Spatial Ecology and Population Genetics of Cats (Felis Catus) Living in Or Near Conservation-Sensitive Areas

Spatial ecology and population genetics of cats (Felis catus) living in or near conservation-sensitive areas

Clare Colette Cross

A thesis submitted for the degree of Master of Science Department of Zoology University of Otago, Dunedin, New Zealand

February 2016

In memory of Caroline Carter.

Abstract

Human-mediated dispersal of organisms across the world has resulted in species introductions into many vulnerable ecosystems. Invasive mammalian predators have had detrimental impacts on native island biota, leading to declines and extinctions of many endemic prey species. Humans have transported cats (Felis catus) across the world as mousers on ships and as companion animals. The role cats (especially feral) have played in the decline and extinction of several island species is clear; however, different types of cats classified by their associations with humans has an influence on the public perception of cat impacts on wildlife and acceptance of appropriate management strategies. I studied the spatial ecology of two different types of cats in two different conservation-sensitive areas (Te Anau Basin and Canterbury/North Otago) in the South Island of New Zealand. I conducted this research to gain an insight into companion cat spatial ecology and feral cat population genetics. Specifically, to investigate individual cat movement patterns and population level movements to discover putative geographic barriers to movement. Additionally, I intended to aid formulation and reinforcement of appropriate and current management strategies with respect to conservation-sensitive areas that support high levels of native biodiversity. In the Te Anau Basin, the township of Te Anau lies on the edge of Lake Te Anau, directly adjacent to Fiordland National Park. The Kepler Mire conservation area, also situated in the Te Anau Basin, is a nearby wetland that supports a diverse range of fauna. I GPS tracked 32 local companion cats (11F:21M) for a maximum of 10 to 14 days over the austral spring/summer. I recorded a total of 19,157 locations prior to filtering data for erroneous locations. Home range and habitat analysis were performed on a filtered dataset of 13,241 locations using 100% minimum convex polygons (MCP) and Objective-Restricted-Edge Polygons (OREP). Dispersal barriers might be acting to prevent movement of tracked cats into Fiordland National Park, but not the Kepler Mire conservation area. I found males (mean MCP: 22.13 ha, OREP: 1.05 ha) exhibited larger movements (home range and distance travelled from home) than females (mean MCP: 8.83 ha, OREP: 0.45 ha) and rural-living cats (mean MCP: 32.54 ha, OREP 1.33 ha) exhibited larger movements than urban-living cats (mean MCP: 5.90 ha, OREP: 0.46 ha). Cats showed a tendency to preferentially select Built, Cover and Sealed habitat features.

iii

Although there was great individual variation in the ranging behaviour, there was no sex or age-related difference observed in the cats’ resource selection. To infer population movements, I used 10 microsatellite loci and a sex- identification marker, in a multiplex framework, to infer population structure of 157 feral cats in the upper Waitaki Basin (Tasman Valley, Ohau River and Ahuriri Valley) and Macraes Flat. I found some evidence of population connectivity between the sites based on migration rates and low FST values, indicating features in the landscape that act to facilitate dispersal. Bayesian clustering analysis noted the presence of three separate clusters; however, assignment rates were low for the Ohau River, Tasman Valley and Macraes Flat sites. Spatial autocorrelation and Mantel tests indicated rough terrain (i.e. mountain ranges) might limit dispersal. Macraes Flat and Ohau River might function as man-made sinks due to lower relatedness scores. Lower relatedness, genetic differentiation scores, and proximity to human habituation suggested there might be genetic input from nearby stray and companion cat populations. Due to large movements exhibited by feral cats in these areas, reinvasion into trapped areas seems likely; however, the Tasman Valley might be able to be managed as an eradication unit, if movement out of the Ohau River and surrounding area is reduced. Continued genetic monitoring of these sites and sampling of local stray and companion cats might help to identify if there is connectivity between different types of cats (i.e. companion, stray and feral). Additionally, continued genetic monitoring might be able to determine if genetic differentiation increases between each site in response to trapping operations. Tighter regulations regarding companion cat management might aid New Zealand conservation efforts by reducing and restricting movement and cat interactions with native wildlife. Stricter companion and stray cat regulations might also benefit feral cat control efforts; however, this aspect requires further analysis.

Acknowledgements

I am incredibly grateful for the input my supervisors, Dr Yolanda van Heezik, Dr Bruce Robertson and Dr Phil Seddon, have contributed towards this project. Thank you for your patience, advice and support throughout the last few years. I have learnt so many valuable skills that I can take with me into the future. This research was funded by the Miss E.L. Hellaby Indigenous Grasslands Trust and DOC and would not have been possible otherwise. In Te Anau, this research was made enjoyable by the many friendly cats, their owners and families. Thank you to the Te Anau DOC office, the Kids Restore the Kepler management team and the local schools (Fiordland College, Te Anau Primary, Southern Stars, Fiordland Kindergarten and Mararoa Primary School) for their incredible support throughout my ‘field’ work. Many thanks to Caroline Carter, Michelle Crouchley, Jo Marsh, Tina Perry, Vaughn Filmer, Peter McMurtrie, Tarn Willans, Claire Shaw, and Tracey Braven. Also to Brendan McBride and the Fiordland Advocate, thank you for raising awareness of the project. To Paula Collinson and all the staff at the YHA, without you I wouldn’t have had such nice accommodations, and the opportunity to meet so many amazing people from all walks of life. To everyone who came to my talks, I appreciated having such an attentive audience – hopefully I can come back! I felt very involved and at ease with the community and am very grateful for being so welcomed. Thank you to Cayley Coughlin, Kate Hand, Mariano Recio, Scott Jarvie and Luke Easton for their time and advice on cats, home range and resource selection analyses. Thank you to Aurelien Vivancos and again Luke Easton for all the statistical advice. In Twizel, the logistic support and guidance provided by the Twizel DOC office were invaluable. The first few days went unexpectedly smoothly thanks to support from Shaun Aitecheson and Rory McNamara. Thank you to all of the DOC trappers who collected samples for me at the Tasman Valley, Ohau River and Macraes Flat Sites. Thank you to Mike Aviss at Blenheim DOC for the loan of your traps and to Richard Maloney for trapping advice. I would also like to thank Murray Mackenzie, for fixing up the traps and Kim Garret for helping out with field work. Cameron Knox – cheers for the field assistance and for being such a great friend. When shall we next go to the Ahuriri!? Thank you also to DOC and the land owners who let me trap on their land in the Ahuriri Valley.

Thanks for the emotional support from my friends, flatmates and office mates (Luke Easton, Danielle Jones, Stacey Bryan, Cat Edwards, Genevieve Coffey, Sam Haultain and the Tae-Kwon Do crew) who over the last few years helped to keep my stress levels in check! Thank you to my Mum and Dad and brothers for your continued support.

Contents

Abstract...... iii Acknowledgements ...... v List of Tables ...... x List of Figures...... xi List of Appendices ...... xiii Chapter One: General Introduction ...... 1 1.1 Organism dispersal and invasive species ...... 2 1.2 Designation of cats: companion, stray, feral, wild ...... 3 1.3 From wild cats to companion cats: History ...... 4 1.4 Biology of companion cats ...... 5 1.4.1 Cat reproduction ...... 5 1.5 Cats as an invasive species ...... 6 1.5.1 Cats as introduced mammals in New Zealand ...... 7 1.6 Methods employed to understand wildlife spatial ecology: Home Range ...... 9 1.6.1 Resource selection ...... 10 1.6.2 Protocols for resource selection ...... 12 1.6.3 Resource selection functions (RSF) ...... 13 1.6.4 Understanding cat home range and resource use ...... 13 1.7 Genetic sampling and conservation ...... 14 1.7.1 Microsatellite genotyping ...... 15 1.7.2 Population genetics ...... 15 1.7.3 Cat genetics ...... 16 1.8 Aims, objectives and importance of sampling sites ...... 17 Chapter Two: Companion cat spatial ecology...... 22 2.1 Introduction ...... 23 2.2 Materials and Methods ...... 25 2.2.1 Study Area ...... 25 2.2.2 Cat recruitment ...... 27 2.2.3 Weighted GPS collar ...... 28 2.2.4 Data collection ...... 29 2.2.5 Data Processing ...... 30 2.2.6 Speed-angle algorithm ...... 30 2.2.7 Home range size estimation ...... 31

vii

2.2.7.1 Determinants of home range size ...... 32 2.2.7.2 Distance moved from home ...... 35 2.2.7.3 Diurnal and nocturnal home ranges ...... 35 2.2.8 Resource Selection Function ...... 36 2.2.8.1 Habitat Map ...... 36 2.2.8.2 Classification of habitat categories ...... 37 2.2.8.3 Defining orders of selection (home range and buffer) ...... 38 2.2.8.4 Resource selection study design ...... 39 2.2.8.5 Resource analysis of cats outside the owner’s home (outside) ...... 39 2.2.8.6 Resource selection statistical analyses ...... 42 2.3 Results ...... 44 2.3.1 Survey results ...... 44 2.3.2 Incremental area analysis (IAA) and fix success rate (FSR) ...... 44 2.3.3 Movement into conservation-sensitive areas ...... 45 2.3.4 Home range size estimation ...... 45 2.3.4.1 Determinants of home range size ...... 49 2.3.4.2 Diurnal and nocturnal ranges ...... 51 2.3.4.3 Distance moved from the owner’s home ...... 54 2.3.5 Prey record ...... 55 2.3.6 Resource selection function results ...... 55 2.3.6.1 Resource Selection notes ...... 55 2.3.6.2 Age and sex ...... 56 2.3.6.3 Second-order selection: Complete ...... 57 2.3.6.4 Third-order selection: Complete ...... 58 2.3.6.5 Second-order selection: Outside ...... 60 2.3.6.6 Third-order selection: Outside ...... 62 2.3.6.7 Time spent in habitats ...... 64 2.4 Discussion...... 69 2.4.1 Movements into conservation sensitive areas ...... 69 2.4.2 Factors influencing companion cat movements ...... 70 2.4.2.1 Overview ...... 70 2.4.2.2 Sex-related effects ...... 71 2.4.2.3 Property type ...... 73 2.4.2.4 Time of day...... 74 2.4.3 Resource selection ...... 75

viii

2.4.4 Risk posed to prey species ...... 76 2.4.5 Sources of Error ...... 77 2.4.5.1 Device size...... 78 2.4.5.2 Fix success rate (FSR) ...... 78 2.4.5.3 Accuracy/interference of GPS devices ...... 79 2.4.5.4 Filtering of inaccurate locations ...... 80 2.4.5 Limitations of study design ...... 80 2.4.6 Management recommendations and future directions ...... 81 Chapter Three: Feral cat population genetics ...... 85 3.1 Introduction ...... 86 3.2 Materials and methods ...... 89 3.2.1 Study Sites ...... 89 3.2.2 Feral cat trapping and sample collection ...... 92 3.2.3 Multiplex genotyping ...... 94 3.2.4 Data analysis ...... 95 3.3 Results ...... 99 3.3.1 Population structure of feral cats ...... 99 3.3.2 Demography of feral cats ...... 106 3.4 Discussion...... 108 3.4.1 Genetic diversity ...... 108 3.4.2 Population connectivity ...... 109 3.4.3 Limitations of study design ...... 111 3.4.4 Management implications ...... 112 Chapter Four: Final Discussion ...... 114 4.1 Definition and perception of cats ...... 115 4.2 What are acceptable management strategies? ...... 116 4.3 Legislation ...... 118 References: ...... 120 Appendices:……………………………………………………………………..…....145

List of Tables

TALBE 2.1 Variables used for 100% MCP and OREP analyses ...... 34 TABLE 2.2 Variables used for maximum and average Euclidean distance analyses .. 35 TALBE 2.3 Diurnal and nocturnal models...... 36 TABLE 2.4 Sample size used for resource selection analyses ...... 38 TABLE 2.5 Resource selection analyses ...... 43 TABLE 2.6 Companion cat tracking period information ...... 44 TABLE 2.7 Linear mixed effect model for 100% MCP home ranges ...... 49 TABLE 2.8 Linear mixed effects model for OREP home ranges ...... 50 TABLE 2.9 Linear mixed effects model for max. ED moved from home…...... 54 TABLE 3.1 Recorded feral cat age ...... 93 TABLE 3.2 Feral cat population genetic information ...... 99

