely to be highly variable and context-dependent (Bradstock et al., 2005). The demonstrated habitat requirements of species provide potential guidelines for fire management (Kelly et al., 2017). Indeed, the present study provides an example of how the link between fire mosaics and persistence of an animal species apparently depends on fire age and its effect on vegetation. Further work clearly needs to be undertaken on the Blackdown population and potential further populations of A. argentus to determine if this conclusion applies at the population or species level. Nevertheless, the frequency and intensity of wildfires are expected to increase as the warming effects of climate change accelerate (Dale et al., 2001). It is therefore unlikely that conservation managers will be able to avoid undertaking prescribed burns to reduce fuel load, and furthermore fires are likely to be important to the ecosystem function as a whole. It is suggested that burning regimes for areas

82

that encompass A. argentus habitat should utilize cool fires that produce patchy burns, maximizing spatial variation in severity (as per Martin and Sapsis, 1992, Leahy et al., 2016). Designated unburnt patches should be maintained for 10+ years (as evident at the unburnt section of the Lookout). The present study corroborates other recent work that recommends burning regimes that retain vegetation for habitat and important refuge for small mammals from introduced predators (McGregor et al., 2014, Legge et al., 2011b, Leahy et al., 2016, Andersen et al., 2012, McGregor et al., 2017). The present study did not quantitatively investigate the effects of feral animals on A. argentus, but over the course of data collection, cattle (Bos taurus), pigs (Sus scrofa), horses (Equus caballus), cats (Felis catus) and dogs/dingos (Canis sp.) were observed at the study sites (pers. obs.), and are well documented at Kroombit Tops both from cage and camera trapping (QPWS and QUT, unpublished data). Ongoing work should specifically examine the impacts of these species on A. argentus. Nevertheless, the links between the declining Australian mammal fauna and both predation by introduced carnivores (Dickman, 1996) and grazing by introduced herbivores (Legge et al., 2011a) are well documented. It is recommended that control measures be substantially intensified for feral animals at Kroombit Tops – particularly cats, cattle, horses and pigs as they are almost certainly having a detrimental effect on A. argentus. A. argentus has been listed as vulnerable in Queensland (Queensland Government, 2015b). The species is currently being considered for a threatened species listing at the federal level. Under the Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act), five criteria are assessed in order to list a species as threatened, as follows: Criterion 1. Population size reduction (reduction in total numbers) Although the present study observed a decline in abundance of an A. argentus population over two years, the recent discovery of the species means that it does not meet this criterion, which requires long-term data to demonstrate population reduction.

83

Criterion 2. Geographic distribution as indicators for either extent of occurrence AND/OR area of occupancy The present study conducted detailed searches for A. argentus, indicating that the species has a restricted geographic range, occurring at two known, isolated locations in conservation areas. Under this criterion, the species is eligible to be listed as Endangered, which requires the extent of occurrence to be < 5,000 km2 (estimated to be 1,008 km2) and the area of occupancy to be < 500 km2

(estimated to be 12 km2) (details of occurrence/occupancy estimates not shown). Criterion 3. Population size and decline A. argentus occurs in very low abundance at Kroombit Tops, and declined over two years following a fire. Ongoing monitoring at the Kroombit Tops Lookout site also indicates decline (2014-2017 in June, number of individuals: 8, 2, 1, 0) (2016-2017 is QUT, unpublished data). Furthermore, the species only occurs in a highly vulnerable ecosystem (restricted montane habitat) subject to ongoing drying under climate change scenarios. Although the present study did not undertake long-term trapping at Blackdown Tableland, the population there exhibited a comparable capture rate to Kroombit Tops. Based on these data, the species is eligible for listing as Endangered under this criterion, which requires < 2,500 mature individuals in existence, and a projected continuing decline. Criterion 4. Number of mature individuals The present study’s capture rates at both Kroombit Tops and Blackdown Tableland suggest that A. argentus is eligible for listing as Endangered for this criterion, which states that at any given time < 250 mature individuals exist. While numbers of a fauna species are notoriously difficult to predict, adult Antechinus numbers approximately halve annually at the male die-off (for A. argentus, late June/early July, Chapter 3), which would place considerable pressure on small existing populations of females. Criterion 5. Quantitative analysis

84

This criterion requires that a population viability analysis be undertaken for the species, but as yet there are insufficient data for this. Under the above criteria from the EPBC Act, the present research suggests that A. argentus is eligible for a federal threatened species as Endangered. The Queensland listing should subsequently be raised to Endangered, and this should likely be extended to the international level, seeing the species placed on the IUCN Red List.

Conclusion

Taken together, the present study has granted a study of the autecology of a species that was previously poorly understood. This work hopefully provides important foundational knowledge for ongoing and necessary conservation management of the species. A. argentus is undoubtedly at risk of extinction. The two populations that are known at present are isolated, genetically fragmented and occur in relatively small areas of conservation land. Several processes are likely threatening both of these populations, namely inappropriate fire frequency, severity and patch size, and the impacts of introduced animals. Appropriate management actions must be undertaken urgently if this threatened species is to persist in an increasingly threatening environment.

85

References

Abbott, I., Peacock, D. & Short, J. 2014. The new guard: the arrival and impacts of cats and foxes. Carnivores of Australia: past, present and future. Collingwood: CSIRO Publishing, 69-104.

Aberton, J. G. 1996. Succession of small mammal communities after fire and reintroduction of the swamp antechinus ‘Antechinus minimus’. Deakin University.

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

Andersen, A., Cook, G. & Williams, D. 2012. Savanna burning: the ecology and economy of fire in tropical savannas. Austral Ecology, 37, 633-633.

Anderson, M. J. 2001. A new method for non‐parametric multivariate

analysis of variance. Austral Ecology, 26, 32-46. Andren, H. 1994. Effects of habitat fragmentation on birds and mammals in

landscapes with different proportions of suitable habitat: a review. Oikos, 355-366.

Andrew, N., Rodgerson, L. & York, A. 2000. Frequent fuel‐reduction

burning: the role of logs and associated leaf litter in the conservation of ant biodiversity. Austral Ecology, 25, 99-107.

Archer, M. 1976. The dasyurid dentition and its relationships to that of didelphids, thylacinids, borhyaenids (Marsupicarnivora) and peramelids (Peramelina: Marsupialia). Australian Journal of Zoology, 24, 1-34.

Archer, M. Carnivorous marsupials. 1982. Royal Zoological Society of New South Wales.

