Copyright and Use of This Thesis This Thesis Must Be Used in Accordance with the Provisions of the Copyright Act 1968

Copyright and use of this thesis This thesis must be used in accordance with the provisions of the Copyright Act 1968. Reproduction of material protected by copyright may be an infringement of copyright and copyright owners may be entitled to take legal action against persons who infringe their copyright. Section 51 (2) of the Copyright Act permits an authorized officer of a university library or archives to provide a copy (by communication or otherwise) of an unpublished thesis kept in the library or archives, to a person who satisfies the authorized officer that he or she requires the reproduction for the purposes of research or study. The Copyright Act grants the creator of a work a number of moral rights, specifically the right of attribution, the right against false attribution and the right of integrity. You may infringe the author’s moral rights if you: - fail to acknowledge the author of this thesis if you quote sections from the work - attribute this thesis to another author - subject this thesis to derogatory treatment which may prejudice the author’s reputation For further information contact the University’s Director of Copyright Services sydney.edu.au/copyright LIFE ON THE EDGE: POPULATION AND BEHAVIOURAL RESPONSES OF THE NATIVE BUSH RAT TO INVASIVE SPECIES AT THE URBAN EDGE.

Wendy K. Gleen

A thesis submitted in fulfilment of the requirements for the degree of Master of Science

School of Biological Sciences

University of Sydney

Submitted April 2013

This research was conducted under ethics approval number: ACEC 09/74A

Statistical analysis for the generalised linear mixed model in Chapter 4 was undertaken with the guidance of Associate Professor Peter Thomson.

Declaration

I hereby declare that this thesis contains no material which has been accepted for the award of any other degree or diploma at any University or equivalent institution and that, to the best of my knowledge and belief, this thesis contains no material previously published or written by another person, except where due reference is made in the text of the thesis.

(Signed) ______

Date ______

This thesis is dedicated firstly to all the bush regenerators I know for their deep knowledge of and care for the Australian bush. Weed control processes are vital for biodiversity conservation and I hope that is sufficiently clear in this thesis.

Secondly it is dedicated to the native rodents of Australia, a taxa often forgotten by the general community. May these wonderful creatures prevail over accelerating environmental problems.

Bush rat Rattus fuscipes Photo by W. Gleen, image courtesy of Taronga Zoo.

Weedy urban edge. Photo by W. Gleen

iii

ABSTRACT

As human population growth in Australia drives urban sprawl, the urban edges created by this process have the potential to impact an increasing area of native habitat. Invasive species are a common feature of urban edges and pose a significant threat to biodiversity globally. However, while it is well known that urban edges provide a point of incursion for invasive species, their impact on native wildlife at urban edges, whether positive or negative, remains poorly understood. In this thesis I examined the responses of native bush rats (Rattus fuscipes) to two invasive alien species that are especially common at urban edges along Australia’s eastern coast; the weed lantana (Lantana camara) and the black rat (Rattus rattus). I first compared habitat quality for bush rats in weedy urban edge, weedy core and ‘pristine’ core areas by examining differences in population abundance and demographics. I used a multi scaled and mechanistic approach to explore the causative factors driving the differences in habitat quality across these areas, also applying and testing current edge and invasion ecology theory.

I found that habitat quality for bush rats was relatively poor at urban edges, however, lantana density was positively correlated with bush rat abundance indices at edges, suggesting that it may ‘seal’ the edge, ameliorating other deleterious edge effects. Lantana appeared to provide an ecosystem service for bush rats which was evident at several scales.

Plant architecture, measured as the evenness of vegetation in the vertical plane, was an important predictor of small mammal fine-scale habitat use, an aspect which has seldom been considered in detail for ground dwelling mammals.

I then used a Giving Up Density (GUD) experiment to test the hypothesis that bush rats use lantana because of lowered perceived predation risk. I compared GUDs of feeding trays set in lantana to those in native vegetation and clear ground. I found that vegetation structure reduces bush rat perceived predation risk irrespective of species (lantana or native) supporting the concept that plant structure is more important than floristic composition in predicting habitat use by small mammals.

Finally, I used camera traps to explore patterns in the abundance and activity of bush and black rats across edge, core weedy and core macrohabitats to test whether competition

iv from black rats explains poor quality habitat for bush rats at the urban edge. Being of equivalent size, bush rats and black rats show symmetrical competition based on residency (Stokes et al., 2012, Stokes et al., 2009b, Stokes et al., 2009a) (the residency hypothesis) which might advantage black rats at urban edges due to the strong potential for rapid reinvasion from urban areas where they are common. Contrary to my predictions, there were no negative associations in population abundance between the two species which coexisted at several sites. Patchy distribution of black rats may explain the lack of support for the residency hypothesis, or the presence of lantana at edge and core weedy sites may have mediated competition by providing visual and physical barriers, thus facilitating co existence.

Overall, my results demonstrate that invasive species do not always have a negative impact on fauna at urban edges in Australia, and in fact may enhance habitat quality at certain ecological scales. This result suggests that further work into the mechanistic basis for the nature of the interactions between native and alien species at urban edges is needed in order to balance the competing consequences of the management of entrenched alien species.

TABLE OF CONTENTS ABSTRACT ...... iv ACKNOWLEDGEMENTS ...... 1 CHAPTER 1 Thesis introduction ...... 4 STUDY CONTEXT ...... 4 INTRODUCTION TO THE TARGET SPECIES ...... 4 EDGE MECHANISMS AND PREDICTIVE MODELS ...... 6 URBAN EDGES ...... 9 INVASION ECOLOGY ...... 10 WEEDS AS HABITAT ...... 11 SPECIES INTERACTIONS AT THE URBAN EDGE ...... 12 THESIS AIMS AND STRUCTURE ...... 12 CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales ...... 14 ABSTRACT ...... 14 INTRODUCTION ...... 14 METHODS ...... 21 DATA ANALYSIS ...... 35 RESULTS ...... 41 DISCUSSION ...... 52 CHAPTER 3 Lantana and bush rat perceived predation risk...... 62 ABSTRACT ...... 62 INTRODUCTION ...... 62 METHODS ...... 69 DATA ANALYSIS ...... 74 RESULTS ...... 75 DISCUSSION ...... 77 CHAPTER 4 Does competition from black rats lower habitat quality for the bush rat at the urban edge? ...... 82 ABSTRACT ...... 82 INTRODUCTION ...... 82 METHOD ...... 87 DATA ANALYSIS ...... 88 RESULTS ...... 92

DISCUSSION ...... 96 CHAPTER 5 General discussion ...... 101 SUMMARY OF KEY FINDINGS ...... 101 FUTURE DIRECTIONS ...... 105 MANAGEMENT IMPLICATIONS OF MY RESULTS ...... 107 REFERENCES ...... 109

vii

ACKNOWLEDGEMENTS

Peter: I was so honoured to be accepted as a student, and that honour has grown along with gratitude for your supervisory talents. You have the happy knack of steering with a light hand and a strong safety net, all the while quietly boosting the confidence of someone feeling at times very out of depth. You have been unfailingly kind and patient, your quiet confidence in my abilities has been a ‘lifeline’ at times, and inspired me to reach way beyond what I thought was the limit of my skills. It is hard to adequately express my thanks, this process has wildly exceeded any expectations I had when starting out, I’ve learnt more than I could have imagined, and I’ve loved every minute!! This has largely been due to the joy of being so well supported and encouraged, not to mention the inspiration of learning from an ecologist of such stature. I am particularly taking away an enhanced appreciation of good scientific writing as I hadn’t quite grasped the true extent of artistry and skill it requires! I will strive to develop that skill, I realise it is of paramount importance. Thankyou.

There are so many people that I have shamelessly depended on over the last 4 years and who have been so generous with their time and expertise:

Grainne: you enabled an idle thought to become reality, for which I shall be forever grateful. Your exuberant enthusiasm for science (and life in general) is infectious, and I hope it never dulls. When you presented such an awesome project with such encouragement for me to get involved, how could I not grab hold? Your staunch support and help has always been so gratefully appreciated, and you are a true inspiration. I look forwards to watching a bright future continue to unfold, and hope I can continue to be a part of it. Bill: the consummate gentleman, ever riding to the rescue; I owe you so much too. Your generosity of spirit is amazing, for which I am extremely grateful. I SO look forwards to sharing some chilled out days of good music with you guys!

Wendy K: the thesis dedication is largely for you. Thank you from the bottom of my heart for your grace in sharing your thesis topic with me. I was at my wits end for a while there – if it wasn’t for your kindness there was a strong chance I wouldn’t have made it through the first few months. I’ve enjoyed this project so much, not least the field work in the stunning bushland of the Central Coast, and the wonderful people I’ve met and it’s due to your generosity. I know you’ll produce some outstanding work, and I DO hope we can collaborate in some way, somehow (if you’ll have me!)……

My lab mates, I’m so grateful to everyone, and honoured to meet such wonderful and inspiring people, and to associate with such intellectual capability. Alex –my buddy in pain – thank you SO much for all your help – when you were flat out yourself you helped me with the mysteries of ANOVAs, GUDs and other bizarre concepts. It was only due to being able to piggy back your pioneering GUD work that chapter 3 was possible. I will forever be in awe of your work ethic, and do hope you never have to go through that intensity again! You inspired me to work harder than I thought possible towards the end. I can’t thank you enough for your words of support when it all seemed too much. May the whales lie quiescent for a while.

Cath, a truly principled, kind and outstanding human being, not to mention outstanding scientist it’s been an honour and an inspiration to learn from you. Jenna, I’m so grateful for all your help, you’ve always cheerfully helped me out with everything from small practicalities, to expert reshaping of recalcitrant chapters. You’re a gem, thank you SO much for your patience!!! Mal: I have missed your cheeky presence of late. I love the fact that the mischievous irreverence hides such a sharp mind. I learnt a lot from you, not least how to combine excellent science and fun (and whales of course). Caragh thanks for starting me off with multivariate analyses, (you’re the ants pants!) and Helen for sound advice on structure, and also for the excellent adventure stories! It’s been an honour to associate with such intellectual capability. Likewise Viyanna, thanks so much for your help, support and friendship.

I would like to thank NSW National Parks and Wildlife (Gosford and Northern Beaches offices), Gosford council, NSW State Forests and Port Stephens Council for assistance with field sites and Taronga Zoo for some study leave. Rebecca Spindler thank you for the most

2 wonderful encouragement and support, Doug Beckers for seeding the idea of postgrad research, and for inspiration in so many ways. Deb, Liz and others from Gosford NPWS, Phil Wood and Anthony Marchment for hugely appreciated assistance with field sites. Special mention to Paul Malligan for advice on the bush regeneration aspect, and other insightful suggestions. The wonderful field work volunteers were too numerous to list, but I am exceedingly grateful to everyone; it would not have been possible without your help. I would especially and sincerely like to thank the Australasian Society of Zookeeping for the Des Spittal scholarship which paid for some much needed field equipment. I also thank Peter McGee and Michael O’Shea for sharing their expertise in such interesting fields and adding some important background to the thesis.

Finally my long suffering friends and family – thank you so much for your support and encouragement. Claire and Angie, thanks so much for putting me up in Port Stephens, and for sterling help and excellent company. The Dixon- Valks for saving my bacon in so many ways – from ute bail outs to peanut butter balls and proof reading, thank you SO much, I honestly don’t know what I’d do without you. By the way you’re in the ‘family’ section I’m afraid! I hope Saskia enjoyed working with Radish fuscipes, as I did. Mum, Dad, Penny and Simon, Lillian and John, what can I say? You make it all possible. Thanks so much Penny and Simon for putting me up at Lovett Bay in your magic tree house, and for EVERYTHING. Mum and Dad have my gratitude for everything from data entry to keeping me very well fed and for service above and beyond the call of duty i.e. a trip to Homebush bay IKEA on a Saturday morning for emergency field supplies. Uncle Keith, thanks for the IT help, you shouldn’t have come all the way from the UK for it though! Lastly, my darling, long suffering husband: thank you so much, mainly for putting up with me, and the things I take on. Also for assuming my share of the tasks, and for endless reassurance and support. I obviously couldn’t do any of it without you. Last but not least my beloved pooch ‘Spodes’ for unconditional love and unlimited hugs. I’m sorry for kicking you out of the house for snoring. You may now rattle the rafters.

This research was undertaken using sustainable principles, and by treading as lightly as possible in the natural sections of the study areas. All fauna records were entered into the NSW NPWS wildlife atlas.

CHAPTER 1 Thesis introduction

STUDY CONTEXT The urban matrix consists of built landscapes characterised by impervious surfaces (MacGregor-Fors, 2011), interspersed with gardens and remnant patches which are usually intensely modified from the original ecosystem. The matrix is inhospitable to most wildlife (McKinney, 2002), additionally edge effects radiate from its boundary into the surrounding natural areas. Edge effects have been studied extensively and continued development of edge theory and its application potential is important firstly because edge effects are key factors that drive biodiversity loss both globally and in Australia (Lindenmayer and Fischer, 2006) and also because there are recent suggestions that the scale and scope of edge effects may have been underestimated (Banks-Leite et al., 2010, Ries and Sisk, 2009). It is also vital to understand how bush rats respond to the urban edge as they are a key species for reintroduction programs into fragmented bushland surrounded by urban settlement. I aim to inform a larger study reintroducing the bush rats to Sydney harbour foreshore to block reinvasion by black rats (Banks et al., 2010-2012). The role of invasive species, particularly fauna, at the urban edge is still poorly understood; additionally the general hypotheses which frame current understanding of invasion ecology are not always empirically supported (Jeschke et al., 2012). By investigating bush rat responses to the urban edge with a focus on invasive species in the context of these frameworks, I aim to increase our understanding of edge and invasion theory.

INTRODUCTION TO THE TARGET SPECIES

The bush rat The Bush Rat is an Australian native rodent (Order Rodentia, Family Muridae), common in Australian coastal ecosystems where there is dense undergrowth (Lunney 2008). Their ecology has been researched extensively (e.g. Banks, 1999, Woodside, 1983, Wood, 1971, Warneke, 1971, Robinson, 1987, Taylor and Calaby, 1988, Peakall et al., 2003, Lindenmayer et al., 2005, White et al., 1996, Banks and Dickman, 2000), and they are known to be habitat generalists. Breeding may be most prevalent in spring but populations near Sydney may breed all year round (Taylor, 1961, White et al., 1996), especially if conditions are good, as they were in this study. Females are natally philopatric whereas males disperse more widely (Robinson, 1987, Woodside, 1983, Macqueen et al., 2008). The species is regularly used for

CHAPTER 1 Thesis introduction

field research as they are common, easily trapped and a model species for testing ecological theory. Despite this, their sensitivity to urban edges is poorly understood.

The black rat The black rat Rattus rattus is an archetypal synanthrope which has invaded the globe, it’s path of invasion closely following that of human settlement (Konečný et al., 2013). It probably arrived in Australia with the first fleet in 1788 and populations have subsequently established around the coast of Australia and throughout Tasmania (Watts and Aslin, 1981). The species thrives in urban areas (Dickman and Watts, 2008), as well as in adjacent bushland (Cox et al., 2000, Weerakoon, 2011, Watts and Aslin, 1981) and along transport corridors such as road edges (Downes et al., 1977). It has also been described as ‘an animal of the forest edge’ (Watts and Aslin, 1981), so the urban edge may represent good quality habitat which could potentially provide a springboard for further invasion into the bushland interior. Black rats negatively impact native wildlife globally in many ways, such as by predation, competition, spread of disease and parasites, however, there is a lack of understanding of these effects on Australian biota (reviewed in Banks and Hughes, 2012). The black rat has recently been identified as a potential competitor of the bush rat, and this potential may be increased at the urban edge if the black rat is indeed more prevalent there.

Lantana Lantana Lantana camara was the most abundant weed species present at urban edges of my study area, and allowed consistency when comparing weed and native plant structure. Lantana is a shrub native to Central and South America, which forms dense thickets that smother other plants; a process possibly aided by allelopathy (Achhireddy and Singh, 1984, Muyt, 2001). Lantana is an aggregate species able to hybridise (Johnson, 2007) and has more than 29 different varieties (which differ morphologically). The most common variety of Lantana in NSW was used in this study. It usually expresses pink flowers, has strong lightly barbed 4-sided stems with a tendency to grow in a horizontal, interlocked pattern with thickets growing 2-4 m high, or taller if supported by other vegetation such as trees (Johnson, 2007, Muyt, 2001). Leaves are opposite and relatively small and the plant has a pungent odour if bruised or broken. Lantana is known to be toxic if consumed by cattle, rodents and other mammals (Sharma et al., 1981, Sharma et al., 1988, Pass et al., 1979,

CHAPTER 1 Thesis introduction

Thorp and Wilson, 1998 onwards), and can affect rodent reproductive capability (Mello et al., 2003, Mello et al., 2005).

Lantana was introduced to Australia as an ornamental plant in the 1840’s (Thorp and Wilson, 1998 onwards). It now covers over four million hectares of land, mainly in Eastern NSW and Queensland (NSW Department of Primary Industries, 2013). Lantana impacts agriculture by invading cultivated land and causing stock death, pasture loss and increased productivity costs (Thorp and Wilson, 1998 onwards). It severely impacts biodiversity, with an estimated 1321 native plant and 158 native animal species threatened by Lantana in Australia. Conversely, 142 animal species are estimated to be advantaged by this weed (NSW Department of Primary Industries, 2013) and lantana has been reported to provide shelter for fauna (Coutts-Smith 2006), however, the vast amount of evidence of this positive role of lantana is anecdotal or informal, with a few exceptions such as Goth and Vogel (2002) and Virkki (2012).

EDGE MECHANISMS AND PREDICTIVE MODELS Edges are defined in ecology as where any two (or more) structurally different ecosystems are juxtaposed, creating an interface that is characterised by a different combination of ecological and microclimatic features to those of either ecosystem interior (Matlack, 1993, Murcia, 1995, Lindenmayer and Burgman, 2005). Edge effects were first studied in the early 1900’s (for a review see Ries et al. 2004b) and it was found that the interface between different ecosystem types could create advantages for some species due to access to a wider range of resources, for example grazing and shelter where grassland abuts forest (Johnson et al., 1995, Yahner, 1988). Since then, the focus of edge ecology has largely been on anthropogenically created edges, due to their increasing extent and impact on remaining natural systems as global urbanisation expands. These types of edges generally have a negative impact on species distribution and biodiversity (Murcia, 1995, Ries et al., 2004b, Lindenmayer and Fischer, 2006, Laurance, 1997).

Edge effects can be biotic (e.g. a change to predation risk), or abiotic (e.g. a change to the microclimate (Murcia, 1995)). They can be primary, which are changes that occur immediately upon creation of an edge e.g. increased light due to canopy removal, or secondary which arise as a consequence of the primary changes e.g. weed growth due to

CHAPTER 1 Thesis introduction

increased light (Harper et al., 2005). Different types of edge effects are usually interactive, and the properties of edges can be a mix of those of each adjoining ecosystem or they can consist of emergent properties i.e. those that are unique to that habitat (Liddicker, 1999). Edge effects can cause a response by biota that is commonly measured as abundance, but other variables can be used. An increase of the measured variable as it approaches the edge is called a positive response, no change in the variable is described as neutral and a decrease is called a negative edge response (Ries et al., 2004b). Although ‘edge effects’ is a collective term for a group of related effects on biota, the responses to an edge by flora and fauna are species specific (Schlaepfer and Gavin, 2001, Bock et al., 2002, Youngentob et al., 2012), and can vary even at small scales in time and space (Laurance et al., 2007). The impacts of edges are most often reported to occur within 50 - 150 m from the boundary of the interface (Murcia, 1995, Laurance et al., 2007, Matlack, 1993), although cascading effects can resonate for much greater distances. Indeed there is increasing evidence that the scope of edge effects is greater than previously realised and it is suggested that anthropogenically created edges can affect ecological processes for several kilometres beyond their immediate vicinity (Laurance et al., 2002, Laurance, 2000). It is also suggested that edge effects drive changes in fauna and flora that were previously attributed to other processes such as habitat fragmentation (Ries and Sisk, 2009, Banks-Leite et al., 2010)..

An increasingly mechanistic approach has been employed for the study of edge effects including a focus on developing frameworks to explain variation of edge effects and species responses to them. Fundamental processes that drive population responses to edges include species movement or ecological flow (including cross boundary subsidies), mortality, access to spatially separate resources, resource mapping, and interactions of species or processes (Fagan et al., 1999, Ries et al., 2004b). Other principles are also recognised, for example, the importance of the landscape matrix, and contrast between the component patches in affecting the strength of species responses at the edge (Brady et al., 2011, Franklin and Lindenmayer, 2009, Santos-Filho et al., 2012, Laurance, 1997, Harper et al., 2005). Using this framework of understanding, models can be generated to predict the species specific responses to an edge. For example, Lidicker (1999) proposed that species responses may fit into one of two fundamental models; the ecotone or matrix, and I use this concept to guide and interpret my investigation of bush rat responses to the urban edge.

CHAPTER 1 Thesis introduction

The matrix model reflects a response to the properties of the adjoining ecosystems, whereas the ecotone model reflects emergent properties that are different to the properties or combination of properties inherent to the adjoining ecosystems (Figure 1.2).

Figure 1.2 Representation of a species response to an edge (after Lidicker 1999). See text for explanation.

The matrix effect generates a clearly demarcated response to each ecosystem (Figure 1.2; 1b and d), except for 1a where there is no response and 1c where there is a response to a mix of the ecosystem properties. In this case the extent of the response would correspond

CHAPTER 1 Thesis introduction

to the degree of the mix (Liddicker, 1999). Where there is an ecotone response (Figure 1.2; 2a, b, c, d) the emergent quality of the ecotone is revealed in the increase or decrease of the response variable, which is not attributable to either ecosystem or a mix of their properties (Liddicker, 1999). Examples of both response types were found in a study on birds and mammals at an urban edge using measures of abundance and independently derived knowledge of habitat associations (Kristan III et al., 2003). The negative response of some species was predictable from habitat changes at the edge (a matrix response), whereas the negative response of other species could not be predicted by habitat changes and it was concluded that emergent properties were driving an ecotone response for these species (Kristan III et al., 2003).

URBAN EDGES Australia’s human population is growing at a rate of 1.5 % per annum and is predicted to reach 44.7 million by 2101 (medium projection, Australian Bureau of Statistics, 2012). Population growth drives an increase in dwellings (McDonald, 2008), causing fragmentation of remaining bushland (Garden et al., 2006) and consequently an increase in the ratio of bushland edge to interior. In the northern sector of the study area (the Central Coast of New South Wales (NSW)) alone, 1,900 ha will be developed to accommodate a population increase of 37,400 people requiring 17,000 new dwellings (NSW Department of Planning and Infrastructure, 2012). It is increasingly important to understand how resulting urban edges change ecological functions.

Urban edges may be characterised by different habitat features than natural edges due to anthropogenic influence. Urban driven abiotic and biotic primary changes may include air, water, land, noise and artificial light pollution (Leishman et al., 2004, Smith and Smith, 2010). Secondary changes may include soil compaction and changes to the fauna assemblage. For example distribution patterns of black bears Ursus americanus at the urban edge in North America were changed due to the extent of edible rubbish at the urban edge, which increased habitat quality there (Beckmann and Berger, 2003). Urban edges can also facilitate establishment of alien species, particularly weeds. A weed is any plant that requires management to control its effect on the environment or economy (Department of sustainability, 2013). Weeds are invasive and usually alien species, they have the potential to threaten biodiversity and profoundly change ecosystem processes (Downey et al., 2009).

CHAPTER 1 Thesis introduction

Once a weed has established at an edge it is well placed to invade the interior of bushland or spread along the boundary (e.g. Holle and Simberloff, 2005, Alston and Richardson, 2006, Smith and Smith, 2010, Brooks, 2007).

INVASION ECOLOGY Invasive species are one of the most important drivers of biodiversity loss globally (International Union for the Conservation of Nature, 2013, Sala et al., 2000, Davis, 2010). In NSW, weeds are implicated in the threatening process of 45% of threatened species, and invasive fauna affect 38% (Coutts-smith and Downey, 2006a). Three factors influence the success of biotic invasion. Firstly suitable vectors or pathways are required to transport the species to the invasion point outside their natural range. Secondly the manner of the introduction is important for successful invasion– there needs to be sufficient numbers of individuals or propagules to sustain pressure and support establishment (Kolar and Lodge, 2001, Lockwood et al., 2005), and separate sources are required to ensure genetic diversity which supports persistence (Kolbe et al., 2004, Konečný et al., 2013). Thirdly, invasive species and the ecosystems they invade may have certain traits that predispose them to invade successfully or be invaded respectively. For the invader these may include but are not restricted to: mode of reproduction (Kolar and Lodge, 2001), phenotypic plasticity (Davidson et al., 2011), superior competitive or predatory ability (Sax and Brown, 2000), possession of novel biochemical ‘weapons’ such as allelopathy (Hufbauer and Torchin, 2007, Callaway and Ridenour, 2004), pre-adaptation to conditions in the new environment (Sax and Brown, 2000, Hufbauer and Torchin, 2007) including an ability enter a mutualistic relationship with another invasive species there (invasion meltdown) (Simberloff and Von Holle, 1999). Hufbauer and Torchin (2007) draw a distinction between processes that give invaders an advantage relative to native species at the point of introduction, and processes that give invaders an advantage relative to conditions they experience in their native range; the latter promotes strong invasion. In a key-lock manner, characteristics of ecosystems can enable such invader traits (Cornelius et al., 1990, Heger and Trepl, 2003), for example where there are insufficient natural enemies to contain populations (enemy release hypothesis), where resources are underutilised by native species (empty niche hypothesis) or where good quality habitat which advantages native species also advantages invasive species (biotic acceptance). Alternately invaders can be blocked by resistance from resident

CHAPTER 1 Thesis introduction

communities (biotic resistance). It is important that such hypotheses are supported by quantitative data, which is sometimes lacking (Jeschke et al., 2012). Consequently understanding of some mechanistic processes and invader-ecosystem interactions remain poorly understood (Jeschke et al., 2012, Levine et al., 2003). There is also little understanding of the effects of invasive species at urban edges on native fauna, despite the fact that many studies document the incursion of invaders from such edges.

WEEDS AS HABITAT Vegetation provides a range of resources for small mammals, such as food (Banks and Dickman, 2000, Robinson, 1987), visual and structural protection from predators (Strauß et al., 2008, Arthur et al., 2004), and mitigation of adverse microclimate (Murcia, 1995, Didham and Lawton, 1999, Young and Mitchell, 1994). Whilst studies examining the impacts to fauna where weeds replace native vegetation are limited, they indicate that the process can have an important and proximate effect on faunal responses by altering ecosystem composition (e.g. Dutra et al., 2011, Gosper and Vivian-Smith, 2006). Effects may be greatest where changes to key influences on small mammal ecology such as food resources and predation risk occur (Batzli et al., 1999).

Weeds are a common feature of the urban edge (Alston and Richardson, 2006) as they can be advantaged by increased light (Thorp and Wilson, 1998 onwards), soil moisture and nutrients (especially nitrogen) (Bidwell et al., 2006). These conditions are a product of physical removal or disturbance of native vegetation (Hobbs and Yates, 2003), urban water runoff or air pollution (National Trust of Australia, 1999, Smith and Smith, 2010, Alston and Richardson, 2006). Propagules are spread into the bushland close to urban centres by the dumping of garden clippings or escape of garden plants (Timmins et al., 2010, Raloff, 2003, Smith and Smith, 2010), and by vectors such as humans, domestic pets and vehicles (Von Der Lippe and Kowarik, 2007), all of which are more common at the edge than the interior of bushland.

There are now more introduced plant species in Australia (approximately 27,500) than there are native (approximately 24,000) (Thuilier, 2012), and according to the ‘tens rule’ (Williamson and Fitter, 1996), 10% of these alien species are likely to appear in the wild and 10% of those are likely to become established. In New South Wales (NSW) only 9% of native

CHAPTER 1 Thesis introduction

vegetation is now unaffected by disturbance and weeds (Australian State of the Environment Committee, 2012). Clearly it is important to understand how weeds change the landscape and the impacts such changes have on fauna. There has been a recent focus on this topic (e.g. Dutra et al., 2011, Virkki et al., 2012, Mattos and Orrock, 2010) which indicates that weed structure can advantage small fauna but there is scope for further exploration of how ecosystem processes are altered when weeds replace native plants (Johnson, 2007, Grice, 2004, Vidler et al., 2004, Coutts-smith and Downey, 2006a). For example, weeds may affect the fine scale habitat use and movement patterns of small fauna (Hitchen et al., 2011).

SPECIES INTERACTIONS AT THE URBAN EDGE Examination of species interactions is important to understand how ecological processes are shaping species responses. Lantana and black rats are two invasive species that may cause a negative response by bush rats at the urban edge. The black rat and the bush rat are sympatric species with similar diet, body form, size, and habit (Stokes et al., 2009a, Watts and Aslin, 1981). They are both small nocturnal rodents, opportunistic omnivores (both are mycophagous to a degree (Vernes and McGrath, 2009, Vernes and Dunn, 2009)), and r- strategists inhabiting native forests along the east coast of Australia. This similarity is reflected in competitive outcomes, where residency tips the balance in predicting success (Stokes et al., 2012, Stokes et al., 2009a, Stokes et al., 2009b). This may be an important process at the urban edge where the black rat may be advantaged. Inter-specific competition is a fundamental driver of species composition which has been poorly studied in the urban context (Shochat et al., 2010). Weeds may also advantage black rats, through the process of invasion meltdown. I will examine whether this lowers habitat quality for the bush rats.

THESIS AIMS AND STRUCTURE In this study I aim to use a mechanistic approach to explore bush rat responses to the urban edge with a focus on the impact of invasive species, thereby firstly filling gaps in our understanding of invasive species impact at Australian urban edges, secondly supporting current theoretical frameworks of edge effects and invasion biology and thirdly assisting predictions of the outcomes of reintroduction programs. In this study I will examine the inter relationships between an invasive plant species, an invasive animal species and a

CHAPTER 1 Thesis introduction

native animal species at the urban edge to test which mechanisms are shaping their responses.