TABLE 3.3 Pairwise Slatkin lineralised FST values...... 103 TABLE 3.4 Migration rates of feral cats from four sites ...... 104 TABLE 3.5 Assignment population rates using Bayesian clustering analyses………106 TABLE 3.6 Molecular and morphological features recorded for feral cat sex ...... 107

List of Figures

FIGURE 1.1 Feral cat capture rates in the Tasman Valley ...... 20 FIGURE 2.1 Map of Fiordland Nation Park in relation to Te Anau and Manapouri…26 FIGURE 2.2 Location of companion cat owners’ homes ...... 28 FIGURE 2.3 Charlie wearing GPS collar...... 29 FIGURE 2.4 Example of buffer habitat maps ...... 41 FIGURE 2.5 Munchkin home range encroaching Kepler Mire conservation area ...... 45 FIGURE 2.6 Smallest and largest 100% MCP home ranges ...... 47 FIGURE 2.7 Smallest and largest OREP home ranges ...... 48 FIGURE 2.8 Differences between sex and property type for 100% MCP and OREP home ranges ...... 51 FIGURE 2.9 Difference between diurnal and nocturnal 100% MCP and OREP home ranges ...... 52 FIGURE 2.10 Example of diurnal and nocturnal 100% MCP and OREP home ranges ...... 53 FIGURE 2.11 Differences between sex and property type for maximum distance moved from home...... 55 FIGURE 2.12 Coefficient plot for second-order complete resource selection for rural-living cats ...... 57 FIGURE 2.13 Coefficient plot second-order complete resource selection for urban- living cats...... 58 FIGURE 2.14 Coefficient plot third-order complete resource selection for rural- living companion cats ...... 59 FIGURE 2.15 Coefficient plot third-order complete resource selection for urban- living cats...... 60 FIGURE 2.16 Coefficient plot second-order outside resource selection for rural- living cats...... 61 FIGURE 2.17 Coefficient plot second-order outside resource selection for urban- living cats...... 62 FIGURE 2.18 Coefficient plot third-order outside resource selection for rural- living cats...... 63 FIGURE 2.19 Coefficient plot third-order outside resource selection for urban- living cats...... 64 FIGURE 2.20 Time spent in habitats at second-order selection ...... 66

FIGURE 2.21 Time spent in habitats at third-order selection ...... 68 FIGURE 3.1 Location of captured feral cats in the upper Waitaki Basin and Macraes Flat ...... 91 FIGURE 3.2 Photograph of trap placement ...... 94 FIGURE 3.3 Mean pairwise genetic relatedness for feral cats ...... 100 FIGURE 3.4 Male and female feral cat mean pairwise genetic relatedness ...... 101 FIGURE 3.5 Principal co-ordinate analysis plot for feral cat genetic distance………101 FIGURE 3.6 Isolation by distance (Mantel test) plot ...... 102 FIGURE 3.7 Correlogram for spatial autocorrelation ...... 103 FIGURE 3.8 Bayesian clustering analysis results plots ...... 105

xii

List of Appendices

APPENDIX 1: Companion cat flier to find participants ...... 145 APPENDIX 2: Information given to study participants ...... 146 APPENDIX 3: Companion cat owner survey ...... 147 APPENDIX 4: Prey record sheet ...... 148 APPENDIX 5: Community involvement ...... 149 APPENDIX 6: Examples of incremental area plots ...... 151 APPENDIX 7: Data tables for Chapter Two ...... 152 APPENDIX 8: Tables of model outputs ...... 158 APPENDIX 9: Example habitat maps ...... 164 APPENDIX 10: Proportions for time spent in habitats ...... 166 APPENDIX 11: Map of Te Anau golf course and companion cat locations ...... 168 APPENDIX 12: Mammals caught in traps ...... 169 APPENDIX 13: Microsatellite loci information ...... 170 APPENDIX 14: Genotyping scores ...... 171

xiii

Chapter One: General Introduction

Chapter One:

General Introduction

Chapter One: General Introduction

1.1 Organism dispersal and invasive species

The dispersal and movement of organisms across the landscape contributes to the natural distribution, variation and expression of biological diversity (Allen & Lee 2006, Nentwig 2007, Nislow et al. 2011). Unrestricted organism dispersal, however, has the potential to negatively impact native biodiversity, leading to uniformity of the invaded ecosystem and altered ecosystem functioning (Townsend & Winterbourn 1992, Allen & Lee 2006, Nentwig 2007). Intentional (e.g. for food, sport, nostalgia, fur trades, and novelties (McLintock 1966, Nentwig 2007)) or accidental (Daniel et al. 1996, Capowiez et al. 2000) human-mediated species dispersal has overcome many natural geographic barriers, leading to species introductions across the world. Introduced species can potentially compete for resources, exert predation pressure, hybridise with natives and introduce novel pathogens into a naïve ecosystem ultimately leading to biodiversity loss (Nowell & Jackson 1996, Allen & Lee 2006). Unfortunately, these consequences might only become noticeable following adaptation of the invader and through extensive research (Norbury et al. 2002, Nentwig 2007). Invasive species are considered a major threat to biodiversity throughout the world (Nentwig 2007, Clout & Williams 2009). Specifically, invasive mammalian predators have been identified as the one of the primary causes of global biodiversity loss, extinctions, and environmental degradation (Gurevitch & Padilla 2004, Mulongoy et al. 2006). Insular systems, such as New Zealand, are especially sensitive to invasion, as indigenous biota have evolved within an isolated system (Mulongoy et al. 2006, Clout & Williams 2009). With little development of defensive mechanisms (e.g. anti-predator behaviours), isolated species are often unable to cope with competitors, predators, and novel pathogens (Blackburn et al. 2004, Mulongoy et al. 2006, Jamieson & Ludwig 2012). Exploitable available resources, reduced predation pressure, disease and competition from native species in isolated systems, such as islands, have contributed to the establishment and spread of invasive mammalian predators (Shea & Chesson 2002, Nentwig 2007). Management programmes that aim to control and/or eradicate invasive mammalian predators have been established to alleviate their detrimental effects and restore native ecosystems (Clout & Williams 2009). However, management of invasive mammalian predators is often long and difficult and reinvasion into managed areas can be a continuing problem (Cleland et al. 2013). At both a fine and coarse scale, an understanding of invasive mammalian predator spatial ecology can aid management

Chapter One: General Introduction decisions, by identifying movements (e.g. home range and resource selection) and invasion pathways (Robertson & Gemmell 2004, Adams et al. 2014a, Adams et al. 2014b). Cats (Felis catus) have been transported throughout the world and have become particularly successful invaders of mainland and insular systems. As concern about their impact on novel ecosystems grows, so too does evidence providing support for the detrimental impacts cats have on native fauna, especially on islands (See: section '1.5 Cats as an invasive species'; Courchamp et al. 1999, Lowe et al. 2000, Galbreath & Brown 2004).

1.2 Designation of cats: companion, stray, feral, wild

This thesis focuses on the domestic cat, a versatile carnivore that exhibits various lifestyles. I have specified the terms below, which are used throughout this thesis, and refer to the different types and degree of association with humans cats can have (O’Hara 2007).

1) A companion cat is defined as a domestic cat that is dependent on humans for its welfare, including food and shelter.

2) A stray cat is defined as a domestic cat which has been lost or abandoned and could be living individually or in a colony. Stray cats live around areas of human habituation and are likely to interbreed with the unneutered companion cat population; therefore, the needs of stray cats are indirectly supplied by humans.

3) A feral cat is defined as a domestic cat which is neither a stray nor a companion as none or very little of its needs are provided for by humans. Feral cats generally do not live around centres of human habituation and the feral cat population size can fluctuate independent of human population size. Populations are self- sustaining and are not dependent upon input from the companion or stray populations. Feral cat populations derive originally from companion cats.

4) A wildcat (F. silvestris lybica) is the hypothesised undomesticated relative of domestic cats.

Chapter One: General Introduction

1.3 From wild cats to companion cats: History

Morphological and behavioural similarities between the wildcat subspecies, F. s. lybica and the companion cat allude to their ancestral relationship (Alderton 1983, Serpell 1988, Bradshaw et al. 1992, Edwards & Turner 1999). All companion cats sampled across the U.K, U.S. and Japan clustered within the F. s. lybica lineage, supporting the relationship between companion cats and the F. s. lybica wildcat ancestors (Driscoll et al. 2007, Driscoll et al. 2009b). However, the location and timing of true domestication is still contentious (Serpell 1988, Vigne et al. 2004, Driscoll et al. 2007, Linseele et al. 2007, 2008). The presence of rodents (Rodentia sp.) around waste and granaries potentially contributed to wildcat domestication by attracting wildcat ancestors to early human settlements, and resulting in the establishment of an initial commensal relationship (Driscoll et al. 2009b, Bateman & Fleming 2012, Hu et al. 2014). While genetic evidence supports the taming of cats in the Fertile Crescent from F. s. lybica, the origin of true domestication is still widely debated (Serpell 1988, Vigne et al. 2004, Driscoll et al. 2007, Linseele et al. 2007). This debate is, in part, due to the dissimilar interpretations of domestication, taming, and commensalism based on the archaeological evidence (Rothwell 2004, Vigne et al. 2004, Linseele et al. 2007, 2008, Bar-Oz et al. 2014, Hu et al. 2014, Van Neer et al. 2014). The earliest reported relationship between humans and cats was an adult human and F. s. lybica burial in Cyprus dated to around 7500 BC (Vigne et al. 2004). As Cyprus has no endemic cat population, this burial provides evidence for human-mediated transportation and early cat taming (Edwards & Turner 1999, Vigne et al. 2004, Driscoll et al. 2009b). However, the absence of burial artefacts (e.g. collar and cage) suggests the cat might have only been a wild commensal (Rothwell 2004). Based on archaeological evidence, Egypt appears to be the centre of taming and domestication. Companion cat ancestors might have originated in the Near East, possibly migrating to Egypt through trade routes (Kurushima et al. 2012). Depictions of cats in Egyptian artwork date to around 2000 BC, providing the earliest evidence of true domestication (e.g. catching mice, Mus spp., Clutton-Brock 1987 and wearing collars, Ginsburg et al. 1991). Since then, through human-mediated transportation and a remarkable ability to adapt, cats have dispersed to every corner of the globe (Bateman & Fleming 2012), becoming one of the most popular pets (Driscoll et al. 2009a, Plantinga et al. 2011) and one of the world’s worst invasive species (Lowe et al. 2000).

Chapter One: General Introduction

1.4 Biology of companion cats

Companion cats have a wide variety of coat colours, textures and patterns, with some retaining variations on the tabby coat pattern of their ancestors (Sunquist & Sunquist 2002). Similarly, the companion cat musculature and skeleton still resembles that of its wildcat ancestors, enabling cats to be quick, agile and powerful for climbing and capturing prey (Bradshaw et al. 1992, Sunquist & Sunquist 2002). Cats can hear tones of up to 65 kHz, as their pinnae function as directional amplifiers, allowing cats to hear the ultrasonic vocalisations of prey (Heffner & Heffner 1985, King 2005). Their well- developed binocular vision allows cats to accurately judge distances for prey capture and jumping (Bradshaw et al. 1992, Sunquist & Sunquist 2002, King 2005). Additionally, their tapetum lucidum provide cats with scotopic vision (sight in almost complete darkness) (Bradshaw et al. 1992, Sunquist & Sunquist 2002, King 2005). Vibrissae (whiskers) around the muzzle, eyes, chin and wrists provide cats with a tactile sense to detect changes in air currents moving around objects, permitting them to detect changes in prey movement to ensure capture (Bradshaw et al. 1992, Sunquist & Sunquist 2002). These adaptations enable companion cats to easily adopt a feral lifestyle.