Atlas of Living Australia. 2015. Available: http://www.ala.org.au [Accessed 3

March 2015. Baker, A., Mutton, T., Mason, E. & Gray, E. 2015. A taxonomic assessment of

the Australian dusky antechinus complex: a new species, the Tasman peninsula dusky antechinus (Antechinus vandycki sp. nov.) and an

86

elevation to species of the mainland dusky antechinus (Antechinus swainsonii mimetes (Thomas)). Memoirs of the Queensland Museum–Nature, 59, 75-126.

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. M., Mutton, T. Y. & Hines, H. B. 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. M., Mutton, T. Y., Hines, H. B. & 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. M. & Van Dyck, S. 2012. Taxonomy and redescription of the Yellow-footed Antechinus, Antechinus flavipes (Waterhouse)(Marsupialia: Dasyuridae). Zootaxa, 3649, 1-62.

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

Banks, S., Finlayson, G., Lawson, S., Lindenmayer, D., Paetkau, D., Ward, S. & Taylor, A. 2005. The effects of habitat fragmentation due to forestry plantation establishment on the demography and genetic variation of a marsupial carnivore, Antechinus agilis. Biological Conservation, 122, 581-597.

Barker, I., Beveridge, I., Bradley, A. & Lee, A. 1978. Observations on spontaneous stress-related mortality among males of the dasyurid marsupial Antechinus stuartii Macleay. Australian Journal of Zoology, 26, 435-447.

Barlow, J., Lennox, G. D., Ferreira, J., Berenguer, E., Lees, A. C., Mac Nally, R., Thomson, J. R., De Barros Ferraz, S. F., Louzada, J. & Oliveira, V. H. F. 2016. Anthropogenic disturbance in tropical forests can double biodiversity loss from deforestation. Nature, 535, 144-147.

87

Barnosky, A. D., Matzke, N., Tomiya, S., Wogan, G. O., Swartz, B., Quental, T. B., Marshall, C., Mcguire, J. L., Lindsey, E. L. & Maguire, K. C. 2011. Has the Earth's sixth mass extinction already arrived? Nature, 471, 51-57.

Bates, D., Mächler, M., Bolker, B. & Walker, S. 2014. Fitting linear mixed-effects models using lme4. arXiv preprint arXiv:1406.5823.

Bennett, A. 1993. Microhabitat use by the long-nosed potoroo, Potorous tridactylus, and other small mammals in remnant forest vegetation, south-western Victoria. Wildlife Research, 20, 267-285.

Berry, L. E., Lindenmayer, D. B. & Driscoll, D. A. 2015. Large unburnt areas, not small unburnt patches, are needed to conserve avian diversity in

fire‐prone landscapes. Journal of Applied Ecology, 52, 486-495.

Bielby, J., Mace, G. M., Bininda-Emonds, O. R., Cardillo, M., Gittleman, J. L., Jones, K. E., Orme, C. D. L. & Purvis, A. 2007. The fast-slow continuum in mammalian life history: an empirical reevaluation. The American Naturalist, 169, 748-757.

Black, H. L. 1972. Differential exploitation of moths by the bats Eptesicus fuscus and Lasiurus cinereus. Journal of Mammalogy, 53, 598-601.

Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen, J. R., Stevens, M. H. H. & White, J.-S. S. 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends in Ecology & Evolution, 24, 127-135.

Borsboom, A. 2005. Xanthorrhoea: A review of current knowledge with a focus on X. johnsonii and X. latifolia, two Queensland protected plants-in-trade, Environmental Protection Agency.

Boulinier, T., Nichols, J. D., Sauer, J. R., Hines, J. E. & Pollock, K. 1998. Estimating species richness: the importance of heterogeneity in species detectability. Ecology, 79, 1018-1028.

Bowman, D. M. 2000. Australian rainforests: islands of green in a land of fire, Cambridge University Press.

Bradley, A., Mcdonald, I. & Lee, A. 1980. Stress and mortality in a small marsupial (Antechinus stuartii, Macleay). General and Comparative Endocrinology, 40, 188-200.

88

Bradstock, R., Bedward, M., Gill, A. & Cohn, J. 2005. Which mosaic? A landscape ecological approach for evaluating interactions between fire regimes, habitat and animals. Wildlife Research, 32, 409-423.

Bradstock, R. A., Williams, J. E. & Gill, M. A. 2002. Flammable Australia: the fire regimes and biodiversity of a continent, Cambridge University Press.

Braithwaite, R. W. 1979. Social dominance and habitat utilization in Antechinus stuartii (Marsupialia). Australian Journal of Zoology, 27, 517-528.

Braithwaite, R. W. & Lee, A. K. 1979. A mammalian example of semelparity. The American Naturalist, 113, 151-155.

Bray, J. R. & Curtis, J. T. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs, 27, 325-349.

Brown, J. H. 1984. On the relationship between abundance and distribution of species. The American Naturalist, 124, 255-279.

Burbidge, A. A. & Mckenzie, N. 1989. Patterns in the modern decline of Western Australia's vertebrate fauna: causes and conservation implications. Biological Conservation, 50, 143-198.

Burnham, K. P. & Anderson, D. R. 2003. Model selection and multimodel inference: a practical information-theoretic approach, Springer Science & Business Media.

Burwell, C. J., Geiser, F., Barritt, M., May, K. & Pavey, C. R. 2005. Additional records of prey of the fat-tailed false antechinus Pseudantechinus macdonnellensis from central Australia. Australian Mammalogy, 27, 227-229.

Byrne, M. 2008. Evidence for multiple refugia at different time scales during Pleistocene climatic oscillations in southern Australia inferred from phylogeography. Quaternary Science Reviews, 27, 2576-2585.

Byrne, M., Steane, D. A., Joseph, L., Yeates, D. K., Jordan, G. J., Crayn, D., Aplin, K., Cantrill, D. J., Cook, L. G. & Crisp, M. D. 2011. Decline of a biome: evolution, contraction, fragmentation, extinction and invasion of the Australian mesic zone biota. Journal of Biogeography, 38, 1635-1656.

89

Byrne, M., Yeates, D., Joseph, L., Kearney, M., Bowler, J., Williams, M., Cooper, S., Donnellan, S., Keogh, J. S. & Leys, R. 2008. Birth of a biome: insights into the assembly and maintenance of the Australian arid zone biota. Molecular Ecology, 17, 4398-4417.

Carroll, R. L. 1988. Vertebrate paleontology and evolution, Freeman. Ceballos, G. & Brown, J. H. 1995. Global patterns of mammalian diversity,

endemism, and endangerment. Conservation Biology, 9, 559-568. Ceballos, G. & Ehrlich, P. R. 2002. Mammal population losses and the

extinction crisis. Science, 296, 904-907. Ceballos, G., Ehrlich, P. R., Barnosky, A. D., García, A., Pringle, R. M. &

Palmer, T. M. 2015. Accelerated modern human–induced species losses: Entering the sixth mass extinction. Science Advances, 1, e1400253.