In Chapter 2 I used capture-mark-recapture and spool and line techniques to examine habitat use by bush rats in edge, core weedy and core habitats. These macrohabitat types were used in order tease apart the effect of weeds from other edge effects. I aimed to determine whether edges are poor quality habitat for bush rats and if so whether weeds (lantana) are a causative factor. The study was at two scales; macrohabitat, where I compared bush rat distribution and demography, and microhabitat where I examined their fine scale use of habitat, including vegetation architecture as well as more traditional measures of habitat structure. Results were discussed in the context of the matrix/ecotone model (Lidicker, 1999) and current edge theory. Results indicated that urban edges are poor quality habitat for bush rats and that lantana is an important predictor of bush rat habitat use at two scales. This understanding was used as a basis to explore causative mechanisms for the reduction in bush rat habitat quality at the urban edge in following chapters.

In Chapter 3 I tested the hypothesis that the mechanism driving bush rat use of lantana at two scales was mediation of perceived predation risk. I used a Giving Up Density (GUD) experiment to compare bush rat responses under lantana, native vegetation (represented by bracken Pteridium esculentum) and clear ground, with the two vegetation types of the same density. The implications of the results and several challenges encountered in the implementation of the technique were discussed.

In Chapter 4 I investigate whether black rats are more abundant at the urban edge due to mechanisms such as ‘spill over’ from the urban matrix, and whether their presence alters habitat quality for bush rats. Using transects of video cameras at the same macrohabitat types used in Chapter 2 I examined whether partitioning in time, space or use of habitat features occurred. Results were discussed in the context of recent invasion and competition hypotheses.

In Chapter 5 I draw together the results of each chapter and look at where it supports existing frameworks of edge and invasion ecology theories, and where it raises further questions or highlights areas of poor understanding.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

ABSTRACT As urban development continues to expand in Australia, bushland is increasingly impacted by urban edge effects. The consequences for native mammals and the mechanisms driving their responses are relatively poorly understood. In this chapter I examine the effect of weedy urban edges on the native bush rat Rattus fuscipes at both a landscape and fine scale using trapping and spool and line techniques. I begin to tease apart the mechanisms causing edge effects in the context of current edge and fauna distribution theories, with a focus on the weed lantana Lantana camara. Bush rats were shown to avoid the urban matrix at a fine scale. Comparison of abundance indices indicated that the urban edge was poorer quality than core or weedy core habitats, with bush rat response to the edge fitting an ecotone rather than matrix model (Liddicker, 1999). A demographic snapshot of the population across different habitat types indicated that an ideal free distribution pattern was being followed so results were not attributable to despotic influences. At edge habitats a positive correlation between bush rat abundance indices and lantana density may indicate that lantana functions as an emergent characteristic of urban edges by improving habitat quality. This hypothesis was supported by the fact that the rats travelled within the weed thickets more than in native vegetation due to a combination of plant density and architecture. Bush rat preference for a path within lantana over rode preference for other habitat features known to be important to bush rats such as the structural complexity represented by small logs and leaf litter depth. This multi-scaled examination revealed how a weed species alters habitat quality and use by native fauna. Studying the response of bush rat populations in this context provided insight into what factors might be important in causing negative small mammal population responses at the urban edge.

INTRODUCTION This chapter examines bush rat sensitivity to urban edges and begins to tease apart the effect of lantana Lantana camara on bush rat Rattus fuscipes habitat use from other edge effects. Bush rat habitat use was studied at two scales to gain greater insights into mechanisms affecting populations in weedy edge habitats. Fauna response to habitat is scale dependent, and a multi-scaled approach is more likely to reveal habitat associations

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

(Haythornthwaite and Dickman, 2006, Holland et al., 2004, Leiner et al., 2010). At the macro scale I compared differences in bush rat populations and at the micro scale I examined the use of habitat elements in weedy edge, weedy core and core areas. In this study urban edges are comprised of two types of adjacent ecosystems: Australian native bushland and the urban matrix, and also the gradient between them which is defined here as the ‘ecotone’. The ecotone was visually distinct as a mix of bushland, areas of open disturbed ground such as walking paths, and weedy patches. The urban matrix was inhabited by >10 people per ha., with services such as power, water, drainage (MacGregor-Fors, 2011) and roads, with associated gardens and small patches of remnant bushland. Edge effects specific to and obvious within urban/bushland ecotones in the greater Sydney area include increased rubbish, anthropogenic disturbance (in the form of increased noise, movement and light), increased presence of exotic vertebrates, such as domestic cats Felis catus and dogs Canis lupus familiaris (likely to be perceived by the bush rats as predators) as well as black rats, Rattus rattus, (potential competitors of bush rats) and weed growth (Smith and Smith, 2010, Old and Wolfenden, 2010, Weerakoon, 2011).

Bush rat use of the urban matrix In order to understand a species’ response to a habitat edge in the context of current model frameworks, the species’ response to the adjacent ecosystems must be defined (Brady et al., 2011, Lidicker, 1999, Ries and Sisk, 2004). Bush rats are known to inhabit core habitat throughout the study area (NSW National Parks and Wildlife Service, 2000), but their response to the urban matrix is unknown. Garden et al. (2006) define three responses of wildlife to anthropogenic development; (a) matrix occupying: species with the ability to occupy and persist in the urban matrix; (b) matrix sensitive: species which do not occupy or persist in the urban matrix, but which can persist in remnant bushland within the urban landscape; and (c) urban sensitive: species that do not persist in the urban landscape. As Bush rats can persist in some bushland remnants within the urban landscape (NSW National Parks and Wildlife Service, 2000, Garden et al., 2007, van der Ree and McCarthy, 2005) it is likely that they do not fit the urban sensitive definition, however, they may be matrix sensitive. Only one study has explicitly studied bush rat propensity to enter the urban matrix, although in this case sample size was too small for formal analysis (n<10). No bush rats were found in the urban matrix, though they were present in core habitat (Brady et al.,

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

2011). It is likely that bush rats avoid the urban environment as they generally avoid anthropogenic features such as roads and pastures (Barnett et al., 1978, Goosem, 2000) and are commonly reported not persisting in urban areas (Mosman Council, 2012, Rockdale City Council, 2012, RATSAK, 2010, Vet HQ, 2011, Seebeck and Menkhorst, 2000, Australian Museum, 2009-2012). This is despite the bush rats’ biological and ecological similarity with the black rat which thrives in the urban environment utilising the abundant resources of food and shelter it provides (Garden et al., 2006, Beckmann and Berger, 2003, NSW National Parks and Wildlife Service, 2000), which will be discussed further in Chapter 4.

Bush rat edge response The majority of studies indicate that bush rats are negatively affected by anthropogenic edge effects, however, the edge areas studied so far have been created by plantation, pasture (Bentley, 2008), roads (Barnett et al., 1978, Ryan, 1999) and campsites (Stokes et al., 2009a), rather than the urban matrix, with one exception (Brady et al., 2011). This study found fewer rats in urban edges (n = 2) than core habitats (n = 6) though sample sizes were too small for formal analysis. These results suggest that bush rats are likely to be negatively affected by urban edges.

The urban/bushland interface being studied here is likely to be composed of emergent qualities as several of the visible edge effects were not visibly characteristic of either adjoining ecosystem (e.g. weeds), or were different in character (e.g. piles of garden or building rubbish as opposed to litter) also the effects are likely to be interactive and possibly synergistic (Murcia, 1995). This is suggestive of an ecotonal rather than a matrix effect (Lidicker, 1999), so it is likely that the bush rat’s edge response fits the ecotone model.

Assessing species response to habitat quality A species response to edges is a response to issues of habitat quality. Population density used alone may be a misleading indicator of habitat quality (Van Horne, 1983) because it may be partly or wholly a response to source/sink dynamics (Mosser et al., 2009), the species’ distribution pattern (Fretwell and Lucas, 1969), or other social constructs (Skagen and Yackel Adams, 2011). For example, although the population density of yellow-legged gulls Larus michahellis was similar between two distinct and neighboring habitat patches, indicators of breeding and survival were lower in one patch, indicating poorer quality there (Oro, 2008). As long as population density indicators are used in combination with other

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales measures (such as population fitness, structure or dynamics), patterns of (inferred) habitat quality can be mapped on a spatial or temporal scale, revealing the mechanisms and processes driving population changes (Mosser et al., 2009, Skagen and Yackel Adams, 2011, Beckmann and Berger, 2003).

Two main theories describe how animals distribute themselves in habitats of differing quality; namely the ideal free and ideal despotic distributions (Fretwell and Lucas, 1969, Fretwell, 1972). Under an ideal free distribution, all individuals first choose the optimal habitat until population pressure depletes its value, then habitats of lower quality are occupied. This theory assumes no restrictions on movement, such as dispersal barriers or territoriality, and that individuals are able to discern differences in habitat quality. It predicts that population density is positively correlated with habitat quality, and that all individuals have equal fitness. Under the ideal despotic distribution, some form of territoriality or dominance disconnects the correlation; good quality patches are defended by dominant animals and resources are underutilized; excluded individuals are distributed within sub optimal patches. Fitness differs between the dominant individuals living in optimum habitat and those forced in to inferior patches. Population density may or may not vary with the difference in quality between the patches and their carrying capacity (Van Horne, 1983) e.g. a good quality patch may support two very fit dominant individuals, a poorer quality patch may still support two animals, but they will be of lesser fitness. In order to accurately interpret the meaning of different abundance patterns of a species across habitat types it is important to consider which distribution pattern is in operation.

Assuming that habitat quality is lower at edge sites, it is unknown whether bush rats are more likely to follow an ideal free or ideal despotic distribution, as it depends largely on the extent of territoriality exhibited by the population, and many aspects of this species’ sociobiology still remain poorly understood (Breed and Ford, 2007). If a despotic distribution is followed, the pattern may be most evident within the female demographic, and/or between age class (adult and juvenile) distribution. This is because female territoriality during breeding season has been inferred from partitioning of home ranges, and also from captive behavioral studies (Robinson, 1987, Woodside, 1983). In this case, females of greater fitness may dominate the higher quality habitat of core areas forcing less fit females into the poorer quality edge sites. Likewise, juvenile bush rats, especially males, may be more

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales abundant at edge sites if they are, like several other Rattus species (Berdoy and Drickamer, 2007), subordinate to adults and consequently pushed into poorer habitat. However, Wood (1971) did not find any separation of female home ranges during bush rat breeding season, and some degree of sociality during winter has been recorded including nest sharing (Woodside, 1983, Sanecki et al., 2006b). This indicates that like other rodent populations (Krebs et al., 2007) bush rat social structure may be quite plastic, and a female based despotic breeding structure may only occur under some conditions.

Is lantana a proximate cause of edge effects on bush rats? If the urban edge is poorer habitat for bush rats, which habitat features reduce its quality? Lantana is an obvious characteristic of the urban edge locally but its effect on habitat quality for small mammals such as the bush rat is unknown. It’s toxicity and impact on reproduction (Chapter 1 ‘Introduction’) may have a negative impact on rodent populations and it may impede movement due to its dense strong stem architecture, and reduce food resources directly by replacing palatable plant species or indirectly by reducing biodiversity (Coutts- smith and Downey, 2006a). Additionally there is some evidence of an association between lantana infestation and a loss of mychorrhizal fungi (Fernandes et al 2001), which is an important food source for bush rats. Several studies (Martin and Banks, 2010 unpublished, Umetsu and Pardini, 2007, Jellinek et al., 2004) have found that endemic small fauna tends to be distributed within native, rather than non native vegetation.

Nevertheless there is increasing evidence that lantana may provide an important function for small mammals and birds. In some areas survival of brush turkey Alectura lathami chicks is at least partially dependent on lantana thickets, which appear to provide decreased predation risk, particularly from cats and birds of prey (Goth and Vogel, 2002). Long-nosed bandicoots Perameles nasuta have been reported to nest under lantana (Chambers and Dickman, 2002) and long-nosed potoroos Potorous tridactylus may also utilise this resource (Burgman and Lindenmayer, 1998). A recent overseas study found that perceived predation risk for endemic rodents was reduced where a shrubby weed invades an ecosystem (Mattos and Orrock, 2010). Martin (2010) also found that weedy patches (including lantana) harboured greater numbers of black rats than patches regenerated with endemic vegetation. These observations suggest that the structure of vegetation can be more

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales important for small mammals than the floristic composition (MacArthur et al., 1966, Barnett et al., 1978, Garden et al., 2007).

Plant architecture There is an abundance of evidence that bush rats prefer ‘structurally complex habitat’, which is usually defined as a prevalence of habitat features such as logs, stems, trees and dense vegetation cover (Warneke, 1971, Holland and Bennett, 2010, Barnett et al., 1978, Maitz and Dickman, 2001, Braithwaite et al., 1978, Holland and Bennett, 2007, Robinson, 1987). This term is distinct from ‘vegetation structure’ which is a general term describing vegetation properties other than floristic composition (Barkman, 1979, van der Maarel, 2005), and which will be used here to refer collectively to vegetation density, cover and architecture. The term ‘vegetation density’ refers to the extent to which the plant biomass visually obscures an object or the ground. The term ‘vegetation cover’ refers to the area occupied by the plant as a percentage of the surface area of the sample plot (van der Maarel, 2005). ‘Vegetation architecture’ is a component of habitat structure, however, there is little understanding of how it relates to habitat use by small mammals. While its definition can vary, vegetation architecture usually refers to, and will be used in this study as; the three dimensional physical arrangement of plants and their parts (Godin, 2000, Barthélémy and Caraglio, 2007, Reinhardt and Kuhlemeier, 2002). Leaf or stem connectivity (Marquis and Lill, 2010, Hannunen, 2002), stem orientation (Rudgers and Whitney, 2006) or extent of branching (Reinhardt and Kuhlemeier, 2002, RiihimÄKi et al., 2006) are among expressions of vegetation architecture which are well known to influence invertebrate ecology (e.g. Hannunen, 2002, Lawton, 1983, Haysom and Coulson, 1998, Pearson, 2009, Langellotto and Denno, 2004). Whether such aspects of vegetation architecture influence mammal habitat use is not well understood; it may be important since stem orientation can predict black rat habitat use (Cox et al., 2000) and foliage height diversity (FHD) can influence patterns of small mammal distribution (Bennett, 1993, August, 1983, Braithwaite et al., 1989, Els and Kerley, 1996). FHD was defined in the 1960s as a measure of the evenness of vegetation density in the vertical plane (MacArthur and MacArthur, 1961) and following Sutherland et al. (2004) I include FHD under the definition of plant architecture since it is an expression of vegetation arrangement. I will be exploring whether bush rats respond to differences in the evenness of vegetation strata in the vertical plane. Lantana

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales appears to have a complex stem architecture and dense growth that may fulfill the bush rats requirements of structural habitat complexity, thus acting as a structural proxy for native vegetation.

Aims Using trapping and spool-and-line techniques I aim to determine bush rat sensitivity to the urban edge in the context of the ecotone/matrix model (Lidicker, 1999). I will examine population differences in edge and non-edge habitat in the context of distribution pattern theory (Fretwell and Lucas, 1969, Fretwell, 1972) using indices of population structure, distribution and fitness, to assess and compare habitat quality. Population structure is defined as the relative abundance of demographic classes such as age and gender. Distribution is defined as how the population is arranged over the different macrohabitat types. Population fitness is assessed from the average body weight and reproductive output, measured by the number of juveniles and adults in breeding condition. Higher body weights and greater reproductive output are indicative of higher population fitness. If bush rat response to the edge is negative, I aim to determine whether the weed lantana functions as a structural proxy for native vegetation, in which case it is unlikely to be a driver of negative bush rat edge responses. A supplementary aim is to explore bush rat response to vegetation architecture in lantana and native vegetation.

Hypotheses and Predictions

Bush Rat use of the urban matrix Hypothesis 1: Bush rats are urban sensitive

Prediction 1: Rats trapped in bushland at the urban edge will not cross in to the urban matrix.

Habitat quality Hypothesis 2: The bush rat response to the urban edge fits the ecotone model (Lidicker, 1999) with the ecotone being poorer quality habitat for bush rats than the core due to edge effects.

Hypothesis 3: The distribution patterns will reflect a female and/or age class based despotic structure based on current understanding of bush rat social structure.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Prediction 2 and 3: Fitness (measured by demographic indicators) will be better in core habitats compared to edges, and despotism will be reflected by a greater proportion of females and/or adults in the optimal core patches. Population abundance indices may or may not be different.

Is lantana a proximate edge effect on bush bats? Hypothesis 4: Lantana provides a structural alternative to native vegetation for bush rats, and does not function as an edge effect.

Prediction 4: At macro scale, populations of bush rats in core weedy and core areas are not different. At fine scale there is no difference in the paths of random and animal spools with respect to scores of abundance (% cover or density) of lantana, since it is a structural proxy for the extant native vegetation, and structure is more important than floristic composition.

Vegetation architecture Hypothesis 5: vegetation architecture is a predictor of bush rat habitat use.

Prediction 5: At fine scale certain artificial categories, developed to describe frequently occurring vegetation architecture patterns based on the evenness of the horizontal strata, will be featured along rats path either more or less than expected given availability.

METHODS

Study Area Study sites were within bushland in National Parks, State Forests and Council Reserves on the Central Coast and Northern Sydney regions of New South Wales (NSW) Australia, in areas within the core distribution range of bush rats (Lunney 2008) (Figure 2.1). All sites were within a large area of contiguous bushland, (> 10ha) to control for the effects of habitat fragmentation, as fauna populations in fragmented habitats may have different density and demography to those in un-fragmented habitats (Holland and Bennett, 2010, Cox et al., 2003, Dunstan and Fox, 1996, Collinge, 2009).

Macrohabitat types Three macrohabitat types (edge, core weedy, and core; n=18, six replicate sites) were sampled using capture-mark-recapture, to determine the minimum number of bush rats known to be alive. Sampling was undertaken between 02/09/2009 and 30/02/2010. Edge macrohabitats were sampled with transects <50m from the urban boundary on the

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales bushland side and were characterised by lantana and anthropogenic disturbance. The urban boundary was defined as the visible demarcation between bushland and the urban environment, specifically the point at which anthropogenic development or clearing covered ≥ 80% of the ground. The demarcation was usually clear, e.g. lawn, backyard fence, road or building abutting the bush land. Core weedy macrohabitats were situated in the interior of the National Park or Reserve, and were characterised by the presence of lantana, surrounded by native bushland. Core sites were characterised by ‘pristine’ bushland (undisturbed and non weedy). Weedy core and core sites were as far away from an edge (urban or agricultural) as possible, given logistical constraints. Minimum distance to edge for either was 200m, mean core site distance to edge was 466m (n = 18, SD = 405), mean core weedy site distance to edge was 413m (n = 18, SD = 206). Most studies on edge effects show they penetrate less than 150m (Laurance et al., 2007, Murcia, 1995, Matlack, 1993). In Sydney bushland, measurable changes to vegetation due to the urban edge have predominated in the first 30 - 60 m (Rose and Fairweather, 1997, Smith and Smith, 2010). As aspect can influence the strength of edge effects (Murcia, 1995) only edges facing North- East or South-East sectors were selected.

Attempts were made to locate non-weedy edges in order to create an orthogonal two factor experiment design (weeds and edges), however, most non-weedy edges were new (only months or weeks old) or under construction. New edges are impacted by different effects to old edges; they are characterised by strong primary and microclimatic effects (Harper et al., 2005, Matlack, 1993, Oosterhoorn and Kappelle, 2000). They were not commensurable with weedy edges, which are older; weed growth is an example of secondary edge effects (Harper et al., 2005). In addition there was either increased human activity at non weedy edge sites due to construction, or decreased activity due to the hiatus between completion of the housing estate, and occupation.

The sites were within one of three forest types: dry open forest, open woodland and temperate rainforest. These were represented evenly across edge, weedy core and core sites to control for effects of patch quality. Sites were considered independent as all were > 800m apart by land, well above the 100-200m2 home range of bush rats (Maitz and Dickman, 2001, Wood, 1971) and also outside of typical dispersal distance (Macqueen et al.,

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

2008, Warneke, 1971). During trapping no rats were recorded crossing between sites, supporting the assumption of site independence.

One representative of each macrohabitat type was randomly chosen and grouped into a trio which was sampled at the same time, and which constituted one replicate where environmental conditions such as moon phase and weather were standardised. One exception to the random nature of grouping was a trio of sites with difficult access which were sampled together for logistical reasons (Figure 2.1 trio designated ‘E’).

Field procedure, trapping. Aluminum folding Elliott traps (33 x 10 x 8cm) were used to sample bush rat populations. Non absorbent cotton wool was placed at the back of the trap for bedding and insulation, and then a bait ball made from a mix of peanut butter, oats and honey was placed on top of the bedding. During dry warm weather conditions a piece of rockmelon approximately 2cm2 was also added to provide water and was often eaten in preference to the bait ball. A plastic bag enclosed the back half of the trap to weather proof it. The trap was always placed with the open end slightly downhill, to prevent the plastic bag filling with rainwater.

Sites were sampled with a transect line comprising 15 traps set 12m apart set for a total of 6 consecutive nights. Pre-baiting was undertaken for the first 3 nights to reduce or negate neo-phobia which can reduce trap success (Tasker and Dickman, 2001). During pre baiting, baited traps with bedding were left upside down with the door open which deactivated the trigger mechanism. All traps were nonetheless checked daily during pre-baiting in case of accidental trapping (which was rare). After the third night of pre-baiting, the traps were re baited, activated, and replaced where soiled by urine or faeces. Capture-mark-recapture procedures were undertaken over the next 3 nights. Processing of traps began at dawn. Non target animals were identified to species level prior to release. All captured bush rats were transferred to a calico bag, then weighed, sex recorded, assessed for reproductive status (see below) and marked by clipping a small amount of fur on the left flank so that over following days recaptured animals could be distinguished from new ones. They were then released. Traps that had captured an animal were replaced with clean ones containing fresh bait and bedding, to reduce the chance of trap success being influenced by a response to odor (Daly et al., 1980, Boonstra and Krebs, 1976). Dirty traps were later cleaned with warm

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales water and detergent, rinsed and dried before re use. Traps were left open during the day, and were checked in the evening to release any diurnal captures (a rare occurrence). All animal processing was undertaken 5m away from the actual trap sites to avoid disturbance of the vegetation structure.

Assessing Population Structure Animal sex was determined by observing the presence of teats, an open vagina or the puckering of tissue signifying an imperforate vagina indicating that it was a female, or the presence of testes or epidydimes indicating it was a male. For non reproductive animals such as juvenile sex was determined using comparative ano-genital distance (Watts and Aslin, 1981). Juveniles were defined as animals ≤ 60g and as being pre dispersal, following Warneke (1971), Wood (1971) and Robinson (1987). Adults were defined as animals > 100g: all animals of this weight range in this study showed reproductive development, in that males had testes descended or palpable, females had an open vagina, and teats visible or palpable. Sub adults were defined as animals 61-100g.

The reproductive development of individuals was ranked as follows, adapted from Aplin, Brown et al (2003): Females were classed as 1: teats not visible/palpable or very small and hidden by fur, 2: teats easily visible, but with fur at the base, or 3: actively reproductive if they had raised swollen teats and/or teats furless at the base with or without discolouration. Males were classed as 1: no testes visible or palpable and epididymal pouch indistinct, 2: testes partially descended and epididymal pouch indistinct or 3: testes fully descended and epididymal pouch full and distinct.

Microhabitat use To determine whether microhabitat features affect trap success, 15 microhabitat variables known from previous studies to influence patterns of habitat use by bush rats (Stokes et al., 2009b, Maitz and Dickman, 2001, Cox et al., 2000) were scored for a 1m radius around each trap (Table 1). At edge and core weedy sites, distance to the edge of the lantana thicket was included as a variable, as well as distance to the urban edge at edge sites. The sample unit was photographed so that consistency of categorisation of variables over time could be maintained. Method and units of measurement for each variable is delineated in Table 1.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

At edge and core weedy sites, traps were placed under lantana where it was within 10m of the transect line; otherwise they were placed under the extant vegetation on the line. Thus lantana, native vegetation and combinations of both were sampled at all transects. Vegetation density was measured following Cox (2000) using a wooden cover-board 50 x 50cm scored with a grid 10cm2 (Table 2.1 and Figure 2.2). Due to the great variation of vegetation structures across and within sites (e.g. difficulties comparing grasses to the complex stem structure of lantana), and the fact that the cover-board could not be used to obtain an aerial view, an additional measure of vegetation density was developed (the ’pictorial reference’ method (Table 2.1 and Figure. 2.3)) to incorporate a different perspective. This method was felt to be a more ‘genuine’ evaluation of predator/prey perspectives although it was more subjective than the cover-board method. It also had an added advantage of being quick to process. To compare the differences in vegetation architecture, categories of were developed (Table 2.2 and Figure 2.4) based on the FHD concept (MacArthur et al., 1962), according to the evenness of the vegetation strata in the vertical plane, i.e. whether there was a gap in the horizontal vegetation layers or not, and the relative height of the gap. These categories were an artificial construct, but were required as time did not permit a more rigorous measurement due to the large number of sample units (n = 929). The categories were common across all macrohabitats. A value of 1 was assigned if the category was represented at a sample unit and 0 if it was absent.

Spool and line tracking (Miles, 1976) was used to obtain a fine grained assessment of bush rats’ use of microhabitat during periods of free movement and to determine whether species respond to vegetation structure and architecture or floristic composition. This method tracks the animals’ pathway by use of a fine thread that unravels from the inside of a nylon bobbin attached to the animal’s back. The thread is fine and light so tends to catch on any rough surface or protuberance in the environment such at vegetation, leaf litter, or rocks. This secures the line so that changes of direction are preserved. This is a commonly used method for mapping habitat use at a fine spatial scale (Leiner et al., 2010, Cox et al., 2000, Bennett, 1993, Boonstra and Craine, 1986), and so was an appropriate choice to sample Bush Rats’ use of microhabitat components for the purposes of this study.

The bobbins used for spooling were 32mm x 10mm and comprised approximately 160 metres of thread. They were wrapped in a thin layer of masking tape so that the glue used

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales for application was not directly applied to the thread itself. This was to increase the potential length of the spool and minimise injury to or entanglement of the animal when the spool detached. This also allowed the unit to be tapered towards the anterior of the animal to create a smooth gradient over the edge of the spool, to reduce the chance of it catching on vegetation (Figure 2.5). All units were less than 3g, and less than 5% of the animal’s body weight.

Following the third night of trapping, two individuals >85g were randomly chosen for spool tracking at each site. The cocoon bobbin package was glued (using super glue, cyanoacrylate) to fur on the animal’s rump, in line with the tail, where it was least likely to impede their movement through the dense lateral stem structure of lantana bushes. Long outer guard hairs on the rump were trimmed so that the spool attached more securely.

Once the spool unit was fitted, the loose end of the line was secured to a solid structure such as a tree trunk and the animal was released at the point of capture. In order to minimise flight response (e.g. see Cox, Dickman et al (2000); Tulloch and Dickman (2006)), rats were left in the calico bag with the head covered, but the rump and spool unit uncovered to prevent entanglement. The bag was placed under a bush, dense foliage or hollow log to reduce real and perceived predation risk, the latter of which could exacerbate flight response. The researcher then left the patch. Where possible the bag was observed from a distant vantage point, and in all cases the rodent was seen to leave the bag slowly, often looking around before moving off relatively slowly and deliberately. Bags were checked after 1 hour to confirm exit was successful. The following day the thread was followed, and microhabitat variables as described above were measured and recorded at every 5m, for a 1m radius from the spool line, until the line was lost or ended. Where burrows or refuges occurred along the line, microhabitat variables were also scored, in the same manner. Refuges were defined as any point where the spool package was found discarded. It was assumed that the package was removed by the carrier and thus the microhabitat at that point was perceived as suitable as a refuge to spend time and energy on removal. The spool line attached to the rat was termed the ‘animal’ line. A control spool line was generated for comparison by taking a random compass bearing from the start of the treatment line; this was called the ‘random’ line. Using a compass, a tape measure was laid in a straight line in the designated direction and variables were measured or scored

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales using the same method as described for the animal line. Measurements or scores from the random line were defined as representing available habitat, as they sampled a random snapshot of the microhabitat variables at that site. At edge sites no rats entered the urban matrix, thus this was taken to represent non viable habitat, and so if the random line met the edge it was angled back into the ecotone following the next random bearing on the list.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Figure 2.1 Site locations on the NSW Central Coast. Red = edge, orange = core weedy, green = control macrohabitats. Sites were grouped into replicates as indicated by the lettering.

Figure 2.2: cover board. Photo by W. Gleen

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Microhabitat Unit of Definition Method of measurement variable measurement Leaf Litter Depth centimeters Distance between the ground and the top Ruler used at 4 points within the sample unit, and the average value recorded. of the leaf litter layer. Leaf Litter Cover % The % of the ground covered by leaf litter Estimate by eye Ground stems Count Stick or stems < 15cm diameter, < 60cm Count. Scored as: converted to from the ground. 1: <10, Score. 2: 11-40, 3: >40 Logs/boulders/ Count 15-30cm diameter, counted if any part of Count. solid objects – the object was within the radius. small Logs/boulders/ Count >30cm diameter, counted if any part of Count. solid objects - the object was within the radius. big Trees small Count Trunk 15-30cm diameter, counted if any Count. part of the object was within the radius. Count Trunk>30cm diameter, counted if any part Count. of the object was within the radius.

Trees big

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

% Cover-board method: % of cover-board Method adapted from Cox (2000): A wooden board 50cm x 50cm, marked with a obscured by vegetation. 10cm grid was placed vertically on the ground at the trap or spool line point. Existing vegetation was used to support the board, or it was held by an assistant. It was viewed from a point on the circumference of the sample unit, at kneeling height (80cm). The number of squares covered by vegetation was used to estimate the % of the cover-board that was obscured. Measurements were taken from 4 points on the circumference and the average value was recorded as the measure of density. Understorey Score 0-3 Pictorial reference method: categorisation 3 categories were defined and a pictorial reference created, adapted from Walker Density based on pictorial reference adapted from and Hopkins (1983). See Figure 2.3. The vegetation was scored by comparison with Walker (1983) the reference, from eye level (1.65m height), from 4 points on the circumference, and from directly above. This was done in the context of how well a bush rat would be obscured from view, following Glen, Sutherland et al (2010). The average value was taken and rounded to fit the closest category 0=no undergrowth or cover 1= low density (rat easily visible) 2= medium density (rat visible with difficulty) 3= complete density (rat completely obscured). Understorey Centimeters Distance between the ground and the top A tape measure was used and where different vegetation types were different height of the understorey, defined as the shrub, heights, the average measurement was taken. grass or ‘fallen tree’ (see Table 2) layer. Lantana Cover % % The proportion (%) of the vegetation Estimate, by eye. within the radius that was Lantana.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Lantana Density, Score 0-3 Pictorial reference method. As above under ‘understorey density’, pictorial reference method.