1.4.1 Cat reproduction

Male and female companion cats typically reach maturity around 7-10 months of age (Alderton 1983, Tsutsui & Stabenfeldt 1992), with sexual maturity likely linked to nutrition and body mass, especially in the feral environment (Jones & Coman 1982). Females will use olfactory cues to advertise their sexual receptivity, to ensure males are present during the oestrus period (Sunquist & Sunquist 2002). Cats are induced ovulators, with ovulation occurring once the female has mated (Sunquist & Sunquist 2002). Gestation lasts on average 66 days (range: 64 – 67; Tsutsui & Stabenfeldt 1992), resulting in an average of three or four young, with a range from one to 10 (Robinson & Cox 1970, Alderton 1983). While the female is able to produce two to three litters throughout the year, peak production is during very early spring and again during summer and autumn (Robinson & Cox 1970). Female oestrus follows a cyclic pattern, but males can be sexually active all year round (Germain et al. 2008). Therefore, breeding cats, specifically feral cats, could produce approximately 12 kittens per year. Due to this relatively high reproductive rate, cat numbers can increase rapidly. Consequently, many companion cat

Chapter One: General Introduction owners opt to sterilize their pets (neutering/spaying) (Kustritz 2007). In addition to preventing unwanted kittens, sterilisation also reduces other sexual-related behaviours in both males and females, such as howling and spraying (Alderton 1983, Kustritz 2007). In the wild, there is no population-wide natural phenomenon to prevent cat breeding, leading to high population numbers. However, many localities with high numbers of stray cats, such as the city of Rome, employ trap-neuter-return (TNR) programmes in an attempt to reduce cat population size in a publicly acceptable manner (Natoli et al. 2006).

1.5 Cats as an invasive species

Cats are considered one of the top 100 worst invasive species and are widely believed to have contributed to the decline and extinction of numerous island bird species (Courchamp et al. 1999, Lowe et al. 2000, Galbreath & Brown 2004). However, the role of companion cats in the decline and loss of many native species across the world is under continued debate (Beckerman et al. 2007). Several authors argue that urban cat predation pressure is insufficient to negatively affect urban songbird populations, as cats tend to catch either old, sick, or very young individuals, or generally prefer to hunt rodents (Leyhausen 1979, Courchamp et al. 1999, Bonnaud et al. 2011). However, the combined predation pressure of millions of cats on prey populations might have a substantial effect on wildlife (Woods et al. 2003), and exceed that of avian annual productivity (Baker et al. 2005, 2008). Modelling shows that predation of urban bird populations by companion cats has the potential to lead to local extinctions within 50 to 100 years (van Heezik et al. 2010). Even under low predation pressure, cat presence and exploratory behaviours are thought to contribute to sub-lethal effects, such as reduced fecundity of prey populations (Beckerman et al. 2007, Bonnington et al. 2013, Gaby 2014). Predation pressure on avian and bat species might be increased over spring and summer, due to increases in the numbers of vulnerable young and breeding female prey (Langham 1990, van Heezik et al. 2010, Ancillotto et al. 2013). Small prey items, such as lizards, are also consumed in large numbers, resulting in drastic annual population production declines (e.g., 5.1 lizards cat-1 day-1; Middlemiss 1995). Cats (feral, stray, companion) living in close proximity to, or spending time in, conservation-sensitive areas or natural bush are of particular concern as they might catch more prey and could have access to rare or endangered prey (Churcher & Lawton 1987, Meek 2003, Ferreira et al. 2011).

Chapter One: General Introduction

Cats are hyper-carnivores, requiring the majority of their diet to consist of protein for normal functioning (Kok & Nel 2004, Eisert 2011, Plantinga et al. 2011). They are highly adaptable, modifying their diet depending on prey availability to continue to meet nutrient demands (Bateman & Fleming 2012). Unlike companion cats, feral cats rely solely on wild-caught prey and scavenging to meet their demanding energetic needs (Bradshaw et al. 1999). A review of feral cat diet of 27 studies in different locations essentially found mammals (78%), birds (16%), reptiles/amphibians (3.7%), invertebrates (1.2%), fish (0.3%) and plant material were all included in the feral cat diet (Plantinga et al. 2011). However, when broken down into continent or island sites, the relative importance of prey items differed. Rats (Rattus spp.) and rabbits (Oryctolagus cuniculus) were the most commonly consumed mammals on continents (Alterio & Moller 1997, Fitzgerald & Turner 2000, Plantinga et al. 2011). High levels of rodent predation by cats on continents might act to regulate rodent prey numbers, providing protection for birds from rodent predation (Courchamp et al. 1999). On islands, birds contributed substantially more to cat diet (Plantinga et al. 2011). On Marion Island (South Africa), for example, nesting seabirds contributed to 81.3% (Bloomer & Bester 1990) and 96.6% (van Aarde 1980) of feral cat diet. On Port-Cros Island in the Mediterranean, feral cat scat analysis indicated rats (70%), rabbits (17.2%), wood mice (Apodemus sylvaticus, 5.1%) and the yelkouan shearwater (Puffinus yelkouan, 6.1%) made up the majority of the diet (Bonnaud et al. 2007). Similarly, companion cats around the world also hunt and consume a variety of prey including rodents, birds, reptiles, amphibians, fish and arthropods (Leyhausen 1979, Woods et al. 2003, Bonnaud et al. 2011, Loss et al. 2013). Companion cats are also thought to hunt regardless of hunger status or regular feeding (Barratt 1997, Fitzgerald & Turner 2000, Loyd et al. 2013b).

1.5.1 Cats as introduced mammals in New Zealand

Cats were successfully introduced to New Zealand as early as the 1830s (Fitzgerald 1990, Allen & Lee 2006) and are the most popular companion animal in New Zealand, with a total owned cat population of approximately 1,419,000 (Bernstein 2007, Mackay 2011). Unlike dogs, companion cats are not legally required to be registered or microchipped in New Zealand (Mackay 2011); therefore, owners are not usually linked to their pets, which may contribute to ownership irresponsibility. However, some local councils have

Chapter One: General Introduction implemented legislation in order to restrict cat ownership as an attempt to reduce overpopulation problems (Invercargill City Council 2013). Local bylaws and Dr Gareth Morgan’s ‘Cats to Go’ campaign (https://garethsworld.com/catstogo) have sparked a social and political debate surrounding the public perception of different types of cats and the impact cats have on native wildlife in New Zealand. In New Zealand, a number of companion cat diet studies have been carried out in different cities (Gillies & Clout 2003, Morgan et al. 2009, Metsers et al. 2010, van Heezik et al. 2010) with comparable results, where differences likely relate to the level of urbanisation of the study area, the proximity to natural areas of the urban habitat and the diversity of available prey (van Heezik et al. 2010). For example, in two Auckland suburbs (fully urban and urban fringe), the most frequently caught prey items differed. In the fully urban area, rodents dominated the prey assemblage, followed by birds, lizards, then invertebrates. In the urban fringe area, invertebrates made up the bulk of prey items followed by birds, lizards, then rodents (Gillies & Clout 2003). In Christchurch, rodents (38%), followed by insects (22%), birds (20%), skinks (18%) and whistling frogs (Litoria ewingii), goldfish (Carassius auratus) and juvenile stoats (Mustela erminea) (2%) were recorded as prey (Morgan et al. 2009). In Dunedin, birds (37%), followed by rodents (34%), invertebrates (20%), common skinks (Oligosoma nigriplantare polychrome, 8%) and ‘other’ mammals (0.7%) were recorded as prey (van Heezik et al. 2010). Regardless, such unregulated hunting by companion cats might exert high predation pressure on prey species, leading to significant and devastating declines. Together, these studies illustrate the wide range of prey taken, which reflects the cats’ generalist diet and ability to use available resources regardless of preference. Despite a number of large-scale published studies, the scientific evidence presented to the public is not strong enough to convince companion animal owners that companion cats do have impacts on wildlife populations and action needs to be taken (Meek 2003). Studies of feral cats in NZ indicate mammals make up a large portion of the diet, followed by birds, then insects, depending on seasonal availability (Fitzgerald & Karl 1979, Langham 1990). Birds constitute a larger portion of cat diet over spring and summer, indicative of seasonal prey abundances (Langham 1990). Lagomorphs (predominantly rabbits) can be a staple prey item of feral cats at some localities, such as on the Otago Peninsula (Dickman 1996), followed by birds (Alterio & Moller 1997). Whereas, in Central Otago, skinks (52%) were heavily consumed by feral cats (Middlemiss 1995, Gillies 2001). On offshore islands, rats and ground-nesting birds often 8

Chapter One: General Introduction make up a larger portion of the diet (Karl & Best 1982, Fitzgerald et al. 1991, Harper 2004). While the inclusion of seabirds in the diet of feral cats on Raoul Island was minimal due to their current low abundance, Fitzgerald et al. (1991) suggests the consumption of ground-nesting seabirds might have been much greater when cats first arrived on the island. Where small mammal prey are absent, feral cats subsist largely on sea- and land-bird species, lizards (Leiolopisma spp.), and invertebrates (e.g. on Herekopare Island) (Karl & Best 1982, Fitzgerald & Veitch 1985, Harper 2004). While feral cats rely solely on wild-caught prey to meet energetic requirements, companion cat populations rely on food provided for by their owners, and are not limited by declines of wild-caught prey (Woods et al. 2003, van Heezik et al. 2010). Although diet studies can provide direct evidence of the types of prey at risk of invasive species predation, an understanding of an organism’s spatial ecology can also help to infer potential impacts. Space and habitat use and putative geographic barriers that limit movements can be identified to infer potential predatory and non-predatory impacts invasive species might impose on prey species. In addition, spatial ecology studies might also provide support for various management decisions. For example, hedgehog (Erinaceus europaeus) spatial ecology studies have been integral in discovering preferential habitats for trap placement to aid management (Jones & Norbury 2006), and ferret (Mustela furo) management has benefit from spatial ecology studies by identifying preferential trapping periods, to reduce densities while preventing reinvasion (Byrom 2002). These studies highlight the importance of spatial ecology to implement the most effective and efficient management strategies for invasive species.

1.6 Methods employed to understand wildlife spatial ecology: Home Range

Global Positioning System (GPS) devices are employed in spatial ecology studies to reveal animal spatial/movement information and to produce an estimate of the tracked animal’s home range (i.e., the “normal” area used by an animal for food gathering, reproduction and survival) (Burt 1943, White & Garrott 1990). Traditional home range estimation methods involve constructing a minimum convex polygon (MCP) whereby the outermost locations of a tracked animal are connected and the area inside is calculated (Mohr 1947, White & Garrott 1990). The ease of calculation, simplicity and flexibility

Chapter One: General Introduction of calculating MCPs has resulted in their continued and wide use. However, recognition of several problems with MCP methods, such as overestimation of home range size and absence of spatial use information (Anderson 1982, Burgman & Fox 2003, Hemson et al. 2005), has led to the development of more advanced home range estimation techniques. Unlike MCPs, probabilistic methods aim to assess the probability of finding an animal at a particular location to produce a density function or “utilization distribution” (Seaman & Powell 1996). While there are several probabilistic methods available, most commonly used is the kernel density estimation method (KDE). KDE methods calculate the relative amount of time an animal spends at a particular geographical location, equating to the intensity of use (Steiniger & Hunter 2013). However, they can be biased depending on the smoothing parameter and bandwidth value employed (Simonoff 1996). Both MCPs and KDEs can fail to identify hard boundaries or irregular structures (e.g. rivers, cliff edges, rocky outcrops, reserve boundaries), inferring impossible space use within an animal’s home range (Getz et al. 2007, Lichti & Swihart 2011). The local convex hull (LoCoH) and Objective-Restricted-Edge Polygon (OREP) methods, developed by Getz & Wilmers (2004) and Kenward et al. (2014) respectively, resolve this discrepancy. LoCoH and OREP home range estimators are less sensitive to outlying points which excludes unused areas, converging on the true distribution as the number of locations increases (Getz & Wilmers 2004, Getz et al. 2007, Lichti & Swihart 2011). Generally, LoCoH and OREP methods apply MCP constructions to produce local convex polygons (i.e. local hulls) of a subset of location data restricted in space (Getz & Wilmers 2004, Kenward et al. 2014). These can determine both animal location and intensity of spatial use. Regardless of the effectiveness of LoCoH and OREP methods for estimating an animal’s home range and spatial use, the simplicity of MCP’s and KDEs has led to their continued use in many spatial ecology studies. However, these methods are not mutually exclusive and can be used in conjunction with each other, especially for comparisons with previous studies.