Chapple, D. G., Hoskin, C. J., Chapple, S. N. & Thompson, M. B. 2011. Phylogeographic divergence in the widespread delicate skink (Lampropholis delicata) corresponds to dry habitat barriers in eastern Australia. BMC Evolutionary Biology, 11, 191.

Charnov, E. L. & Schaffer, W. M. 1973. Life-history consequences of natural selection: Cole's result revisited. The American Naturalist, 107, 791-793.

Chia, E. K., Bassett, M., Leonard, S. W., Holland, G. J., Ritchie, E. G., Clarke, M. F. & Bennett, A. F. 2016. Effects of the fire regime on mammal occurrence after wildfire: Site effects vs landscape context in fire-prone forests. Forest Ecology and Management, 363, 130-139.

Christensen, P. & Kimber, P. Effect of prescribed burning on the flora and fauna of south west Australian forests. Proceedings of the Ecological Society of Australia, 1975.

Clarke, K. & Gorley, R. 2015. PRIMER v7: User Manual/Tutorial; PRIMER-E: Plymouth, UK, 2015.

Cockburn, A. 1997. Living slow and dying young: senescence in marsupials. Marsupial Biology: Recent Research, New Perspectives. University of New South Wales Press, Sydney.

90

Cockburn, A., Scott, M. P. & Scotts, D. J. 1985. Inbreeding avoidance and male-biased natal dispersal in Antechinus spp.(Marsupialia: Dasyuridae). Animal Behaviour, 33, 908-915.

Crowther, M. S. & Blacket, M. 2003. Biogeography and speciation in the dasyuridae: why are there so many kinds of dasyurids. Predators with pouches: the biology of carnivorous marsupials, 124-32.

Czechura, G. V. 1986. A new species of Taudactylus (Myobatrachidae) from southeastern Queensland, Australia. Memoirs of the Queensland Museum, 22, 299-307.

Dale, V. H., Joyce, L. A., Mcnulty, S., Neilson, R. P., Ayres, M. P., Flannigan, M. D., Hanson, P. J., Irland, L. C., Lugo, A. E. & Peterson, C. J. 2001. Climate change and forest disturbances: climate change can affect forests by altering the frequency, intensity, duration, and timing of fire, drought, introduced species, insect and pathogen outbreaks, hurricanes, windstorms, ice storms, or landslides. BioScience, 51, 723-734.

Daubenmire, R. 1959. A canopy-coverage method of vegetational analysis. Northwest Science, 33, 43-64.

Denny, E. A. & Dickman, C. 2010. Review of cat ecology and management strategies in Australia. Invasive Animals Cooperative Research Centre, Canberra.

Department of the Environment, W., Heritage and Arts, 2008. Threat abatement plan for predation by feral cats. Canberra.

Dickman, C. 1982. Some observations of the behaviour and nest utilization of free-living Antechinus stuartii (Marsupialia: Dasyuridae). Australian Mammalogy: Journal of the Australian Mammal Society, 5, 75.

Dickman, C. 1985. Effects of photoperiod and endogenous control on timing of reproduction in the marsupial genus Antechinus. Journal of Zoology, 206, 509-524.

Dickman, C. 1989. Demographic responses of Antechinus stuartii (Marsupialia) to supplementary food. Austral Ecology, 14, 387-398.

Dickman, C. 1996. Impact of exotic generalist predators on the native fauna of Australia. Wildlife Biology, 2, 185-195.

91

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

Dickman, C., Mahon, P., Masters, P. & Gibson, D. 1999. Long-term dynamics of rodent populations in arid Australia: the influence of rainfall. Wildlife Research, 26, 389-403.

Dirzo, R., Young, H. S., Galetti, M., Ceballos, G., Isaac, N. J. & Collen, B. 2014. Defaunation in the Anthropocene. Science, 345, 401-406.

Doherty, T. S., Davis, R. A., Etten, E. J., Algar, D., Collier, N., Dickman, C. R., Edwards, G., Masters, P., Palmer, R. & Robinson, S. 2015. A

continental‐scale analysis of feral cat diet in Australia. Journal of

Biogeography, 42, 964-975. Dotta, G. & Verdade, L. M. 2007. Trophic categories in a mammal assemblage:

diversity in an agricultural landscape. Biota Neotropica, 7, 0-0. Doughty, C. E., Wolf, A. & Malhi, Y. 2013. The legacy of the Pleistocene

megafauna extinctions on nutrient availability in Amazonia. Nature Geoscience, 6, 761-764.

Elith, J., Leathwick, J. R. & Hastie, T. 2008. A working guide to boosted regression trees. Journal of Animal Ecology, 77, 802-813.

Engler, R., Guisan, A. & Rechsteiner, L. 2004. An improved approach for predicting the distribution of rare and endangered species from

occurrence and pseudo‐absence data. Journal of Applied Ecology, 41,

263-274. Ewer, R. F. 1973. The carnivores, Cornell University Press. Fensham, R. J., Fraser, J., Macdermott, H. J. & Firn, J. 2015. Dominant tree

species are at risk from exaggerated drought under climate change. Global Change Biology, 21, 3777-3785.

Fisher, D. & Dickman, C. 1993. Diets of insectivorous marsupials in arid Australia: selection for prey type, size or hardness? Journal of Arid Environments, 25, 397-410.

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

92

Fisher, D. O., Double, M. C., Blomberg, S. P., Jennions, M. D. & Cockburn, A. 2006. Post-mating sexual selection increases lifetime fitness of polyandrous females in the wild. Nature, 444, 89-92.

Fisher, R. A. 1922. On the interpretation of χ2 from contingency tables, and

the calculation of P. Journal of the Royal Statistical Society, 85, 87-94. Flannery, T. 2012. After the future: Australia's new extinction crisis. Quarterly

Essay, 1. Foley, J. C. 1957. Droughts in Australia: Review of records from earliest

years of settlement to 1955, Director of Meteorology. Fox, B. & Archer, E. 1984. The diets of Sminthopsis murina and Antechinus

stuartii (Marsupialia: Dasyuridae) in sympatry. Wildlife Research, 11, 235-248.

Fox, B. J. 1982. Fire and mammalian secondary succession in an Australian coastal heath. Ecology, 63, 1332-1341.

Fox, B. J., Fox, M. D. & Mckay, G. 1979. Litter accumulation after fire in a eucalypt forest. Australian Journal of Botany, 27, 157-165.

Friedman, J. H. 2001. Greedy function approximation: a gradient boosting machine. Annals of Statistics, 1189-1232.