Other weeds % % The proportion (%) of the vegetation Estimate, by eye. within the radius that was comprised of weeds other than Lantana. Other weeds Score 0-3 As above under ‘understorey density’, As above under ‘understorey density’, pictorial reference method. Density pictorial reference method. Canopy Density % Extent of shade and cover provided by the Estimate was made using pictorial reference developed by Walker (1983) canopy layer. Penetration of Metres (+ve A >2m2 extent of Lantana of medium or Cloth tape measure, or estimate, if the area was impassable weed thicket value if within dense growth. the thicket, - ve value if outside it.) Shortest Metres (+ve Urban edge = Where there is a visible Cloth tape measure, or estimate if impassable. distance to value if on demarcation between bushland and urban edge bushland urban environments; the point at which side, -ve anthropogenic development or clearing value if on covers ≥ 80% of the ground. suburban side.)

Table 2.1 Microhabitat variables. Variables were measured within a 1 metre radius surrounding the trap or point on the spool line.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Category 0 No vegetation

Category 1 Light

Category 2 Medium

Category 3 Dense

Figure 2.3 Vegetation density categories, pictorial reference. Photos by W. Gleen

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Figure 2.4 Vegetation architecture categories. Photos by W. Gleen.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Architecture Architecture definition name cavity: tall Undergrowth forming a tent-like structure above the ground with the foliage or stem layer beginning >30cm above ground. A cover board reading of less than 20% within the cavity. cavity: low A visually impenetrable raft of stems and/or leaf mulch and grasses 5 - 30cm off the ground, (e.g. horizontal stem layer, matted bark litter on top of stem or other structure), with a cavity underneath to ground level. Cover-board reading less than 20% within the cavity. Note often it was not possible to place the cover-board without destroying the vegetation structure. Where this was undesirable the reading was estimated by eye. homogeneous Homogeneous undergrowth with no discrete cavity within the structure fallen tree Horizontal sticks or stems, as in a fallen shrub/tree > 60cm above the ground. No cavity. Table 2.2 Vegetation architecture categories

Figure 2.2 Spool line fitted to bush rat. Photo by W. Gleen.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

DATA ANALYSIS Data from one core weedy site were excluded from all population level analyses as data were confounded by the abundance of a native competitor, the swamp rat Rattus lutreolus, which are known to socially dominate bush rats thus may have biased results. For all parametric analyses in this section unless otherwise specified data met required assumptions; upon visual inspection of the data, no large departure from normality was found, and variance was sufficiently homogeneous (Quinn and Keough, 2002).

Minimum number known alive and demography The minimum number known alive (MNKA: Krebs, 1999)was used as an index of abundance and was calculated for each replicate (site) by adding the number of captures on the first night of trapping to the numbers of new animals (identified by lack of markings) captured on the subsequent two nights. Differences in the MNKA and composition of age structure between macrohabitats were investigated using a two way split plot ANOVA with number of rats as the dependent variable, macrohabitat, age class and site as factors with site a nested (within macrohabitat) random factor. A post hoc Tukey-Kramer HSD was used to identify which terms were significantly different to each other.

Differences in numbers of males and females in each macrohabitat were investigated by a two way split plot ANOVA in JMP version 9 (SAS Institute, 1989-2010), with number of rats as the dependent variable, gender, macrohabitat and site as factors with site a nested (within macrohabitat) random factor and gender crossed with macrohabitat. The frequency of each reproductive score (1-3) was counted for males and females at each site then data pooled across macrohabitat. Male and female data were analysed separately. This generated a 2 by 3 contingency table for females, and a 2 by 2 contingency table for males (male data was binomial as there were no adults with a score of 1). Both sets of data resulted in over 20% of cells having an expected count of <5, so an exact contingency table analysis was undertaken (Kirkman, 1996). Body weight was analysed for males and females separately using a nested ANOVA with site as a random factor nested within macrohabitat. Where results were statistically non significant, power analyses were run using Russ Lenth’s Java applets

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales for Power and Sample size (Lenth, 2006-9) to confirm that the experiments had the power to detect biologically meaningful differences. The detectable contrast was expressed as a percentage difference of the core mean. It was not possible to analyse the results for reproductive status in this way due to the nature of the statistical test used (an exact test).

Correlation of vegetation density measurement methods Data sets were tested for correlation using the Spearman correlation in JMP version 9 (SAS Institute, 1989-2010) to determine whether the novel pictorial reference method of measuring vegetation density was comparable with the more objective and traditional cover-board method, so that results from the former method could be used with confidence.

Does extent of Lantana predict Bush Rat abundance? To examine the relationship between lantana extent and bush rat abundance indices, the mean score for lantana density and % cover (representing two different measures of Lantana extent) were calculated for each edge site from random lines. These were compared with the score for MNKA at each respective site using a scatter plot of Bivariate fit and tested using Pearson correlation in JMP version 9 (SAS Institute, 1989- 2010). The same analyses were repeated for core weedy sites.

Effect of microhabitat variables on trap success The relationships between trap success and microhabitat variables were examined by stepwise regression in JMP version 9 (SAS Institute, 1989-2010). Trap success was defined as the number of bush rat captures in a trap over the three trap nights. Data were adjusted to account for unavailable traps due to occupation by other species, trap closure or bait burglary following Simonetti (1989) as follows

Trap Success = A / 100(TU-NA) Equation 1 where A = number of rat captures, TU = number of trapping units, NA = unavailable traps.

Only data from core weedy and edge sites was used, as core sites lacked lantana. Correlations between variables were also inspected using scatter-plots. The model with the smallest AIC value was chosen. As the vegetation category data was nominal

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales whereas other microhabitat variables consisted of continuous data, it was analyzed separately using the same method. The bivariate association between trap success and distance to the edge of the suburb edge was also tested separately with the Pearson Correlation, as this value was only applicable to edge sites. As the result was not statistically significant, a power analysis using Russ Lenth’s Java applets for Power and Sample size (Lenth, 2006-9) was run to confirm that the experiment had sufficient power to detect a biologically meaningful difference. The result was expressed as a percent difference to the core mean. It was not possible to run a power analysis for the data analysed using stepwise regression as there is no appropriate test for this statistic type.

Habitat use along path in weedy habitats To determine which microhabitat variables were preferentially used at weedy patches by bush rats in comparison with availability during periods of free movement, random and animal data from edge and core weedy sites was compared using mean values of each microhabitat feature for each spool line. Data were treated as unpaired given that the random spool was designed to be an independent measure of the site due to its random setting. Sample sizes of less than 4 (due to lines breaking or being lost in burrows) were excluded to improve the precision of means following visual examination of means of each microhabitat variable (where data was continuous) plotted against the coefficient of variation (Figure 2.7). Variation visibly declined where sample size was greater than 4.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

3000

2500

2000 logs Canopy density 1500 litter depth litter cover veg density

Coefficient Coefficient ofvariation 1000 stems veg ht

500

0 0 5 10 15 20 25 Sample size

Figure 2.7. The effect of sample size on the coefficient of variation. The demarcation of variance where n ≤ 4 is highlighted with a box.

Similarities between the suites of microhabitat variables in each macrohabitat were examined using PRIMER 6 (Clarke and Gorley, 2006) using spool lines as replicates. The variable ‘distance to weed thicket’ was transformed to positive by +1 prior to pre treatment in PRIMER so that the values could be analysed using the same resemblance coefficient (Bray-Curtis).

A multidimensional scaling (MDS) plot was used to represent the similarity between mean microhabitat variables within each macrohabitat based on a Bray-Curtis similarity matrix of microhabitat data that was standardised (due to the different measures that were used e.g. % cover and counts) and square root transformed (to more evenly weight contributions from dominant and rarer variables (Clarke and Gorley, 2006)). Fifty restarts were used. Differences were tested with a three -way

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

PERMANOVA with 9999 permutations using a two-way nested design with factors path (random) nested in macrohabitat (fixed). Site was not included in the model as there was no significant site effect (p > 0.1). SIMPER was used to identify which variables best explained the differences between macrohabitat. Difference in ‘distance to suburb edge’ between random and animal data was examined separately by an independent t-test as this variable was not applicable to core weedy sites. This data was log transformed prior to analysis to conform to required assumptions of normality and homogeneity of variance.

Anthropogenic disturbance Anthropogenic disturbance may affect vegetation structure (Oosterhoorn and Kappelle, 2000), and it is possible that this has occurred at edge and core weedy sites. This could confound results, for example it may render bush rats more likely to use alternative vegetation such as lantana. Considering only non- weedy portions of random spool lines, data points were compared across macrohabitat type to determine whether native vegetation differed. Means were not used as data points from random spool lines were considered to be independent. Where there was no understorey (a score of zero), the sample unit was assigned to the vegetation type that preceded and followed it on the line and it was considered to be patchiness within that vegetation type. If a different vegetation type preceded and followed, the value was excluded. Results were then compared using multivariate techniques. A Bray-Curtis similarity matrix was generated from standardised data which was also square root transformed. Differences were tested with a two -way PERMANOVA with 9999 permutations using a one-way design. Factors were site (random) nested in macrohabitat (fixed).

Habitat use along path within lantana To see if the rats preferred certain features within lantana, which may indicate which aspects of the plant are important to them, animal and random paths within weedy portions of the line only were compared. Data was selected where lantana cover exceeded or equaled 50% (use of a 100% criteria resulted in a small sample size). Measures of lantana extent (density, % cover and distance to the edge of the weed thicket) were excluded, as they were not relevant to this question. Data was filtered

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales as random path only weed species 100% (lantana) and zero for native vegetation. PERMANOVA, SIMPER and representation by MDS plot were undertaken using the same procedures used in ‘Habitat use along path in weedy habitats’; again there were no site effects (p > 0.1) so site was not included in the model.

Differences in microhabitat characteristics between Lantana and native vegetation To see how lantana alters the natural environment at a fine scale, similarities between the suites of microhabitat variables of lantana and native vegetation were examined using PRIMER 6 (Clarke and Gorley, 2006) using data from edge and core weedy sites only. Only variables representing physical properties of the vegetation were included (i.e. ‘canopy cover’ and distance measures were excluded). Means were not used as data points from random spool lines were considered to be independent. Data was filtered as random path only weed species 100% (Lantana) and zero (native vegetation). PERMANOVA, SIMPER and representation by MDS plot were undertaken using the same procedures used in ‘Habitat use along path in weedy habitats’; again there were no site effects (p > 0.1) so site was not included in the model. Factor was vegetation type (fixed).

Habitat use along path in non-weedy habitats To see which microhabitat variables were used by the bush rats significantly more than expected in comparison with availability at core patches, random and animal data from core sites were compared. PERMANOVA and representation by MDS plot were undertaken using the same procedures used in ‘Habitat use along path in weedy habitats’, except in this case factors were path (random) nested within site (random) (as PERMANOVA revealed a non significant result, SIMPER was not used). As the non significant result for path was unexpected, the difference between the random and treatment spool lines was then examined for each variable independently by univariate analysis to clarify whether there were any differences at all. Where data conformed to the assumptions required for a parametric analysis, a paired t- test was used in JMP version 9 (SAS Institute, 1989-2010) to compare the means for Animal and Random spool lines. Where the data did not comply with the required assumptions, a (non-parametric) Wilcoxon test was undertaken using Wilcoxon-Mann- Whitney Test (v1.0.4) in Free Statistics Software (v1.1.23-r7)(Holliday, 2012).

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

RESULTS

Minimum number known alive and demography

There were significant differences in the MNKA between macrohabitat types F (2, 28) = 7.2, P = 0.007, n = 17, Figure 2.8. Tukeys HSD revealed that edge sites had fewer bush rats than core weedy and core sites. There were more adults than juveniles or sub adults captured (F (2, 28) = 24.02, P < 0.0001) but there was no interaction between macrohabitat and age class; F (4, 28 ) = 1.11, P = 0.37.

14 N=17 a a

10 b

MNKA 6

0 Edge Core Weedy Core Macrohabitat

Figure 2.8. Mean MNKA ± SE at each macrohabitat type. Differences in the means are denoted by letters a and b.

There was no difference between gender F (1, 14) = 2.31, P = 0.15, or macrohabitat F (2,

14) = 2.75, P = 0.1, nor any interaction between them; F (2, 14) = 1.31, P = 0.31 (n = 34). A power analysis showed that the experiment had the power to detect a 67.5% difference from the core mean (a biologically meaningful result). Similarly there was no difference in the mean score of reproductive status between macrohabitats for females; P = 0.455 n = 9, or males; P = 0.119 n = 4. There were also no differences in mean body weight between macrohabitat for either males; F (15, 46) = 0.23, P = 0.80 n =

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

16 or females; F (11, 32) = 1.9, P = 0.98, n = 12. A power analysis showed that the experiment had the power to detect a 9.8% and 10.3% difference from the core mean respectively (both biologically meaningful results).

Correlation of vegetation density measurement methods The two methods of measuring vegetation density (cover-board and pictorial reference) were positively correlated; rs = 0.855, p < 0.0001, n = 928. The pictorial reference method was used for all further analyses.

Does the extent of Lantana at the urban edge predict Bush Rat abundance? There was no significant correlation between lantana % cover and MNKA at edge sites; r = 0.55, P = 0.26 n = 6 but there was a significant positive correlation between mean Lantana density and MNKA at edge sites; r = 0.81, P = 0.049, n = 6 Figure 2.9. At core weedy sites there was no significant correlation between lantana % cover and MNKA; r = -0.42, P = 0.49, or lantana density and MNKA; r = -0.43, P = 0.47.

5 R2 = 0.66

MNKA 3 MNKA Linear (MNKA) 2

0 0 0.2 0.4 0.6 0.8 1 1.2 Mean lantana density

Figure 2.9 Correlation between MNKA at sites and lantana density, at edge sites only. N = 6.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Effect of microhabitat variables on trap success The stepwise regression of microhabitat data and trap success (excluding vegetation categories) produced a final model that explained only 15% of the variance with three significant terms. Mean trap success was significantly affected by distance to weed thicket n = 165, F (1, 153) = 5.86, p = 0.02, lantana density; F (1, 153) = 7.44, p = 0.007, and lantana % cover; F (1, 153) =9.58, p = .002. Both measures of lantana density were negatively correlated with extent of ‘other weeds’ (mainly annuals), which was probably due to suppression of these species by lantana. It is unlikely that this confounded the result. The Pearson correlation test revealed there was no significant correlation between distance to the suburb edge and trap success; n = 165, r = 0.20, p = 0.06. A power analysis showed that the experiment had the power to detect a 21% difference from the core mean (a biologically meaningful result). The stepwise regression of vegetation categories and trap success only explained only 2% of the variance. No significant terms were identified from the model.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Standardise Samples by Total Transform: Square root Resemblance: S17 Bray Curtis similarity 2D Stress: 0.23 path C T

Random Animal

Figure 2.10 Multidimensional scaling plot of Bray-Curtis similarity coefficients showing differences in path between animal and random spool lines, based on microhabitat variables, for edge and core weedy sites. Each point represents a spool line N= 40.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Random Animal Variable Av.Abund Av.Abund Av.Diss Diss/SD Contrib. % Cum.% lantana density 1.98 3.16 2.50 1.57 10.04 10.04 litter depth 4.48 4.11 1.92 1.36 7.71 17.75 small logs 1.10 1.05 1.80 1.13 7.22 24.98 cavity: low 0.77 1.48 1.75 1.35 7.04 32.01 lantana % cover 1.29 1.93 1.63 1.37 6.55 38.57 small trees 0.97 0.48 1.47 1.08 5.92 44.48 big logs 0.58 0.92 1.40 1.26 5.64 50.12 canopy cover 2.03 1.37 1.34 1.21 5.39 55.51 other weed 0.59 0.37 1.21 0.75 4.86 60.38 density cavity: tall 0.55 0.53 1.21 0.95 4.86 65.24 fallen tree 0.69 0.47 1.18 1.08 4.75 69.99 understorey 3.88 4.20 1.16 1.18 4.68 74.66 density no. ground stems 3.77 3.85 1.15 1.38 4.62 79.28 homogeneous 2.56 2.20 1.03 1.49 4.14 83.43 big trees 0.53 0.50 0.99 1.08 3.98 87.41 litter % cover 3.03 2.44 0.99 1.40 3.96 91.37

Table 2.3 Results of SIMPER analysis showing dissimilarity in microhabitat variables between random and animal lines at edge and core- weedy sites.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Habitat use along path in weedy habitats The MDS plot showed distinct separation in the microhabitat features of random and animal path albeit with a high stress value, 0.23 (Figure 2.10). PERMANOVA confirmed the difference was significant Pseudo- F (2, 36) = 2.34, p = .008 n = 40, and that there were no significant differences between macrohabitats; Pseudo F = 0.47, p = 0.67. The dissimilarity of microhabitat variables between random and animal paths was explained by SIMPER (Table 3). The independent t test showed that there was no significant difference between random and animal spool lines for ‘distance to suburb edge’; t = 0.49, p = 0.63, n = 20.

Vegetation disturbance Microhabitat components of native vegetation did not differ between macrohabitat types edge, core weedy and core; Pseudo F (2, 218) = 0.91, p = 0.51 n = 239, although there were significant site differences (Pseudo-F (15, 218) = 7.2, p = 0.0001). This indicates that the native vegetation at edge and core weedy sites was not impacted with respect to microhabitat variables by anthropogenic disturbance, thus results were not confounded.

Habitat use along path within lantana Within the lantana (i.e. by only considering the portion of the spool line characterized by Lantana), the MDS plot showed separation in the microhabitat features of random and animal path with a low stress value, though it was visible only in the 3D representation (Figure 2.11). The difference was confirmed with PERMANOVA; n = 31,

Pseudo- F (2, 27) = 2.36, p = .023. The dissimilarity between random and animal paths was explained by SIMPER (Table 4)

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Standardise Samples by Total Transform: Square root Resemblance: S17 Bray Curtis similarity 3D Stress: 0.11 path C T

Random Animal

Figure 2.11 Multidimensional scaling plot of Bray-Curtis similarity coefficients showing differences in path between animal and random spool lines, based on microhabitat variables, within Lantana only n = 31. Each point represents a point along a random spool line (one sample unit).

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Random Animal Av. Av. Av. Diss. Diss./SD Contrib. % Cum.% Abund. Abund. litter % cover 6.84 6.39 1.86 0.88 15.96 15.96 Distance to edge of 7.04 7.28 1.47 0.99 12.64 28.60 weed thicket litter depth 0.87 1.12 0.96 1.30 8.24 36.84 cavity: low 0.11 0.39 0.95 1.27 8.15 44.99 ground stems 0.94 1.11 0.82 1.31 7.05 52.05 cavity: tall 0.24 0.19 0.77 1.16 6.62 58.66 lantana density 0.91 1.13 0.72 1.41 6.16 64.82 big logs 0.11 0.21 0.70 1.06 6.01 70.83 homogeneous 0.49 0.61 0.68 0.99 5.82 76.65 small logs 0.03 0.22 0.59 0.80 5.04 81.70 big trees 0.10 0.17 0.57 0.81 4.94 86.63 fallen Tree 0.18 0.11 0.56 0.97 4.83 91.46

Table 2.4 Results of SIMPER analysis showing dissimilarity in microhabitat variables between random and animal lines within lantana only.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Standardise Samples by Total Transform: Square root Resemblance: S17 Bray Curtis similarity 2D Stress: 0.07 vegtype native lantana

Figure 2.12 Multidimensional scaling plot of Bray-Curtis similarity coefficients showing differences in microhabitat variables between lantana and native vegetation. Each point represents a point along a random spool line (one sample unit).

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Native Lantana Av. Av. Av. Diss. /SD Contrib. % Cum. % Abund. Abund. Diss. litter cover 7.75 8.86 6.23 0.79 30.47 30.47 understorey 1.03 1.37 3.67 0.82 17.95 48.42 density no. ground stems 0.85 1.12 2.80 1.07 13.70 62.12 homogeneous 0.52 0.81 2.02 1.13 9.88 72 litter depth 1.17 1.06 1.43 1.07 7.06 73.97 cavity: low 1.10 0.29 1.21 0.74 6.00 79.98 cavity: tall 0.18 0.15 1.17 0.35 5.78 85.75 fallen tree 0.09 0.14 0.78 0.43 3.86 89.61 small logs 0.18 0.02 0.71 0.47 3.50 93.11

Table 2.5 Results of SIMPER analysis showing dissimilarity in microhabitat variables between lantana and native vegetation lines at edge and core- weedy sites.

Differences in microhabitat characteristics between lantana and native vegetation The MDS plot showed distinct separation in the microhabitat features of lantana and native vegetation (Figure 2.12). PERMANOVA confirmed the difference was significant

Pseudo- F (1, 152) = 10.25, p = 0.001 n = 153. The dissimilarity of microhabitat variables between random and animal paths was explained by SIMPER (Table 3).

Habitat use along path in non-weedy habitats Within core sites, there was no difference in microhabitat features between the animal and random paths; n = 16, Pseudo-F (5, 6) = 1.12 p = 0.39 although there were significant site differences; Pseudo-F (6, 64) = 4.39, p = 0.0003. Despite this, univariate analyses showed that the animal path was characterised by denser vegetation than the random; n = 16 t = 2.19, p = 0.04 although univariate analyses did not reveal any other significant differences between animal and random spool lines.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

Burrows and Refuges Three burrows were discovered in the weedy habitats. In all cases the spool line disappeared into the burrow and was not able to be followed further. All three burrows were under Lantana, two were at the base of large (>30cm diameter) logs or trees. Two out of four refuges were under the cavity: low layer (Figure 2.13). No analyses were undertaken due to small sample size.

Figure 2.13 Refuge under cavity: low layer (in native vegetation), spool unit discarded shredded underneath. Photo by W. Gleen

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

DISCUSSION As predicted, these results indicate that the response by bush rats to the urban edge fits the ecotone model, and that populations follow an ideal free distribution over habitats of different quality, with urban edges being poorer quality than core habitats. Bush rat abundance indices were lower at edge sites, and an ideal free distribution pattern is indicated because no significant differences were found in the demography (adult gender composition, adult reproductive status, adult body weight and age class) of rats caught in the different macrohabitat types. The lower abundance indices at edge sites are likely to be a true indication of habitat quality, and not a consequence of despotism.

Empirical testing of distribution patterns has mainly focused on vole populations in Europe or America, and both ideal free (Morris and MacEachern, 2010, Pusenius and Schmidt, 2002, Lin and Batzli, 2001, Morris, 1997) and ideal despotic patterns have been reported (Messier et al., 1990, Morris, 1989, Pusenius and Schmidt, 2002, Lin and Batzli, 2001), along with variation between or within species (Pusenius and Schmidt, 2002, Lin and Batzli, 2001) and according to patch dynamics such as the position of the patch in the landscape (Morris, 1997). Whilst female territoriality is common in rodent populations (Wolff, 2003) and was expected in this study, very little information is available on distribution patterns of Australian rodents, especially at a landscape scale, although it can be inferred from a recent study on bush rat populations in a logged forest landscape (Lunney et al., 2009) that distribution followed an ideal free distribution, which supports the conclusions made here. It was also noted that there were differences in the proportion of juveniles across macrohabitat type; 19.6% at core sites, 6.3% at core weedy, and 0% at edge sites, indicating that perhaps core sites supported greater reproductive potential. Despite the trend these differences were not significant, probably due to small sample size in age classes, but are consistent with the idea that edge sites are poorer quality habitat.

However, it is possible that the apparent difference between edge and core sites was driven by sampling bias. Trap transects or grids draw animals from the effective trapping area (Aplin et al., 2003), which needs to be the same in the patches being compared. If a portion of one effective trapping area overlaps an ecosystem that is

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales not viable habitat for the target species then a disproportionate measurement of abundance may be taken there, since the area sampled will be less (Figure 2.6). This is another reason it is important to define the species response to each of the ecosystems that contribute to the ecotone, as well as the response to the ecotone itself (Brady et al., 2011).

140m 140mm m A. B.

Light fill: the effective trap area of the transect

Darker fill: non-viable habitat

__ Transect

Figure 2.6 Representation of how sampling is biased where part of the draw area of the transect is non-viable habitat for the target species such as the urban matrix. A. represents an edge site where the effective trap area is less than B. due to the area of non viable habitat.

To address this, a polygon was drawn surrounding each transect to a distance of 100m (±1m) from any point for females and 200m (±1m) for males using Google earth (imagery dated within 12 months of the survey dates). The area of the polygon was calculated (in m2 ) using an online tool ‘Zonum Solutions KLM area and Length’ (2006 - 2012). This area was called the ‘effective trapping area’ and was based on the assumption that Bush Rat home range length is approximately 100m for females and 200m for males (Maitz and Dickman, 2001, Watts and Aslin, 1981, Wood, 1971). This approach accounts for animals whose home range perimeter may overlap the transect (Warneke, 1971). Another polygon was then overlaid on any urban development

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales within the effective trapping area and the same calculation was used. This area was called the ‘non-viable area’ and was subtracted from the effective trapping area to give the adjusted, effective trapping area. Using this value the MNKA for each transect was expressed as rats per 10 ha, and was called the adjusted MNKA. It is recognised that rodent behaviour is notoriously plastic (Lacey and Sherman, 2007), and that home ranges and probabilities of individual encountering the transect are probably varied, so this method should be interpreted with caution. The result was analysed by a one way ANOVA in JMP version 9 (SAS Institute, 1989-2010), with adjusted MNKA as the dependent variable and macrohabitat the predictor variable.

When the data was adjusted to account for sampling area (by expressing trap success as rats per 10 ha of viable habitat) statistical difference between macrohabitats disappeared as the numbers of captures dropped, F (2, 14) =1.41, p =0.28 n = 17, although the pattern of fewer animals being captured at edges remained. It seems unlikely that the patterns are solely driven by sampling bias and the underlying method of the data adjustment relied heavily on a number of important assumptions including home range estimates from other populations, times and locations (Peakall et al., 2006, Lunney, 1983, Wheeler, 1970). These assumptions may have been too rigorously applied and it is recommended that the data adjustment method is tested against known data or any future studies on fauna responses to the urban edge control for this phenomenon a priori.

Spooling indicated that at the fine scale bush rats do not enter the urban matrix, which supports other evidence that this species is urban sensitive (Brady et al., 2011, Garden et al., 2010, Barnett et al., 1978, Australian Museum, 2009-2012). It is possible that separation between the urban and bush environments was due to the boundary being used as a demarcation of home ranges. However, small mammals regularly cross other habitat and territorial boundaries in heterogeneous environments (Batzli et al., 1999), and examples of bush rats crossing between different natural ecosystem types exist (Bentley, 2008, Maitz and Dickman, 2001). Therefore it is likely that the response can be attributed to an avoidance of the urban matrix, which indicates that the species is urban sensitive. However, further investigation at a larger scale is needed to confirm this conclusion.

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

If it is accepted that there was an ecotone response by bush rat populations, lantana can be eliminated as a negative edge effect because the bush rat index of abundance was the same for core and core weedy sites. No difference was detected in native vegetation between the three macrohabitat types, so the bush rats’ preference for lantana cannot be attributed to the fact that the native vegetation at weedy or edge sites was suboptimal. Instead the index of abundance at urban edges was positively correlated with lantana density. While cause and effect should not be assumed from a correlation, I suggest that dense lantana increases the habitat quality of urban edges for the target species, and is consequently an emergent property of the urban edge for bush rats (Figure 2.14). The most likely alternate explanation is that a higher soil moisture content at some sites caused covariance of both measures, as soil moisture is known to advantage both lantana (Thorp and Wilson, 1998 onwards) and bush rats (Braithwaite et al., 1978, Warneke, 1971). This is improbable as the aspect of all edges, which determines the drying effect of sun and wind, was similar and there was no similar pattern at core weedy sites.

Figure 2.14 schematic representation of ecotonal response of the bush rat at the urban edge ameliorated by lantana, based on Lidicker (1999) a: response when lantana is of high density b: response when lantana is of low density

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales

This provides a clue for the identification of the main mechanism/s driving the bush rats’ edge response. Whatever they are responding to it is at least partially ameliorated by lantana density. Dense vegetation is known to have a ‘sealing’ effect on edges, and a similar positive response by bush rats to sealed edges has been reported in native Wet Tropics habitats (Goosem, 2000). Dense vegetation at edges may reduce to some extent the strong primary microclimatic impacts (Laurance et al., 2007, Harper et al., 2005, Matlack, 1993). It may also reduce predation risk that is sometimes reported as an edge effect (Poulin and Villard, 2011, Storch et al., 2005, Wilcove, 1985), and anthropogenic disturbance, which bush rats may not tolerate. Although lantana is recognised as a buffer of forest edges and thought to be useful for small mammals (Thorp and Wilson, 1998 onwards), this is the first empirical evidence to support these views.

However, it is important to consider limitations of scale (Peles et al., 1999) as the weedy patches sampled were smaller than the likely home range size for the species and in all cases native elements also existed within the catchment area of the transect. Consequently these results should be interpreted in the context of the surrounding landscape. Small mammals in heterogeneous environments generally depend on movement across patch boundaries to obtain resources (Batzli et al., 1999), and persistence depends on landscape as well as patch variables (Law and Dickman, 1998, Santos-Filho et al., 2012). This study has revealed that lantana is a useful resource for bush rats, however, it is likely that lantana is useful by adding to the heterogeneity of the landscape, rather than by existing as a complete resource in itself. Weeds, by definition, progress to a monoculture thus homogenizing the landscape whereas small mammals generally benefit from a patchy heterogeneous landscape (Law and Dickman, 1998, Foster and Gaines, 1991), and it is unlikely that lantana would provide sufficient resources for the bush rat at a landscape scale given the complex and changing requirements of vertebrate species within a home range (Law and Dickman, 1998, Bennett, 1993, Burgman and Lindenmayer, 1998, Holland and Bennett, 2007, Stevens and Tello, 2011).

The lack of significant predictors of bush rat habitat use at core sites other than vegetation density was unexpected, since many other studies have identified

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales important microhabitat predictors for this species in native habitat (Strauß et al., 2008, Barnett et al., 1978, Holland and Bennett, 2007, Warneke, 1971). The most likely explanation was to do with site effect. As weedy sites fitting the required criteria such as distance to edge and independence were rare, they were selected from a variety of surrounding forest types, as discussed in the introduction. Control sites were then selected to match this disparity, thus there was a large degree of between site variation, as revealed by the strong site effect in the PERMANOVA result. Consequently, unified trends would have been harder to detect.