1.6.1 Resource selection

Wildlife tracking can also provide information pertaining to the resources and habitats an animal exploits (Manly et al. 2002). Animals are thought to prefer resources that

Chapter One: General Introduction increase survivability; thus, space use can be used to extrapolate which resources are important to an animal’s survival (Buskirk & Millspaugh 2006, Ciarniello et al. 2007, Aarts et al. 2008). Resource selection analyses are commonly used to identify and predict potential resource use and preferred resources. As well as discovering which resources might be important for survival, resource selection analyses might also help to identify potential barriers to movement and species’ interactions (Aarts et al. 2008). In order to quantify an animal’s preferred space use, an estimate of use is compared to an estimate of what an animal has available to it (Boyce et al. 2002, Manly et al. 2002). A resource is considered preferred when it is used disproportionally more than what is available (Manly et al. 2002). To classify use, animal locations can be categorized into the habitat types in which they occurred, and proportions of points in or distances to each habitat type can be calculated (Johnson 1980, Aebischer et al. 1993, Conner et al. 2003). While identifying used resources in this way is often fairly straightforward, identifying available resources, can be more difficult. Availability can be estimated by calculating the area of different habitat types or by calculating proportions of random points that fall within each potential available habitat type (Aebischer et al. 1993, Conner et al. 2003). Each animal might have an unequal opportunity to encounter each defined habitat type, or the measure used to quantify availability might not capture what is realistically available to an animal (Fieberg et al. 2010). When conducting resource selection studies, several factors must be taken into account to prevent unnecessary bias entering the data. An animal’s spatial behaviour and subsequent resource selection might be subject to external forces within the spatial and temporal environment (such as seasonal availability of resources) (Arthur et al. 1996, Hebblewhite & Merrill 2008). Territoriality by conspecifics, human activity and noise common in highly modified environments, such as the urban landscape, might influence an animal’s movements and their consequent resource use (Kitts-Morgan et al. 2015). In addition, animals might be detected disproportionally in different habitat types, leading to an over or under estimation of use. For example, buildings and vegetative cover might disrupt GPS satellite signals, resulting in inaccurate locations and missing data (Recio et al. 2011, Hubert et al. 2015). Resource selection, therefore, can be studied based upon different sampling methods and at different scales depending on the study.

Chapter One: General Introduction

1.6.2 Protocols for resource selection

The study system and the specific research questions dictate one of three frameworks researchers can employ to sample resource units in resource selection studies. These three different protocols proposed by Manly et al. (2002) are presented below (referred to as Sampling Protocol A, B, or C):

A) Available resource units are randomly sampled and used resource units are randomly sampled

B) Available resource units are randomly sampled and a sample of unused resource units are taken

C) Unused and used resource units are independently sampled

In addition to the three sampling protocols above, one of the three study designs which researchers can employ to evaluate resource selection proposed by Manly et al. (2002) are presented below (referred to as Study design I, II or III):

I) Use and availability are measured at the population level and individual animals are not identified

II) Use is measured at an individual level while availability is measured at the population level

III) Use and availability are measured at an individual level

As well as understanding sampling protocols and study designs and any potential sources of bias, researchers must also understand the role of spatial scale in influencing resource selection (McLoughlin et al. 2002, Ciarniello et al. 2007). To evaluate resource selection at different spatial scales, Johnson (1980) proposed four orders of selection that occur in a hierarchical manner:

1) Selection of a general physical or geographical range (first-order)

2) Selection of home range within the first order (second-order - also referred to as Buffer in this thesis)

3) Usage of particular habitat components within the home range (third-order – also referred to as HR in this thesis)

Chapter One: General Introduction

4) Attaining particular food resources within chosen habitat components (fourth- order)

1.6.3 Resource selection functions (RSF)

Statistical advances in calculating resource selection functions (RSFs) have improved the performance of modelling to estimate an animal’s probability of use as a function of availability (Boyce et al. 2002, Manly et al. 2002). Categorical and continuous variables, non-linear terms, and autocorrelation can be incorporated into multi-scale generalized linear model designs that reflect the hierarchical nature of habitat selection (Johnson 1980; Nielsen et al 2002). Mixed-effects modelling can also improve the predictive power of generalized linear models by explicitly defining individual-level selection within the model (Gillies et al. 2006, Hebblewhite & Merrill 2008).

1.6.4 Understanding cat home range and resource use

Studies companion cat home ranges in different environments provide insights into habitats used and the great variation in home range size. Companion cat home range sizes have been recorded to be as little as 0.05 ha in residential Stewart Island (Wood et al. 2016) and 0.4 ha in rural France (Germain et al. 2008), to over 100 ha in rural Otago (Metsers et al. 2010). Home range size might be a function of various factors including: conspecific density, prey density and weather (Molsher et al. 2005, Germain et al. 2008). Companion cats have been found to exhibit larger movements to source preferred prey types (Meek 2003) which might reside within highly utilized habitats. Natural bushland is often used by cats as it provides cover for successful hunting than open habitats (Langham & Porter 1991, Metsers et al. 2010). Consequently, green areas in urban landscapes might provide important refuge for cats, where abundant prey can be found (Bateman & Fleming 2012, Thomas et al. 2014). However, open forest and grassland habitats might be preferentially used if inhabited by abundant prey species (e.g. rabbits, Molsher et al. 2005) or if they are transition habitats between focal points, such as the owners home or hunting and resting sites. Wildlife interactions in urban areas occur frequently as the urban sprawl moves towards natural landscapes (Meek 2003). For example, companion cats living adjacent to a wetland in an urban suburb of Christchurch

Chapter One: General Introduction exhibited skewed movements towards the wetland and brought home a larger quantity and greater variation of prey than companion cats living further away (Morgan et al. 2009). Consequently, the close proximity of urban environments to conservation- sensitive areas to might lead to urban-dwelling cats exerting predation pressure and sub- lethal effects on native biota.

1.7 Genetic sampling and conservation

The application and importance of genetics in species conservation and the incorporation of defined strategies for conservation genetics were identified and developed in the 1970s (Frankel 1970, Frankham et al. 2002). Molecular genetics provide the tools required to describe within and among population diversity related to changes in allele frequency (Hartl & Clark 2007). Collection and preservation of genetic variability and ‘gene pool reserves’ were important ideas highlighted by Frankel (1970) for species management (Frankham et al. 2002). Endangered species management has benefited from the use of molecular genetics, facilitating the identification of inbreeding and low genetic diversity of small populations. Individuals can be identified for increasing genetic variability and reducing inbreeding to improve fitness (e.g. via genetic introgression - introducing individuals from closely related sub-species, Hedrick 1995, Frankham et al. 2002, Frankham 2003). In addition, molecular information can be beneficial for translocation projects by identify appropriate individuals to translocate to maximise genetic variability and reduce inbreeding (Keller & Waller 2002, De Barba et al. 2010). Important sites for reintroduction or protection to facilitate gene flow and dispersal pathways between patchily distributed species can be identified (Loeschcke et al. 1994, Frankham et al. 2002). In these contexts, molecular genetics are a useful tool to understand species’ movements, dispersal and gene flow patterns to aid prevention of population extinction (Loeschcke et al. 1994, Frankham et al. 2002). In addition, population genetics are rapidly becoming a useful tool to understand invasive species ecology and population dynamics, to assist the formulation and implementation of appropriate control methods (e.g. Robertson & Gemmell 2004, Hansen et al. 2007, Adams et al. 2014a).

Chapter One: General Introduction

1.7.1 Microsatellite genotyping

The development of microsatellite markers for genotyping was an important advancement of molecular genetics for understanding population dynamics (Queller et al. 1993, Jarne & Lagoda 1996). The high levels of polymorphism (large numbers of alleles) exhibited by microsatellite loci can be measured to infer genetic diversity and population dynamics (Frankham et al. 2002, Zane et al. 2002). Microsatellite genetic markers are short tandem repeats (STR) of 1-6 nucleotides (typically 5-40 repeats at each microsatellite locus) that are abundant within the nuclear genome of most taxa (Goldstein & Schlotterer 1999, Zane et al. 2002, Selkoe & Toonen 2006). Microsatellites can be exclusively amplified by using specific DNA sequences (primers) that bind to sequences surrounding the locus of interest (Goldstein & Schlotterer 1999, Sunnucks 2000, Allendorf et al. 2013). This amplification of loci-specific microsatellite markers (using polymerase chain reaction (PCR)) and subsequent genotyping can be performed on multiple loci simultaneously in a “multiplex” (Chamberlain et al. 1988, Allendorf et al. 2013). Multiplexes vastly improve the efficiency and cost of genotyping multiple individuals at many loci simultaneously by using florescent dyes to identify different loci (Schuelke 2000, Markoulatos et al. 2002, Allendorf et al. 2013). Advancements from mitochondrial DNA and allozyme analysis into the development of microsatellite genetic markers have resulted in greatly improved accuracy of population identification (Randi & Ragni 1991, Hubbard et al. 1992, Queller et al. 1993), structure and connectivity, allowing movement and dispersal patterns, even at fine local scales, to be identified (Frankham et al. 2002, Say et al. 2003, Selkoe & Toonen 2006).

1.7.2 Population genetics

Population composition can be described through the use of population genetics; the study of alleles and genetic principles at a population level (Hartl & Clark 2007, Nielsen & Slatkin 2013). Deviations from the Hardy-Weinberg equilibrium (HWE) can identify population structure based on varying allele frequencies of subpopulations within a larger population (Hedrick 2005, Hartl & Clark 2007, Nielsen & Slatkin 2013). F-statistics (FIT,

FIS and FST) are also used to quantify population structure by describing the genetic variation in allele frequencies between the total population and subpopulations (Jarne & Lagoda 1996, Hedrick 2005, Hartl & Clark 2007, Nielsen & Slatkin 2013). Restricted

Chapter One: General Introduction gene flow, genetic drift, non-random mating and inbreeding due to geographical barriers can result in a naturally structured population (Whitlock et al. 2000, Hedrick 2005, Hartl & Clark 2007, Nielsen & Slatkin 2013). This has enabled movement patterns such as dispersal pathways, barriers to dispersal and reinvasion risk of invasive species to be identified (e.g. Calmet et al. 2001, Robertson & Gemmell 2004, Hansen et al. 2007, Adams et al. 2014a).

1.7.3 Cat genetics

A genetic database of F. catus microsatellite loci has been created for use in forensic science, leading to the development of an 11 microsatellite loci multiplex, “MeowPlex” and a male-specific marker for sex identification (Butler et al. 2002, Menotti-Raymond et al. 2005, 2012). The MeowPlex has subsequently been implemented in Australia to identify the human-mediated introduction and subsequent establishment of feral cats primarily from Europe, and potentially later, Asia (Koch et al. 2015, Spencer et al. 2015). The MeowPlex has also been used to infer the population structure and movement pathways of feral cats between mainland Australia and offshore islands (Koch et al. 2014). Koch et al. (2014) discovered past multiple colonisation events to Dirk Hartog Island (a national park) via human-mediated dispersal to the insular system, due to the presence of recent genetic differentiation between mainland and island feral cats. However, each of the sites sampled on the island showed high levels of gene flow between them, indicative of a single population over the entire island (Koch et al. 2014). On Hawai’i Island, Hansen et al. (2007) also employed microsatellite genotyping to determine that long-distance dispersal of cats were not limited by putative geographical barriers (lava flows). Unfortunately, island-wide management must be used in order to manage feral populations (Hansen et al. 2007). On Grande Terre, however, more recent genetic differentiation implies geographical barriers limit feral cat dispersal, facilitating the local eradication of feral cat populations (Pontier et al. 2005). Geographical barriers prevent dispersal and subsequently, connectivity (gene flow) between populations, resulting in subpopulations with large FST values of sampled loci (Robertson & Gemmell 2004, Hedrick 2005, Hartl & Clark 2007, Nielsen & Slatkin 2013). Consequently, suspected barriers to dispersal can be identified to determine if a population can be managed as separate ‘eradication units’ or must be managed as a whole

Chapter One: General Introduction

(Hansen et al. 2007, Koch et al. 2014). For example, removal of a small portion of a larger population will result in reinvasion from the surrounding areas (Koike et al. 2006); however, if invasion pathways can be identified through population genetics, management can be focused to significant areas (Hampton et al. 2004). Where populations are differentiated, reinvasion risk into managed areas will be low, allowing each population to be treated as an effective ‘eradication unit’ with little risk of reinvasion (Hampton et al. 2004, Robertson & Gemmell 2004, Adams et al. 2014a). Hence, genetic sampling is important to identify population dynamics to inform successful control strategies for management of invasive species, such as feral cats. The use of population genetic information can aid managers to make informed decisions about appropriate management strategies. Whole-island eradication on Grande Terre for example, will be difficult, although local eradication might be possible due to the discovered population genetic structure (Pontier et al. 2005). However, whole island feral cat management on Hawai’i and Dirk Hartog Islands is likely required as dispersal into managed areas is highly likely (Hansen et al. 2007, Devillard et al. 2011, Koch et al. 2014).