Friend, G. R. & Taylor, J. A. 1985. Habitat preferences of small mammals in

tropical open‐forest of the Northern Territory. Austral Ecology, 10,

173-185. Gallus, S., Janke, A., Kumar, V. & Nilsson, M. A. 2015. Disentangling the

relationship of the Australian marsupial orders using retrotransposon and evolutionary network analyses. Genome Biology and Evolution, 7, 985-992.

Garrard, G. E., Bekessy, S. A., Mccarthy, M. A. & Wintle, B. A. 2015. Incorporating detectability of threatened species into environmental impact assessment. Conservation Biology, 29, 216-225.

Gill, A. M., Stephens, S. L. & Cary, G. J. 2013. The worldwide “wildfire” problem. Ecological Applications, 23, 438-454.

Gill, J. L., Williams, J. W., Jackson, S. T., Lininger, K. B. & Robinson, G. S. 2009. Pleistocene megafaunal collapse, novel plant communities, and enhanced fire regimes in North America. Science, 326, 1100-1103.

93

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

Gotelli, N. J. 1995. A primer of ecology, Sinauer Associates Sunderland, MA. Gray, E. 2013. A field study of breeding biology in a new species of

carnivorous marsupial, the buff-footed antechinus A. mysticus. Unpublished honors thesis, Queensland University of Technology.

Gray, E. L., Baker, A. M. & 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.

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

Grueber, C., Nakagawa, S., Laws, R. & Jamieson, I. 2011. Multimodel inference in ecology and evolution: challenges and solutions. Journal of Evolutionary Biology, 24, 699-711.

Gu, W. & Swihart, R. K. 2004. Absent or undetected? Effects of non-detection of species occurrence on wildlife–habitat models. Biological Conservation, 116, 195-203.

Guisan, A. & Thuiller, W. 2005. Predicting species distribution: offering more than simple habitat models. Ecology Letters, 8, 993-1009.

Hale, S., Nimmo, D. G., Cooke, R., Holland, G., James, S., Stevens, M., De Bondi, N., Woods, R., Castle, M. & Campbell, K. 2016. Fire and climatic

extremes shape mammal distributions in a fire‐prone landscape.

Diversity and Distributions, 22, 1127-1138. Hall, S. 1980. The diets of two coexisting species of Antechinus (Marsupialia:

Dasyuridae). Wildlife Research, 7, 365-378. Hindmarsh, R. & Majer, J. D. 1977. Food requirements of mardo (Antechinus

flavipes (Waterhouse)) and the effect of fire on mardo abundance. Research paper, West Australian Forestry Department.

Hines, H. 2014. Presidential address 2014: Kroombit tops: Endemism and outliers. Queensland Naturalist, 52, 4.

94

Hobbs, R. J. 2005. Landscapes, ecology and wildlife management in highly modified environments–an Australian perspective. Wildlife Research, 32, 389-398.

Holland, G. J. & Bennett, A. F. 2007. Occurrence of small mammals in a fragmented landscape: the role of vegetation heterogeneity. Wildlife Research, 34, 387-397.

Hoskin, C. J., Hines, H. B., Meyer, E., Clarke, J. & Cunningham, M. 2013. A new treefrog (Hylidae: Litoria) from Kroombit Tops, east Australia, and an assessment of conservation status. Zootaxa, 3646, 426-446.

Hutto, R. L., Keane, R. E., Sherriff, R. L., Rota, C. T., Eby, L. A. & Saab, V. A. 2016. Toward a more ecologically informed view of severe forest fires. Ecosphere, 7.

Ibm 2012. SPSS statistics version 21.0 for Microsoft Windows. Chicago, IL, USA.

Iucn. 2016. The IUCN Red List of Threatened Species. Version 2016-3 [Online]. Available: http://www.iucnredlist.org [Accessed 13 April 2017.

James, C. & Moritz, C. 2000. Intraspecific phylogeography in the sedge frog

Litoria fallax (Hylidae) indicates pre‐Pleistocene vicariance of an

open forest species from eastern Australia. Molecular Ecology, 9, 349-358.

Johnson, C. 2006. Australia's mammal extinctions: a 50,000-year history, Cambridge University Press.

Johnson, J. B. & Omland, K. S. 2004. Model selection in ecology and evolution. Trends in Ecology & Evolution, 19, 101-108.

Johnson, S. & Valentine, P. S. 2004. Notes on the biology and distribution of Trapezites taori Atkins (Lepidoptera: Hesperiidae). The Australian Entomologist, 31, 13.

Jones, M., Dickman, C. R. & Archer, M. 2003. Predators with pouches: the biology of carnivorous marsupials, CSIRO PUBLISHING.

Jones, M. E. & Barmuta, L. A. 1998. Diet overlap and relative abundance of sympatric dasyurid carnivores: a hypothesis of competition. Journal of Animal Ecology, 67, 410-421.

Kelly, L. & Brotons, L. 2017. Using fire to promote biodiversity. Science, 355, 1264-1265.

95

Kelly, L. T. & Bennett, A. F. 2008. Habitat requirements of the yellow-footed antechinus (Antechinus flavipes) in box–ironbark forest, Victoria, Australia. Wildlife Research, 35, 128-133.

Kelly, L. T., Brotons, L. & Mccarthy, M. A. 2017. Putting pyrodiversity to work for animal conservation. Conservation Biology.

Kelly, L. T., Dayman, R., Nimmo, D. G., Clarke, M. F. & Bennett, A. F. 2013. Spatial and temporal drivers of small mammal distributions in a

semi‐arid environment: The role of rainfall, vegetation and life‐

history. Austral Ecology, 38, 786-797. Kelly, L. T., Nimmo, D. G., Spence-Bailey, L. M., Clarke, M. F. & Bennett, A. F.

2010. The short-term responses of small mammals to wildfire in semiarid mallee shrubland, Australia. Wildlife Research, 37, 293-300.

Kelly, L. T., Nimmo, D. G., Spence‐Bailey, L. M., Haslem, A., Watson, S. J.,

Clarke, M. F. & Bennett, A. F. 2011. Influence of fire history on small

mammal distributions: insights from a 100‐year post‐fire

chronosequence. Diversity and Distributions, 17, 462-473.

Kelly, L. T., Nimmo, D. G., Spence‐Bailey, L. M., Taylor, R. S., Watson, S. J.,

Clarke, M. F. & Bennett, A. F. 2012. Managing fire mosaics for small mammal conservation: a landscape perspective. Journal of Applied Ecology, 49, 412-421.

Knight, E. H. & Fox, B. J. 2000. Does habitat structure mediate the effects of forest fragmentation and human-induced disturbance on the abundance of Antechinus stuartii? Australian Journal of Zoology, 48, 577-595.