Use of habitat features in weedy and native vegetation The ‘pictorial reference’ method of measuring vegetation density that I developed had the advantages of being quick to implement and incorporating an aerial perspective, it was felt to represent a more realistic ‘predator eye view’. It was shown to correlate well with the more objective traditional cover-board method, which supports the findings of Glen, Sutherland et al. (2010) who used a similar method and also found a correlation with the cover-board method.

The macrohabitat results were supported at a fine scale; bush rat trap success was greater within lantana thickets and where lantana had a higher coverage and density score, indicating that the weed is advantageous. Lantana density also explained the most difference between random and animal spool lines. Other habitat variables that featured along the animal path more than expected given their availability included lantana cover, big logs and the cavity: low and homogeneous architectures. The result for big logs supports other research showing that such structures are preferentially used by Rattus species for shelter and ease of movement (Cox et al., 2000, Strauß et al., 2008). As predicated, vegetation architecture was also an important fine scale habitat feature. The cavity: low layer would probably provide a combination of visual and structural protection from avian and mammalian predators whilst preserving a clear path for prey movement. This is likely to reduce both perceived predation risk and the energetic costs of locomotion. Runways in the vegetation have been shown to be important for bush rat movement (Woodside, 1983), and a similar but artificial structure has been shown to reduce perceived predation risk in Australian small mammals (Stokes et al., 2004). There may also be parallels between the cavity: low

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales layer and the subnivean layer which bush rats use above the snow line (Sanecki et al., 2006a) as do small mammals in the northern hemisphere as it obstructs most predators (Korslund and Steen, 2006, Pruitt, 1984 IN ) and moderates microclimate due to the insulation of snow (Happold, 1998). The use of cavities within native vegetation (Hibbertia riparia and Lepidosperma semiteres) by bush rats has been reported and was attributed to the value as a refuge (Frazer and Petit, 2007), but has not been empirically tested. The’ homogeneous’ category of architecture was also preferred by the bush rats within the lantana; this arrangement probably reduced perceived and real predation risk as it provided relatively even vertical and horizontal cover. Even cover equates to a high FHD measurement (MacArthur et al., 1962), which is thought to affect perceived predation risk of small prey (Els and Kerley, 1996, Barkman, 1988). The cavity: tall structure by contrast represents uneven cover, equivalent to a low FHD measurement, and although it only explained 4.86 and 6.62% respectively of the difference between animal and random path, the rats used the cavity: tall structure less than it was available both at a site scale and when within lantana only. FHD can affect microclimate (Barkman, 1979) and this structure would presumably facilitate air convection and an unstable microclimate whereas the cavity: low layer would be more likely to have a stabilizing effect. However, perceived predation risk is known to impose a greater cost on prey populations than the metabolic energy required for foraging (Brown and Kotler, 2007), so it is likely to be the main function.

Variables that featured along animal path less than expected due to their availability included leaf litter, small logs and trees. This was unexpected since such features comprise elements of structural complexity which is known to be a predictor of bush rat habitat use (Maitz and Dickman, 2001, Holland and Bennett, 2007, Spencer et al., 2005) Rodents may prefer deep leaf litter (Cox et al., 2000) as it harbors food resources such as invertebrates (Dugdale, 1996, Bultman and Uetz, 1982) and fungi (Fuhrer, 2005). The anomalous result may be explained by the fact that the leaf litter was observed during field work to be occasionally held above ground by the characteristic horizontal stem structure of lantana, especially where it formed the cavity: low layer, and it is likely that there are fewer trees, and thus fewer small logs in

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales lantana thickets, due to either suppression by the weed or the prior anthropogenic clearing that allowed lantana to establish in the first place, whereas big logs persist for longer. Examination of characteristics of animal path within lantana only was illuminating; the rats were in fact selecting deeper patches of leaf litter and areas with small logs when within lantana. Data analysis supported this; leaf litter was shallower and there were fewer small logs within lantana than native vegetation (Table 5). Thus the bush rats’ principal preference for lantana overrode the selection of habitat features such as deep leaf litter and small logs, which are traditionally described as being elements of structural complexity. Mammal habitat selection is scale dependent (Leiner et al., 2010, Haythornthwaite and Dickman, 2006, Holland et al., 2004, Garden et al., 2010), this shows that examination of very fine scale habitat associations i.e. subsets of microhabitat type can be informative.

The use of deep leaf litter under lantana is likely to indicate foraging behaviour. Scats were opportunistically collected from an edge site during this study. More than 20 separate collections of scats were washed through a sieve, 95 um grid, to recover coarse dietary components in a supplementary study (P.A. McGee, unpublished data). He found the spores of arbuscular mycorrhizal (AM) fungi were unusually infrequent, possibly as a consequence of site disturbance. Fungi are a critical winter food source for bush rats (Tory et al., 1997, Watts and Braithwaite, 1978). If disturbance limits the diversity and abundance of AM fungi, their absence may contribute to the poor quality of habitat for bush rats at edge sites. Plant material but not insects were well represented in the scats. The low energy content of plant material compared to insects (Breed and Ford, 2007) and lack of fungi in winter may provide an explanation for the poor quality of edge habitat for bush rats. Although these informal findings indicate that food may be a limiting resource at edge sites it is unlikely that this explains bush rat use of lantana as habitat features relating to safety (i.e. cover and density) were more important than those relating to foraging (i.e. leaf litter) for bush rats. Lantana was deemed a priory more likely to reduce than increase food resources (Chapter 2 Introduction).

Vegetation density was universally important for path whether the bush rats were within lantana or native vegetation, and the fact that lantana tended to be denser than

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales native vegetation (Table 5) is likely to be a reason that it is preferred. This result (that the weed was denser than native vegetation) supports Bennett (1993) who found a negative correlation between floristic richness and vegetation density.

As predicted, vegetation architecture also influenced the bush rats’ path, in particular a greater abundance of the cavity: low structure which was important both at site level and within lantana. This cavity develops due to the tendency of lantana to take root from horizontal stems, forming layers (Muyt, 2001). It also occurred in native vegetation, for example where a raft of stems and leaf litter accumulated above a sparse low shrub layer. An abundance of ground stems was also important, although this was probably due to a positive correlation with the cavity: low layer which was usually composed of horizontal stems; though they were suspended above ground they often fell within the height restriction specified for the count of ground stems.

Lantana had different microhabitat characteristics to native vegetation, in addition to being denser and creating a shallower but more consistent coverage of leaf litter it had a greater tendency towards an even growth pattern (‘homogeneous’ architecture), apart from a low cavity whereas native vegetation was more inclined to have taller gaps in the horizontal layers. There were more ‘fallen tree’ architectures in lantana, possibly from areas where it was encroaching on the surrounding vegetation and causing tree mortality (Laurance et al., 2007). In summary, results indicate that lantana is a predictor of bush rat habitat use, due to a combination of the cavity: low architecture and density. Lantana mitigated urban edge effects on bush rats and I next examine the hypothesis that the causative mechanism for this is perceived predation risk.

Future Directions Understanding the dynamics of these populations is somewhat constrained due to the short term nature of this study. Indices of survival and production over time would be more likely to reveal mechanisms, such as source/sink dynamics (Mosser et al., 2009) across edge and core habitats. Investigation of the distribution patterns followed by this species would also inform future studies on bush rat ecology. Further evidence is required to confirm the effect suggested here that lantana or dense vegetation at the

CHAPTER 2 Habitat use by the bush rat Rattus fuscipes at the weedy urban edge: a study at two scales urban edge advantages bush rat populations, and to determine whether it applies to other small mammal species.

CHAPTER 3 Lantana and bush rat perceived predation risk

CHAPTER 3 Lantana and bush rat perceived predation risk.

ABSTRACT Weeds are widely and profoundly changing the structure of the Australian bushland. It is important to understand how this change affects the processes driving fauna responses in order to inform land management decisions. In Chapter 2 I showed that lantana (Lantana camara) is an important habitat feature for bush rats, and lantana density was positively correlated with bush rat abundance indices at the urban edge. In this Chapter I use a giving up density experiment to test the hypothesis that lantana structure reduces perceived predation risk for bush rats foraging in a mixed predator landscape. Bush rat perceived predation risk, as judged by the Giving Up Density (GUD) method, were compared between patches with no vegetation cover (clear ground), lantana and native vegetation (bracken Pteridium esculentum) with the two vegetation types of equal density. Both lantana and bracken lead to lower GUDs suggesting that vegetation structure is generally perceived as safer than open ground irrespective of the particular plant species involved. Lantana appears to be a structural proxy for native vegetation in reducing perceived predation risk and, to a degree, advantages bush rat populations by this mechanism.

INTRODUCTION The fear of predation can be a powerful influence on the behaviour of individuals (Ydenberg et al., 2007), and consequently has the capacity to shape population dynamics and distribution of potential prey (Arthur and Pech, 2003). In Chapter 2, I showed that bush rats use the weed lantana at both a fine and landscape scale. Also, bush rat abundance indices were positively correlated with lantana density at the urban edge which may indicate that lantana mitigates important urban edge effects. This correlation may at least partially be explained if lantana reduces their perceived predation risk which is likely to be an important and negative edge mechanism for native rodents (Bock et al., 2002). Bush rats are at risk of predation from a suite of native predators in the study area, including snakes, birds of prey, lace monitors Varanus varius and possibly tiger quolls Dasyurus maculatus (Guarino, 2001, Banks,

CHAPTER 3 Lantana and bush rat perceived predation risk

1999, NSW National Parks and Wildlife Service, 2000, Strauß et al., 2008). Potential alien predators of bush rats were identified in a survey of the study area undertaken shortly before this study commenced (Old and Wolfenden, 2010). Cats Felis catus, dogs Canis familiaris and foxes Vulpes vulpes were common, dogs particularly so, and foxes were seen to successfully predate small mammals. Human activity (e.g. movement, noise) is also likely be perceived by bush rats as a predation risk, as has been shown for gerbils (Randall and Rogovin, 2002). In this chapter I use artificial foraging patches to test whether lantana lowers bush rat perceived predation risk which may explain their use of lantana and the apparent advantage the weed confers to bush rat populations at the urban edge.

Measuring small mammal perceived predation risk Giving Up Density (GUD) (Brown, 1988) is a well established technique used to measure perceived predation risk. The value of food gained in a patch is offset against the cost of predation risk, missed opportunity and the energetic costs of foraging, as represented in the equation H = C + P + O; H = quitting harvest rates, C = metabolic cost of foraging, P the predation cost of foraging, O missed opportunity costs. This technique uses a food reward that is progressively more difficult to obtain, classically in the form of seeds mixed through a substrate such as sand, so that the forager can obtain initial reward easily but then has to invest increased time and energy to locate buried items, creating diminishing returns with additional foraging. When the perceived risks and costs outweigh the benefit, the forager will ‘give up ’the patch. The missed opportunity, energetic and metabolic costs can be assumed to be the same and cancelled out if the experiment units are run at the same time, placed close enough so that environmental conditions are standard (other than the variable being tested),and if the substrate and food resource are identical. The amount of food remaining can therefore be used to quantify perceived predation risk. GUD experiments are regularly used to measure rodent responses to habitat or microhabitat variables (e.g. Arthur et al., 2004, Dutra et al., 2011, Mattos and Orrock, 2010, Brown et al., 1988, Strauß et al., 2008). In this chapter GUD’s were tested between open ground, lantana and bracken. Bracken, which like lantana, is toxic to mammals (Fenwick, 1989, Alonso-Amelot and Avendana, 2002) was used to represent native vegetation to achieve consistency.

CHAPTER 3 Lantana and bush rat perceived predation risk

Does vegetation structure mediate perceived predation risk? Vegetation structure refers collectively to vegetation density, cover and architecture (as defined in Chapter 2). Vegetation density refers to the extent to which the plant biomass provides visual obstruction of the ground or an object, vegetation cover refers to the area occupied by the plant as a percentage of the surface area of the sample plot and vegetation architecture refers to the three dimensional physical arrangement of plants and their parts. In Chapter 2 I found that all structural aspects of lantana, cover, density and architecture, predicted bush rat habitat use. The concept that vegetation structure can be more important than floristic composition for habitat use by small animals is not new (e.g. MacArthur et al., 1962) and recent studies also indicate that weeds can provide structural habitat for rodents, with weeds being more important as a refuge from predation risk than a food source (Dutra et al., 2011, Mattos and Orrock, 2010, Simone et al., 2012). For example, a hickory–oak deciduous forest system in the USA was invaded by a weedy shrub, Amur honeysuckle Lonicera maackii, which created an under storey layer that was previously lacking. Endemic white-footed mice Peromyscus leucopus perceived the novel shrub structure as safer (less of a predation risk) than the original habitat (Mattos and Orrock, 2010, Dutra et al., 2011) which was shown to be a more important attribute than its food value for mice in the Peromyscus genus (2011). In Australia dense cover of weedy thickets can reduce brush turkey Alectura lathami chick predation rate (Goth and Vogel, 2002), and lantana structure has recently been found to be important in maintaining reptile communities (Virkki et al., 2012). Thus weeds can advantage small prey by providing a refuge from predation due to improved structural cover.

Although in some cases vegetation structure can advantage prey by providing a refuge, in others it can disadvantage prey by hindering visual detection of predators (Embar et al., 2011) and escape (Schooley et al., 1996, Simonetti, 1989). Whether vegetation structure is a help or hindrance can depend on the hunting strategy of the predator. According to the risk allocation hypothesis (Lima and Bednekoff, 1999), prey species are able to distinguish different hunting strategies from the cues presented by predators and adapt their response accordingly, making sophisticated choices in risk management. For example, kangaroo rats Dipodomys deserti and D. merriami avoided

CHAPTER 3 Lantana and bush rat perceived predation risk

vegetated patches because their main predator used an ambush technique, which is less effective over clear ground. However, when the main predator operated by aerial attack, their preference reversed (Bouskila, 1995). Locally, bandicoots Perameles nasuta at North Head near Sydney, preferred open ground which was attributed to the ability to detect and avoid predators more easily (Chambers and Dickman, 2002). Dense undergrowth is also thought to provide an advantage for cats and foxes as they hunt using ambush techniques (Edwards et al., 2002, Hopcraft et al., 2005, White et al., 2006), and these feral predators pose a strong predation risk to small mammals in this study area (Kutt, 2011, Meek 2003, Old and Wolfenden, 2010). The death adder Acanthopis antarcticus is also common in the study area. It also hunts by ambush, often hiding at the base of shrubs (Cogger, 1992), so is also likely to be advantaged more by vegetation structure than open ground. If perceived predation risk is in fact higher under lantana, then bush rats must be accessing superior resources there given their apparent preference for this vegetation type.

In this mixed predator landscape, bush rats must be involved in a very complex game as they are at risk from a wide range of predators, including alien species, with a variety of hunting styles. Other complicating factors include the fact that different predator species may directly or indirectly mediate each others’ hunting success (Finke and Denno, 2004, Denno et al., 2005, Ritchie and Johnson, 2009, Ylonen and Brown, 2007) and recent research indicates that cues from relatively recently arrived feral predators, such as cats, may not be understood by native species (Carthey and Banks, 2012). Given this complexity, at any given time the individual bush rats’ information state is likely to be poor. Despite this, and the fact that several of their predators may be ambush style hunters advantaged by undergrowth, studies on small Australian mammals indicate that they perceive vegetation structure as safer than open ground when foraging (Stokes et al., 2004, Strauß et al., 2008).

Does vegetation architecture mediate perceived predation risk? A variant hypothesis on the ‘structure vs floristics’ idea is that fauna will recognise differences in plant species architecture. The target plant species in this chapter have different architecture, lantana has a strong barbed horizontal interwoven stem

CHAPTER 3 Lantana and bush rat perceived predation risk

architecture (Thorp and Wilson, 1998 onwards, Muyt, 2001) whereas bracken is characterised by flexible vertical stems and leathery foliage (Fairley and Moore, 1989).

Two important ways vegetation can affect predation risk are by influencing the detection and accessibility of prey (Rieder et al., 2010, Butler and Gillings, 2004, Bennetts et al., 2006), and the two mechanisms are not necessarily linked. For example whilst ground foraging bats Myotis myotis were able to detect prey in dense vegetation, their ability to successfully capture items was hampered due to obstruction of wing envelopment (Rainho et al., 2010). Wire mesh netting placed at ground level, which presumably provided a physical but not visual impediment to predators, resulted in lower GUDs for two small mammal species Antechinus stuartii and Sminthopsis murina than in the open (Stokes et al., 2004). A meta analysis of vertebrate GUD studies in terrestrial systems revealed that the role of vegetation structure is more important to prey for facilitating evasion from predators rather than preventing detection by them (Verdolin, 2006). This implies that prey should be able to distinguish between vegetation structures which impede predators versus structures which merely obstructs detection. I suggest that the impediment of bush rat predator movement (which determines prey accessibility) is primarily dependent on architecture and vegetation strength. For example, invertebrates will seek refuge in complex, within-plant architecture which impedes predators (Denno et al., 2005).

A predictor of bush rat habitat use identified in Chapter 2 was cavity: low architecture, an artificial category developed to describe frequently occurring vegetation architecture patterns, based on the evenness of the horizontal layers (Table 2.2). The cavity: low layer may reduce perceived predation risk by providing a clear space to facilitate escape (Schooley et al., 1996, sensu Ylonen and Brown, 2007) whilst also obstructing detection and access by predators due to the covering layer of strong stems, and this may partly explain bush rat’s use of lantana. However, most studies indicate that vegetation structure mediates prey perceived predation risk use measures of density or cover (Wilkinson et al., 2013, Strauß et al., 2008, Yunger et al., 2002, Orrock et al., 2004); architecture is more likely to be species specific, so available evidence suggests that plant architecture may be less important than density and cover.

CHAPTER 3 Lantana and bush rat perceived predation risk

Perceived predation risk at the urban edge Several studies suggest that habitat edges can be riskier for mobile prey. For example edge effects at the sea grass and sand edge for King George whiting Sillaginoides punctata were explained by predation by salmon, Arripis spp., (Smith et al., 2011) and a negative correlation between skinks and predatory birds at the peri-urban and woodland edge combined with higher strike rates there suggested increased predation of skinks (Anderson and Burgin, 2008). Urban edges may be perceived as a more dangerous environment than bushland interiors by small mammals because they are likely to be impacted by wandering domestic dogs and cats (Smith and Smith, 2010, Barratt, 1998, Old and Wolfenden, 2010). These human companion animals live in the urban matrix and are often uncontrolled (Old and Wolfenden, 2010) and therefore likely to have a ‘spill over’ effect on the adjacent edge habitat (Manolis et al., 2002, Oksanen et al., 1992, Liddicker, 1999). Such predators can use habitat edges as a pathway (Andren, 1995) and cats in the urban environment prefer to travel along vegetation boundaries (Meek 2003), such as would be provided by the abuttal of bushland to the urban boundary. Additionally, I found that domestic dogs were regularly walked along or near the boundary between the bushland and the urban matrix. Snakes may also prefer edges to facilitate thermoregulation (Blouin-Demers and Weatherhead, 2001). Urban edges are also likely to be characterised by increased human activity which may be perceived as a risk by small mammals. I regularly observed that human disturbance appeared greater in this area e.g. children playing in backyards, joggers and cyclists. Light from buildings and infrastructure may be an indirect cue of predation risk since moonlight is known to be (Brown and Kotler, 2007, Mattos and Orrock, 2010). If the urban edge is perceived as particularly risky by bush rats, lantana may provide a valuable service in this macrohabitat type by providing a refuge to mitigate perceived predation risk there.

Aims Using a GUD experiment I aim to determine whether bush rat perceived predation risk is lower under lantana and native vegetation structure as opposed to open ground. Experiment modification was required following the pilot study, as it was not possible to compare different vegetation architecture patterns in a crossed design as originally planned.

CHAPTER 3 Lantana and bush rat perceived predation risk

Hypothesis Vegetation structure is more important than species for reducing bush rat perceived predation risk when foraging in a mixed predator landscape.

Predictions Bush rat GUD will be lower under both native vegetation and lantana compared to open ground.

Study Sites The study areas were within National Parks or Nature Reserves on the Central Coast of NSW, Australia. Sites for the pilot study were characterised by the weed lantana interspersed within a landscape of native vegetation. Sites for the main study (Figure 2.1) were characterised by native bracken Pteridium esculentum as well as lantana.

Sites were spatially independent, being separated by >400m, a distance twice that of the home range for the species (Maitz and Dickman, 2001, Robinson, 1987), or a minimum of 200m apart and also separated by a movement barrier such as a road (Barnett et al., 1978, McDonald and St Clair, 2004).

Pilot trial Pilot trials were undertaken to assess the feasibility of testing Bush rat GUD response to differences in architecture categories cavity: low and homogeneous (Table 2.2) under native vegetation and lantana in a crossed design, where the density of vegetation types being compared is the same. Artificial food patches placed 5-10m apart under five different microhabitat types; native (homogeneous), native (cavity: low), lantana (homogeneous), lantana (cavity: low) and no vegetation cover. Two replicates were sampled concurrently. Five important issues were identified from the pilot study. Firstly, non-target species compromised results at one site via excessive interference (such as 100% visitation rate) especially black rats Rattus rattus, which almost completely depleted the food resource. Secondly it was very difficult to place the trays under the vegetation without altering the architecture that was being measured. Thirdly, it was not possible to find sites providing natural occurrence of H and CL architectures under similar-structured native vegetation, there was very high variation between native species in stem and foliage size, strength and orientation which would have confounded results. Fourthly, whilst the second site was not

CHAPTER 3 Lantana and bush rat perceived predation risk

affected by interference from non-target species, only one or two out of five trays were foraged by the target species per night. Interpretation of GUD’s depends on all resource patches being foraged within the same event (one night) so that conditions are constant and results commensurable. The failure to record a complete event may have been due to the number of trays and their relatively large degree of spatial separation (required due to juxtaposition of the relevant vegetation types) which may have raised the costs of search and or transit beyond the value of the food resource. It was unlikely to have been due to excess resource quantity as the trays that were foraged were depleted to a GUD lower than the species usually attains locally (Carthey, 2011). Alternatively the acclimatisation period may have been insufficient for the bush rats to have discovered the location of all the trays. Lastly, manipulation of vegetation was attempted to replicate the CL architecture where it did not occur naturally in order to facilitate a closer arrangement of trays, but the result was not considered to be sufficiently similar structurally to the natural growth pattern, so this idea was discarded. The problems that were encountered in this pilot study necessitated modification of experiment design.

METHODS A power analysis was run a priori to estimate the number of replicates required to detect significant differences in GUD’s between vegetation types, using values extracted from data collected for a comparable experiment using bush rats in a similar geographic location (Carthey, 2011). Russ Lenth’s Java applets for Power and Sample size (Lenth, 2006-9) for a 1-way ANOVA was used for this calculation. Using SD = 5.82; Power was set at 0.8 and α at 0.5 by convention, a sample size of n=12 would be able to detect an effect size of 6.5 (or an approximate 20 % difference between GUDs), which was similar to the effect size found by Carthey (2011). Twelve sites were sampled between 7th July and 2nd November 2012.

A replicate consisted of three artificial food patches, placed under bracken, lantana and one surrounded by clear ground. Following the pilot study, comparisons of vegetation architecture categories based on the evenness of the horizontal layers were removed from the design; instead GUD’s were tested only between open ground, lantana and bracken where both vegetation types were of the same height, fit the H

CHAPTER 3 Lantana and bush rat perceived predation risk

architecture type and were highly dense (category 3 pictorial reference method, Table 2.1). Food patches consisted of plastic trays (30cm length x 24cm width x 11cm height) that contained an inedible substrate of 1kg (± 20g) sterile Sydney sand (kiln dried), with 30g (±0.09g) peanut fragments mixed through in a random distribution. Fragments were weighed by digital scales to nearest 0.1g and were mixed into the sand by hand. A clear lid covered each tray with a 15cm hole allowing rat access and visibility of prey/predator, but preventing excess spillage of sand. Trays were monitored by remote sensing infra red camera (Scoutguard SG550V) in order to determine which species foraged there. Cameras were set approximately 50cm from the tray, supported by a 2m long 20mm2 surveyors peg. Cameras were set to video mode with 3M pixel 640*480 quality for 30 sec with 0 sec gap between recordings and high sensitivity. A clear perspex sheet 50 x 50cm was suspended 50-90cm above the tray in order to prevent wetting or flooding by rain, avoiding any disturbance of the vegetation. As the perspex sheeting was transparent and obstruction from the support posts negligible (< 5% of the tray obscured as measured by cover-board), small variation in cover height was not considered likely to affect perceived predation risk.

To ensure that vegetation density was constant between the lantana and bracken units, a clear perspex cover-board 50 x 50cm scored with a grid 10cm2 was used to check that the food patch was visible to a similar degree (Figure 3.2). This method was adapted from the trapping/spooling study (Chapter 2), and from Freitas, Cerqueira et al (2002). The tray was observed through the cover-board from a distance of 1.5m and a height of approximately 80cm (kneeling height, to more approximate a fox or cat eye view) whereby the view of the forage tray filled 2 squares of the cover-board. The proportion of the squares in which view of the tray was obscured by vegetation was estimated for the native vegetation. This was done from the point of researcher access to the tray (called the front of the tray), from above, and from each side and the back of the tray where this was possible without trampling the vegetation, if not an estimate was made. The same process was then undertaken for the lantana section and then, where necessary, the lantana was manipulated to match the native vegetation in terms of visual obstruction of the tray and also height, by adding or subtracting branches which were woven into the existing structure in a similar angle to

CHAPTER 3 Lantana and bush rat perceived predation risk

the natural growth pattern. Thus native and weedy sample units were of similar height and vegetation density, therefore any significant difference in results would be attributable to other features such as architecture or strength of the vegetation.

The three trays were placed 3-6m apart (Figure 3.2) which was the tightest configuration possible given the difficulty in finding a juxtaposition of the requisite vegetation types. Other similar GUD studies have successfully used distances within this range (Stokes et al., 2004, Yunger et al., 2002), or greater (Strauß et al., 2008) and spacing that varies by a factor of 2 has been used successfully (Brown, 1988). The vegetation category extended for a radius of at least 1m from the food patch on three sides other than the front of the tray for the treatment unit (trays were pushed under the structure about 30cm from the front), and for a continuous radius of 1m for the open tray.

An acclimatisation period of four nights was used to increase the chance of the rodents finding the tray and learning how to forage there successfully although in two cases it was extended for an extra night where a complete event had not occurred by night 3. During this period the remaining nut fragments were sieved from each sand tray each morning and 30g were replaced and mixed. On the fifth night the experiment commenced and repeated for three nights to account for camera failure, interference from non target species or non appearance by the target species, in order to increase the chance of attaining at least one successful result per site. Again, each morning the remaining nut fragments were sieved from each sand tray and weighed and 30g were replaced and mixed. Also, camera batteries were changed and SD cards containing the video footage were downloaded and replaced. Videos were viewed to determine which species foraged there.

In the pilot study and the experiment proper there was a dichotomy in that there was either considerable interference from non-target species, such as Antechinus stuartii or other Rattus species, or very little to none. Results were accepted where bush rats comprised >90% of the video footage, and were the last species to forage at a tray, that is, they were the species to which the decision to give up could be attributed (Brown and Kotler, 2007).

CHAPTER 3 Lantana and bush rat perceived predation risk

Because weather and moon phase may provide indirect cues of predation (Brown et al., 1988) and consequently determine use of vegetation cover by prey (Mattos and Orrock, 2010, Dutra et al., 2011), weather type was recorded each night according to terms from the Australian Bureau of Meteorology (2011), and then placed into one of three categories: precipitation, cloudy, or fine; as all weather types that occurred during the experiment fitted one of these categories. Moon phase was considered to be a site effect, since change in illumination would be negligible over a three day period. Where possible, three sites were run concurrently (which was the maximum possible logistically) so that weather could be included as a factor in the analysis.

Figure 3.1 GUD site map, showing the position of each replicate on the Central Coast of NSW, Australia.

CHAPTER 3 Lantana and bush rat perceived predation risk

Figure 3.2 One experimental replicate, showing the three different vegetation types sampled. Photo by W. Gleen.

CHAPTER 3 Lantana and bush rat perceived predation risk

Figure 3.3 Use of cover board to measure vegetation density and ensure parity between vegetation types. Photo by C. Marsh.

DATA ANALYSIS Differences in GUD were investigated using a two factor ANOVA without replication in Microsoft excel, following visual inspection of distribution of data in JMP v. 9 (SAS Institute, 1989-2010) which revealed no violations of the assumptions of normality and homogeneity of variance (Quinn and Keough, 2002). GUD was the dependent variable, factors were site and vegetation type. Where a significant difference was detected by ANOVA, independent t-tests were used to determine where the differences lay. Due to the high degree of site failure and the fact that several sites had to be run asynchronously, it was not possible to include weather in the model because different types of weather occurred between sites meaning the model would have been incomplete. Instead, effect of weather on GUD was analysed for each vegetation type separately by one way ANOVA in JMP version 9 (SAS Institute, 1989-2010), with data pooled across site.

CHAPTER 3 Lantana and bush rat perceived predation risk

An event was considered complete where at least two trays were foraged by the target species. Due to their proximity, if two trays were foraged it was assumed that the Bush rats had knowledge of the existence of the third tray therefore non-visitation was counted as a foraging decision. Three sites had only one complete event recorded, the other sites had two or three complete events recorded. In total, sixteen complete events were recorded from eight sites. The sample size was smaller than had been planned, again due to site failure (n = 4 sites where there were non target visits, lack of bush rat visit or flooding). A two factor ANOVA using the same method as above was used to analyse differences in GUD using the whole data set (i.e. including GUDs from non-target species) to gain an indication of the response of the whole mammal community, given that all non-target species (Rattus and Antechinus species) were within a similar weight range to Bush rats and susceptible to the same range of predators. One instance of brush-tail possum interference was excluded from this analysis.