1.8 Aims, objectives and importance of sampling sites

This study aimed to evaluate the potential impact of cats in or near conservation-sensitive areas by quantifying: 1) the spatial ecology of companion cats living adjacent to FNP and the KMCA in the Te Anau Basin (see Figure 2.1 in Chapter Two for map), and 2) the dispersal patterns of feral cats between valley systems in the upper Waitaki Basin with Macraes Flat as an outgroup (see Figure 3.1 in Chapter Three for map). Conservation-sensitive areas in New Zealand are plentiful and provide protection for many native and endemic species from invasive mammalian predators such as cats. Cat management is a multifaceted problem which requires a multivariate approach to investigate different aspects of cat ecology presented by human associations. In this study, I have assessed two different levels of spatial ecology of two different types of cats. In the Te Anau Basin, individual companion cat movements in relation to Fiordland National Park (FNP) and the Kepler Mire conservation area (KMCA) were of interest (wildlife tracking), while feral cat population movements (population genetics) throughout the upper Waitaki Basin were of interest.

Chapter One: General Introduction

My objectives in the Te Anau Basin were to: 1) determine the home range of companion cats living within Te Anau and in close proximity to FNP and the KMCA, 2) identify the factors that influence home range and resource selection by companion cats, 3) confirm if the tracked cats enter FNP or the KMCA and pose a risk to biota living within. In the Te Anau Basin, Fiordland National Park is a conservation-sensitive area with an eclectic mix of some of New Zealand’s rare and endangered species, such as the kea (Nestor notabilis), kaka (N. meridionalis), takahē (Porphyrio hochstetteri), yellowhead/Mōhua (Mohoua ochrocephala), short-tailed bats (Mystacina sp.) and long- tailed bats (Chalinolobus tuberculatus) (O’Donnell et al. 1999, Department of Conservation 2010). The “Kids Restore the Kepler” (KRTK) project aims to re-establish birdsong within the Kepler area of FNP by involving kids to help manage pests and restore the native vegetation (Kids Restore the Kepler 2013). Access into the Kepler area of FNP is restricted to two bridges across the Waiau River, the control gates, which regulate water flows from Lake Te Anau into Lake Manapouri for hydro-electric power generation and the swing bridge at Rainbow Reach (New Zealand Government 2007). Cats have been recorded on camera using the control gates to gain access to the Kepler area (Carter 2013, Kids Restore the Kepler 2013, Brimecombe et al. 2014 - unpublished student report). However, it was not possible to confirm the ownership status of the cats. The local DOC office and KRTK participants were concerned that local companion cats might have been entering nearby conservation-sensitive areas. Another nearby conservation-sensitive area, the Kepler Mire conservation area (KMCA), lying east of FNP and south of the Te Anau township, is a large wetland set between low moraine hills and is surrounded by patches of dense scrub and diverse wetland vegetation (Seppala & Koutaniemi 1985, Dickinson et al. 2002, Boffa Miskell Limited 2006). Much of the fauna consists of members of the Anatidae family as well as many other avian species (Cromarty & Scott 1995). An assessment of companion cat home range in close proximity to these conservation-sensitive areas can aid in determining if companion cats are entering FNP or the KMCA. In the upper Waitaki Basin my objectives were to: 1) describe feral cat population structure and the dispersal patterns of feral cats living in the river and valley systems using Macraes Flat as an outgroup, and 2) explore migration rates, source-sink dynamics, relatedness of individuals, gene flow and genetic diversity to inform management decisions. 18

Chapter One: General Introduction

The braided rivers and wetlands of the upper Waitaki Basin are threatened by habitat degradation and invasive mammalian predators (Woolmore et al. 2012). Within the basin, feral cats have been identified as important predators of the resident ground- nesting birds with predation occurring at nests and on adults (Pierce 1996; BirdLife International 2000; Keedwell et al. 2002b, Sanders & Maloney 2002). Subsequently, feral cats are the target of various trapping operations (Woolmore et al. 2012, Cleland et al. 2013). The Project River Recovery (PRR) project actioned by the Department of Conservation and funded by Meridian Energy Limited and Genesis Energy, aims to reduce any adverse impacts of hydro-electric power generation and predators affecting braided river ecosystems (Woolmore et al. 2012). Maintenance of habitat and ecological communities in the riverbeds and wetlands has been the focus of the Project since its commencement in 1991, with an objective to implement and test the efficacy of large- scale predator control. By removing predators from the system, depleted and threatened populations might recover and large populations might remain stable (Woolmore et al. 2012). Through the PRR, the upper Ohau River has been a site of intensive multi-year predator control since 2010, to protect the nationally endangered black-fronted tern (Chlidonias albostriatus) colony (Woolmore et al. 2012, Robertson et al. 2013). North of the Ohau River is the Tasman Valley, fed by the Tasman River and managed by the Kaki Recovery Programme and Tasman Valley Predator Control Projects (Cleland et al. 2013). The Kaki Recovery Programme primarily involves captive management for wild release of the critically endangered kaki (black stilt, Himantopus novaezelandiae) (van Heezik et al. 2005) in conjunction with wide-scale predator control to release birds in a predator-free environment (Cleland et al. 2013). Since commencement of the Tasman Valley Predator Control Project in 2004, a total of 1,175 feral cats have been caught in kill traps placed along the Tasman River (11% of all target captures) (Cleland et al. 2013). Capture rates appear to follow a seasonal distribution with fewer captures during spring/summer (Keedwell & Brown 2001) and peaks during autumn (Cleland et al. 2013; Figure 1.1). Captures from the true right of the valley are often greater than the true left (S. Aitechson pers. comm.).

Chapter One: General Introduction

Figure 1.1: Cat capture rates in the Tasman Valley from March 2005 to February 2013 (included as 2012). Figure modified from: Cleland et al. 2013.

Even with extensive trapping implemented by Project River Recovery along the Ohau River and the Tasman Valley Predator Control Project in the Tasman Valley, reinvasion into trapped areas is a continuing problem (Woolmore et al. 2012, Cleland et al. 2013). Feral cats in the Tasman Valley and at Ohau River have been found to occupy large home ranges (Tasman Valley: 178 - 2486 ha, Ohau River: 16 – 6753 ha) (Recio et al. 2010, Cruz et al. 2014); however, the connectivity between populations is currently unknown. The Ahuriri Valley, lying south-west of the Tasman Valley and Ohau River, is a currently un-managed braided river site that might provide some insight into population movement of feral cats in the upper Waitaki Basin (Kitson & Thiele 1910). The Ahuriri Valley was previously managed solely for Kaki; however, logistic and maintenance costs led to the termination of predator control within the valley even though it is still frequented by Kaki and other endemic bird species (G. Currall pers. comm., pers. obs.). Being situated ~130 km from the upper Waitaki Basin, Macraes Flat provides an outgroup for the feral cat population genetic analysis. The Macraes Flat site, in North Otago, is characterised by a variety of native and exotic grasslands providing habitats for the critically endangered Grand (Oligosoma grande) and Otago (O. otagense) skinks, as well as Hoplodactylus and other Oligosoma species (Norbury et al. 2006). An ongoing feral cat predator control programme operates at Macraes Flat to protect these threatened lizards (Norbury et al. 2006; Spitzen-van der Sluijs et al. 2009). Genetic sampling of cats at these four sites separated by putative geographic barriers will provide evidence of the

Chapter One: General Introduction level of population connectivity to inform management decisions to reduce reinvasion and aid identification of eradication units. Information gathered in my study will be important to inform owners and community members of owned and unowned cat’s movements and elucidate potential cat impacts on native wildlife. Specifically, provide important spatial information on cats residing near FNP and the KMCA and provide a better understanding of population movements of feral cats in the upper Waitaki Basin, to ultimately, reinforce or adapt current management strategies.

Chapter Two: Companion cat spatial ecology

Chapter Two:

Home range and resource selection of companion cats living near Fiordland National Park and the Kepler Mire conservation area

Bear wearing the GPS collar.

Chapter Two: Companion cat spatial ecology

2.1 Introduction

With the introduction of invasive species to vulnerable ecosystems, spatial ecology studies have provided valuable ecological information required for management (Byrom 2002; Jones and Norbury 2006). Animal tracking using VHF and GPS has provided a comprehensive view of an animal’s movement patterns and space use – commonly, home range and resource selection (Burt 1943, White & Garrott 1990, Manly et al. 2002). Tracking of companion cats has been undertaken to determine potential impacts on vulnerable prey species. Companion cats are one of the most popular companion animals (Bernstein 2007, Mackay 2011) but also an invasive species in New Zealand, capable of travelling long distances, occupying large areas and selecting preferentially habitats where their prey species are found (Morgan et al. 2009, Metsers et al. 2010, Thomas et al. 2014). Previous research suggests cat presence around conservation-sensitive areas might be more problematic than previously thought. This part of my research focussed on the spatial ecology of companion cats (Felis catus) living in close proximity to Fiordland National Park (FNP) and the Kepler Mire conservation area (KMCA) (see Figure 2.1). Lake Te Anau and the Waiau River separate the Kepler area of FNP from the Te Anau township (population of ~ 1900, Statistics New Zealand 2013). Two bridges crossing the Waiau River allow unrestricted access into the Kepler area of FNP (Figure 2.1). Movement of cats across the control gates into the Kepler area has been observed at night via trail cameras placed by “Kids Restore the Kepler” students (Carter 2013; unpublished student prepared report - Brimecombe et al. 2014). However, the ownership status of these cats was unidentifiable. These bridges might, therefore, reduce the efficacy of the river as a barrier to cat movement as well as permitting access by brush-tail possums (Trichosurus vulpecula) and dogs (Canis familiaris). “Kids Restore the Kepler” (KRTK) is a Te Anau-based conservation project largely coordinated by children at the local kindergartens, primary and secondary schools and funded by the Fiordland Conservation Trust (Fiordland Conservation Trust 2013, Kids Restore the Kepler 2013). Many projects undertaken by KRTK involve predator trapping and habitat restoration to improve the survivability of the large variety of native species living within the area. Five minute bird counts conducted by Kids Restore the Kepler in the Kepler area in 2013 detected an average abundance of 14.2 birds/count (up from 8.4 in 2012), with native and endemic species detected, including grey warblers

Chapter Two: Companion cat spatial ecology

(Gerygone igata), tomtits (Petroica macrocephala), bellbirds (Anthornis melanura melanura), brown creepers (Mohoua novaeseelandiae) and riflemen (Acanthisitta chloris) (Marsh 2013). Fiordland National Park is also home to short-tailed (Mystacina sp.) and long-tailed bats (Chalinolobus tuberculatus), native fish, lizards and FNP endemic invertebrates (O’Donnell et al. 1999, New Zealand Government 2007). Consequently, KRTK have a vested interest in understanding the movements and habitat selection of companion cats in relation to the Kepler area of FNP. The aim of this study was to discover and evaluate the potential impacts of companion cats on nearby conservation-sensitive areas by assessing their ranging behaviour and resource selection. To achieve this, my objectives were to: 1) define the home range of cats living in close proximity to FNP and the KMCA; 2) identify factors that influence home range; 3) confirm if any tracked cats enter FNP or the KMCA, thereby potentially posing a risk to biota living within; and 4) quantify resource use of tracked cats at the second- and third- orders of selection. Home ranges of companion cats in rural areas can be very large (> 100 ha, Metsers et al. 2010) and individual cats can be very idiosyncratic in their behaviour. Information gathered in my study will be important to inform conservation groups, the public and owners of their cat’s movements and to elucidate their potential impacts on the wildlife residing within the park.