Kraaijeveld‐Smit, F., Ward, S. & Temple‐Smith, P. 2003. Paternity success

and the direction of sexual selection in a field population of a semelparous marsupial, Antechinus agilis. Molecular Ecology, 12, 475-484.

Krajewski, C., Painter, J., Buckley, L. & Westerman, M. 1994. Phylogenetic structure of the marsupial family dasyuridae based on cytochrome b DNA sequences. Journal of Mammalian Evolution, 2, 25-35.

96

Krajewski, C. & Westerman, M. 2003. Molecular systematics of Dasyuromorphia. Predators with pouches: the biology of carnivorous marsupials. Melbourne: CSIRO Publishing.

Krajewski, C., Woolley, P. A. & Westerman, M. 2000a. The evolution of reproductive strategies in dasyurid marsupials: implications of molecular phylogeny. Biological Journal of the Linnean Society, 71, 417-435.

Krajewski, C., Wroe, S. & Westerman, M. 2000b. Molecular evidence for the pattern and timing of cladogenesis in dasyurid marsupials. Zoological Journal of the Linnean Society, 130, 375-404.

Kutt, A. & Gordon, I. 2012. Variation in terrestrial mammal abundance on

pastoral and conservation land tenures in north‐eastern Australian

tropical savannas. Animal Conservation, 15, 416-425. Kutt, A. S. & Woinarski, J. C. 2007. The effects of grazing and fire on

vegetation and the vertebrate assemblage in a tropical savanna woodland in north-eastern Australia. Journal of Tropical Ecology, 95-106.

La Sorte, F. A. & Jetz, W. 2010. Projected range contractions of montane biodiversity under global warming. Proceedings of the Royal Society of London B: Biological Sciences, rspb20100612.

Laurance, W. F., Dell, B., Turton, S. M., Lawes, M. J., Hutley, L. B., Mccallum, H., Dale, P., Bird, M., Hardy, G. & Prideaux, G. 2011. The 10 Australian ecosystems most vulnerable to tipping points. Biological Conservation, 144, 1472-1480.

Lazenby‐Cohen, K. A. & Cockburn, A. 1991. Social and foraging components

of the home range in Antechinus stuartii (Dasyuridae: Marsupialia). Austral Ecology, 16, 301-307.

Leahy, L., Legge, S. M., Tuft, K., Mcgregor, H. W., Barmuta, L. A., Jones, M. E. & Johnson, C. N. 2016. Amplified predation after fire suppresses rodent populations in Australia’s tropical savannas. Wildlife Research, 42, 705-716.

Legge, S., Kennedy, M. S., Lloyd, R., Murphy, S. A. & Fisher, A. 2011a. Rapid recovery of mammal fauna in the central Kimberley, northern

97

Australia, following the removal of introduced herbivores. Austral Ecology, 36, 791-799.

Legge, S., Murphy, S., Kingswood, R., Maher, B. & Swan, D. 2011b. EcoFire: restoring the biodiversity values of the Kimberley region by managing fire. Ecological Management & Restoration, 12, 84-92.

Letnic, M., Dickman, C., Tischler, M., Tamayo, B. & Beh, C.-L. 2004. The responses of small mammals and lizards to post-fire succession and rainfall in arid Australia. Journal of Arid Environments, 59, 85-114.

Lomolino, M. V. & Channell, R. 1995. Splendid isolation: patterns of geographic range collapse in endangered mammals. Journal of Mammalogy, 76, 335-347.

Lopes, G. P. & Leiner, N. O. 2015. Semelparity in a population of Gracilinanus agilis (Didelphimorphia: Didelphidae) inhabiting the Brazilian cerrado. Mammalian Biology-Zeitschrift für Säugetierkunde, 80, 1-6.

Lucas, C., Hennessy, K., Mills, G. & Bathols, J. 2007. Bushfire weather in southeast Australia: recent trends and projected climate change impacts. Report to the Climate Institute of Australia Bushfire Cooperative Research Centre, Australian Bureau of Meteorology and CSIRO Marine and Atmospheric Research, Canberra.

Lunney, D., Ashby, E., Grigg, J. & O’connell, M. 1986. Food availability and habitat selection of Sminthopsis leucopus (Gray)(Marsupialia: Dasyuridae) in logged forest on the south coast of New South Wales. Australian Mammalogy, 9, 105-110.

Lunney, D., Cullis, B. & Eby, P. 1987. Effects of logging and fire on small mammals in Mumbulla State Forest, near Bega, New South Wales. Wildlife Research, 14, 163-181.

Lunney, D., Matthews, A. & Grigg, J. 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.

Luo, Z.-X. 2007. Transformation and diversification in early mammal evolution. Nature, 450, 1011-1019.

Magnusdottir, R., Wilson, B. A. & Hersteinsson, P. 2008. Dispersal and the influence of rainfall on a population of the carnivorous marsupial

98

swamp antechinus (Antechinus minimus maritimus). Wildlife Research, 35, 446-454.

Marchesan, D. & Carthew, S. M. 2004. Autecology of the yellow-footed antechinus (Antechinus flavipes) in a fragmented landscape in southern Australia. Wildlife Research, 31, 273-282.

Marlow, B. 1961. Reproductive behaviour of the marsupial mouse, Antechinus flavipes (Waterhouse)(Marsupialia) and the development of the pouch young. Australian Journal of Zoology, 9, 203-218.

Martin, H. 2006. Cenozoic climatic change and the development of the arid vegetation in Australia. Journal of Arid Environments, 66, 533-563.

Martin, R. E. & Sapsis, D. B. Fires as agents of biodiversity: pyrodiversity promotes biodiversity. Proceedings of the conference on biodiversity of northwest California ecosystems. Cooperative Extension, University of California, Berkeley, 1992.

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

Mason, E. D., Firn, J., Hines, H. B. & Baker, A. M. 2017. Breeding biology and growth in a new, threatened carnivorous marsupial. Mammal Research, 62, 179-187.

Mcallan, B. & Dickman, C. 1986. The role of photoperiod in the timing of reproduction in the dasyurid marsupial Antechinus stuartii. Oecologia, 68, 259-264.

Mcallan, B. & Geiser, F. 2006. Photoperiod and the timing of reproduction in Antechinus flavipes (Dasyuridae: Marsupialia). Mammalian Biology-Zeitschrift für Säugetierkunde, 71, 129-138.

Mcallan, B. M., Dickman, C. R. & Crowther, M. S. 2006. Photoperiod as a reproductive cue in the marsupial genus Antechinus: ecological and evolutionary consequences. Biological Journal of the Linnean Society, 87, 365-379.