RESULTS

Giving up Densities

There was a significant difference in GUD’s between vegetation types F (2, 30) = 4.71, p = 0.02, n = 16, Figure 3.4. Post hoc T-tests revealed that Bush rats foraging in open vegetation types left higher GUD’s than when in weedy types; t = 2.74, p = 0.008. Non- target species that investigated or foraged at the trays were brown antechinus Antechinus stuartii, dusky antechinus Antechinus swainsonii, swamp rat, Rattus lutreolus, black rat, Rattus rattus, brush-tailed possum Trichosurus vulpecula, long- nosed bandicoot Perameles nasuta (in the pilot study only) and currawong, Strepera graculina.

Bush rat predators identified in the footage were boobook owl Ninox novaeseelandiae (an unsuccessful capture of an Antechinus stuartii by this species was recorded on camera), lace monitor Varanus varius and cat Felis catus, both of which inspected a tray but did not forage. A fox scat collected opportunistically at one site was analysed for prey identification; bush rat fur was identified (B. Triggs 2010 pers. comm.,17 April), indicating active predation was occurring. Where A.stuartii was the last forager at a tray, GUDs were as low as 1.9g (mean =14.5, SD = 8.82, n = 12). The result from all

CHAPTER 3 Lantana and bush rat perceived predation risk

visitors (34 events recorded over 12 sites) revealed a similar pattern; significant differences in GUD’s between vegetation types F (2, 66) = 7.43, p = 0.001, n =34, with highest GUDs in the open vegetation compared to either the native or weedy vegetation; t = -3.14, p = 0.002, n = 34 and t = 2.85, p = 0.004 n = 34. Notably for both bush rat only data and all visitor data, GUDs did not differ between native and weedy vegetation (bush rats only t = 1.47, p = 0.08 all visitors t = -0.63, p = 0.27 n = 34).

Post hoc T test results (in red where significant):

A-B: T=-3.14, p = 0.002; B-C: T=2.85, p = 0.048, A-C: T=-0.63, p = 0.27.

D-E: T=-1.75, p= 0.0503, E-F: T=2.74, p = 0.008, D-F: T = 1.46, p = 0.08.

D E F

A B C

Figure 3.4 Mean GUD for different vegetation types ± SE. Bush rat only n=16, small mammal n = 34.

Weather (precipitation, cloudy or fine) had no significant effect on GUD for any vegetation type, either for bush rat responses; F (2, 13) = 0.91 to 2, p = 0.1 to 0.42, or all visitors; F (2, 31) = 0.05 to 2, p = 0.15 to 0.95.

CHAPTER 3 Lantana and bush rat perceived predation risk

DISCUSSION The GUD experiment indicated that bush rats perceive that foraging within lantana is safer than in open conditions, as predicted. Although GUDs in native vegetation were not significantly lower than open ground, they tended to be lower and may have been significantly so had so many replicates not been excluded due to my strict criteria for inclusion. When data from all visitors was included GUDs were indeed significantly lower under native vegetation than in open ground and no different from the animals’ use of lantana. There are two main implications of these results. Firstly they suggest that vegetation structure is more important than floristics (species) for reducing perceived predation risk for small foragers. While the concept that vegetation structure is more important than floristic composition for determining fauna responses is longstanding (Haering and Fox, 1995, e.g. Bennett, 1993, Vernes, 2003, Garden et al., 2007), mechanistic explanations have not generally been explored and are poorly understood. These results suggest that perceived predation risk is one reason small fauna responds to structural rather than floristic habitat characteristics, which accords with recent studies indicating that weeds can advantage small prey by providing a refuge from predation due to improving structural cover. (Dutra et al., 2011, Mattos and Orrock, 2010, Simone et al., 2012). Whether the risk is genuinely higher in open areas is poorly known and there is scope to investigate what the survival consequences are of foragers’ use of structural cover (irrespective of species) rather than open habitat. That mortality is correlated with perceived predation risk is an assumption that requires testing if we are to fully understand predator-prey dynamics. Dickman (1992) revealed that predator-experienced house mice Mus musculus shifted their habitat use to dense vegetation cover in response to predation risk, which increased their survival 2.5 times above that of predator-naive counterparts. Survival of snowshoe hares Lepus americanus and fathead minnows Pimephales promelas from predation has also been shown to be greater where vegetation cover is available and used for predator avoidance by prey, as opposed to open habitat (Gazdewich and Chivers, 2002, Rohner and Krebs, 1996). These studies indicate that there are real survival benefits from the availability and use of vegetation structure in a risky environment, but examples where this is investigated are few.

CHAPTER 3 Lantana and bush rat perceived predation risk

Secondly, in accord with other studies of small forager perceived predation risk (Stokes et al., 2004, Strauß et al., 2008, Dutra et al., 2011), vegetation structure was perceived by bush rats as safer than open ground even though open ground may afford the rats an advantage against ambush predators such as cats and death adders. This may also support the hypothesis that the risk from cats may be poorly understood by the native species such as the bush rat (Carthey and Banks, 2012), since dense vegetation may be riskier for the prey of this species.

The architecture of lantana and bracken are dissimilar, particularly the strength and alignment of stems. Since there was no difference in bush rat GUD between the two species it suggests that architecture is not perceived as an important safety feature of the vegetation. However, bush rats apparent preference for lantana may have been explained had I been able to test the effect of the cavity: low architecture pattern on perceived predation risk as this feature may provide a dual advantage of escape facilitation as well as visual and physical protection from predators. There is scope for further understanding of whether vegetation architecture changes perceived predation risk. Garden (2007) suggests that variation in prey responses to vegetation structure is related to how it obstructs their movement and impedes escape from predators, for example animals of larger body size will be impeded by taller vegetation than smaller animals. However, predators, which are often larger than their prey, are likely to be restricted by the same structure. If the role of vegetation structure in facilitating evasion of predators by foraging prey is more important than its role in facilitating avoidance of detection by predators, as suggested by Verdolin (2006), it raises the question of whether prey can distinguish between vegetation structures which impede predators versus structures which merely obstruct detection. Vegetation architecture may play an important but as yet unexplored role in this regard.

Bush rat preference for lantana may be because they use lantana for reasons other than safety. In a seminal study on the bush rats’ life history, Warneke (1971) postulated that dense cover may be important due to its correlation with greater food resources, conditions for burrow establishment and also as a buffer from weather conditions, though this discussion was derivative rather than empirically based.

CHAPTER 3 Lantana and bush rat perceived predation risk

Lantana does not appear to preclude burrow development as all three burrows found in weedy sites during previous research (Chapter 2) were situated underneath lantana. However, this is unlikely to provide an alternate hypothesis for bush rat use of lantana, as burrow placement is influenced primarily by edaphic qualities (Taylor, 1961, White et al., 1996). The horizontal stem structure of lantana may also provide a stable microclimate and a clear pathway allowing ease of movement and aiding the conservation of energy. Again this is unlikely to fully explain their use of lantana since energetic costs of foraging for rodents are generally considered to be lower than predation costs (Brown et al., 1994), especially in warm temperate environments such as the NSW Central Coast.

When travelling under lantana, bush rats select deeper patches of leaf litter (Chapter 2), which may indicate foraging and provide an alternate explanation for their use of lantana. Weeds can advantage invertebrate species by providing food resources and structures for ovopositing (Graves and Shapiro, 2003). Snails, known to be eaten by black rats Rattus rattus (Pers. com. Shea, 2012) a species with similar dietary habits as the bush rat (Dickman and Watts, 2008, Lunney, 2008) may also benefit from the microclimate underneath lantana thickets and may therefore be a food source for bush rats (Stokes et al 2007). However, lantana plants are unlikely to be eaten by bush rats as they are toxic to rodents and other mammals (Sharma et al., 1981, Sharma et al., 1988, Pass et al., 1979). There is also some limited evidence to suggest that there may be a loss of or change to the soil fungal component in association with lantana infestations (Australian Government, 2009, Shaukat and Siddiqui, 2001) which is an important food source for bush rats (Vernes and Dunn, 2009, Tory et al., 1997), and overall lantana reduces biodiversity (Coutts-smith and Downey, 2006a, Thorp and Wilson, 1998 onwards) which may reduce food availability through the loss of plant and insect diversity. The apparent advantage lantana gives to bush rat populations, especially where it grows densely at the urban edge, is at least partially explained by its capacity to reduce perceived predation risk via its structure. Other resources such as food and ease of access may also be of some benefit, which further investigation (for example faecal analysis) could confirm.

CHAPTER 3 Lantana and bush rat perceived predation risk

GUD Challenges Three challenging issues were encountered during this experiment, which exemplify some of the potential limitations of GUD techniques identified in a recent evaluation of GUD processes (Bedoya-Perez, 2013). Firstly, although I set up to test bush rat use of food trays, my trays had visits by seven other different species. Non-target use of trays is a potentially common problem in the field use of GUD techniques that is only rarely acknowledged. GUD techniques have been used extensively since the 1980’s, but only recently has the advent of remote sense-ing camera technology enabled a more accurate assessment of exactly which species are visiting and which species leaves the GUD that is measured. Older methods such as tracking boards (Stokes et al., 2004, Strauß et al., 2008), or use of circadian timing to target certain species (e.g. assuming isolation of species due to diurnal vs nocturnal habit (Brown, 1988)) are likely to be inaccurate. For example, tracking boards might not record bird species landing directly on the sample unit. By using remote sense-ing cameras I was able to detect non-target animal visitation and exclude results attributable to them.

Secondly, the state of the forager is another important issue to consider in GUD experiments (Bedoya-Perez, 2013). An individual’s response to predation risk is known to be influenced by social factors (Randall and Rogovin, 2002, Owings and Coss, 2007, Hersek and Owings, 1993, Halle, 1988) demographic status (age, gender, reproductive state) (Ylonen and Brown, 2007, Bertram and Leggett, 1994) and morphology and fitness (Schwanz et al., 2012, Clark, 1994). In field experiments it is generally not possible to control gender, age or size, which will vary within a real population. In this study in order to maintain independence, sample units were spatially separated within a heterogeneous landscape, consequently the different populations would have been shaped by unique circumstances; the profile in terms of forager state was thus likely to be different in each case. Also the resource richness of the surrounding matrix can also affect GUD (Searle et al., 2008) due to increasing or decreasing missed opportunity costs, and this would be differ at each site.

Lastly the ‘preference’ (lowest GUD) for vegetation type varied within site from night to night, which provided ‘noise’ in the data. This could have been a reflection of the risk allocation hypothesis (Lima and Bednekoff, 1999), where changing types of threat

CHAPTER 3 Lantana and bush rat perceived predation risk

elicit different responses from well informed prey, however, this oscillation occurred 62.5% of the time, indicating that perhaps something is driving the phenomenon. It may be a consequence of the fact that the resource is artificially provided and is predictable in space and time, which may affect predator-prey interactions (Bedoya- Perez, 2013). In the context of the predator-prey game possibly the predictability of returning to the same forage tray on consecutive nights is perceived as risky by prey, especially as any odour left on the first night may inform predators of their presence (Hughes and Banks, 2010). However, this would contradict other evidence suggesting that predators do not respond to fine scale prey behaviour (Searle et al., 2008). Behavioural traits of the forager have been identified as a potential limitation of GUD techniques (Bedoya-Perez, 2013); investigation of small mammal patch use over longer sequential foraging periods would address this issue.

The problems encountered during the design and implementation of this experiment do not invalidate the result, but potentially decreases the chance of detecting a significant difference due to increasing the variation of between site responses. They also support Bedoya-Perez et als’ (2013) findings which highlight the need for further research to resolve some of the questions these issues raise regarding the application of Giving Up Density techniques in the field.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rat at the urban edge?

ABSTRACT In this chapter I investigate whether the presence of black rats (Rattus rattus) alters the habitat quality for bush rats (Rattus fuscipes) at urban edges. The black rat, an invasive synanthropic species, may be more prevalent at the urban-bushland edge due to processes such as spill over from the urban matrix. Recent hypothesis predict that these processes would lead to competitive dominance of black rats over bush rats as competition between the two species is symmetrical except where residency predicts dominance (the residency hypothesis) (Stokes et al., 2009b, Stokes et al., 2009a, Stokes et al., 2012). I explored this possibility using camera traps to examine patterns in abundance and activity of both species across urban edge, core weedy and core habitats. Although the absence of black rats at all core sites may have been due to biotic resistance from resident bush rats, there was no clear evidence that higher abundance of black rats resulted in competitive dominance at the urban edge. Consequently it is unlikely that black rats are a cause of lower habitat quality for bush rats at urban edges. Lack of support for the residency hypothesis may have been due to the patchy nature of black rat distribution which lead to gaps in the data set. Alternately lantana, which is denser than native vegetation, may have mediated competition by providing visual and physical barriers to aggressive encounters. These results should be considered with caution, however, as due to the short term nature of the study they reflect this inter specific relationship at only one point in time. Longer term studies and examination of demographic indicators are required for better understanding of black and bush rat associations at the urban edge.

INTRODUCTION In Chapter 2 I found evidence to support the idea that weedy urban edges were of poorer quality habitat for bush rats compared to core patches, based on both population and demographic properties of the trapped population. In Chapter 3 I

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? showed that bush rats perceive weeds (Lantana) as having lower predation risk, which is usually a property enhancing habitat quality rather than reducing it. In this Chapter I investigate whether weedy urban areas were of lower quality because of inter-specific competition by the introduced and invasive black rat Rattus rattus, a symmetrical competitor of the bush rat (Stokes et al., 2009a) known to be associated with urban development (Watts, 1983, Dickman and Watts, 2008).

Competition between species is a fundamental ecological mechanism that shapes and limits populations, drives community dynamics and defines biodiversity as species adapt to gain an advantage or survive in the struggle for resources. An often-cited key trait of many invasive species is an ability to outcompete similar endemic species (Mooney and Cleland, 2001, Sax and Brown, 2000, Human and Gordon, 1996). Where two coexisting species have the same requirements, the competitive exclusion principle (Gause, 1934) dictates that one species will displace the other. This might be because one is more efficient than the other, termed exploitation competition (Petren and Case, 1996, Bøhn and Amundsen, 2001, e.g. Alatalo et al., 1987), or they may be better placed to win aggressive interactions, termed interference competition (Rowles and O’Dowd, 2007, van Riel et al., 2007, e.g. Human and Gordon, 1999, Piper and Catterall, 2003). Over evolutionary time frames, the disadvantaged species will counter this pressure by adapting physically or behaviorally in order to access alternate resources leading to niche separation (Rosenzweig, 1981). However, within ecological time frames, encompassing periods not long after invasion by alien species, competition has the potential to be intense.

It is only recently that attention has been focused on the role of the invasive black rat in mediating ecological processes in Australia (for a review see Banks and Hughes, 2012). In the past its impact on native species in terms of displacement was thought to be relatively benign (Watts, 1983, Watts and Aslin, 1981). However, recent investigations (Stokes et al., 2009b, Stokes et al., 2009a, Stokes et al., 2012) into spatial separation of the two species found that it was the resident species that was competitively dominant. I will henceforth refer to this as the ‘residency hypothesis’. Stokes (2009a) found that black rats were associated with disturbed habitats such as camp sites; this is characteristic of the species which is known to thrive in

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? anthropogenically disturbed or built environments, whereas bush rats, though capable of persisting in disturbed habitats, were associated with undisturbed areas.

The urban edge can be a point of incursion for invasive species, though the process is best understood for plants. Alien plant species that originate in the urban matrix often have the opportunity and conditions to invade bushland from the urban edge. For example, plant propagules can be generated in artificially large volumes in the urban matrix (Holle and Simberloff, 2005) and can be spread by transport to or along the urban-bushland edge(Brooks, 2007). Disturbance of natural vegetation and changed abiotic conditions at the urban edge can then create conditions that favour establishment of alien weed species (Scherer-Lorenzen et al., 2007, Levine et al., 2003), which consequently are often characteristic of such edges (Alston and Richardson, 2006, Hansen and Clevenger, 2005, Smith and Smith, 2010).

By changing the physical properties of the edge weeds can become an inherent part of edge processes, prompting cascading effects that may alter ecosystem processes (Charles and Dukes, 2008) such as the mechanics of seed dispersal (Cadenasso and Pickett, 2001), or behaviour of fauna (Dutra et al., 2011). While there are fewer reports of invasive fauna causing or mediating edge effects, one well-documented example is the effect of the argentine ant Linepithema humile. This aggressive ant species alien to southern California USA was shown to invade natural areas there from the urban matrix, displacing several native ant species by interference and exploitation competition, and as such functioning as a biotic edge effect (Suarez et al., 1998, Bolger, 2007). The invasive black rat may compete with bush rats at urban edges since both species have similar dietary and habitat requirements.

Four key processes may lead to increased abundance of black rats at the urban edge. Firstly access to spatially separate resources (Ries et al., 2004a) may advantage the species given that they can utilise urban as well as bushland resources (Dickman and Watts, 2008, Weerakoon, 2011) such as artificial nest sites (for example roofs or buildings), or supplementary food (Beckmann and Berger, 2003). Secondly the black rat, being a synanthropic species, may be pre adapted to characteristics of the urban edge such as disturbed ground (Stokes et al., 2009b), and an ability to utilise urban

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? resources. For example, using a bait marker and trapping study, Weerakoon (2011) found that 70% of black rats caught in edge habitats of bushland areas in Sydney (NSW) had visited the urban matrix within the last two weeks. Thirdly the presence of lantana, another invasive species, could advantage them by a mutualistic relationship known as invasion meltdown (Simberloff and Von Holle, 1999). Fourthly, populations may simply spill over (Rowley, 1994) from the urban matrix into edge habitat.

Spill over of mobile organisms has been documented across human altered and natural habitats such as the agricultural – bushland boundary (Blitzer et al., 2012, Rand et al., 2006) and is a well known effect in marine sanctuaries where aquatic species spill over to and supplement adjacent fishing zones (Russ et al., 2003, McClanahan and Mangi, 2000). Urban areas may provide reservoir populations of black rats which act as a source for edge habitats whereas bush rats are seldom reported from the urban matrix and in Chapter 2 I found that bush rat abundance was greatest in core areas at least 200m from the edge. Therefore if recruitment to the edge occurs for bush rats it is from a more dispersed source. This unequal reinvasion potential driven by spill over of black rats from the urban matrix into the adjacent bushland may allow black rat residency to be established and defended at the urban edge.

If the black rat is likely to prevail over rush rats at the urban edge, the converse is likely to be true in core habitats as native species are advantaged by their long term incumbency and niche establishment (Shea and Chesson, 2002).

In this chapter I use remote sensing video cameras to census bush and black rat populations at weedy edge, weedy core and core sites to test predictions of the competition hypothesis as an explanation for the apparent poor quality of urban edges for bush rats. These macrohabitat types were used in order to tease apart fauna responses to weeds as opposed to other edge properties (as discussed in Chapter 2). Camera traps have become widely established as a method for estimating abundance (e.g.Garrote et al., 2012, Silver et al., 2004, Rowcliffe et al., 2008, Karanth and Nichols, 1998) and reliably predict black rat abundances in habitat similar to mine (Weerakoon, 2011). For example, an index created for black rat populations in Sydney from remote sensing camera footage reliably revealed population differences associated with

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? macrohabitat types (Martin and Banks, 2010). Cameras also allow a temporal aspect to be explored which is important when examining inter specific competition since temporal partitioning of the same resource can indicate interference competition (Kronfeld-Schor and Dayan, 2003, Harrington et al., 2009, Carothers and Jaksić, 1984) and/or explain co-existence (Kronfeld-Schor and Dayan, 1999, Kronfeld-Schor and Dayan, 2003). They are also advantageous when aiming to compare the abundances of different species that would otherwise require different sampling techniques (as black rats tend to avoid Elliott traps while bush rats use cage traps less (Stokes, 2013)). I will assume here that frequency of independent visits correlates with abundance, and that this correlation is standard between the replicates (sites) being compared. Between site variation caused by extrinsic factors is not as important though if the aim is to measure relative rather than absolute abundance. In this study one site from each macrohabitat type was sampled concurrently to standardise these factors as much as possible. Another important assumption is that that population densities are sufficiently large, or resources sufficiently scarce to engender competition (Wiens, 1977).

Aim My specific aim is to test the Bush/Black Rat residency hypothesis to see whether it explains poor habitat quality at the urban edge for bush rats due to inter specific competition from black rats which are advantaged by urban edge effects. I also aim to examine the role that the weed Lantana plays in the relationship. I will do this by examining the association of abundance indices between the two species at macro scale at the weedy urban edge, core weedy and core habitats to look for negative associations supporting the residency hypothesis. Competition manifests at different scales, so I will also look for evidence of fine scale habitat partitioning and temporal partitioning which may indicate interference competition.

Hypothesis The residency hypothesis is that black bats, being synanthropic, will have a competitive advantage over bush rats in gaining residency at the urban edge, partly because they are subsidised by and will spill over from the urban matrix.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

Prediction I predict that there will be a negative association of abundance indices for black and bush rats at macrohabitat scale, with relatively high black rat abundance indices and relatively low bush rat abundance indices at the urban edge. The converse will be true in core habitats, and either may occur in core weedy habitats. Where both species are present in a patch, inter-specific competition should be reflected by fine scale spatial, habitat or temporal partitioning, especially in weedy patches.

METHOD Three macrohabitat types were sampled weedy edge, weedy core, and core (defined in Chapter 2). The target weed species was lantana Lantana camara. The same sites were used as for earlier trapping and spooling (Figure 2.2). Six replicates (n = 18 sites) were sampled in the austral winter at each site five baits were placed at 12m intervals along a transect; a bait consisted of a cardboard square (20 x 30cm) soaked in peanut oil anchored to the ground using two golf tees. Each bait was monitored by a remote sensing infra red camera (Scoutguard SG550v) set approximately 1m away at a height of approximately 1m, held by a wooden stake or extant vegetation stems (Figure 4. 1). Variation was necessary to avoid disturbance of the vegetation. For examination of fine scale habitat use, eleven microhabitat variables were recorded at each camera trap for a 1m radius of the bait following the methods described in Chapter 2 (Table 2.1); vegetation architecture type: cavity: low, cavity: tall, homogeneous, fallen tree, leaf litter %cover, vegetation density (0-3method), vegetation height, lantana %cover, lantana density, big logs (termed ‘logs’), big trees (termed ‘trees’), distance to suburb edge (at edge sites only). The set of variables sampled represents a condensed sample from that used in Chapter 2 to target those most likely to predict presence. Cameras were set to video mode and to record for 60 seconds when triggered by movement with a 0 second gap between recordings, 3M pixel 640*480 quality and high sensitivity and recording was run over three nights. A trio of sites, one of each macrohabitat type was sampled concurrently. Sampling started and ended at midday. Each video was labeled by site, macrohabitat type, position on the transect and night (N1, N2, N3). Videos were scored by recording the presence or absence of fauna which was identified to species, and the number of individuals if more than one of the same species was recorded in one video. Bush and black rats are physically similar and were

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? identified according to the criteria developed by Martin and Banks (unpublished (2010) which uses video recordings to fully observe key diagnostic features such as tail length, thickness, face shape, ear position and behaviour. I scored 100% in a blind test conducted a priori where I viewed videos of randomly presented pre-identified black and bush rats.

Figure 4.1 Position of bait (peanut oil soaked cardboard) and camera (Scoutguard 550v) for camera trapping surveys of bush rat and black rat activity in edge core and weedy habitats. Photo by W. Gleen.

DATA ANALYSIS

Method To generate an index of abundance, it was necessary to define an independent visit by an animal from a long foraging event (i.e. consecutive videos of the same animal repeatedly triggering the camera). The time intervals between consecutive con specific

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? visits were examined for both species separately as a frequency histogram (Figure 1). Visual examination indicated there was a demarcation following the two minute interval for both species. Consequently a time interval of three minutes or more was chosen as representing an independent visit.

300

250

200

150 Bush Rat Black Rat 100

3 1 2 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18 19 20 >30

Figure 4.2 Histogram of the frequency of time intervals between visits to the bait (in minutes). A demarcation following the two minute interval is visible.

An index of abundance for black and bush rats respectively was derived for each replicate (site) by pooling the number of independent visits for all 5 cameras on the transect. Only first night data was used as it may be the most accurate measure of true abundance over consecutive rodent activity periods (Weerakoon, 2012).

Bush and Black Rat abundance indices and macrohabitat type Black rat indices of abundance for each site were compared across macrohabitat type using a Wilcoxon test in JMP v. 9 (SAS Institute, 1989-2010), and Bush Rat indices of abundance were compared using ANOVA and the same software. Separate analyses were used due to the patchy nature of the black rat data which could not be transformed to suit the requirements of a parametric test. Data adjusted to account

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? for non-trappable area (Chapter 2), was also analysed, however, the result did not differ in any meaningful way from unadjusted data. Until this method can be tested against known data, use of unadjusted data may be more accurate, and was used henceforth. Where results were non significant power analyses were run using Russ Lenth’s Java applets for Power and Sample size (Lenth, 2006-9) to confirm that the experiments had the power to detect biologically meaningful differences. The detectable contrast was expressed as a percentage difference of the core mean.

Relationship between black and bush rat abundance indices Interference competition can often lead to patterns of association that can be difficult to identify with simple linear regression, especially when a subordinate species may only show responses to the dominant competitor when it is at higher densities, while at lower densities, the numbers of the subordinate species responds to other environmental features. Thus to test association between black rat abundance indices and Bush Rat abundance indices (y) I used a quantile regression using the rq() function, part of the quantreg package in R (R Core Team, 2012). As explained and demonstrated by Johnson and VanDerWal (2009), significant associations between species may be missed by examining simple linear relationships whereas a quantile regression yields more detailed information by accounting for associations at either end of the abundance scale.

Habitat associations of each species To examine the habitat associations of each species in weedy habitats I fitted a Generalised Linear Mixed Model (GLMM) to the bush rat and black rat abundance indices from edge and core weedy sites only. A Poisson model with logarithmic link was selected allowing for over-dispersion of the count data. A random term for ‘site' was included in the model. The initial fixed effects in the model were vegetation architecture type (cavity: low, cavity: tall, homogeneous, fallen tree), leaf litter cover, vegetation density, vegetation height, lantana cover, lantana density, logs, trees, macrohabitat type. Backwards elimination was used to eliminate redundant variables. A significance threshold of P < 0.05 was used. The GLMM was fitted using GenStat Release 15 (www.vsni.co.uk). The variable ‘distance to edge boundary’ was analysed separately as it was only relevant to edge sites; the association was examined using a

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? scatter plot of bivariate fit in JMP v. 9 (SAS Institute, 1989-2010) and tested using Pearsons correlation. Where results were non significant power analyses were run using Russ Lenth’s Java applets for Power and Sample size (Lenth, 2006-9) to confirm that the experiments had the power to detect biologically meaningful differences. The detectable contrast was expressed as a percentage difference of the core mean. It was not possible to run a power analysis for data analysed by quantile regression since there is no appropriate test for this statistic type.

Temporal separation If one species was competitively dominant over the other, the subordinate species might delay visits to places with the dominant in order to avoid potentially dangerous encounters. To test this and determine whether a black rat visit delayed bush rat visits or vice versa, each black rat visit was used as a reference point, then the time (in minutes) to the last previous bush rat visit (indicating black rat response) and the time to the next bush rat visit (indicating bush rat response) was compared. This method of considering temporal separation accounts for the comparative rarity of black rat visits. Only the first night of recording was examined due to low visitation rates over the following nights. Where there were two animals of the same species in the frame it was recorded as one event. The results were compared with a paired t-test in JMP v. 9.

Species might also partition space use over a longer time frame so as to avoid competitive interactions. To investigate this I examined activity patterns - the number of visits in each hour interval - for both species separately, pooling results across site for the full 72 hr sampling period. A Kolmogorov-Smirnov two sample test for large samples (Sokal and Rohlf, 1995) was used to determine whether distributions of times of activity for each species were from populations with the same distribution. Where results were statistically non significant, power analyses were run using Russ Lenth’s Java applets for Power and Sample size (Lenth, 2006-9) to confirm that the experiments had the power to detect biologically meaningful differences. The detectable contrast was expressed as a percentage difference of the core mean.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

RESULTS

Bush and Black Rat abundance indices and macrohabitat type Black rats were patchily distributed across all macrohabitats and only occurred at only one third of all sites. As a result there were no significant differences in black rat camera visits between macrohabitat types Chisquare (2) = 1.77, p = 0.41, n = 18, despite the fact that there were no black rats at core sites with one exception. Visits by bush rats also didn’t differ significantly between macrohabitat types; F (2) = 3.25, p = 0.07. A power analysis showed that the experiment had the power to detect a 135% difference from the core mean for bush rats, which is a biologically meaningful result however the experiment had the power to detect a 2,580% difference from the core mean for black rat visits which is not a biologically meaningful result.

Relationship between black and bush rat activity There was high variation in bush and black rat activity across the suite of sites but no obvious pattern in their relationship as the quantile regression did not result in any significant regression lines, Table 4.1. Bush rats were present at all sites and both species were present at six out of eighteen sites.

Regression line T P 0.05 0.49 0.63 0.1 0.39 0.7 0.25 0.24 0.81 0.5 0.09 0.93 0.75 -0.28 0.78 0.9 -0.40 0.70 0.95 -0.73 0.48

Table 4.1 Summary statistics for a quantile regression of the number of independent bush rat and black rat visits per site (for the first night of sampling only). N = 18.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

Habitat associations of each species There was a positive association with bush rat activity and lantana cover and the vegetation architectures cavity: low, fallen tree and cavity: tall and a negative association with the number of logs, Table 4.2. Black rat activity was also positively associated with the cavity: low vegetation architecture and also with lantana, but with ‘density’ instead of ‘cover’. Habitat features different to those used by bush rats were leaf litter cover and the number of logs (a positive instead of negative association), Table 4.3.

There was no significant association between bush rat activity and distance to the suburb edge r = 0.12, P = 0.56, N = 27, or black rat activity and distance to the suburb edge r = - 0.03, P = 0.89, N = 27.

Variable d.f. SE SED F Fpr Remission Back transformed coefficient means (on the original scale) cavity: 1 NA 0.38 5.71 0.021 NA 0: 0.00 low 1: 0.91 logs 1 0.68 NA 5.61 0.022 -1.614 NA cavity: 1 NA 0.4 5.61 0.022 NA 0: 0.00 tall 1: 0.96 fallen 1 NA 0.005 7.69 0.008 NA 0: 0.00 tree 1: 1.51 lantana 1 0.005 NA 5.78 0.020 0.01168 NA cover

Table 4.2 Summary of GLMM statistics on habitat associations of bush rats. N = 56

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

Variable d.f. SE SED F Fpr Remission Back transformed coefficient. means (on the original scale) cavity: 1 NA 0.50 16.0 <0.00 NA 0: 0.00 low 9 1 1: 1.99 litter 1 0.74 NA 27.0 <0.00 0.73 NA 6 1 logs 1 0.74 NA 13.0 <0.00 2.66 NA 2 1 lantana 3 17301 NA 8.53 <0.00 NA 0: -19.75 density (Avera 1 1: 0.00 ge) 2: 1.99 3: 3.69

Table 4.3: Summary of GLMM statistics on habitat associations of black rats. N = 56.