Chapter Two: Companion cat spatial ecology

2.2 Materials and Methods

2.2.1 Study Area

I tracked companion cats within the Te Anau Basin, New Zealand (Figure 2.1, 45.4167⁰ S, 167.7167⁰ E), which lies near the Kepler Peninsula, a 12,000 ha area of Fiordland National Park (FNP) currently managed largely by the Kepler Challenge committee and KRTK (Fiordland Conservation Trust 2013). The Kepler Peninsula is surrounded by one of New Zealand’s Great Walks, ‘The Kepler Track’ and is, in part, bounded by Lake Te Anau, Lake Manapouri and the Waiau River (Figure 2.1; Fiordland Conservation Trust 2013). Direct access into the Kepler area is via the control gates near Te Anau and the swing bridge at Rainbow Reach over the Waiau River (Figure 2.1; Government 2007, Department of Conservation 2010). The control gates regulate the flow of water from Lake Te Anau into the Waiau River (New Zealand Government 2007, Department of Conservation 2010). The Kepler Mire conservation area (KMCA) protects the Kepler Mire, a ‘string-bog’ characterised by having a parallel pool (flark) and ridge (string) pattern and is the largest of the wetlands in the Te Anau Basin (Seppala & Koutaniemi 1985, Dickinson et al. 2002, Boffa Miskell Limited 2006). The fauna is dominated by members of the Anatidae family (e.g. Paradise Shelduck (Tadorna variegata), Mallard (Anas platyrhynchos), Grey Duck (Aythya superciliosa) and the New Zealand Scaup (A. novaeseelandiae)), while also being frequented by the Australasian Bittern (Botaurus poiciloptilus), Australasian Harrier (Circus approximans) and South Island Fernbird (Bowdleria punctata punctate) (Cromarty & Scott 1995).

Chapter

Two

catCompanion spatial ecology

Figure 2.1: Map of the Te Anau Basin to display the location of Fiordland National Park, the access points into Fiordland National Park (Control gates labelled CG, Rainbow Reach swing bridge labelled RR), and the Kepler Mire conservation area. Inset: A map of NZ displaying the Te

26 Anau Basin.

Chapter Two: Companion cat spatial ecology

2.2.2 Cat recruitment

I obtained 33 companion cats for my study by advertising in the local newspaper (The Fiordland Advocate), through the Kids Restore the Kepler programme, in local school newsletters and letterbox flier drops (Figure 2.2; Appendix 1). As owner permission was required to track companion cats, it was not possible to ensure companion cat recruitment was truly random. Many households in the Te Anau area own more than one cat as is common throughout New Zealand (van Heezik et al. 2010); therefore, six cases occurred where two companion cats were tracked from the same household. Some companion cats in the same household were not tracked as they were < 1 year old or the owners were not interested in volunteering those cat(s). Companion cats were included in the study if they were healthy, > 1 year old (fully grown) and heavier than 3 kg. Prior to the study, I gave each owner a synopsis of the study (Appendix 2), asked them to fill out a survey with details about their cat(s) (Appendix 3; modified from Metsers 2008), and to record any prey the cat caught while wearing the GPS collar (Appendix 4).

Chapter Two: Companion cat spatial ecology

Figure 2.2: Aerial photographs of the Te Anau Basin displaying the access points, Fiordland National Park and the Kepler Mire conservation area with the location of owners’ homes included within the study.

2.2.3 Weighted GPS collar

I fitted each companion cat included in the study with a 26 g GPS (Mobile Action Technology® i-gotU GT120 USB Travel Logger; 44.5 x 28.5 x 13 mm; accuracy of up to approximately 60 m, Coughlin & van Heezik 2014) enabling the cat to be tracked. Each collar was also fitted with a 50 g lead weight to act as a counter balance to reduce collar movement and ensure the i-gotU device remained dorsally positioned so as to increase location acquisition (fix success) rates (Figure 2.3; Coughlin & van Heezik 2014). I attached the GPS devices and weights to each collar using cable ties covered

Chapter Two: Companion cat spatial ecology with electrical tape to reduce movement and any discomfort experienced by the cats.

Each collar was fitted as snugly as possible to reduce rubbing. All collars were <3% body mass of companion cats. I programmed each i-gotU device to obtain locations every 15 minutes to enable the assumption of independence between locations (Recio et al. 2010). If owners did not have a collar, I supplied them with a Vedante Super Reflective Break- Away Cat Collar (Vedante; Reflective Safety Cat Collar, NZ Pet Supplies) for the duration of the study. Cats that did not wear a collar before participating in the study underwent a three-day acclimation period involving wearing the collar without the GPS or weight. Safety (break-away) collars were used as the break-away mechanism allowed the collar to stretch and the cat to free itself if caught or stuck. If owners preferred, the owner’s collar was used.

Figure 2.3: Photograph of Charlie wearing the weighted GPS collar to demonstrate the dorsal position of the i-gotU and ventral position of the lead weight.

2.2.4 Data collection

I tracked each cat for a minimum of 10 and a maximum of 14 days over November and December 2014 (01/11/2014 - 08/12/2014), during austral late spring and early summer. A tracking period of greater than six days is recommended to characterise companion cat spatial behaviour (Metsers et al. 2010, van Heezik et al. 2010). One battery change occurred over this period for each cat (at approximately seven days). Cats were tracked at this time of year to reduce influences of rainfall and reduced activity in cold temperatures over winter (Churcher & Lawton 1987, Barratt 1997, Barratt 1998,

Chapter Two: Companion cat spatial ecology

Fitzgerald & Turner 2000, Kays & DeWan 2004, van Heezik et al. 2010). Rainfall data

for each day of the study period were obtained through ‘The National Climate Database’ (cliflo.niwa.co.nz). Any day with > 0.2 mm of rain was considered a ‘rain day’. Prey might also be more vulnerable to predation at this time of year due to an increase in nestlings and fledglings (Langham 1990, Germain et al. 2008, van Heezik et al. 2010). Conducting my research over November – December also provided me with the opportunity to present my proposal, and findings, to the local schools and the public, and to encourage school children and the public to be more aware of companion cat movements (Appendix 5). Cat details obtained from the owner-filled survey included: age, breed, sex, property characteristics, neutered status, general feeding times, general activity patterns and collared status. The body mass of each cat was calculated by subtracting my body mass (obtained prior to weighing each cat) from the combined mass of myself and the cat as many owners were unsure of their cat’s body mass. Body mass was recorded to the nearest 100 g using human scales. Prey captures (number and species where possible) were recorded by cat owners over the study period.

2.2.5 Data Processing

I imported GPS data from each i-gotU device onto my computer and exported it as CSV files using ‘@trip PC’, the software supplied by Mobile Action Technology®. Using the LINZ online co-ordinate conversion utility (Land Information New Zealand; www.linz.govt.nz), I converted the GPS data from WGS1984 (World Geodetic System 1984) to NZTM2000 (New Zealand Transverse Mercator 2000) co-ordinates.

2.2.6 Speed-angle algorithm

To remove any erroneous GPS locations from each cat’s dataset, I filtered the 19,157 NZTM2000 GPS locations though a speed-angle algorithm to remove locations considered unusual (e.g. unlikely speeds or angles observed between consecutive locations, Recio et al. 2010, Augé et al. 2011, Adams et al. 2014b, de Weerd et al. 2015). I plotted the speeds and angles between each location for each cat to determine speeds or angles that were unlikely to occur in reality. Locations were removed if speed from the

Chapter Two: Companion cat spatial ecology previous point was > 0.2 ms-1 and the turning angle was > 160⁰, resulting in a total filtered

dataset of 13,241 (giving a total of 5,916 erroneous GPS locations that I removed prior to further analysis). I visually assessed filtered locations on ortho-rectified aerial photographs of Te Anau (taken 2013 - 2014, NZTM2000 map projection, 0.4 m pixel resolution, 1:5,000 layout) and Manapouri (taken 2005 - 2011, NZTM2000 map projection, 0.75 m pixel resolution, 1:1000,000 layout) (www.linz.govt.nz) in ArcGIS 10.1 (ESRI 2014) to determine if locations occurred within the Kepler area of the FNP or within the KMCA. To determine diel activity patterns, I assigned filtered locations to either ‘diurnal’ or ‘nocturnal’ classifications in relation to dawn and dusk over the tracking period (the ‘diurnal’ period was considered as 30 minutes before dawn and 30 minutes after dusk).

2.2.7 Home range size estimation

I imported filtered locations into Ranges 9.0 (Kenward et al. 2014) to estimate companion cat home ranges. To determine the extent to which each cat’s home range was revealed, I carried out incremental area analyses (IAA). Examples of revealed and unrevealed incremental area analyses can be found in Appendix 6. To quantify space use by the sampled companion cat population, I used two different home range estimators: (1) minimum convex polygon (100% MCP) and (2) Objective-Restricted-Edge Polygons (OREP). 100% MCPs describe space use at a relatively broad-scale, whereby the outer-most locations recorded for an animal are constructed into a convex polygon (Mohr 1947, White & Garrott 1990). Although simple to use, the MCP estimator will often over-estimate spatial use by including areas not used by an animal (e.g. water bodies) (Getz & Wilmers 2004, Getz et al. 2007, Kenward et al. 2014). I have used 100% MCPs here to enable comparisons with values reported in other home range studies. The OREP home range estimator is the Ranges equivalent to the local convex hull method (Getz & Wilmers 2004, Getz et al. 2007, Kenward et al. 2014) and is essentially a concave polygon capable of estimating non- linear and multi-modal home ranges, therefore excluding space not used by an animal (c.f. Kernel and MCP estimators Meek 2003)). Because I was interested in determining any movements of cats into conservation-sensitive areas, I opted to use the broad 100% MCPs accompanied with OREPs (which are much more restricted) home ranges to capture occasional forays as 31

Chapter Two: Companion cat spatial ecology well as normal daily movements. As excursions outside normal movements are often

not considered part of an animal’s home range and not reported, I also included values for 95% MCPs (see Table 2.8 in section ‘2.2.7 Home range size estimation’), to enable comparisons with earlier studies. To calculate OREPs, I set the edge-restriction distance to a nearest-neighbour- based outlier exclusion distance (NNED) (c.f. adaptive LoCoH (Getz et al. 2007, Kenward et al. 2014). Home range estimations were conducted using all filtered locations (diurnal/nocturnal together) and diurnal/nocturnal locations separately. One cat (Houdini) was removed from the diurnal/nocturnal analysis as she was actively confined indoors at night. In addition, using the ‘Near’ function in ArcGIS (ESRI 2014), I calculated the distance of each location to the owner’s home to determine maximum and average Euclidean distance moved from home. Each of the following statistical analyses carried out within this and the following section (‘2.2.7 Home range size estimation’ and ‘2.2.8 Resource Selection Function’) were performed in R 3.0.2 (R Core Team 2014), unless otherwise stated.