Mcardle, B. H. & Anderson, M. J. 2001. Fitting multivariate models to

community data: a comment on distance‐based redundancy analysis.

Ecology, 82, 290-297.

99

Mcdonald, W. & Sharpe, P. 1986. The plant communities and flora of Kroombit Tops. The Queensland Naturalist, 27, 3-10.

Mcgregor, H. W., Cliff, H. B. & Kanowski, J. 2017. Habitat preference for fire scars by feral cats in Cape York Peninsula, Australia. Wildlife Research, 43, 623-633.

Mcgregor, H. W., Legge, S., Jones, M. E. & Johnson, C. N. 2014. Landscape management of fire and grazing regimes alters the fine-scale habitat utilisation by feral cats. PLOS One, 9, e109097.

Mckinney, M. L. & Lockwood, J. L. 1999. Biotic homogenization: a few winners replacing many losers in the next mass extinction. Trends in Ecology & Evolution, 14, 450-453.

Mess, A. & Carter, A. M. 2007. Evolution of the placenta during the early radiation of placental mammals. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 148, 769-779.

Mitchell, K. J., Pratt, R. C., Watson, L. N., Gibb, G. C., Llamas, B., Kasper, M., Edson, J., Hopwood, B., Male, D. & Armstrong, K. N. 2014. Molecular phylogeny, biogeography, and habitat preference evolution of marsupials. Molecular Biology and Evolution, 31, 2322-2330.

Monamy, V. & Fox, B. J. 2010. Responses of two species of heathland rodents to habitat manipulation: vegetation density thresholds and the habitat accommodation model. Austral Ecology, 35, 334-347.

Monteith, G. 1986. Insects from Kroombit Tops, Queensland, with some results of a site survey of Coleoptera. The Queensland Naturalist, 27, 27-34.

Moore, B. & Monteith, G. 2004. A second Australian species of the Gondwanan stag beetle genus Sphaenognathus buquet (Coleoptera: Lucanidae). Memoirs of the Queensland Museum, 49, 693-700.

Moritz, M. A., Batllori, E., Bradstock, R. A., Gill, A. M., Handmer, J., Hessburg, P. F., Leonard, J., Mccaffrey, S., Odion, D. C. & Schoennagel, T. 2014. Learning to coexist with wildfire. Nature, 515, 58-66.

Moritz, M. A., Hurteau, M. D., Suding, K. N. & D'antonio, C. M. 2013. Bounded ranges of variation as a framework for future conservation and fire management. Annals of the New York Academy of Sciences, 1286, 92-107.

100

Mutton, T. Y. 2017. Evolutionary biology of the Australian carnivorous marsupial genus Antechinus. Unpublished PhD thesis, Queensland University of Technology.

Mutton, T. Y., Gray, E. L., Fuller, S. J. & Baker, A. M. 2017. Life history, breeding biology and movement in a new species of carnivorous marsupial, the buff-footed antechinus (Antechinus mysticus) and a sympatric congener, the subtropical antechinus (Antechinus subtropicus). Mammal Research, 62, 373-385.

Nakagawa, S. & Schielzeth, H. 2013. A general and simple method for

obtaining R2 from generalized linear mixed‐effects models. Methods

in Ecology and Evolution, 4, 133-142. Newell, G. R. & Wilson, B. A. 1993. The relationship between cinnamon

fungus (Phytophthora cinnamomi) and the abundance of Antechinus stuartii (Dasyuridae: Marsupialia) in the Brisbane Ranges, Victoria. Wildlife Research, 20, 251-259.

Nicholls, J. & Austin, J. 2005. Phylogeography of an east Australian wet‐

forest bird, the satin bowerbird (Ptilonorhynchus violaceus), derived from mtDNA, and its relationship to morphology. Molecular Ecology, 14, 1485-1496.

Panfil, S. N. & Harvey, C. A. 2015. REDD+ and biodiversity conservation: A review of the biodiversity goals, monitoring methods, and impacts of 80 REDD+ projects. Conservation Letters.

Parrott, M. L., Ward, S. J., Temple-Smith, P. D. & 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.

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

Pearce, C. 2016. Diet, breeding and growth in a new carnivorous marsupial, the buff-footed antechinus, Antechinus mysticus, at the northern and southern limits of its range. Queensland University of Technology.

101

Pearson, D. E. & Ruggiero, L. F. 2003. Transect versus grid trapping arrangements for sampling small-mammal communities. Wildlife Society Bulletin, 454-459.

Pearson, K. 1900. X. On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, 50, 157-175.

Peet, R. K., Wentworth, T. R. & White, P. S. 1998. A flexible, multipurpose method for recording vegetation composition and structure. Castanea, 262-274.

Penn, A. M., Sherwin, W. B., Lunney, D. & Banks, P. B. 2003. The effects of a low-intensity fire on small mammals and lizards in a logged, burnt forest. Wildlife Research, 30, 477-486.

Peres, C. A. & Palacios, E. 2007. Basin‐wide effects of game harvest on

vertebrate population densities in Amazonian forests: implications for

animal‐mediated seed dispersal. Biotropica, 39, 304-315.

Ponisio, L. C., Wilkin, K., M'gonigle, L. K., Kulhanek, K., Cook, L., Thorp, R., Griswold, T. & Kremen, C. 2016. Pyrodiversity begets plant–pollinator community diversity. Global Change Biology, 22.5, 1794-1808.

Ponniah, M., Hughes, J. M. & Wallis, G. 2004. The evolution of Queensland spiny mountain crayfish of the genus Euastacus. I. Testing vicariance and dispersal with interspecific mitochondrial DNA. Evolution, 58, 1073-1085.

Predavec, M. & Dickman, C. R. 1994. Population dynamics and habitat use of the long-haired rat (Rattus villosissimus) in south-western Queensland. Wildlife Research, 21, 1-9.

Price, S. A., Hopkins, S. S., Smith, K. K. & Roth, V. L. 2012. Tempo of trophic evolution and its impact on mammalian diversification. Proceedings of the National Academy of Sciences, 109, 7008-7012.

Promislow, D. E. & Harvey, P. H. 1990. Living fast and dying young: A

comparative analysis of life‐history variation among mammals.

Journal of Zoology, 220, 417-437.

102

Purdie, R. W. & Slatyer, R. 1976. Vegetation succession after fire in sclerophyll

woodland communities in south‐eastern Australia. Austral Ecology,

1, 223-236. Queensland Government. 2015a. Department of Natural Resources and

Mines pluviometer data [Online]. Available: https://www.dnrm.qld.gov.au/water/water-monitoring-and-data/portal

[Accessed 10 November 2015. Queensland Government 2015b. Nature Conservation Act 1992. Queensland Herbarium. 2013. Regional Ecosystem Description Database

(REDD). Version 6.1 (February 2013). Queensland Department of Science [Online]. [Accessed 2 October 2014.