Temporal Separation Where both species coexisted at a site, distribution in a scale of hours was different between the two population samples; unsigned difference D = 1.82 > D.05 = 0.038. N = 72 (Figure 4.3), these different nightly activity patterns suggest avoidance behaviour within site which can be an indication of interference competition.

At a fine scale (examining time elapsed between con-specific visits) there was also temporal separation between species, which was absolute; i.e. there were no instances where bush and black rat visits overlapped. There was no difference in time interval between bush rat visits following black rat presence or black rat visits following bush rat presence; t (18) = 0.4, p = 0.69 suggesting that there wasn’t any clear dominance of one species over the timing of a later visit by the other species. A power analysis showed that the experiment had the power to detect a 46.49% difference from the core mean (a biologically meaningful result)

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

8 bush rat black rat

6 Percentageoftotal activity

1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 Start time of hour interval

Figure 4.3: Percentage of total activity (independent visits) across 72 hours that occurred in each hour interval for bush and black rats. There were a total of 210 bush rat independent visits to cameras at 18 sites and 79 black rat visits at 6 sites.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

DISCUSSION The camera trap results did not provide clear support for the hypothesis that Black Rat populations are subsidised by urban resources at urban-bushland edges, or that spill over is operating strongly from the urban matrix. This hypothesis predicted significant difference in black rat abundance indices between the urban edge and other macrohabitat types but black rats were patchily distributed and not more abundant at the urban edge. As a result, the experiment lacked the power to detect a biologically meaningful difference in black rat abundance indices and there was no clear test of the idea that higher abundance of black rats resulted in competitive dominance at the urban edge. In fact there was no association between black and bush rat abundance indices at macrohabitat scale and at several sites both species were present.

Edge and core weedy sites in this study all had a history of disturbance as did the one core site at which black rats were present (it was discovered post hoc to have had an old picnic site nearby). Notably, there were no black rats in undisturbed core sites, which supports existing understanding that establishment of black rat populations is facilitated by disturbance events (Watts and Aslin, 1981, Cox et al., 2000, Martin and Banks, 2010 unpublished). Alternately, as bush rats were present at all sites, it is possible that black rat absence at undisturbed core sites reflects biotic resistance by resident bush rats, supporting the residency hypothesis (Stokes et al., 2012). It is possible that core weedy habitats were actually the best quality habitat for black rats (as mean abundance indices were greater there), as was the case for bush rats, but that the difference was not detected due to low frequency of site presence, black rats being present at less than one third of all sites.

Fine scale temporal separation of the two species was absolute; while there were occasions where several conspecific individuals were active at the bait concurrently; there were no recordings of the two different species together. This separation does not seem to be a consequence of asymmetrical competitive interactions as there was no difference in the interval following an inter-specific visit for bush or black rats, but rather supports the premise that they are symmetrical competitors. However, this result is another indication that there was no clear residents competitive advantage as Stokes (Stokes et al., 2012) found clear competitive dominance of the resident species.

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge?

Moreover, while the two species had different nightly activity patterns, there was no clear shift in circadian patterns of activity e.g. black rats becoming diurnal in areas with bush rats or vice versa, as can occur under intense inter-specific competition (Di Bitetti et al., 2009, Harrington et al., 2009).

Both rodent species were associated with similar habitat features, but also had some unique preferences, which might also explain their ability to coexist. Both black and bush rats used cavity: low architecture and lantana, although they used lantana slightly differently. It was found bush rat activity was associated with lantana % cover and black rats with lantana density. This supports earlier results (Chapter2, Chapter 3) that indicate that lantana, and also vegetation architecture (particularly the cavity: low category) are important habitat features for bush rats and suggests that they may be universally important for small mammals. Black rat activity was positively associated with leaf litter cover, supporting existing evidence that it is an important predictor of habitat use for this species (Cox et al., 2000). Similarly there was a positive association with logs whereas bush rat activity was negatively associated with logs. This may indicate habitat partitioning, although apparent avoidance of logs by bush rats can be a consequence of a negative correlation with lantana which is an overriding habitat preference (Chapter 2). Bush rat use of ‘fallen tree’ and ‘cavity: tall’ architecture was unexpected since neither architecture influenced bush rat path in the previous study (Chapter 2). It is likely that the association with fallen trees was due to the process of placing the bait which by necessity created a small gap in the vegetation, so this category may have been perceived by the bush rats as cavity: low. Their use of cavity: tall architecture contrasted with Chapter 2 where it was used less than predicted from availability, which was attributed to unstable microclimate due to increased air convection. The cavity: tall category was broadly defined and included any cavity where the vegetation layer began ≥ 30cm above ground level, meaning some cavities within this category were much larger than others. The category was imprecise partly as a consequence of time constraints prohibiting individual measurement of each cavity and partly because there was insufficient occurrence of intermediate strata to warrant definition of finer categories. Function may differ with size of cavity for example while increased air convection may have caused an adverse microclimate,

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? smaller cavities may have been more similar to cavity: low in function consequently there was variation in response. Response differences to cavity height could be clarified with more precise measurement in future.

In summary, while some habitat features were shared, partitioning in time and of some microhabitat features occurred between the two species. Microhabitat partitioning could be also explained by differences in biology driving different habitat selection, but this is unlikely since both species are habitat generalists (Braithwaite et al., 1978, Dickman and Watts, 2008, Lunney, 2008, Pereoglou et al., 2013). It is more likely that the partitioning is a reflection of interference competition (Di Bitetti et al., 2009, Harrington et al., 2009), a process which allows species with similar requirements to co exist (Morris et al., 2000).

It is possible that where the two species coexist, habitat partitioning occurred vertically, but the study design did not allow examination of this aspect as the focus was at ground level due to bush rats’ ground dwelling habit. Black rats are adept climbers (Watts and Aslin, 1981), so it is possible that they focused more on arboreal resources than the bush rat although a local study of black rat habitat use showed that they were most active on the ground (Cox et al., 2000). If vertical partitioning did occur it should not invalidate these results as the arboreal resources are not available to terrestrial bush rats thus are not open to competition, and it is therefore black rat impact at ground level that is most relevant to consider.

Given that bush and black rats appear to co-exist in areas with a disturbance history, it is not immediately clear why the interference competition between these two species noted by Stokes et al. (2012) was not at play. It is possible that differences in temporal and spatial patterns of activity meant that the two species avoided aggressive interactions, although this scenario would also have been a possibility for Stokes’ study system. It is also possible that external factors, such as predation pressure at urban edges, kept rat densities below levels at which competition would become important to necessitate aggressive area defence. Competition is a density-dependent phenomenon, including competition between species, and a competitively inferior species might be tolerated if there is little to be gained from aggressive exclusion. High

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? turnover in edge sites with high predation pressure might also mean that neither species establishes for long enough to gain a residents advantage.

It is also possible that co existence of the two species at edge and core weedy sites is explained by the presence of lantana, which is denser than native vegetation and may provide a physical and visual barrier restricting aggressive encounters. Changes in habitat structure can mediate aggressive inter specific encounters (Brown, 1971) and an increase in habitat complexity reduced aggression between clones of a parthenogenetic gecko Lepidodactylus lugubris (Short and Petren, 2008). If lantana does indeed reduce aggressive encounters this would in turn facilitate the partitioning of time and some habitat features that was observed, which allows co existence (Kotler et al., 1993, Ziv et al., 1993).

Conclusions and implications for future study Neither the spill -over hypotheses (to explain black rat use of urban edges) nor the residency hypotheses (to explain lower bush rats numbers at urban edges) were clearly supported by the camera trapping results. Therefore it appears that poor quality habitat at the urban edge for bush rats (Chapter 2) is not explained by competition from black rats. Lack of clear support (at macro scale) for the residency hypothesis may have been due to the patchy distribution of black rats, specifically a lack of representation in the data set of sites with moderate to high black rat activity (Figure 4.5). Black rat distribution may be characteristically patchy locally since it also occurred in a recent study of the species along Sydney urban edges (Weerakoon, 2011). When studying this species locally in future, compensation for patchy distribution by using larger sample sizes, or prior determination of black rat presence may be helpful. Recognition of site history (which was beyond the scope of this study) may be important in revealing underlying causes for such patchiness, which has important implications for understanding urban black rat ecology.

Additionally, although methods similar to those used in this study have been used successfully to detect patterns in species abundance (Martin and Banks, 2010, Karanth and Nichols, 1998), short term use of camera traps may not give reliable estimates of population density. Moreover, as highlighted in Chapter 2, density can be a misleading indicator of habitat quality (after Van Horne, 1983) and assessment of the fitness of 99

CHAPTER 4 Does competition from black rats lower habitat quality for the bush rats at the urban edge? individuals in the different habitats would give a better idea of the relative quality of habitats for the two species. Demographic data (Chapter 2) suggests that edge habitats were poorer quality for bush rats- if these habitats were of relatively high quality for black rats this result would support predictions of the resident’s advantage hypothesis

Investigation of demographic and temporal aspects (e.g. Stokes et al., 2009b), or manipulation experiments (e.g. Stokes et al., 2012, Maitz and Dickman, 2001) (particularly targeted removal of black rats using a vertical rodenticide system such as developed by the Invasive Animals Conservation and Research Centre (Meek 2012)) would allow greater insight into the black rats role in mediating ecological functions at the urban edge, a subject with many knowledge gaps but which is important to develop (Banks and Hughes, 2012).

These findings support results from Chapter 2 and Chapter 3 demonstrating that invasive species are not necessarily directly detrimental to native fauna at urban edges. The question of what drives loss of habitat quality at urban edges for bush rats thus remains unclear and will be addressed in the thesis general discussion.

100

Chapter 5 General discussion

CHAPTER 5 General discussion

SUMMARY OF KEY FINDINGS My overall objectives in this thesis were to firstly fill gaps in our understanding of invasive species impact at Australian urban edges, secondly to test current theoretical frameworks of edge effects and invasion biology and thirdly to assist predictions of the outcomes of reintroduction programs.

I assessed habitat quality for the bush rat at urban edges on the Central Coast NSW, with a focus on the impact posed by two invasive species, the weed lantana, and the introduced black rat. I compared bush rat population and demographic differences between edge, core weedy and core habitats and examined their habitat use at two scales. I found that that the urban edge was relatively poor quality habitat for bush rats and my results fitted predictions of the ecotone/matrix model (Lidicker, 1999) of an ecotone response, which reflects lantana’s role as an emergent property of the edge. Rather than reducing habitat quality, lantana density was positively correlated with bush rat abundance indices at edges indicating that it may ‘seal’ the edge, ameliorating deleterious edge effects. Further, core weedy habitats were of equal habitat quality to undisturbed core habitats for bush rats, and lantana was a predictor of habitat use at a fine scale, supporting the concept that lantana maintains or increases bush rat habitat quality.

I also found that plant architecture, measured as the evenness of vegetation on the vertical plane, was an important predictor of small mammal fine scale habitat use, particular the cavity: low architecture. Homogeneous architecture (vertically even cover) was also preferred and vegetation with taller cavities was used less than predicted from availability. Lantana altered habitat structure by increasing density and by reducing leaf litter depth but generating more consistent leaf litter cover. It also increased the evenness of vegetation growth on the vertical plane. Bush rats preferred to use denser patches and the cavity: low architecture when moving through lantana.

I then used a GUD experiment to test the hypothesis that bush rat perceived lower predation risk under lantana, native vegetation as compared to open ground. I found that vegetation structure reduces bush rat perceived predation risk irrespective of

101

Chapter 5 General discussion species (lantana vs native), which explains why plant structure is often more important than floristics in predicting habitat use by small mammals (Bennett, 1993, Garden et al., 2007, Friend and Taylor, 1985, Spencer et al., 2005).

Finally, I tested predictions of the residency hypothesis (Stokes et al., 2012, Stokes et al., 2009b, Stokes et al., 2009a) that inter- specific competition from resident black rats would lower habitat quality at the urban edge for bush rats. I used remote sensing video cameras to compare bush and black rat abundance indices and to look for evidence of temporal, spatial or habitat partitioning in the same three macrohabitat types used previously (Chapter 2). Contrary to my predictions, there were no strong indicators of competition such as negative associations of population abundance and the two species co existed at several sites. Patchy distribution of black rats may explain the lack of support for the residency hypothesis, or the presence of lantana at edge and core weedy sites may have mediated competition by providing visual and physical barriers, thus facilitating co existence.

Together, the sets of results point to three general issues about the study of invasive species and small mammal habitat use.

1. Spatial considerations in the assessment of habitat quality The apparent contradiction that urban edges (with abundant lantana) were poor quality habitat for bush rats (based on their demography) yet lantana was preferred by bush rats in their movements (based on fine scale spooling data) highlights the importance of scale in determining habitat quality. Consideration of scale is vital in ecological studies (Haythornthwaite and Dickman, 2006, Leiner et al., 2010) and a detailed understanding of bush and black rat habitat use could not have been gained without the use of multiple scales of measurement. Indeed I found that important habitat preferences were revealed at a finer scale even than microhabitat. When I examined path within one microhabitat feature (lantana), a masking effect was discovered as the preference for a path within lantana over rode preference for other habitat features known to be important to bush rats such as the structural complexity represented by small logs and leaf litter depth. In chapter 3 I showed that this preference was likely due to the reduced perception of predation risk. This multi- scaled examination revealed how a weed species increased one (spatial) aspect of

102

Chapter 5 General discussion habitat quality (i.e. safety from predation) and use by native fauna while at another scale it reduces it.

2. The importance of plant architecture A close examination of bush rat movement patterns identified that plant architecture was likely to be important in their choice of habitat, something that has seldom been considered in detail for ground dwelling mammals. The concept of foliage height diversity (FHD) was developed as an explanation for bird habitat use and it was suggested that the positive correlation between FHD and bird diversity was due increased foraging space and facilitation of habitat partitioning (MacArthur and MacArthur, 1961). Mechanisms to explain how FHD explains species distribution and habitat use have since rarely been tested, particularly for non-bird taxa. Use of such habitat structures by small mammals has been reported previously (Frazer and Petit, 2007, Sanecki et al., 2006a) but the specific role of architecture has not been empirically tested. It is likely that bush rats use complex plant architecture because it provides both a clear pathway for escape with visual and physical protection, which suggests that it may be an important mediator of small mammal perceived predation risk. If architecture is the general cue small mammals use to assess safety, it is possible that the alien/native status of the plant species proving such architecture is not relevant to this aspect of habitat use.

3. Taking a mechanistic approach to understanding positive alien-native interactions That bush rats made preferential use of lantana, one of Australia’s worst weeds, suggests a positive ecosystem role for this alien species at certain ecological scales, an idea which runs against the entrenched management dogma that all alien species are bad and need to be eradicated in all situations. Similarly, as bush rats were able to coexist with alien black rats, vertebrate pest control may not always be the most urgent imperative when aiming to conserve local endemic fauna. Lantana may be helpful to certain fauna at the urban edge, and by sealing the edge may prevent or slow the establishment of worse environmental weeds; however, at a landscape scale it threatens biodiversity (Coutts-Smith and Downey, 2006b). Investigation of invasion processes at the urban edge has been well studied (particularly where they apply to plant species), however, the flow on effects such as impacts to native fauna are less well known, apart from a few well documented examples (e.g. Suarez et al., 1998, 103

Chapter 5 General discussion

Bolger, 2007). My results indicated that invasive species do not necessarily cause negative faunal responses at urban edges which is perhaps surprising given the level of threat invasive species pose to endemic flora and fauna (International Union for the Conservation of Nature, 2013). This conclusion is strengthened by the fact that the focus species represented disparate taxa: plant (lantana) and animal (black rat). What was most important, however, was that my results tested the mechanistic basis for the nature of the interactions between native and aliens.

I found the ecotone/matrix model (Lidicker, 1999) to be a useful predictor of bush rat edge response. Lantana was revealed to be an emergent property of the urban edge, with density improving habitat quality there. I used a manipulative experiment to explore the mechanism driving this process and demonstrated that lantana acts as a ‘structural proxy’ for native vegetation in reducing bush rat perceived predation risk. These results add to the body of evidence from American (Mattos and Orrock, 2010, Dutra et al., 2011), and Australian studies (Gosper and Vivian-Smith, 2006, Goth and Vogel, 2002, Virkki et al., 2012) that weeds can advantage some small fauna and have important ecological function. Use of existing theoretical frameworks to guide methods and interpret results enhanced the understanding of which mechanistic processes might be explaining bush rat responses. Conversely, using empirical results to support such frameworks strengthens their development and will ultimately lead to greater predictive power. I advocate this approach for future studies. For example, the ecotone/matrix model used in this study would be strengthened by further studies which use mechanistic approaches such as manipulative as well as observational methods in order to untangle the complex processes which drive population abundance patterns.

An understanding of the causative mechanisms for the lack of habitat quality at the urban edge is required if bush rats are to be returned to urban bush land fragments where they have become locally extinct. I demonstrated that invasive species per se do not explain the drop in quality, so while the cause remains unknown, the weed effect can be discounted. Three key findings from this study point to a likely causative mechanism of negative edge effects for bush rats; firstly lantana grows more densely than native vegetation, secondly lantana density is positively correlated with bush rat abundance at poor quality urban edges (i.e. lantana density appears to reduce edge 104

Chapter 5 General discussion effects), and thirdly dense lantana reduces perceived predation risk. From these findings it can reasonably be inferred that predation risk may be why edges are low habitat quality for bush rats, and it is also possible that lower food quality is also implicated. An initial motivation for this study was to inform a larger study reintroducing the bush rats to Sydney harbour foreshore to block reinvasion by black rats (Banks et al., 2010-2012). My results show that while habitat quality may be low at urban edges for bush rats, they do persist there, and weeds (which are common in the Sydney harbour foreshore) are not necessarily the direct cause of the decline in habitat quality and may in fact improve habitat quality, at least in terms of refuge from predation.

FUTURE DIRECTIONS Every research project inevitably brings new questions to the fore and based on my research I suggest three areas where our understanding is poor and future research would be beneficial.

Firstly, changes to ecosystem function usually result in flow- on effects such as modifying species interactions or resource flow (Fagan et al., 1999, Ewers et al., 2012, Laurance, 2000, Ries and Sisk, 2004). For example amur honeysuckle Lonicera maackii, a species native to Asia, is an established weed in the United States, which thrives on edge environments. This invasive plant species has in turn advantaged an avian seed disperser, the American robin Turdus migratorius , and the composition of dispersed seed was consequently altered for up to 1.5 km away (Watling and Orrock, 2010). The first suggested avenue of future investigation relates to the importance of understanding the wider effects of weed infestation on ecosystem function in Australia. By changing the structure of the environment and supporting bush rat populations (Chapter 2) lantana may cause flow- on effects which may cascade beyond the immediate environment.

For example, I demonstrated that bush rats travel more under lantana than native vegetation (Chapter 2). Given that bush rats cache food (Taylor and Calaby, 1988, Woodside, 1983) it is important to determine the potential for negative changes to native plant species propagation if caching of native seeds occurs under an alleleopathic or suppressive plant such as lantana. Predation rate of native seeds can

105

Chapter 5 General discussion be greater under cover of weeds. As native seed stocks deplete, the weed reproduces with less competition+ leading to a cycle whereby the weed is advantaged by apparent competition (Mattos and Orrock, 2010, Meiners, 2007). Another important question is whether the support of bush rat populations by lantana affects the wider faunal community. For example, although bush rat perceived predation risk did not differ between lantana and native vegetation, this may not necessarily reflect predator success and it is possible that bush rat predators could be affected by increased abundance or decreased accessibility of prey in weedy patches, as can occur in invertebrate populations (Pearson, 2009).

A second area of research I suggest is further investigation into the implications of vegetation architecture as an important habitat feature at a fine scale for small mammals. This has significant implications for increasing our understanding of small mammal habitat use. For example, it may offer an additional or alternative explanation for microhabitat use by fauna to measures of habitat complexity used more traditionally in Australian small mammal studies (e.g. Warneke, 1971, Bennett, 1993, Cunningham et al., 2005, Holland and Bennett, 2009). There is also scope to explore the relationship between vegetation architecture and perceived predation risk since an attempt here (Chapter 3) failed for logistical reasons. Had testing been possible, I had predicted that vegetation architecture – either evenness of vegetation or stem strength and orientation – would explain bush rat strong ‘preference’ for lantana due to mediation of perceived predation risk. Another an important question is whether prey can differentiate between factors that obstruct detection by predators, and those that obstruct access by predators. These are separate mechanisms which have different importance for prey (Verdolin, 2006) and the operation of them may be determined by vegetation architecture. Use of artificial structures may better enable some of these questions to be explored as attempts to use live extant vegetation proved too difficult for practical reasons. Such structures have been used successfully to examine other questions relating to perceived predation risk (Stokes et al., 2004, Mattos and Orrock, 2010, Arthur et al., 2004).

A third avenue for research is to redress the skew towards plants in investigations of species invasion at the edge; by comparison there are relatively few studies involving fauna. Australia has an abundance of invasive fauna species, many of which are listed 106

Chapter 5 General discussion as key threatening processes (NSW Office of Environment and Heritage, 1995) and understanding the role of urban edges in the invasion processes of such species is critical for effective management of feral species for example development of threat abatement plans. The most effective approach would incorporate testing of established invasion hypotheses (Jeschke et al., 2012, Catford et al., 2008) so that predictive models can be developed to the same extent as they have been for edge theory.

MANAGEMENT IMPLICATIONS OF MY RESULTS My results have important implications for weed management practices. Weed management in natural areas falls under the auspice of bush regeneration- a practice designed to remove the weeds and restore the native vegetation, assuming that once the native plants are restored, the native wildlife will follow. An important principle of bush regeneration is the avoidance of unintended negative effects to biota (Wright, 1991), although this aspect has only very recently become a research focus (O'Meara and Jack, 2011, Lindenmayer et al., 2013, Martin and Banks, 2010). My findings empirically support the common understanding that weed structure can provide important habitat for wildlife and should be removed slowly and patchily to allow structural replacement to occur by native vegetation. Further, my results highlight the need for attention to structure when planting vegetation at the urban edge; it may be particularly important at the urban edge to replace invasive species with native plants carrying the same traits (Gosper and Vivian-Smith, 2006). As habitat quality is improved by dense plant structure particularly where it grows along the edge, structural augmentation may be a useful consideration where urban sensitive species require protection from edge effects. Fast – growing species, those with a lateral stem structure and tendency to form a low cavity similar to lantana would be most likely to maximise the effect. It is also possible that dense vegetation with strong architecture may ameliorate perceived predation risk by discouraging, or providing a buffer against anthropogenic disturbance.

My results also have implications for the conservation assessment of habitat quality from floristic composition. Vegetation condition assessment methods used for offset programs such as the NSW biodiversity banking and offset scheme (Department of Environment and Climate Change, 2008) incorporate consideration of weed extent 107

Chapter 5 General discussion

(Parkes et al., 2003, Gibbons and Freudenberger, 2006), which is considered to reduce land quality by impacting biodiversity. My result clearly demonstrates that weedy patches may have valuable role acting as a reservoir for native fauna, and that vegetation structure and function are important indicators of land quality as well as composition, and should be included in vegetation assessment methods (Gibbons and Freudenberger, 2006).

108

REFERENCES ACHHIREDDY, N. R. & SINGH, M. 1984. Allelopathic effects of lantana (Lantana camara) on milkweedvine (Morrenia odorata). Weed Science, 32, 757-761. ALATALO, R. V., ERIKSSON, D., GUSTAFSSON, L. & LARSSON, K. 1987. Exploitation competition influences the use of foraging sites by tits: experimental evidence. Ecology, 284-290. ALONSO-AMELOT, M. E. & AVENDANA, M. 2002. Human carcinogenesis and bracken fern: a review of the evidence. Current Medicinal Chemistry, 9, 675-686. ALSTON, K. P. & RICHARDSON, D. M. 2006. The roles of habitat features, disturbance, and distance from putative source populations in structuring alien plant invasions at the urban/wildland interface on the Cape Peninsula, South Africa. Biological Conservation, 132, 183-198. ANDERSON, L. & BURGIN, S. 2008. Patterns of bird predation on reptiles in small woodland remnant edges in peri-urban north-western Sydney, Australia. Landscape Ecology, 23, 1039-1047. ANDREN, H. 1995. Effects of landscape composition on predation rates at habitat edges. In: HANSSON, L., FAHRIG, L., MERRIAM, G. (ed.) Mosaic Landscapes and ecological Processes. London: Chapman and Hall. APLIN, K. P., BROWN, P. R., JACOB, J., KREBS, C. J. & SINGLETON, G. R. 2003. Field methods for rodent studies in Asia and the Indo pacific, Canberra, Australian Centre for International Agricultural Research. ARTHUR, A. D. & PECH, R. P. 2003. The non-lethal impacts of predation on mouse behaviour and reproduction: implications for pest population dynamics In: SINGLETON, G. R., HINDS, L. A., KREBS, C. J. & SPRATT, D. M. (eds.) Rats, Mice and People Rodent Biology and Management. Canberra: Australian Centre for International Agriculture Research. ARTHUR, A. D., PECH, R. P. & DICKMAN, C. R. 2004. Habitat structure mediates the non-lethal effects of predation on enclosed populations of house mice. Journal of Animal Ecology, 73, 867-877. AUGUST, P. V. 1983. The role of habitat complexity and heterogeneity in structuring tropical mammal communities. Ecology, 64, 1495-1507. AUSTRALIAN BUREAU OF METEOROLOGY. 2011. Australian weather calender: moon phases [Online]. Available: http://www.bom.gov.au/calendar/annual/moon.shtml [Accessed 01-12-2011 2011]. AUSTRALIAN BUREAU OF STATISTICS. 2012. Regional Population Growth Australia 2011 [Online]. Available: http://www.abs.gov.au/ausstats/[email protected]/Products/3218.0~2011~Main+Features~M ain+Features?OpenDocument#PARALINK0 [Accessed 25/10/2012 2012]. AUSTRALIAN GOVERNMENT. 2009. Weeds: [online]. Available: http://www.weeds.gov.au. [Accessed 15th October 2009]. AUSTRALIAN MUSEUM. 2009-2012. Animal Species: Bush Rat [Online]. Sydney. Available: http://australianmuseum.net.au/Bush-Ra [Accessed 13-03-2012]. AUSTRALIAN STATE OF THE ENVIRONMENT COMMITTEE. 2012. The State of the Environment 2011: Coasts [Online]. Canberra. Available: The State of the Environment 2011 Committee, 2011) [Accessed 25-10-2012 2012]. BANKS-LEITE, C., EWERS, R. M. & METZGER, J.-P. 2010. Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos, 119, 918-926. BANKS, P. B. 1999. Predation by introduced foxes on native bush rats in Australia: do foxes take the doomed surplus? Journal of Applied Ecology, 36, 1063-1071. BANKS, P. B. & DICKMAN, C. R. 2000. Effects of winter food supplementation on reproduction, body mass, and numbers of small mammals in montane Australia. Canadian Journal of Zoology, 78, 1775 - 1783. BANKS, P. B., DICKMAN, C. R., HAYWARD, M. G., LUNNEY, D., PECH, R. P. & BYROM, A. 2010- 2012. Return of the native: reintroductions, reinvasion and a new paradigm in restoration ecology. Australian Research Council linkage grant. [online]. Canberra.