2.2.7.1 Determinants of home range size

I constructed linear mixed effects models (LMMs) using the ‘lme’ function in the ‘nlme’ package (Pinheiro et al. 2016) to include random factors (i.e. variation at an individual level) as well as fixed factors (i.e. predictors of interest) (Grueber et al. 2011). By using a mixed effects model, I was able to include a random factor (individual level variation: cat ID) within each model. This allowed the model to account for behavioural differences and sample size differences of recorded number of locations between each tracked companion cat (accounting for unbalanced sample sizes between individuals) (Gillies et al. 2006, Grueber et al. 2011). Furthermore, as individual variation is explicitly defined, inferences made from the following analyses can be extrapolated to the whole population (Neter et al. 1996). Random factors are particularly important for interpreting model outputs where subjects (e.g. cats) are known to express highly variable behaviours (Metsers et al. 2010). I constructed a global additive model using all filtered locations for each estimator and all predictors of interest. Response variables included the 100% MCP and OREP home range estimations. Predictor variables included details of each cat (fixed factors: sex, property type, age, period (days cat was tracked for), start date (of the tracking period), rain days (number of days with > 0.2 mm of rain while the cat was being

Chapter Two: Companion cat spatial ecology tracked) and collared status (whether or not the cat wore a collar prior to taking part in

my study). The interaction between sex and property type was found to be non-significant and removed a priori to reduce model complexity. Prior to analysis, I used the ‘cor’ function to compute Pearson’s correlation coefficients between each variable to ensure correlated variables were not included within the same model. Correlations were considered to exist between variables if the correlation coefficient r > 0.6 (Hosmer et al. 2013). As correlations with categorical and continuous variables (such as sex versus body mass) cannot be tested appropriately using a correlation test, I used a two-sided t-test to determine if there was a significant difference of body mass between males and females; body mass was consequently excluded from analyses as males were significantly heavier than females (P < 0.05), indicating a relationship with sex. There was no difference in body mass between urban- and rural-living cats. Body mass was not considered to be correlated with property type (P > 0.05). Age was centred to enable a biological interpretation of the data. I assessed the data for normality by visually inspecting the histogram of each variable. Consequently, prior to model construction, I log transformed the 100% MCP and OREP home range estimations and distances moved from home to conform to assumptions of normality and homoscedasticity (equal variances across predicted variables) (Whitlock & Schluter 2009). Following transformations, I used the ‘grubbs.test’ function to test for the presence of outliers within the data. I used the maximum likelihood estimation method to estimate parameters in the model (Tabachnick & Fidell 2007). I calculated marginal and conditional R2 values using the ‘r.squaredGLMM’ function in the ‘MuMIN’ package which allowed me to determine the amount of variation explained by each model with fixed factors alone (R2m) and with the inclusion of a random factor (R2c) (Nakagawa & Schielzeth 2013, Barton 2015). The amount of variation explained by the random factor in the global models was high (100% MCP R2m = 0.45, R2c = 0.93, OREP R2m = 0.36, R2c = 0.92), indicating there was a high amount of individual variation in the data. In order to select the most appropriate model for the data, I used the ‘dredge’ function from the ‘MuMIn’ package to assemble model sets with different combinations of terms from the global model that also met the following selection criteria (Barton 2015). To select the best models, the Akaike Information Criterion (AIC) adjusted for small sample biases (AICc) and the associated Akaike weights (wi) were used to allow a better interpretation of the relative likelihood of each model given the data (Burnham & 33

Chapter Two: Companion cat spatial ecology

Anderson 2010). The best models were selected based on a delta AICc < 2 and were

those considered to explain the highest proportion of variation with the fewest parameters (Burnham & Anderson 2010). I averaged the best models using the ‘model.avg’ function in the ‘MuMIn’ package (Barton 2015) to produce the ‘final model’ for each home range estimator (Table 2.1). Model averaging allowed for a much more robust inference of the data and reduced model selection bias on coefficient estimates by incorporating sampling variance as well as variance associated with model selection uncertainties (Burnham & Anderson 2010, Nakagawa & Freckleton 2011). Estimates were obtained using the ‘shrinkage’ method if not all predictors were included in the final model. The shrinkage method accounted for any missing data associated with model averaging and reduced bias influencing the interpretation of estimates (Burnham & Anderson 2010, Nakagawa & Freckleton 2011).

Table 2.1: Response (home range estimator) and predictor (fixed and random factors) variables included in each final model for estimating OREP and 100% MCP home ranges of the companion cats sampled in the Te Anau Basin, NZ.

Home Range Fixed factors Random Estimator factor (log transformed)

OREP Sex + Property type + Age Cat ID

100% MCP Sex + Property type + Age + Period + Cat ID Start date + Rain days + Collar

To ensure that model conclusions based on model outputs for the sex variable were appropriate for the data, I used bootstrapping to validate the 100% MCP and OREP LMMs. Bootstrapping enabled me to ensure estimates produced following model selection were not influenced by unequal sample sizes between males and females (males: n = 21, females: n = 11). I subdivided the raw data into two parts: the male and female values for relevant response variables, and stored these in two separate data frames. I used the ‘unique’ function to repeatedly sample the male sub-set to produce every combination of unique data sets corresponding to the number of females. The female sub-set was replicated and combined with each male dataset to produce multiple

Chapter Two: Companion cat spatial ecology complete datasets for subsequent analyses. The model used was the ‘final model’ for

each home range estimator to reduce complexity of the bootstrapping (i.e. predictors were only included in bootstrapping if considered influential on the home range size in the averaged model). I saved model outputs to a text file and visually inspected to determine if the percentage of models producing a significant result (P < 0.05) for the ‘Sex’ predictor variable were over 50%.

2.2.7.2 Distance moved from home

I measured the nearest Euclidean distance from each owner’s home to each GPS location using the ‘Near’ tool in ArcGIS (ESRI 2014), then calculated the average and maximum Euclidean distances to obtain greater insight into each companion cat’s spatial ecology. I constructed a Linear Mixed Effects Model (LMM) as per section ‘2.2.7.1. Determinants of home range size’ with ‘Maximum and average Euclidean distance’ as separate response variables in two different models (Table 2.2). The amount of variation explained by the random factor in the global models was high (Maximum distance model: R2 m = 0.46, R2c = 0.93; Average distance model: R2 m = 0.34, R2c = 0.92), indicating there was a high amount of individual variation in the data.

Table 2.2: Response (home range estimator) and predictor (fixed and random factors) variables included in each final model for maximum (Max. E.D.) and average Euclidean distance (Av. E.D.) each of the companion cats sampled in the Te Anau Basin, NZ moved from home.

Response Variable Fixed factors Random (log transformed) factor

Max. E.D. from home Sex + Property type + Rain Days Cat ID

Av. E.D. from home Sex + Property Type + Period + Age Cat ID

2.2.7.3 Diurnal and nocturnal home ranges

I constructed linear mixed effects models using diurnal and nocturnal locations separately for each home range estimator. Models were constructed using predictor variables considered influential on home range size determined from the complete set of filtered

Chapter Two: Companion cat spatial ecology locations for each cat (see previous section: ‘2.2.7.1: Determinants of home range size’;

Table 2.3). To enable me to compare the diurnal model with the nocturnal model, I used the function ‘predict’ to predict results for each model. I then used a paired t-test to compare the diurnal and nocturnal model predictions for each home range estimator separately. This allowed for the incorporation of multiple variables into the analysis of diurnal and nocturnal home ranges.

Table 2.3: Home range estimator and fixed and random factors included in each final model for estimating diel home ranges of the sampled companion cats in the Te Anau Basin.

Home range estimator Fixed factors Random (log transformed) factor

OREP (diurnal) Sex + Property type + Age Cat ID

OREP (nocturnal) Sex + Property type + Age Cat ID

100% MCP (diurnal) Sex + Property type + Age + Period + Cat ID Start date + Rain days + Collar

100% MCP (nocturnal) Sex + Property type + Age + Period + Cat ID Start date + Rain days + Collar

2.2.8 Resource Selection Function

2.2.8.1 Habitat Map

I imported ortho-rectified aerial photographs of Te Anau (taken 2013-2014, New Zealand Transverse Mercator (NZTM) map projection, 0.4 m pixel resolution, 1:5,000 layout) and Manapouri (taken 2005-2011, New Zealand Transverse Mercator (NZTM) map projection, 0.75 m pixel resolution, 1:10,000 layout) (Land Information New Zealand, www.linz.govt.nz) into ArcGIS 10.1 (ESRI 2014) to use as a template to create a vector habitat map of each cat’s home range and study area. In addition to the aerial photographs, I used ground truthing of various areas, topographic maps (Land Information New Zealand, www.linz.govt.nz) and Google Street View (Google Incorporated, www.google.com/maps/streetview), to visually classify habitat features into broad categories using ArcGIS 10.1.

Chapter Two: Companion cat spatial ecology

2.2.8.2 Classification of habitat categories

I created six habitat categories that have been used previously for companion cat resource selection in urban and rural areas (Metsers et al. 2010; Horn et al. 2011; Thomas et al. 2014). I included the Wetland habitat category due to the ecological diversity wetlands support and previously documented use of wetlands by companion cats in New Zealand (Cromarty & Scott 1995, Morgan et al. 2009).

1. Building – Commercial and public buildings such as shops and schools, and buildings on properties > 0.5 ha (often farm buildings)

2. Urban – Buildings (often homes) and gardens in the residential area of Te Anau (properties < 0.5 ha)

3. Cover – Included both scrub (short woody vegetation, includes plant nurseries) and trees (forested areas, plantations)

4. Grassland – Mixture of native and exotic grassland often composing pasture, playing fields, empty sections, also included lake edge

5. Sealed – Roads, footpaths, tracks, carparks and driveways on properties > 0.5 ha

6. Wetland – Mires, swamps and bogs

Note: Any cat’s home range or buffer that excluded any of the aforementioned habitat categories, not shared with other cats included in the analysis, were removed a priori, resulting in different sample sizes for each analysis (Table 2.4). Consequently, in order to infer resource selection with a larger sample size for third- and second-order selection, Building and Urban habitat categories were combined (Table 2.4).

Chapter Two: Companion cat spatial ecology

Table 2.4: Sample sizes of companion cat resource selection in the Te Anau Basin, NZ. Final sample sizes reported under n (bold) are those used in analyses following the combining of Building and Urban habitat categories into Built. Grass refers to Grassland and Wet refers to Wetland habitat categories.

Habitat category Order of Cat Building Urban Cover Grass Sealed Wet n selection Group Third Rural 14 2 14 14 14 0 14 (HR) Urban 7 18 11 16 18 0 11 Second Rural 14 5 14 14 14 12 12 (Buffer) Urban 18 18 18 18 18 0 18

2.2.8.3 Defining orders of selection (home range and buffer)

I assessed resource selection at two levels: within the 100% MCP home range calculated in section ‘2.2.7 Home range size estimation’ (Johnson’s third-order of selection, HR) and within a larger area (hereafter: buffer) around each cat’s owner’s home (Johnson’s second-order of selection; Johnson 1980). To determine the radius of each cat’s buffer, I separated the companion cats I tracked into two groups based upon the location of their home in either properties < 0.5 ha (“urban”, within the Te Anau township itself) or > 0.5 ha (“rural”, not within the Te Anau township). These two groups were used to identify the major difference in habitat features available to the companion cats (i.e. comprised mostly of Urban or Grassland habitat features respectively). I used the maximum Euclidean distance travelled from the owner’s home by any one cat within each of the rural or urban groups and added approximately 5% to this value to determine the radius of a buffer around each owner’s home. The radius for rural living cats was 2500 m (area: 1963.5 ha) whereas the radius for urban living cats was 650 m (area: 133 ha). Furthermore, as one cat within each of these groups walked these distances (and similar distances have been noted previously, Metsers et al. 2010), I considered it possible for all cats within each group to also walk those distances.

Chapter Two: Companion cat spatial ecology

2.2.8.4 Resource selection study design

I used a “Sampling Protocol A, Study Design III” study design as outlined in Manly et al. (2002) and described in Chapter One. I used a ratio of 1:4 of used-to-available resource units in order to determine all habitat features present within each home range and buffer. Using the ‘clip’ tool in ArcGIS 10.1, I extracted each cat’s 100% MCP home range and buffer habitat map from the larger habitat map. Onto each cat’s habitat maps (buffer and 100% MCP), I overlaid each cat’s used resource units (filtered GPS locations) and available resource units (generated using the ‘Create Random Points’ tool in ArcGIS 10.1). Any points that fell within inaccessible areas within the buffer and home range (i.e. water - e.g. Lake Te Anau, Lake Manapouri, the Waiau River or ponds and FNP where bridge access was also not included, Pollard 1999; Figure 2.4) were removed prior to further analysis. Inaccessible areas were not able to be selected by animals and therefore could not represent habitat features available for selection (Baasch et al. 2010). To implement resource selection analysis, I used a distance approach to account for known location error associated with GPS devices (Conner et al. 2003, Coughlin & van Heezik 2014). Additionally, in contrast to compositional methods, the distance approach creates a continuous measure of habitat use which reduces zero values within the data. Zero values can be problematic for statistical analyses (Conner et al. 2003, Zuur et al. 2009). I calculated the Euclidean distance from each used and available point to the nearest of each habitat category using the ‘Near’ tool in ArcGIS 10.1.

2.2.8.5 Resource analysis of cats outside the owner’s home (outside)

I was interested to deduce the cats’ resource selection without the inclusion of their owner’s home, as companion cats in New Zealand have been found to spend a large portion of time indoors (Gaby 2014). To identify resource selection outside the owner’s home, I removed any points that fell within the owner’s home and removed the owner’s home from the Building habitat category in the rural area in order to infer resource selection without the influence of the owner’s home. Cats have been found to ‘cheat’ on their owners and anecdotal evidence suggests some cats in this study spent time at other people’s homes (M Harcourte, K Aitken, pers. comm.; Block 2010). However, as I was unable to determine if cats were indoors or close by any other buildings other than their home, I based my inferences on assuming that companion cats were not found in any

Chapter Two: Companion cat spatial ecology other buildings, as their presence inside their owner’s home was certain. This analysis is

hereafter labelled ‘outside’.