Queensland Herbarium 2016. Regional Ecosystem Description Database (REDD). Queensland Department of Science, Information Technology and Innovation: Brisbane.

R Core Team 2015. R: a language and environment for statistical computing. Vienna, Austria; 2015.

Read, D. 1987. Diets of sympatric Planigale gilesi and P. tenuirostris (Marsupialia: Dasyuridae): relationships of season and body size. Australian Mammalogy, 10, 11-21.

Recher, H. F., Lunney, D. & 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.

Ridgeway, G. 2006. gbm: Generalized boosted regression models. R package version, 1, 55.

Robinson, N. M., Leonard, S. W., Bennett, A. F. & Clarke, M. F. 2014. Refuges for birds in fire-prone landscapes: the influence of fire severity and fire history on the distribution of forest birds. Forest Ecology and Management, 318, 110-121.

Robinson, N. M., Leonard, S. W., Ritchie, E. G., Bassett, M., Chia, E. K., Buckingham, S., Gibb, H., Bennett, A. F. & Clarke, M. F. 2013. Review:

Refuges for fauna in fire‐prone landscapes: their ecological function

and importance. Journal of Applied Ecology, 50, 1321-1329. Rowe, T., Rich, T. H., Vickers-Rich, P., Springer, M. & Woodburne, M. O.

2008. The oldest platypus and its bearing on divergence timing of the

103

platypus and echidna clades. Proceedings of the National Academy of Sciences, 105, 1238-1242.

Russell‐Smith, J., Ryan, P. G., Klessa, D., Waight, G. & Harwood, R. 1998.

Fire regimes, fire‐sensitive vegetation and fire management of the

sandstone Arnhem Plateau, monsoonal northern Australia. Journal of Applied Ecology, 35, 829-846.

Sale, M., Kraaijeveld-Smit, F. & Arnould, J. 2013. Multiple paternity in the swamp antechinus (Antechinus minimus). Australian Mammalogy, 35, 227-230.

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

Sale, M., Wilson, B. & Arnould, J. 2009. Comparison of life‐history

characteristics of island and mainland populations of the swamp antechinus. Journal of Zoology, 277, 119-125.

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

Sanderson, E. W., Jaiteh, M., Levy, M. A., Redford, K. H., Wannebo, A. V. & Woolmer, G. 2002. The human footprint and the last of the wild: the human footprint is a global map of human influence on the land surface, which suggests that human beings are stewards of nature, whether we like it or not. BioScience, 52, 891-904.

Schulz, M. & De Oliveira, M. 1995. Microchiropteran fauna of Kroombit Tops, central Queensland, including a discussion on survey techniques. Australian Zoologist, 30, 71-77.

Scott, M. P. 1986. The timing and synchrony of seasonal breeding in the marsupial, Antechinus stuartii: interaction of environmental and social cues. Journal of Mammalogy, 67, 551-560.

Short, J. & Smith, A. 1994. Mammal decline and recovery in Australia. Journal of Mammalogy, 75, 288-297.

104

Sibly, R. M. & Brown, J. H. 2007. Effects of body size and lifestyle on evolution of mammal life histories. Proceedings of the National Academy of Sciences, 104, 17707-17712.

Sisk, T. D., Launer, A. E., Switky, K. R. & Ehrlich, P. R. 1994. Identifying extinction threats. BioScience, 44, 592-604.

Smith, G. C. 1984. The biology of the yellow-footed antechinus, Antechinus flavipes (Marsupialia: Dasyuridae), in a swamp forest on Kinaba Island, Cooloola Quensland. Wildlife Research, 11, 465-480.

Smith, K. K. 2001. Early development of the neural plate, neural crest and facial region of marsupials. Journal of Anatomy, 199, 121-131.

Statham, H. 1982. Antechinus stuartii (Dasyuridae, Marsupialia) diet and food availability at Petroi, northeastern New South Wales. Carnivorous Marsupials (Ed. M. Archer).

Stawski, C., Körtner, G., Nowack, J. & Geiser, F. 2015. The importance of mammalian torpor for survival in a post-fire landscape. Biology Letters, 11, 20150134.

Stawski, C., Nowack, J., Körtner, G. & Geiser, F. 2017. A new cue for torpor induction: charcoal, ash and smoke. Journal of Experimental Biology, 220, 220-226.

Stearns, S. C. 1976. Life-history tactics: a review of the ideas. The Quarterly Review of Biology, 51, 3-47.

Stokes, V. L., Pech, R. P., Banks, P. B. & Arthur, A. D. 2004. Foraging behaviour and habitat use by Antechinus flavipes and Sminthopsis murina (Marsupialia: Dasyuridae) in response to predation risk in eucalypt woodland. Biological Conservation, 117, 331-342.

Swinburn, M. L., Fleming, P. A., Craig, M. D., Grigg, A. H., Garkaklis, M. J., Hobbs, R. J. & Hardy, G. E. S. J. 2008. The importance of grasstrees (Xanthorrhoea preissii) as habitat for mardo (Antechinus flavipes leucogaster) during post-fire recovery. Wildlife Research, 34, 640-651.

Tasker, E. M. & Dickman, C. R. 2001. A review of Elliott trapping methods for small mammals in Australia. Australian Mammalogy, 23, 77-87.

Tingley, M. W., Ruiz-Gutiérrez, V., Wilkerson, R. L., Howell, C. A. & Siegel, R. B. Pyrodiversity promotes avian diversity over the decade following forest fire. Proceedings of the Royal Society B, 2016. The Royal Society.

105

Turner, D. & Meister, O. 1988. Hunting behaviour of the domestic cat. Domestic cat: the biology of its behavior (Ed. D. C. Turner and P. Bateson.

Turner, I. 1996. Species loss in fragments of tropical rain forest: a review of the evidence. Journal of Applied Ecology, 200-209.

Ungar, P. S. 2010. Mammal teeth: origin, evolution, and diversity, JHU Press.

Van Dyck, S. 1981. The relationships of Antechinus stuartii and A. flavipes (Marsupialia: Dasyuridae) with special reference to Queensland. Carnivorous Marsupials (Ed. M. Archer). University of New South Wales.

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

Van Dyck, S., Gynther, I. & Baker, A. 2013. Field companion to the mammals of Australia, New Holland Publishers.

Van Dyck, S. & Strahan, R. 2008. The Mammals of Australia, New Holland Pub Pty Limited.