109

Available: http://researchdata.ands.org.au/the-return-of-the-native-reintroductions- reinvasions-and-a-new-paradigm-in-restoration-ecology. [Accessed 01/03/2013]. BANKS, P. B. & HUGHES, N. K. 2012. A review of the evidence for potential impacts of black rats (Rattus rattus) on wildlife and humans in Australia. Wildlife Research, 39, 78-88. BARKMAN, J. J. 1979. The investigation of vegetation texture and structure. In: WERGER, M. J. A. (ed.) The study of vegetation. The Hague: Dr. W Junk bv publishers. BARKMAN, J. J. 1988. A new method to determine some characters of vegetation structure. Vegetation, 78, 81-90. BARNETT, J. L., HOW, R. A. & HUMPHREY, W. F. 1978. The use of habitat components by small mammals in Eastern Australia. Australian Journal of Ecology, 3, 277-85. BARRATT, D. G. 1998. Predation by house cats, (Felis catus), in Canberra, Australia. II. Factors affecting the amount of prey caught and estimates of the impact on wildlife. Wildlife Research, 25, 475-487. BARTHÉLÉMY, D. & CARAGLIO, Y. 2007. Plant architecture: a dynamic, multilevel and comprehensive approach to plant form, structure and ontogeny. Annals of Botany, 99, 375-407. BATZLI, G. O., HARPER, S. J., LIN, Y. K. & DESY, E. A. 1999. Experimental analysis of population dynamics: scaling up to the landscape. In: BARRETT, G. W. & PELES, J. D. (eds.) Landscape Ecology of Small Mammals. New York: Springer-Verlag. BECKMANN, J. P. & BERGER, J. 2003. Using black bears to test ideal-free distribution models experimentally. Journal of Mammalogy, 84, 594-606. BEDOYA-PEREZ, M. A., MELLA, V.S.A, CARTHEY, A.J.R, MCARTHUR, C, BANKS, P.B. 2013. A practical guide to avoid giving up on giving up densities. Behavioral Ecology and Sociobiology, 67, 1541-1553. BENNETT, A. F. 1993. Microhabitat use by the long-nosed potoroo Potorous tridactylous and other small mammals in remnant forest vegetation of South Western Victoria. Wildlife Research, 20, 267-285. BENNETTS, R. E., DARBY, P. C. & KARUNARATNE, L. B. 2006. Foraging patch selection by snail kites in response to vegetation structure and prey abundance and availability. Waterbirds, 29, 88-94. BENTLEY, J. M. 2008. Role of movement, interremnant dispersal and edge effects in determining sensitivity to habitat fragmentation in two forest-dependent rodents. Austral Ecology, 33, 184-196. BERDOY, M. & DRICKAMER, L. C. 2007. Comparative social organization and life history of Rattus and Mus. In: WOLFF, J. O. & SHERMAN, P. W. (eds.) Rodent Societies, an Ecological and Evolutionary Perspective. Chicago: The University of Chicago Press. BERTRAM, D. F. & LEGGETT, W. C. 1994. Predation risks during the early life history period of fishes: sepparating the effects of size and age. Marine ecology Progress series, 109, 105-114. BIDWELL, S., ATTIWILL, P. M. & ADAMS, M. A. 2006. Nitrogen availability and weed invasion in a remnant native woodland in urban Melbourne. Austral Ecology, 31, 262-270. BLITZER, E. J., DORMANN, C. F., HOLZSCHUH, A., KLEIN, A.-M., RAND, T. A. & TSCHARNTKE, T. 2012. Spillover of functionally important organisms between managed and natural habitats. Agriculture, Ecosystems & Environment, 146, 34-43. BLOUIN-DEMERS, G. & WEATHERHEAD, P. J. 2001. Habitat use by black rat snakes (Elaphe obsoleta obsoleta) in fragmented forests. Ecology, 82, 2882-2896. BOCK, C. E., VIERLING, K. T., HAIRE, S. L., BOONE, J. D. & MERKLE, W. W. 2002. Patterns of rodent abundance on open-space grasslands in relation to suburban edges. Conservation Biology, 16, 1653-1658. BØHN, T. & AMUNDSEN, P.-A. 2001. The competitive edge of an invading specialist. Ecology, 82, 2150-2163. BOLGER, D. T. 2007. Spatial and temporal variation in the argentine ant edge effect: implications for the mechanism of edge limitation. Biological Conservation, 136, 295- 305. 110

BOONSTRA, R. & CRAINE, I. T. M. 1986. Natal nest location and small mammal tracking with a spool and line technique. Canadian Journal of Zoology, 64, 1034-1036. BOONSTRA, R. & KREBS, C. J. 1976. The effect of odour on trap response in Microtus townsendii. Journal of Zoology, 180, 467-476. BOUSKILA, A. 1995. Interactions between predation risk and competition: A field study of kangaroo rats and snakes. Ecology, 76, 165-178. BRADY, M. J., MCALPINE, C. A., MILLER, C. J., POSSINGHAM, H. P. & BAXTER, G. S. 2011. Mammal responses to matrix development intensity. Austral Ecology, 36, 35-45. BRAITHWAITE, R. W., COCKBURN, A. & LEE, A. K. 1978. Resource partitioning by small mammals in lowland heath communities of south-eastern Australia. Australian Journal of Ecology, 3, 423-445. BRAITHWAITE, R. W., LONSDALE, W. M. & ESTBERGS, J. A. 1989. Alien vegetation and native biota in tropical Australia: the impact of Mimosa pigra. Biological Conservation, 48, 189-210. BREED, B. & FORD, F. 2007. Native Mice and Rats. Collingwood, Victoria, CSIRO Publishing. BROOKS, M. L. 2007. Effects of land management practices on plant invasions in wildland areas. Biological Invasions, 147-162. BROWN, J. H. 1971. Mechanisms of competitive exclusion between two species of chipmunks. Ecology, 305-311. BROWN, J. S. 1988. Patch use as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology, 22, 37-47. BROWN, J. S. & KOTLER, B. P. 2007. Foraging and the ecology of fear. In: STEPHENS, D., BROWN, J. S. & YDENBERG, R. (eds.) Foraging Behavior and Ecology Chicago: The University of Chicago Press. BROWN, J. S., KOTLER, B. P., SMITH, R. J. & WIRTZ, W. O. 1988. The effects of owl predation on the foraging behavior of heteromyid rodents. Oecologia, 76, 408-415. BROWN, J. S., KOTLER, B. P. & VALONE, T. J. 1994. Foraging under predation-a comparison of energetic and predation costs in rodent communities of the Negev and Sonoran deserts. Australian Journal of Zoology, 42, 435-448. BULTMAN, T. L. & UETZ, G. W. 1982. Abundance and community structure of forest floor spiders following litter manipulation. Oecologia, 55, 34-41. BURGMAN, M. A. & LINDENMAYER, D. B. 1998. Conservation Biology for the Australian Environment, NSW, Surrey Beatty and Sons Pty Ltd. BUTLER, S. J. & GILLINGS, S. 2004. Quantifying the effects of habitat structure on prey detectability and accessibility to farmland birds. Ibis, 146, 123-130. CADENASSO, M. & PICKETT, S. 2001. Effect of edge structure on the flux of species into forest interiors. Conservation Biology, 15, 91-97. CALLAWAY, R. M. & RIDENOUR, W. M. 2004. Novel weapons: invasive success and the evolution of increased competitive ability. Frontiers in Ecology and the Environment, 2, 436-443. CAROTHERS, J. H. & JAKSIĆ, F. M. 1984. Time as a niche difference: the role of interference competition. Oikos, 403-406. CARTHEY, A. 2011. In Prep. Ph. D. thesis, Sydney University. CARTHEY, A. J. R. & BANKS, P. B. 2012. When does an alien become a native species? A vulnerable native mammal recognizes and responds to its long-term alien predator. PLoS ONE, 7, e31804. CATFORD, J. A., JANSSON, R. & NILSSON, C. 2008. Reducing redundancy in invasion ecology by integrating hypotheses into a single theoretical framework. Diversity and Distributions, 15, 22-40. CHAMBERS, L. K. & DICKMAN, C., R. 2002. Habitat selection of the long-nosed bandicoot, Perameles nasuta (Mammalia, Peramelidae), in a patchy urban environment Austral Ecology, 27, 334-342. CHARLES, H. & DUKES, J. S. 2008. Impacts of invasive species on ecosystem services. In: NENTWIG, W. (ed.) Biological Invasions. Berlin Heidelberg: Springer-Verlag. 111

CLARK, C. W. 1994. Antipredator behavior and the asset-protection principle. Behavioral Ecology, 5, 159-170. CLARKE, K. R. & GORLEY, R. N. 2006. PRIMER v6: User Manual / Tutorial, Plymouth, PRIMER-E Ltd. COGGER, H. G. 1992. Reptiles and Amphibians of Australia, Chatswood, Reed Books. COLLINGE, S. K. 2009. Ecology of Fragmented Landscapes, Baltimore, The John Hopkins University Presss. CORNELIUS, R., SCHULTKA, W. & MEYER, G. 1990. On the invasion potential of alien plant species. Verhandlungen der Gesellschaft für Ökologie, 19, 20-29. COUTTS-SMITH, A. & DOWNEY, P. O. 2006. Impact of weeds on threatened biodiversity in New South Wales. CRC for Australian Weed Management [online]. Adelaide. Available: http://www.canberra.edu.au/centres/iae/pdfs/2006_Coutts-Smith-and- Downey_Weeds-CRC-Tech-no-11.pdf. [Accessed 02/05/2012]. COX, M., DICKMAN, C. R. & COX, W. G. 2000. Use of habitat by the black rat (Rattus rattus) at North Head, New South Wales; an observational and experimental study. Austral. Ecology, 25, 375-385. COX, M. P., DICKMAN, C. R. & HUNTER, J. 2003. Effects of rainforest fragmentation on non- flying mammals of the Eastern Dorrigo Plateau. Australian Bioolical Conservation, 115, 175-89. CUNNINGHAM, R. B., LINDENMAYER, D. B., MACGREGOR, C., BARRY, S. & WELSH, A. 2005. Effects of trap position, trap history, microhabitat and season on capture probabilities of small mammals in a wet eucalypt forest. Wildlife Research, 32, 657-671. DALY, M., WILSON, M. I. & BEHRENDS, P. 1980. Factors affecting rodents' responses to odours of strangers encountered in the field: experiments with odour-baited traps. Behavioral Ecology and Sociobiology, 6, 323-329. DAVIDSON, A. M., JENNIONS, M. & NICOTRA, A. B. 2011. Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta‐analysis. Ecology Letters, 14, 419-431. DAVIS, M. A. 2010. Researching invasive species 50 years after Elton: a cautionary tale. Fifty Years of Invasion Ecology: The Legacy of Charles Elton, 267-276. DENNO, R. F., FINKE, D. L. & LANGELLOTTO, G. A. 2005. Direct and indirect effects of vegetation structure and habitat complexity on predator-prey and predator-predator interactions. . In: BARBOSA, P. & I. CASTELLANOS (eds.) Ecology of predator-prey interactions. Oxford, U.K.: Oxford University Press. DEPARTMENT OF ENVIRONMENT AND CLIMATE CHANGE 2008. BioBanking Assessment Methodology.Report number DECC 2008/385. Sydney: Department of Environment and Climate Change NSW. DEPARTMENT OF SUSTAINABILITY, E., WATER, POPULATION AND COMMUNITIES. 2013. Weeds in Australia: about weeds. [Online]. Canberra. Available: http://www.environment.gov.au/biodiversity/invasive/weeds/weeds/what.html [Accessed March 29]. DI BITETTI, M. S., DI BLANCO, Y. E., PEREIRA, J. A., PAVIOLO, A. & PÉREZ, I. J. 2009. Time partitioning favors the coexistence of sympatric crab-eating foxes (Cerdocyon thous) and pampas foxes (Lycalopex gymnocercus). Journal of Mammalogy, 90, 479-490. DICKMAN, C., R. & WATTS, C. H. S. 2008. Black rat Rattus rattus. In: VAN DYCK, S. & STRAHAN, R. (eds.) The Mammals of Australia. 3rd ed. Chatswood: Reed New Holland. DICKMAN, C. R. 1992. Predation and habitat shift in the house mouse, Mus domesticus. Ecology, 73, 313-322. DIDHAM, R. K. & LAWTON, J. H. 1999. Edge structure determines the magnitude of changes in microclimate and vegetation structure in tropical forest fragments. Biotropica, 31, 17- 30. DOWNES, S. J., HANDASYDE, K. A. & ELGAR, M. A. 1977. Variation in the use of corridors by introduced and native rodents in south-eastern Australia. Biological Conservation, 82, 379-383. 112

DOWNEY, P. O., WILLIAMS, M. C., WHIFFEN, L. K., TURNER, P. J., BURLEY, A. L. & HAMILTON, M. A. 2009. Weeds and biodiversity conservation: a review of managing weeds under the New South Wales Threatened Species Conservation Act 1995. Ecological Management & Restoration, 10, S53-S58. DUGDALE, J. S. 1996. Natural history and identification of litter‐feeding Lepidoptera larvae (Insecta) in beech forests, Orongorongo Valley, New Zealand, with especial reference to the diet of mice (Mus musculus). Journal of the Royal Society of New Zealand, 26, 251-274. DUNSTAN, C. E. & FOX, B. J. 1996. The Effects of Fragmentation and Disturbance of Rainforest on Ground-Dwelling Small Mammals on the Robertson Plateau, New South Wales, Australia. Journal of Biogeography, 23, 187-201. DUTRA, H., BARNETT, K., REINHARDT, J., MARQUIS, R. & ORROCK, J. 2011. Invasive plant species alters consumer behavior by providing refuge from predation. Oecologia, 166, 649-657. EDWARDS, G. P., PREU, N. D., CREALY, I. V. & SHAKESHAFT, B. J. 2002. Habitat selection by feral cats and dingoes in a semi-arid woodland environment in central Australia. Austral Ecology, 27, 26-31. ELS, L. M. & KERLEY, G. I. H. 1996. Biotic and abiotic correlates of small mammal community structure in the Groendal Wilderness Area, Eastern Cape, South Africa. Koedoe - African Protected Area Conservation and Science 39, 121-130. EMBAR, K., KOTLER, B. P. & MUKHERJEE, S. 2011. Risk management in optimal foragers: the effect of sightlines and predator type on patch use, time allocation, and vigilance in gerbils. Oikos, 120, 1657-1666. EWERS, R. M., BARTLAM, S. & DIDHAM, R. K. 2012. Altered species interactions at forest edges: contrasting edge effects on bumble bees and their phoretic mite loads in temperate forest remnants. Insect Conservation and Diversity. FAGAN, W. F., CANTRELL, R. S. & COSNER, C. 1999. How habitat edges change species interactions. The American Naturalist, 153, 165-182. FAIRLEY, A. & MOORE, P. 1989. Native plants of the Sydney district: an identification guide. Kenthurst: Kangaroo Press, in association with the Society for Growing Australian Plants-NSW. 432p.-illus., col. illus., maps.. ISBN, 864172613. FENWICK, G. R. 1989. Bracken (Pteridium aquilinum)—toxic effects and toxic constituents. Journal of the Science of Food and Agriculture, 46, 147-173. FINKE, D. L. & DENNO, R. F. 2004. Predator diversity dampens trophic cascades. Nature, 429, 407-410. FOSTER, J. & GAINES, M. S. 1991. The effects of a successional habitat mosaic on a small mammal community. Ecology, 72, 1358-1373. FRANKLIN, J. F. & LINDENMAYER, D. B. 2009. Importance of matrix habitats in maintaining biological diversity. Proceedings of the National Academy of Sciences, 106, 349-350. FRAZER, D. S. & PETIT, S. 2007. Use of Xanthorrhoea semiplana (grass-trees) for refuge by Rattus fuscipes (southern bush rat). Wildlife Research, 34, 379-386. FREITAS, S. R., CERQUEIRA, R. & VIEIRA, M. V. 2002. A device and standard variables to describe microhabitat structure of small mammals based on plant cover. Brazilian Journal of Biology, 62(4B), 795-800. FRETWELL, S. 1972. Populations in a Seasonal Environment, Princeton, Princeton University Press. FRETWELL, S. D. & LUCAS, H. L. 1969. On territorial behavior and other factors influencing habitat distribution in birds. Acta Biotheoretica, 19, 16-36. FRIEND, G. R. & TAYLOR, J. A. 1985. Habitat preferences of small mammals in tropical open- forest of the Northern Territory. Australian Journal of Ecology, 10, 173-185. FUHRER, B. 2005. A Field Guide to Australian Fungi, Hawthorn, Victoria., Blooming Books. GARDEN, J., MCALPINE, C., PETERSON, A. N. N., JONES, D. & POSSINGHAM, H. 2006. Review of the ecology of Australian urban fauna: a focus on spatially explicit processes. Austral Ecology, 31, 126-148. 113

GARDEN, J., MCALPINE, C. & POSSINGHAM, H. 2010. Multi-scaled habitat considerations for conserving urban biodiversity: native reptiles and small mammals in Brisbane, Australia. Landscape Ecology, 25, 1013-1028. GARDEN, J. G., MCALPINE, C. A., POSSINGHAM, H. P. & JONES, N. J. 2007. Habitat structure is more important than vegetation composition for local-level mnagement of native terrestrial reptile and small mammal species living in urban remnants: A case study from Brisbane, Australia. Austral Ecology, 32, 669-6885. GARROTE, G., GIL-SÁNCHEZ, J. M., MCCAIN, E. B., DE LILLO, S., TELLERÍA, J. L. & SIMÓN, M. Á. 2012. The effect of attractant lures in camera trapping: a case study of population estimates for the Iberian lynx (Lynx pardinus). European Journal of Wildlife Research, 1-4. GAUSE, G. 1934. The struggle for existence. Williams and Wilkins, Baltimore. GAZDEWICH, K. J. & CHIVERS, D. P. 2002. Acquired predator recognition by fathead minnows: influence of habitat characteristics on survival. Journal of chemical ecology, 28, 439- 445. GIBBONS, P. & FREUDENBERGER, D. 2006. An overview of methods used to assess vegetation condition at the scale of the site. Ecological Management & Restoration, 7, S10-S17. GLEN, A., SUTHERLAND, D. & CRUZ, J. 2010. An improved method of microhabitat assessment relevant to predation risk. Ecological Research, 25, 311-314. GODIN, C. 2000. Representing and encoding plant architecture: a review. Annals of forest science, 57, 413-438. GOOSEM, M. 2000. Effects of tropical rainforest roads on small mammals: edge changes in community composition. Wildlife Research, 27, 151-163. GOSPER, C. R. & VIVIAN-SMITH, G. 2006. Selecting replacements for invasive plants to support frugivores in highly modified sites: A case study focusing on Lantana camara. Ecological Management & Restoration, 7, 197-203. GOTH, A. & VOGEL, U. 2002. Chick survival in the megapode Alectura lathami (Australian brush-turkey). Wildlife Research, 29, 503-511. GRAVES, S. D. & SHAPIRO, A. M. 2003. Exotics as host plants of the California butterfly fauna. Biological Conservation, 110, 413-433. GRICE, A. 2004. Weeds and the monitoring of biodiversity in Australian rangelands. Austral Ecology, 29, 51-58. GUARINO, F. 2001. Diet of a large carnivorous lizard, Varanus varius. Wildlife Research, 28, 627-630. HAERING, R. & FOX, B. J. 1995. Habitat utilization patterns of sympatric populations of Pseudomys gracilicaudatus and Rattus lutreolus in coastal heathland: a multivariate analysis. Australian Journal of Ecology, 20, 427-441. HALLE, S. 1988. Avian predation upon a mixed community of common voles (Microtus arvalis) and Wood Mice (Apodemus sylvaticus). Oecologia, 75, 451-455. HANNUNEN, S. 2002. Vegetation architecture and redistribution of insects moving on the plant surface. Ecological Modelling, 155, 149-157. HANSEN, M. J. & CLEVENGER, A. P. 2005. The influence of disturbance and habitat on the presence of non-native plant species along transport corridors. Biological Conservation, 125, 249-259. HAPPOLD, D. C. D. 1998. The subalpine climate at smiggin holes, Kosciusko National Park, Australia, and its influence on the biology of small mammals. Arctic and Alpine Research, 30, 241-251. HARPER, K. A., MACDONALD, S. E., BURTON, P. J., CHEN, J., BROSOFSKE, K. D., SAUNDERS, S. C., EUSKIRCHEN, E. S., ROBERTS, D. A. R., JAITEH, M. S. & ESSEEN, P.-A. 2005. Edge influence on forest structure and composition in fragmented landscapes. Conservation Biology, 19, 768-782. HARRINGTON, L. A., HARRINGTON, A. L., YAMAGUCHI, N., THOM, M. D., FERRERAS, P., WINDHAM, T. R. & MACDONALD, D. W. 2009. The impact of native competitors on an

114

alien invasive: temporal niche shifts to avoid interspecific aggression? Ecology, 90, 1207-1216. HAYSOM, K. A. & COULSON, J. C. 1998. The Lepidoptera fauna associated with Calluna vulgaris: effects of plant architecture on abundance and diversity. Ecological Entomology, 23, 377-385. HAYTHORNTHWAITE, A. S. & DICKMAN, C. R. 2006. Distribution, abundance, and individual strategies: a multi-scale analysis of dasyurid marsupials in arid central Australia. Ecography, 29, 285-300. HEGER, T. & TREPL, L. 2003. Predicting biological invasions. Biological Invasions, 5, 313-321. HERSEK, M. J. & OWINGS, D. H. 1993. Tail flagging by adult California ground squirrels: a tonic signal that serves different functions for males and females. Animal Behaviour, 46, 129-138. HITCHEN, D. J., BURGIN, S., RIDGEWAY, P. & WOTHERSPOON, D. 2011. Habitat use by the jacky lizard Amphibolurus muricatus in a highly degraded urban area. Animal Biology, 61, 185-197. HOBBS, R. J. & YATES, C. J. 2003. TURNER REVIEW No. 7. Impacts of ecosystem fragmentation on plant populations: generalising the idiosyncratic. Australian Journal of Botany, 51, 471-488. 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. HOLLAND, G. J. & BENNETT, A. F. 2009. Differing responses to landscape change: implications for small mammal assemblages in forest fragments. Biodiversity and Conservation, 18, 2997-3016. HOLLAND, G. J. & BENNETT, A. F. 2010. Habitat fragmentation disrupts the demography of a widespread native mammal. Ecography, 33, 841-853. HOLLAND, J. D., BERT, D. G. & FAHRIG, L. 2004. Determining the spatial scale of species' response to habitat. BioScience, 54, 227-233. HOLLE, B. V. & SIMBERLOFF, D. 2005. Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology, 86, 3212-3218. HOLLIDAY, I. E. 2012. Wilcoxon-Mann-Whitney Test (v.1.04). Free Statistics Software (v.1.1.23- r7) [Online]. Available: http://www.wessa.net/rwasp_Reddy- Moores%20Wilcoxon%20Mann-Witney%20Test.wasp [Accessed 30 - 05-2012]. HOPCRAFT, J. G. C., SINCLAIR, A. R. E. & PACKER, C. 2005. Planning for success: Serengeti lions seek prey accessibility rather than abundance. Journal of Animal Ecology, 74, 559-566. HUFBAUER, R. A. & TORCHIN, M. E. 2007. Integrating ecological and evolutionary theory of biological invasions. Biological Invasions, 79-96. HUGHES, N. K. & BANKS, P. B. 2010. Interacting effects of predation risk and signal patchiness on activity and communication in house mice. Journal of Animal Ecology, 79, 88-97. HUMAN, K. & GORDON, D. 1996. Exploitation and interference competition between the invasive Argentine ant, Linepithema humile, and native ant species. Oecologia, 105, 405-412. HUMAN, K. & GORDON, D. 1999. Behavioral interactions of the invasive Argentine ant with native ant species. Insectes Sociaux, 46, 159-163. INTERNATIONAL UNION FOR THE CONSERVATION OF NATURE. 2013. Invasive Species [Online]. Malaga. Available: http://www.iucn.org/about/union/secretariat/offices/iucnmed/iucn_med_programm e/species/invasive_species/ [Accessed 01-03-2013 2013]. JELLINEK, S., DRISCOLL, D. A. & KIRKPATRICK, J. B. 2004. Environmental and vegetation variables have a greater influence than habitat fragmentation in structuring lizard communities in remnant urban bushland. Austral Ecology, 29, 294-304. JESCHKE, J. M., GÓMEZ APARICIO, L., HAIDER, S., HEGER, T., LORTIE, C. J., PYŠEK, P. & STRAYER, D. L. 2012. Support for major hypotheses in invasion biology is uneven and declining. NeoBiota, 14, 1-20.

115

JOHNSON, A. S., HALE, P. E., FORD, W. M., WENTWORTH, J. M., FRENCH, J. R., ANDERSON, O. F. & PULLEN, G. B. 1995. White-tailed deer foraging in relation to successional stage, overstory type and management of southern appalachian forests. American Midland Naturalist, 133, 18-35. JOHNSON, C. N. & VANDERWAL, J. 2009. Evidence that dingoes limit abundance of a mesopredator in eastern Australian forests. Journal of Applied Ecology, 46, 641-646. JOHNSON, S. 2007. Review of the declaration of Lantana species in New South Wales. . In: Weeds unit biosecurity, C. A. M. S., NSW Department of Primary Industries. (ed.). Orange. KARANTH, K. U. & NICHOLS, J. D. 1998. Estimation of Tiger densities in India using photographic captures and recaptures. Ecology, 79, 2852-2862. KIRKMAN, T. W. 1996. Statistics to use [Online]. Available: http://www.physics.csbsju.edu/stats/exact_NROW_NCOLUMN_form.html [Accessed 02/02/2012 2012]. KOLAR, C. S. & LODGE, D. M. 2001. Progress in invasion biology: predicting invaders. Trends in Ecology & Evolution, 16, 199-204. KOLBE, J. J., GLOR, R. E., SCHETTINO, L. R., LARA, A. C., LARSON, A. & LOSOS, J. B. 2004. Genetic variation increases during biological invasion by a Cuban lizard. Nature, 431, 177-181. KONEČNÝ, A., ESTOUP, A., DUPLANTIER, J. M., BRYJA, J., BÂ, K., GALAN, M., TATARD, C. & COSSON, J. F. 2013. Invasion genetics of the introduced black rat (Rattus rattus) in Senegal, West Africa. Molecular Ecology, 22, 286-300. KORSLUND, L. & STEEN, H. 2006. Small rodent winter survival: snow conditions limit access to food resources. Journal of Animal Ecology, 75, 156-166. KOTLER, B. P., BROWN, J. S. & SUBACH, A. 1993. Mechanisms of species coexistence of optimal foragers: temporal partitioning by two species of sand dune gerbils. Oikos, 67, 548- 556. KREBS, C. J. 1999. Current paradigms of rodent population dynamics - what are we missing? In: SINGLETON, G., . HINDS, L., LEIRS, H., ZHANG, Z. (ed.) Ecologically-based Rodent Management. Canberra: Australian Centre of International Agricultural Research. KREBS, C. J., LAMBIN, X. & WOLFF, J. O. 2007. Social behaviour and self-regulation in murid rodents. In: WOLFF, J. O. & SHERMAN, P. W. (eds.) Rodent societies. Chicago: The University of Chicago Press KRISTAN III, W. B., LYNAM, A. J., PRICE, M. V. & ROTENBERRY, J. T. 2003. Alternative causes of edge-abundance relationships in birds and small mammals of California coastal sage scrub. Ecography, 26, 29-44. KRONFELD-SCHOR, N. & DAYAN, T. 1999. The dietary basis for temporal partitioning: food habits of coexisting Acomys species. Oecologia, 121, 123-128. KRONFELD-SCHOR, N. & DAYAN, T. 2003. Partitioning of time as an ecological resource. Annual Review of Ecology, Evolution, and Systematics, 153-181. KUTT, A. 2011. The diet of the feral cat (Felis catus) in north-eastern Australia. Acta Theriologica, 56, 157-169. LACEY, E. A. & SHERMAN, P. W. 2007. The Ecology of Sociality in Rodents. In: WOLFF, J. O., SHERMAN, P.W. (ed.) Rodent Societies:An Ecological and Evolutionary Perspective. Chicago: The University of Chicago Press. LANGELLOTTO, G. A. & DENNO, R. F. 2004. Responses of invertebrate natural enemies to complex-structured habitats: a meta-analytical synthesis. Oecologia, 139, 1-10. LAURANCE, W. F. 1997. Responses of mammals to rainforest fragmentation in tropical Queensland: a review and synthesis. Wildlife Research, 24, 603-612. LAURANCE, W. F. 2000. Do edge effects occur over large spatial scales? Trends in Ecology and Evolution, 15. LAURANCE, W. F., LOVEJOY, T. E., VASCONCELOS, H. L., BRUNA, E. M., DIDHAM, R. K., STOUFFER, P. C., GASCON, C., BIERREGAARD, R. O., LAURANCE, S. G. & SAMPAIO, E. 2002. Ecosystem decay of amazonian forest fragments: a 22-year investigation. Conservation Biology, 16, 605-618. 116

LAURANCE, W. F., NASCIMENTO, H. E. M., LAURANCE, S. G., ANDRADE, A., EWERS, R. M., HARMS, K. E., LUIZÃO, R. C. C. & RIBEIRO, J. E. 2007. Habitat Fragmentation, Variable Edge Effects, and the Landscape-Divergence Hypothesis. PLoS ONE, 2, e1017. LAW, B. S. & DICKMAN, C. R. 1998. The use of habitat mosaics by terrestrial vertebrate fauna: implications for conservation and management. Biodiversity and Conservation, 7, 323- 333. LAWTON, J. 1983. Plant architecture and the diversity of phytophagous insects. Annual review of entomology, 28, 23-39. LEINER, N. O., DICKMAN, C. R. & SILVA, W. R. 2010. Multiscale habitat selection by slender opossums (Marmosops spp.) in the Atlantic forest of Brazil. Journal of Mammalogy, 91, 561-565. LEISHMAN, M. R., HUGHES, M. T. & GORE, D. B. 2004. Soil phosphorus enhancement below stormwater outlets in urban bushland: spatial and temporal changes and the relationship with invasive plants. Soil Research, 42, 197-202. LENTH, R. V. 2006-9. Java applets for Power and Sample Size (Computer Software) [Online]. Available: http://www.stat.uiowa.edu/~rlenth/Power/ [Accessed 19/06/11]. LEVINE, J. M., VILA, M., ANTONIO, C. M., DUKES, J. S., GRIGULIS, K. & LAVOREL, S. 2003. Mechanisms underlying the impacts of exotic plant invasions. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270, 775-781. LIDDICKER, W. Z. J. 1999. Responses of mammals to habitat edges: a landscape perspective – Preface to this Special Edition. Landscape Ecology, 14, 331-331. LIDICKER, W. 1999. Responses of mammals to habitat edges: an overview. Landscape Ecology, 14, 333-343. LIMA, S. L. & BEDNEKOFF, P. A. 1999. Temporal variation in danger drives antipredator behavior: the predation risk allocation hypothesis. The American Naturalist, 153, 649- 659. LIN, Y.-T. K. & BATZLI, G. O. 2001. The influence of habitat quality on dispersal, demography, and population dynamics of voles. Ecological Monographs, 71, 245-275. LINDENMAYER, D. & BURGMAN, M. 2005. Practical Conservation Biology, Victoria, CSIRO publishing. LINDENMAYER, D. B., CUNNINGHAM, R. B. & PEAKALL, R. O. D. 2005. The recovery of populations of bush rat Rattus fuscipes in forest fragments following major population reduction. Journal of Applied Ecology, 42, 649-658. LINDENMAYER, D. B. & FISCHER, J. 2006. Habitat Fragmentation and Landscape Change, An Ecological and Conservation Synthesis., Collingwood, CSIRO Publishing. LINDENMAYER, D. B., MACGREGOR, C., DEXTER, N., FORTESCUE, M. & COCHRANE, P. 2013. Booderee National Park management: Connecting science and management. Ecological Management & Restoration, 14, 2-10. LOCKWOOD, J. L., CASSEY, P. & BLACKBURN, T. 2005. The role of propagule pressure in explaining species invasions. Trends in Ecology & Evolution, 20, 223-228. LUNNEY, D. 1983. Bush Rat Rattus fuscipes. In: STRAHAN, R. (ed.) Complete Book of Australian Mammals.Auckland: Cornstalk Publishing. LUNNEY, D. 2008. Bush Rat, Rattus fuscipes. In: VAN DYCK, S., STRAHAN, R. (ed.) The Mammals of Australia, Third Edition. Sydney: Reed New Holland. LUNNEY, D., MATTHEWS, A., EBY, P. & PENN, A. M. 2009. The long-term effects of logging for woodchips on small mammal populations. Wildlife Research, 36, 691-701. MACARTHUR, R., RECHER, H. & CODY, M. 1966. On the Relation between Habitat Selection and Species Diversity. The American Naturalist, 100, 319-332. MACARTHUR, R. H. & MACARTHUR, J. W. 1961. On bird species diversity. Ecology, 42, 594-598. MACARTHUR, R. H., MACARTHUR, J. W. & PREER, J. 1962. On bird species diversity. II. prediction of bird census from habitat measurements. The American Naturalist, 96, 167-174.