Van Loon, A. 1970. All that litter is not old. A look at fuel accumulation in blackbutt forests on the mid-north coast of NSW. A preliminary summary. Forestry Commission of NSW Unpublished Report.

Van Valkenburgh, B. 1988. Trophic diversity in past and present guilds of large predatory mammals. Paleobiology, 14, 155-173.

Wainer, J. 1976. Studies of an island population of Antechinus minimus (Marsupialia, Dasyuridae)[Great Glennie Island, Victoria]. Australian Zoologist (Australia).

Wainer, J. W. 1989. Ecological studies of island and mainland populations of Rattus fuscipes and Antechinus minimus in coastal heathland in southern Victoria. PhD, University of Melbourne.

Wakefield, N. A. & Warneke, R. M. 1967. Some revision in Antechinus (Marsupialia), Reprinted for the Fisheries and Wildlife Department Victoria.

Warren, W. C., Hillier, L. W., Graves, J. a. M., Birney, E., Ponting, C. P., Grützner, F., Belov, K., Miller, W., Clarke, L. & Chinwalla, A. T. 2008.

106

Genome analysis of the platypus reveals unique signatures of evolution. Nature, 453, 175-183.

Watkins, A. F., Mcwhirter, J. L. & King, C. M. 2010. Variable detectability in long-term population surveys of small mammals. European Journal of Wildlife Research, 56, 261-274.

Westerman, M., Krajewski, C., Kear, B. P., Meehan, L., Meredith, R. W., Emerling, C. A. & Springer, M. S. 2016. Phylogenetic relationships of dasyuromorphian marsupials revisited. Zoological Journal of the Linnean Society, 176, 686-701.

Western, D. 1979. Size, life history and ecology in mammals. African Journal of Ecology, 17, 185-204.

Whelan, R., Ward, S., Hogbin, P. & Wasley, J. Responses of heathland Antechinus stuartii to the Royal National Park wildfire in 1994. Proceedings of the Linnean Society of New South Wales, 1996. Linnean Society of New South Wales, 97-108.

Willmott, W. F. & Moore, T. 2006. Rocks and landscapes of the national parks of central Queensland, Geological Society of Australia, Queensland Division.

Wilson, B., Aberton, J. & Reichl, T. 2001. Effects of fragmented habitat and fire on the distribution and ecology of the swamp antechinus (Antechinus minimus maritimus) in the eastern Otways, Victoria. Wildlife Research, 28, 527-536.

Wilson, B., Bourne, A. & Jessop, R. 1986. Ecology of small mammals in coastal heathland at Anglesea, Victoria. Wildlife Research, 13, 397-406.

Wilson, B., Robertson, D., Moloney, D., Newell, G. & Laidlaw, W. Factors affecting small mammal distribution and abundance in the eastern Otway Ranges, Victoria. Proceedings of the Ecological Society of Australia, 1990. 379-396.

Woinarski, J. C. 1999. Fire and Australian birds: An annotated bibliography. Australia’s Biodiversity: Responses to Fire. Canberra: Department of Environment and Heritage.

Woinarski, J. C., Burbidge, A. A. & Harrison, P. L. 2014. A review of the conservation status of Australian mammals. Therya, 6, 155-166.

107

Woinarski, J. C., Burbidge, A. A. & Harrison, P. L. 2015. Ongoing unraveling of a continental fauna: decline and extinction of Australian mammals since European settlement. Proceedings of the National Academy of Sciences, 112, 4531-4540.

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

Woolley, P. Reproduction in Antechinus spp. and other dasyurid marsupials. Symposia of the Zoological Society of London, 1966. 281-294.

Wroe, S. 1997. A reexamination of proposed morphology-based synapomorphies for the families of Dasyuromorphia (Marsupialia). I. Dasyuridae. Journal of Mammalian Evolution, 4, 19-52.

York, A. 1999. Long-term effects of frequent low-intensity burning on the abundance of litter-dwelling invertebrates in coastal blackbutt forests of southeastern Australia. Journal of Insect Conservation, 3, 191-199.

108

Appendix 1. Methods and results of rainfall data

collection and analysis Monthly rainfall data from August 2013 to September 2015 was obtained from the Queensland Government Department of Natural Resources and Mines (Queensland Government, 2015a). This data is collected autonomously by a pluviograph situated approximately 1 km west-northwest of the Lookout site. In addition to actual rainfall data, Foley’s drought index was calculated (actual rainfall for the 3 years preceding each month minus mean 3-yearly rainfall divided by mean annual rainfall). Effectively, this index takes past rainfall into account, allowing for the climate history of the study sites to be represented (Foley, 1957, Fensham et al., 2015). Only rainfall data dating back to 1993 (when pluviograph recording was instigated) were able to be obtained. To test whether monthly rainfall over 2013-2015 differed significantly from the mean monthly rainfall since 1993, a linear regression analysis was applied in R version 3.2.2. (R Core Team, 2015).

Rainfall from July 2013 to September 2015 deviated substantially from mean monthly values (recorded since 1993) (Fig 5.2). Overall, Foley’s index was high (above 0.5) throughout the study period. Based on mark-recapture data, August-October is the period between birth and weaning of A. argentus young (see Chapter 3). Rainfall in August-October 2013 was lower (total 47 mm) than in August-October 2014 (total 221 mm). However, Foley’s index was relatively high in 2013: 2.08 in August 2013 (the preceding 3-year rainfall was 127% of the mean) and 1.85 in October 2013 (121% of the mean). In comparison, Foley’s index in 2014 was markedly lower: 0.76 (93% of the mean) in August and 0.77 (94% of the mean) in October (Fig 5.2). Thus, based on Foley’s index, the rainfall in the second half of 2013 was markedly higher than in the second half of 2014. In all studied Antechinus species, weaning is followed by an approximately three-month period in which free-living juveniles disperse and grow to adult size (Cockburn et al., 1985). For A. argentus, this period falls in November-

109

January. We found there was a lower than typical (since 1993) amount of rainfall recorded in Nov 2013-Jan 2014 (345 mm) followed by a higher than average amount in Nov 2014-Jan 2015 (606 mm) (Fig 5.2). This trend was evident if not as pronounced in light of the climate history, with the Foley’s index for Nov 2013-Jan 2014 decreasing from 1.76-1.03 (119-101% of the mean), while in Nov 2014-Jan 2015 it increased from 0.75-0.81 (94-95% of the mean) (Fig 5.2). Notwithstanding these observed trends, formal statistical analysis did not detect an abnormal amount of rainfall in either year, with mean monthly rainfall significantly predicting actual monthly rainfall (F[1, 24] = 16.86, P < 0.001, adjusted R2 = 0.39).