117

MACGREGOR-FORS, I. 2011. Misconceptions or misunderstandings? On the standardization of basic terms and definitions in urban ecology. Landscape and Urban Planning, 100, 347- 349. MACQUEEN, P. E., NICHOLLS, J. A., HAZLITT, S. L. & GOLDIZEN, A. W. 2008. Gene flow among native bush rat, Rattus fuscipes (Rodentia: Muridae), populations in the fragmented subtropical forests of south-east Queensland. Austral Ecology, 33, 585-593. MAITZ, W. E. & DICKMAN, C. R. 2001. Competition and habitat use in native Australian Rattus: is competition intense, or important? Oecologia, 128, 526-538. MANOLIS, J. C., ANDERSEN, D. E., CUTHBERT, F. J. & DUPLESSIS, M. 2002. Edge effect on nesting success of ground nesting birds near regenerating clearcuts in a forest- dominated landscape. The Auk, 119, 955-970. MARQUIS, R. J. & LILL, J. T. 2010. Impact of plant architecture versus leaf quality on attack by leaf-tying caterpillars on five oak species. Oecologia, 163, 203-213. MARTIN, S. 2010. Habitat use by black rats (Rattus rattus) and their response to bush regeneration. Honours thesis, University of NSW. MATLACK, G. R. 1993. Microenvironment variation within and among forest edge sites in the eastern United States. Biological Conservation, 66, 185-194. MATTOS, K. J. & ORROCK, J. L. 2010. Behavioural consequences of plant invasion: an invasive plant alters rodent antipredator behaviour. Behavioural Ecology, Advance Access Publication. MCCLANAHAN, T. R. & MANGI, S. 2000. Spillover of exploitable fishes from a marine park and its effect on the adjacent fishery. Ecological Applications, 10, 1792-1805. MCDONALD, R. I. 2008. Global urbanization: can ecologists identify a sustainable way forward? Frontiers in Ecology and the Environment, 6, 99-104. MCDONALD, W. & ST CLAIR, C. C. 2004. Elements that promote highway crossing structure use by small mammals in Banff National Park. Journal of Applied Ecology, 41, 82-93. MCKINNEY, M. L. 2002. Urbanization, biodiversity, and conservation. BioScience, 52, 883-890. MEEK , P. 2003. Home range of house cats Felis catus living within a National Park. Australian Mammalogy, 25, 51-60. MEEK, P. 2012. Use of vertical bait stations for rodent control [online]. Invasive animals Conservation and Research Centre, Bruce. Available: http://www.youtube.com/watch?v=bplHcas7OAw [Accessed 7/10/2013] MEINERS, S. J. 2007. Apparent competition: an impact of exotic shrub invasion on tree regeneration. Biological Invasions, 9, 849-855. MELLO, F. B., JACOBUS, D., CARVALHO, K. & MELLO, J. R. B. 2003. Effects of Lantana camara (Verbenacea) on rat fertility. Veterinary and Human Toxicology, 54, 20. MELLO, F. B., JACOBUS, D., CARVALHO, K. & MELLO, J. R. B. 2005. Effects of Lantana camara (Verbenaceae) on general reproductive performance and teratology in rats. Toxicon, 45, 459-466. MESSIER, F., VIRGL, J. A. & MARINELLI, L. 1990. Density-dependent habitat selection in muskrats: a test of the ideal free distribution model. Oecologia, 84, 380-385. MILES, M. A. 1976. A simple method of tracking mammals and locating triatomine vectors of Trypanosoma Cruzi in Amazonian Forest. The American Journal of Tropical Medicine and Hygiene, 25, 671-674. MOONEY, H. A. & CLELAND, E. E. 2001. The evolutionary impact of invasive species. Proceedings of the National Academy of Sciences, 98, 5446-5451. MORRIS, D. 1989. Density-dependent habitat selection: Testing the theory with fitness data. Evolutionary Ecology, 3, 80-94. MORRIS, D. W. 1997. Optimally foraging deer mice in prairie mosaics: a test of habitat theory and absence of landscape effects. Oikos, 80, 31-42. MORRIS, D. W., FOX, B. J., LUO, J. & MONAMY, V. 2000. Habitat-dependent competition and the coexistence of Australian heathland rodents. Oikos, 91, 294-306. MORRIS, D. W. & MACEACHERN, J. T. 2010. Active density-dependent habitat selection in a controlled population of small mammals. Ecology, 91, 3131-3137. 118

MOSMAN COUNCIL. 2012. Bush Rats [Online]. Sydney: Mosman council. Available: http://www.mosman.nsw.gov.au/environment/bushland/bush-rats [Accessed 12-03- 2012]. MOSSER, A., FRYXELL, J. M., EBERLY, L. & PACKER, C. 2009. Serengeti real estate: density vs. fitness-based indicators of lion habitat quality. Ecology Letters, 12, 1050-1060. MURCIA, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution, 10, 58-62. MUYT, A. 2001. Bush Invaders of South-East Australia, Victoria, R.G. and F.J. Richardson. NATIONAL TRUST OF AUSTRALIA 1999. Bush Regenerators Handbook. Sydney. National Trust of Australia. NSW DEPARTMENT OF PLANNING AND INFRASTRUCTURE 2012. North Wyong Shire Structure Plan. Sydney. NSW DEPARTMENT OF PRIMARY INDUSTRIES. 2013. Lantana - Weed of National Significance [Online]. Orange. Available: http://www.dpi.nsw.gov.au/agriculture/pests- weeds/weeds/profiles/lantana [Accessed 01-03-2013 2013]. NSW NATIONAL PARKS AND WILDLIFE SERVICE 2000. Atlas of NSW Wildlife. New South Wales National Parks and Wildlife Service: Sydney. NSW OFFICE OF ENVIRONMENT AND HERITAGE 1995. Threatened species conservation act. O'MEARA, J. & JACK, A. 2011. Measuring success in endangered species habitat management, the Brickpit, Sydney Olympic park 2006-2011. Australasian Plant Conservation: Journal of the Australian Network for Plant Conservation, 20, 20. OKSANEN, T., OKSANEN, L. & GYLLENBERG, M. 1992. Exploitation ecosystems in heterogeneous habitat complexes II: Impact of small-scale heterogeneity on predator- prey dynamics. Evolutionary Ecology, 6, 383-398. OLD, J. M. & WOLFENDEN, J. 2010. Report to Gosford City Council - Extent of feral and ‘pest’ intrusions within the Gosford LGA Reserves. Gosford. OOSTERHOORN, M. & KAPPELLE, M. 2000. Vegetation structure and composition along an interior-edge-exterior gradient in a Costa Rican montane cloud forest. Forest Ecology and Management, 126, 291-307. ORO, D. 2008. Living in a ghetto within a local population: an empirical example of an ideal despotic distribution. Ecology, 89, 838-846. ORROCK, J. L., DANIELSON, B. J. & BRINKERHOFF, R. J. 2004. Rodent foraging is affected by indirect, but not by direct, cues of predation risk. Behavioral Ecology, 15, 433-437. OWINGS, D. H. & COSS, R. G. 2007. Social and antipredator systems: intertwining links in multiple time frames. In: WOLFF, J. O. & SHERMAN, P. W. (eds.) Rodent Societies an Ecological and Evolutionary Perspective. Chicago: The University of Chicago Press. PARKES, D., NEWELL, G. & CHEAL, D. 2003. Assessing the quality of native vegetation: The ‘habitat hectares’ approach. Ecological Management & Restoration, 4, S29-S38. PASS, M. A., FINDLAY, L., PUGH, M. W. & SEAWRIGHT, A. A. 1979. Toxicity of reduced lantadene A (22[beta]-angeloyloxyoleanolic acid) in the rat. Toxicology and Applied Pharmacology, 51, 515-521. PEAKALL, R., EBERT, D., CUNNINGHAM, R. & LINDENMAYER, D. 2006. Mark–recapture by genetic tagging reveals restricted movements by bush rats (Rattus fuscipes) in a fragmented landscape. Journal of Zoology, 268, 207-216. PEAKALL, R., RUIBAL, M. & LINDENMAYER, D. B. 2003. Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes. Evolution, 57, 1182-1195. PEARSON, D. E. 2009. Invasive plant architecture alters trophic interactions by changing predator abundance and behavior. Oecologia, 159, 549-558. PELES, D., BOWNE, D. R. & BARRETT, G. W. 1999. Influence of landscape structure on movement patterns of small mammals. In: BARRETT, G. W. & PELES, J. D. (eds.) Landscape Ecology of Small Mammals. New York: Springer-Verlag.

119

PEREOGLOU, F., LINDENMAYER, D., MACGREGOR, C., FORD, F., WOOD, J. & BANKS, S. 2013. Landscape genetics of an early successional specialist in a disturbance‐prone environment. Molecular Ecology. PETREN, K. & CASE, T. J. 1996. An experimental demonstration of exploitation competition in an ongoing invasion. Ecology, 118-132. PIPER, S. D. & CATTERALL, C. P. 2003. A particular case and a general pattern: hyperaggressive behaviour by one species may mediate avifaunal decreases in fragmented Australian forests. Oikos, 101, 602-614. POULIN, J.-F. & VILLARD, M.-A. 2011. Edge effect and matrix influence on the nest survival of an old forest specialist, the Brown Creeper Certhia americana. Landscape Ecology, 26, 911-922. PRUITT, W. O. 1984. Snow and small mammals. Winter Ecology of Small Mammals. Pittsburgh., Carnegie Museum of Natural History. PUSENIUS, J. & SCHMIDT, K. A. 2002. The effects of habitat manipulation on population distribution and foraging behavior in meadow voles. Oikos, 98, 251-262. QUINN, G. P. & KEOUGH, M. J. 2002. Experimental Design and Data Analysis for Biologists, Cambridge, Cambridge University Press. RAINHO, A., AUGUSTO, A. M. & PALMEIRIM, J. M. 2010. Influence of vegetation clutter on the capacity of ground foraging bats to capture prey. Journal of Applied Ecology, 47, 850- 858. RALOFF, J. 2003. Cultivating weeds: is your yard a menace to parks and wildlands? Science News, 163, 15-18. RAND, T. A., TYLIANAKIS, J. M. & TSCHARNTKE, T. 2006. Spillover edge effects: the dispersal of agriculturally subsidized insect natural enemies into adjacent natural habitats. Ecology Letters, 9, 603-614. RANDALL, J. A. & ROGOVIN, K. A. 2002. Variation in and meaning of alarm calls in a social desert rodent Rhombomys opimus. Ethology, 108, 513-527. RATSAK. 2010. Rats and Mice in Australia [Online]. Sydney. Available: http://www.ratsak.com.au/about-rats-mice/rodents [Accessed 112-03-2012]. REINHARDT, D. & KUHLEMEIER, C. 2002. Plant architecture. EMBO reports. Bern, Switzerland: Institute of Plant Sciences. RIEDER, J., NEWBOLD, T. A. S. & OSTOJA, S. 2010. Structural changes in vegetation coincident with annual grass invasion negatively impacts sprint velocity of small vertebrates. Biological Invasions, 12, 2429-2439. RIES, L., FLETCHER JR, R. J., BATTIN, J. & SISK, T. D. 2004. Ecological responses to habitat edges: mechanisms, models, and variability explained. Annual Review of Ecology, Evolution, and Systematics, 491-522. RIES, L. & SISK, T. D. 2004. A predictive model of edge effects. Ecology, 85, 2917-2926. RIES, L. & SISK, T. D. 2009. Emergent properties of fragmentation when edge effects are extrapolated over complex landscapes. 94th ESA Annual Meeting. Albuquerque, New Mexico. RIIHIMÄKI, J., VEHVILÄINEN, H., KAITANIEMI, P. & KORICHEVA, J. 2006. Host tree architecture mediates the effect of predators on herbivore survival. Ecological Entomology, 31, 227-235. RITCHIE, E. G. & JOHNSON, C. N. 2009. Predator interactions, mesopredator release and biodiversity conservation. Ecology Letters, 12, 982-998. ROBINSON, A. C. 1987. The ecology of the bush rat, Rattus fuscipes (Rodentia:Muridae), in Sherbrooke Forest, Victoria. Australian Journal of Mammalogy, 11, 35-49. ROCKDALE CITY COUNCIL 2012. Unwanted Visitors? Rockdale City Council. [online] Available: http://www.rockdale.nsw.gov.au/animals/Pages/pdf/Animals/a_rat_brochure.pdf [Accessed 13/09/2012]. ROHNER, C. & KREBS, C. J. 1996. Owl predation on snowshoe hares: consequences of antipredator behaviour. Oecologia, 108, 303-310.

120

ROSE, S. & FAIRWEATHER, P. G. 1997. Changes in Floristic Composition of Urban Bushland Invaded by Pittosporum undulatum in Northern Sydney, Australia. Australian Journal of Botany, 45, 123-149. ROSENZWEIG, M. L. 1981. A theory of habitat selection. Ecology, 62, 327-335. ROWCLIFFE, J. M., FIELD, J., TURVEY, S. T. & CARBONE, C. 2008. Estimating animal density using camera traps without the need for individual recognition. Journal of Applied Ecology, 45, 1228-1236. ROWLES, A. & O’DOWD, D. 2007. Interference competition by Argentine ants displaces native ants: implications for biotic resistance to invasion. Biological Invasions, 9, 73-85. ROWLEY, R. J. 1994. Marine reserves in fisheries management. Aquatic conservation: marine and freshwater ecosystems, 4, 233-254. RUDGERS, J. A. & WHITNEY, K. D. 2006. Interactions between insect herbivores and a plant architectural dimorphism. Journal of Ecology, 94, 1249-1260. RUSS, G. R., ALCALA, A. C. & MAYPA, A. P. 2003. Spillover from marine reserves: the case of Naso vlamingii at Apo Island, the Philippines. Marine ecology Progress series, 264, 15- 20. RYAN, B. 1999. The influence of roads on small mammal ecology in the north-west of Sydney. Master of Science (Env) thesis, unpublished, The University of Sydney. SALA, O. E., CHAPIN, F. S., ARMESTO, J. J., BERLOW, E., BLOOMFIELD, J., DIRZO, R., HUBER- SANWALD, E., HUENNEKE, L. F., JACKSON, R. B. & KINZIG, A. 2000. Global biodiversity scenarios for the year 2100. Science, 287, 1770-1774. SANECKI, G. M., COWLING, A., GREEN, K., WOOD, H. & LINDENMAYER, D. 2006a. Winter distribution of small mammals in relation to snow cover in the subalpine zone, Australia. Journal of Zoology, 269, 99-110. SANECKI, G. M., GREEN, K., WOOD, H., LINDENMAYER, D. & SANECKI, K. 2006b. The influence of snow cover on home range and activity of the bush-rat (Rattus fuscipes) and the dusky antechinus (Antechinus swainsonii). Wildlife Research, 33, 489-496. SANTOS-FILHO, M., PERES, C., DA SILVA, D. & SANAIOTTI, T. 2012. Habitat patch and matrix effects on small-mammal persistence in Amazonian forest fragments. Biodiversity and Conservation, 21, 1127-1147. SAS INSTITUTE 1989-2010. JMP©, Verison 9. . Cary, NC. SAX, D. F. & BROWN, J. H. 2000. The paradox of invasion. Global Ecology and Biogeography, 9, 363-371. SCHERER-LORENZEN, M., VENTERINK, H. O. & BUSCHMANN, H. 2007. Nitrogen enrichment and plant invasions: the importance of nitrogen-fixing plants and anthropogenic eutrophication. Biological Invasions, 163-180. SCHLAEPFER, M. A. & GAVIN, T. A. 2001. Edge effects on lizards and frogs in tropical forest fragments. Conservation Biology, 15, 1079-1090. SCHOOLEY, R. L., SHARPE, P. B. & HORNE, B. V. 1996. Can shrub cover increase predation risk for a desert rodent? Canadian Journal of Zoology, 74, 157-163. SCHWANZ, L. E., PREVITALI, M. A., GOMES-SOLECKI, M., BRISSON, D. & OSTFELD, R. S. 2012. Immunochallenge reduces risk sensitivity during foraging in white-footed mice. Animal Behaviour, 83, 155-161. SEARLE, K. R., STOKES, C. J. & GORDON, I. J. 2008. When foraging and fear meet: using foraging hierarchies to inform assessments of landscapes of fear. Behavioral Ecology, 19, 475- 482. SEEBECK, J. & MENKHORST, P. 2000. Status and conservation of the rodents of Victoria. Wildlife Research, 27, 357-369. SHARMA, O. P., MAKKAR, H. P. S. & DAWRA, R. K. 1988. A review of the noxious plant Lantana camara. Toxicon, 26, 975-987. SHARMA, O. P., MAKKARI, H. P. S., DAWIRA, R. K. & NEGI, S. S. 1981. A Review of the Toxicity of Lantana camara (Linn) in Animals. Clinical Toxicology, 18, 1077-1094.

121

SHAUKAT, A. S. & SIDDIQUI, I. A. 2001. Lantana camara in the soil changes the fungal community structure and reduces impact of Meloidogyne javanica on mungbean. Phytopathologia Mediterranea, 40. SHEA, K. & CHESSON, P. 2002. Community ecology theory as a framework for biological invasions. Trends in Ecology & Evolution, 17, 170-176. SHEA, M. 2012. RE: Technical Officer Malacology, The Australian Museum. Personal communication. Type to GLEEN, W. SHOCHAT, E., LERMAN, S. B., ANDERIES, J. M., WARREN, P. S., FAETH, S. H. & NILON, C. H. 2010. Invasion, competition, and biodiversity loss in urban ecosystems. BioScience, 60, 199-208. SHORT, K. H. & PETREN, K. 2008. Boldness underlies foraging success of invasive Lepidodactylus lugubris geckos in the human landscape. Animal Behaviour, 76, 429- 437. SILVER, S. C., OSTRO, L. E. T., MARSH, L. K., MAFFEI, L., NOSS, A. J., KELLY, M. J., WALLACE, R. B., GOMEZ, H. & AYALA, G. 2004. The use of camera traps for estimating jaguar Panthera onca abundance and density using capture/recapture analysis. Oryx, 38, 148-154. SIMBERLOFF, D. & VON HOLLE, B. 1999. Positive interactions of nonindigenous species: invasional meltdown? Biological Invasions, 1, 21-32. SIMONE, I., PROVENSAL, C. & POLOP, J. 2012. Habitat use by corn mice (Calomys musculinus) in cropfield borders of agricultural ecosystems in Argentina. Wildlife Research, 39, 112- 122. SIMONETTI, J. A. 1989. Microhabitat use by small mammals in Central Chile. Oikos, 56, 309- 318. SKAGEN, S. K. & YACKEL ADAMS, A. A. 2011. Potential misuse of avian density as a conservation metric. Conservation Biology, 25, 48-55. SMITH, P. & SMITH, J. 2010. Urban edge effects in the Blue Mountains, New South Wales: implications for design of buffers to protect significant habitats. Pacific Conservation Biology, 16, 92-100. SMITH, T. M., HINDELL, J. S., JENKINS, G. P., CONNOLLY, R. M. & KEOUGH, M. J. 2011. Edge effects in patchy seagrass landscapes: The role of predation in determining fish distribution. Journal of Experimental Marine Biology and Ecology, 399, 8-16. SOKAL, R. R. & ROHLF, F. J. 1995. Biometry: the principles and practice of statistics in biological research. WH Freeman, New York. SPENCER, R. J., CAVANOUGH, V. C., BAXTER, G. S. & KENNEDY, M. S. 2005. Adult free zones in small mammal populations: response of Australian native rodents to reduced cover. Austral Ecology, 30, 868-876. STEVENS, R. D. & TELLO, J. S. 2011. Diversity begets diversity: Relative roles of structural and resource heterogeneity in determining rodent community structure. Journal of Mammalogy, 92, 387-395. STOKES, V., BANKS, P. & PECH, R. 2012. Influence of residency and social odors in interactions between competing native and alien rodents. Behavioral Ecology and Sociobiology, 66, 329-338. STOKES, V. L. 2013. Trappability of introduced and native rodents in different trap types in coastal forests of south-eastern Australia. Australian Mammalogy, 35, 49-53. STOKES, V. L., BANKS, P. B., PECH, R. P. & SPRATT, D. M. 2009a. Competition in an invaded rodent community reveals black rats as a threat to native bush rats in littoral rainforest of south-eastern Australia. Journal of Applied Ecology, 46, 1239-1247. STOKES, V. L., BANKS, P. B., PECH, R. P. & WILLIAMS, R. L. 2009b. Invasion by Rattus rattus into native coastal forests of south-eastern Australia: are native small mammals at risk? Austral Ecology, 34, 395-408. 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. 122

STORCH, I., WOITKE, E. & KRIEGER, S. 2005. Landscape-scale edge effect in predation risk in forest-farmland mosaics of Central Europe. Landscape Ecology, 20, 927-940. STRAUß, A., SOLMSDORFF, K., PECH, R. & JACOB, J. 2008. Rats on the run: removal of alien terrestrial predators affects bush rat behaviour. Behavioral Ecology and Sociobiology, 62, 1551-1558. SUAREZ, A. V., BOLGER, D. T. & CASE, T. J. 1998. Effects of fragmentation and invasion on native ant communities in coastal Southern California. Ecology, 79, 2041-2056. SUTHERLAND, W. J., NEWTON, I. & GREEN, R. E. 2004. Bird ecology and conservation: a handbook of techniques, Oxford University Press, USA. TASKER, E. & DICKMAN, C. R. 2001. A review of Elliott trapping methods for small mammals in Australia. Australian Mammalogy, 23, 77-87. TAYLOR, J. M. 1961. Reproductive biology of the Australian bush rat. University of California Publications in Zoology, 60, 1-66. TAYLOR, J. M. & CALABY, J. H. 1988. Rattus fuscipes. Mammalian Species, 298, 1-8. THORP, J. R. & WILSON, M. 1998 onwards. The Lantana Profile. In: AUSTRALIA, W. (ed.) www.weeds.org.au. THUILIER, C. 2012. Introduced plants outnumber natives. Australian Geographic. Sydney. TIMMINS, S. M., JAMES, A., STOVER, J. & PLANK, M. 2010. Is garden waste dumping really a problem? 17th Australian weeds conference. New frontiers in New Zealand: together we can beat the weeds. New Zealand plant protection society. Christchurch, New Zealand. 455-458. TORY, M. K., MAY, T. W., KEANE, P. J. & BENNETT, A. F. 1997. Mycophagy in small mammals: A comparison of the occurrence and diversity of hypogeal fungi in the diet of the long- nosed potoroo Potorous tridactylus and the bush rat Rattus fuscipes from southwestern Victoria, Australia. Australian Journal of Ecology, 22, 460-470. TULLOCH, A. I. & DICKMAN, C. R. 2006. Floristic and structural components of habitat use by the eastern pygmy-possum (Cercartetus nanus) in burnt and unburnt habitats. Wildlife Research, 33, 627-637. UMETSU, F. & PARDINI, R. 2007. Small mammals in a mosaic of forest remnants and anthropogenic habitats—evaluating matrix quality in an Atlantic forest landscape. Landscape Ecology, 22, 517–530. VAN DER MAAREL, E. 2005. Vegetation ecology–an overview. Vegetation ecology, 1-51. VAN DER REE, R. & MCCARTHY, M. A. 2005. Inferring persistence of indigenous mammals in response to urbanisation. Animal Conservation, 8, 309-319. VAN DYCK, S. & STRAHAN, R. E. S. 2008. The Mammals of Australia, Third Edition, Sydney, Reed New Holland. VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. The Journal of Wildlife Management, 47, 893-901. VAN RIEL, M. C., HEALY, E. P., VAN DER VELDE, G. & BIJ DE VAATE, A. 2007. Interference competition among native and invader amphipods. Acta Oecologica, 31, 282-289. VERDOLIN, J. L. 2006. Meta-analysis of foraging and predation risk trade-offs in terrestrial systems. Behavioral Ecology and Sociobiology, 60, 457-464. VERNES, K. 2003. Fine‐scale habitat preferences and habitat partitioning by three mycophagous mammals in tropical wet sclerophyll forest, north‐eastern Australia. Austral Ecology, 28, 471-479. VERNES, K. & DUNN, L. 2009. Mammal mycophagy and fungal spore dispersal across a steep environmental gradient in eastern Australia. Austral Ecology, 34, 69-76. VERNES, K. & MCGRATH, K. 2009. Are introduced black rats (Rattus rattus) a functional replacement for mycophagous native rodents in fragmented forests? Fungal Ecology, 2, 145-148. VET HQ. 2011. Pet Health and Wellbeing. Vet HQ newsletter [Online]. Available: http://www.edmblast.com.au/email/p/viewemail.aspx?bid=2559&eid=151390 [Accessed 12-03-12].

123

VIDLER, S., SINDEL, B. & JOHNSON, S. Year. Using your cute and furries: the role of threatened species in weed awareness. In: Proceedings Australian Weeds Conference, 2004. VIRKKI, D. A., TRAN, C. & CASTLEY, J. G. 2012. Reptile Responses to Lantana Management in a Wet Sclerophyll Forest, Australia. Journal of Herpetology, 46, 177-185. VON DER LIPPE, M. & KOWARIK, I. 2007. Long-distance dispersal of plants by vehicles as a driver of plant invasions. Conservation Biology, 21, 986-996. WALKER, J. & HOPKINS, M. S. 1983. Vegetation. In: MYERS, K., MARGULES, C.R. AND MUSTO, I. (ed.) Survey Methods for Nature Conservation. Canberra: CSIRO. WARNEKE, R. M. 1971. Field Study of the bush rat (Rattus fuscipes). Wildlife Contributions, Victoria, 14. WATLING, J. I. & ORROCK, J. L. 2010. Measuring edge contrast using biotic criteria helps define edge effects on the density of an invasive plant. Landscape Ecology, 25, 69-78. WATTS, C. & BRAITHWAITE, R. 1978. The diet of Rattus Lutreolus and five other rodents in southern Victoria. Wildlife Research, 5, 47-57. WATTS, C. H. S. 1983. Black Rat Rattus rattus. In: STRAHAN, R. (ed.) The Australian Museum Complete Book of Australian Mammals. Ryde: Cornstalk Publishing. WATTS, C. H. S. & ASLIN, H. J. 1981. The Rodents of Australia, Sydney, Angus and Robertson. WEERAKOON, M. 2012. Unpublished data. The University of Sydney. WEERAKOON, M. K. 2011. The Movements of Black Rats across the Urban-Bushland Interface; a Study using Rhodamine B. Master of Philosophy. Unpublished. University of New South Wales. WHEELER, S. H. 1970. The ecology of Rattus fuscipes greyi on Kangaroo Island, South Australia. Doctor of Philosophy. The University of Adelaide. WHITE, J. G., GUBIANI, R., SMALLMAN, N., SNELL, K. & MORTON, A. 2006. Home range, habitat selection and diet of foxes (Vulpes vulpes) in a semi-urban riparian environment. Wildlife Research, 33, 175-180. WHITE, R. M., KENNAWAY, D. J. & SEAMARK, R. F. 1996. Reproductive Seasonality of the Bush rat (Rattus fuscipes greyi) in South Australia. Wildlife Research, 23, 317-36. WIENS, J. A. 1977. On competition and variable environments: populations may experience" ecological crunches" in variable climates, nullifying the assumptions of competition theory and limiting the usefulness of short-term studies of population patterns. American Scientist, 590-597. WILCOVE, D. S. 1985. Nest Predation in Forest Tracts and the Decline of Migratory Songbirds. Ecology, 66, 1211-1214. WILKINSON, E. B., BRANCH, L. C. & MILLER, D. L. 2013. Functional habitat connectivity for beach mice depends on perceived predation risk. Landscape Ecology, 1-12. WILLIAMSON, M. & FITTER, A. 1996. The varying success of invaders. Ecology, 1661-1666. WOLFF, J. O. 2003. Density-dependence and the socioecology of space use in rodents. In: SINGLETON, G. R., HINDS, L.A., KREBS, C.K., SPRATT, D.M. (eds.) Rats Mice and People;Rodent Biology and Management. Canberra: Australian Centre for Agricultural Research. WOOD, D., H. 1971. The ecology of Rattus fuscipes and Melomys cervinipes (Rodentia:Muridae) in a south-east Queensland rainforest. Australian Journal of Zoology, 19, 371-92. WOODSIDE, D. P. 1983. The Role of Social Behaviour and Spacing in Populations of the Bush Rat, Rattus fuscipes. Doctor of Philosophy. Australian National University. WRIGHT, P. (ed.) 1991. Bush Regenerators’ Handbook. Sydney. The National Trust of Australia, (NSW). YAHNER, R. H. 1988. Changes in wildlife communities near edges. Conservation Biology, 2, 333- 339. YDENBERG, R., BROWN, J. & STEPHENS, D. 2007. Foraging: an overview. In: STEPHENS, D., BROWN, J. & YDENBERG, R. (eds.) Foraging Behavior and Ecology. Chicago: The University of Chicago Press.

124

YLONEN, H. & BROWN, J. S. 2007. Fear and the Foraging, Breeding and Sociality of Rodents. In: WOLFF, J. O. & SHERMAN, P. W. (eds.) Rodent Societies An ecological and evolutionary perspective. Chicago: University of Chicago Press. YOUNG, A. & MITCHELL, N. 1994. Microclimate and vegetation edge effects in a fragmented podocarp-broadleaf forest in New Zealand. Biological Conservation, 67, 63-72. YOUNGENTOB, K. N., YOON, H.-J., COGGAN, N. & LINDENMAYER, D. B. 2012. Edge effects influence competition dynamics: A case study of four sympatric arboreal marsupials. Biological Conservation, 155, 68-76. YUNGER, J. A., MESERVE, P. L. & GUTIÉRREZ, J. R. 2002. Small-mammal foraging behavior: mechanisms for coexistance and implication for population dynamics. Ecological Monographs, 72, 561-577. ZIV, Y., ABRAMSKY, Z., KOTLER, B. P. & SUBACH, A. 1993. Interference competition and temporal and habitat partitioning in two gerbil species. Oikos, 237-246. ZONUM SOLUTIONS 2006 - 2012. KML area and length [online].Available: http://zonums.com/online/kmlArea/ [Accessed 15/09/2012].

125