Chapter Two: Companion cat spatial ecology

Figure 2.4: Examples of buffer and HR habitat maps of two rural- living companion cats in the Te Anau Basin to display available resource units (RU), used filtered GPS locations and habitat categories, (a) removal of RU from water (Jerry), (b) removal of RU from water and FNP in the absence of an access point (Merlot).

Chapter Two: Companion cat spatial ecology

2.2.8.6 Resource selection statistical analyses

The statistical analyses and model diagnostics I used to analyse resource selection were the same as those in section ‘2.2.7.1 Determinants of home range size’, with the exception of the following. To examine resource selection including resource units both within the owner’s home (hereafter: complete) and excluding resource units within the owner’s home (hereafter: outside), I constructed general linear mixed effects models (GLMMs) using the ‘glmer’ function in the ‘lme4’ package. Prior to model construction, I created a total of eight distance datasets using the ‘subset’ function to split each dataset (complete or outside) based upon the level of selection: third- or second-order and the property type the cat lived in: rural or urban (Table 2.5). To analyse each dataset, I coded each used resource unit as 1 and each available resource unit as 0. I used these binary response variables in a logistic regression (GLMM), commonly expressed as:

̂ ̂ ̂ ̂ 푤̂(푥) = exp⁡(훽0 +⁡훽1푥1 +⁡훽2푥2 + ⋯ +⁡훽푛푥푛)

̂ with covariates (푥푛)⁡and coefficients (훽푛) as estimates of the logistic regression (Manly et al. 2002, Gillies et al. 2006). The effect size (푤̂) took a negative value to indicate selection and a positive value to indicate no selection, as the probability of an animal using a resource increases as distance decreases. Predictor variables (fixed factors) included in the models were the habitat features defined in section ‘2.2.8.2 Classification of habitat categories’, with the exception of Wetland which was only included within the ‘third-order: Rural’ analysis for both complete and outside analyses. In addition, cat age and sex were included as covariates as resource selection might be influenced by individual sex and age (Manly et al. 2002). Outliers existed (P < 0.05) within the Wetland habitat variable where available resource units were present on the Wetland habitat feature (Complete – third-order: Rural, Outside – third-order: Rural) and were not removed. I created the models using a binomial family with a ‘logit’ link transformation. Where correlations were found to exist, I created multiple global candidate model sets to ensure each model contained a set of non-correlated variables. To determine the best model(s) to fit my data, I used backwards elimination of the least significant variables with low effect sizes to remove variables from each model, corroborated with model averaging.

Chapter Two: Companion cat spatial ecology

Table 2.5: Resource selection function analyses and sample sizes used to infer companion cat resource selection in the Te Anau Basin, NZ. Complete resource selection refers to selection including all filtered GPS locations and all accessible available resource units. Outside resource selection refers to the removal of any used or available resource units that fell within each cat’s owner’s home and removal of the owner’s home within the Building habitat feature in the rural area. Order of selection refers to Johnson’s (1980) orders of selection and the cat group refers to the property type used within the analysis.

Question Order of selection Cat group Sample size

Complete resource Second (Buffer) Rural 12 selection Urban 18 Third (HR) Rural 14 Urban 11 Outside resource selection Second (Buffer) Rural 12 Urban 18 Third (HR) Rural 14 Urban 11

Chapter Two: Companion cat spatial ecology

2.3 Results

2.3.1 Survey results

A total of 32 (11 females, 21 males) cats were included in the analysis of this study, as one cat lost the GPS collar (Table 2.6).

Table 2.6: Tracking period in days, the number of companion cats tracked for each period and the reasons given for shorter periods of tracking of companion cats GPS tracked in the Te Anau Basin, NZ.

Tracking Period Number of Reasons (days) Cats

< 6 1 Device lost

10 1 Owner removed collar due to chafing

11 4 Owner removed collar due to chafing (2) Owner removed collar due to caught jaw (1) GPS device malfunction (1)

13 23

14 4

Total 33

Details for each cat (breed, age - years/months, approximate body mass - kg, sex, property type, bell wearer, collar wearer, period tracked for, start date of tracking period and the number of rain days over the tracking period) are presented in Appendix 7 - Table 1. All cats included in the study were neutered. Further results from the owner-filled survey can be found in Appendix 7 - Table 2.

2.3.2 Incremental area analysis (IAA) and fix success rate (FSR)

Incremental area analyses were carried out for each individual cat to determine if each home range was fully revealed over the tracking period. Visual inspection of the number of locations vs area plots indicated that eight of 32 were partially revealed. However, as

Chapter Two: Companion cat spatial ecology over 200 locations were recorded for each cat (following filtering, range: 255 – 671), no

home range was a priori excluded from subsequent analyses. Approximately 51% of potential locations were not recorded by the devices, giving a FSR of approximately 49%.

2.3.3 Movement into conservation-sensitive areas

None of the tracked cats crossed the control gates or the Rainbow Reach swing bridge into FNP. One cat (Munchkin) entered the KMCA (Figure 2.5).

Figure 2.5: 100% MCP home range of Munchkin (male) and the Kepler Mire conservation area in the Te Anau Basin, South Island, NZ. The value in brackets indicates home range size.

2.3.4 Home range size estimation

Movement details (home range size, maximum Euclidean distance moved from home) for each cat can be found in Appendix 7 - Table 3. Home range size using the 100% MCP estimator ranged from 0.63 ha (Lulu; Figure 2.6a) to 109.53 ha (Kelvin; Figure 2.6b; mean 17.56 ha; CI: 8.03 – 27.09, Appendix 7 – Table 3). OREP home ranges were much 45

Chapter Two: Companion cat spatial ecology smaller than 100% MCP home ranges, ranging from 0.09 ha (Lulu; Figure 2.7a) to 4.30

ha (Kelvin; Figure 2.7b; mean 0.845 ha; CI: 0.48 – 1.20, Appendix 7 – Table 3).

Chapter Two: Companion cat spatial ecology

Figure 2.6: Examples of 100% MCP home ranges for companion cats living in the Te Anau Basin in the South Island, NZ. (a) smallest: Lulu (female); (b) largest: Kelvin (male). Values in brackets indicate home range size.

Chapter Two: Companion cat spatial ecology

Figure 2.7: Examples of OREP home ranges for companion cats living in the Te Anau Basin in the South Island, NZ. (a) smallest: Lulu (female), (b) largest: Kelvin (male). Values in brackets indicate home range size. 48

Chapter Two: Companion cat spatial ecology

2.3.4.1 Determinants of home range size

As males (mean mass: 5.62 kg, CI: 5.17 – 6.07) were found to be significantly heavier than females (mean: 4.58 kg, CI: 3.96 – 5.21; two sample t-test: t = 2.645, df = 20, P = 0.015; Appendix 7 - Table 2), body mass was removed as a determinant of home range size. The relatively large effect sizes reported in the final models indicated male and rural-living companion cats exhibited larger 100% MCP and OREP home ranges than female and urban-living companion cats respectively (Figure 2.8; Tables 2.7 and 2.8 respectively). Age, previous collar wearer, tracking period, number of rain days and start date were also included in the final 100% MCP model; however, the effect sizes were minimal indicating little influence on home range size (Table 2.7). Similarly, the lower effect size of age, indicated a small influence on OREP home range size (Table 2.8).

Table 2.7: Final linear mixed-effects model for companion cat 100% MCP home ranges living in the Te Anau Basin, NZ. Table includes the estimated coefficient (effect size (Est.)), standard error (SE), and z-values rounded to two decimal places. Predictors in bold were statistically significant (P < 0.01).

Model Predictors Est. SE z-value

Property + Intercept 40.46 163.93 0.24 Sex + Age + Property (Urban) -1.23 0.37 3.17 Collar + Period + Sex (Male) 1.09 0.38 2.73 Rain days + Age -0.03 0.05 0.66 Start date Collar 0.15 0.45 0.32 Period 0.12 0.20 0.59 Rain days 0.04 0.09 0.44 Start date -0.01 0.02 0.39

Chapter Two: Companion cat spatial ecology

Table 2.8: Final linear mixed-effects model for companion cat OREP home range living in the Te Anau Basin, NZ. Table includes the estimated coefficient (effect size Est.)), standard error (SE), and z-values rounded to two decimal places. Predictors in bold were statistically significant (P < 0.01).

Model Predictors Est. SE z-value

Sex + Intercept -0.74 0.29 2.48 Property + Sex (Male) 0.75 0.30 2.43 Age Property (Urban) -0.72 0.28 2.41

Age -0.01 0.03 0.48

Chapter Two: Companion cat spatial ecology

Figure 2.8: Sex and property type against mean ± 2 standard errors for 32 companion cats living in the Te Anau Basin in the South Island, NZ, (a) mean log 100% MCP and (b) mean log OREP home range size.

2.3.4.2 Diurnal and nocturnal ranges

Nocturnal 100% MCP home ranges were significantly larger than diurnal 100% MCP ranges (paired t-test: t = -3.08, df = 31, P = 0.004; Figure 2.9a; see Figure 2.10a for example). In contrast, nocturnal OREP ranges were not significantly larger than diurnal OREP ranges (paired t-test: t = -1.30, df = 31, P = 0.204; Figure 2.9b; see Figure 2.10b for example).

Chapter Two: Companion cat spatial ecology

Figure 2.9: Mean ± 2 standard errors of diurnal and nocturnal home ranges for 31 companion cats living in the Te Anau Basin in the South Island of NZ, (a) log 100% MCP home ranges and (b) log OREP home ranges.

Chapter Two: Companion cat spatial ecology

Figure 2.10: Examples of diurnal/nocturnal home ranges for two male companion cats living in the Te Anau Basin in the South Island, NZ, (a) Jerry’s 100% MCP home ranges, (b) Max’s OREP home ranges. Values in brackets indicate home range size.

Chapter Two: Companion cat spatial ecology

2.3.4.3 Distance moved from the owner’s home

The effect sizes for both sex and property were relatively high, indicating a biologically meaningful influence on the maximum and average Euclidean distance travelled from home of each companion cat (Figure 2.11; Table 2.9). Males travelled further distances from home than females and rural-living cats travelled further than urban-living cats (Table 2.9). The lower effect size of the number of rain days indicates the number of rain days the cat experienced had little influence on the distances travelled from home (Table 2.9). Maximum Euclidean distances moved from home ranged from 71.40 m (Lulu) to 2445.41 m (Felix; mean overall max. ED: 426.06 m; Appendix 7 – Table 3). Mean average Euclidean distance moved from home was 79.92 m (Appendix 7 – Table 3).

Table 2.9: Final linear mixed-effects model for maximum Euclidean distance moved from home by companion cats in the Te Anau Basin in the South Island, NZ. Table includes the estimated coefficient (effect size (Est.)), standard error (SE), and the z–values to two decimal places. Predictors in bold were statistically significant (P < 0.01).

Model Predictors Est SE z-value

Sex + Intercept 5.52 0.37 14.58 Property Sex (male) 0.71 0.23 3.02 + Rain days Property (urban) -0.70 0.22 3.03

Rain days 0.02 0.04 0.43

Chapter Two: Companion cat spatial ecology

Figure 2.11: Property type and sex against average ± 2 standard errors of the log of maximum Euclidean distance (max. ED) moved from home by 32 companion cats in the Te Anau Basin in the South Island, NZ.

2.3.5 Prey record

In total, 15 prey items were reported to be brought home over the tracking period by six cats. Birds were identified as: blackbird (Turdus merula) (2), thrush (Turdus philomelos) (1), house sparrow (Passer domesticus) (2) and unknown (1). Three birds were classified as ‘baby’.

2.3.6 Resource selection function results

2.3.6.1 Resource Selection notes

‘Complete’ selection refers to analyses comprising filtered used GPS locations and available resource units, while ‘outside’ selection refers to analyses for which resource units (used or available) falling within cat owners’ homes have been removed, and the owners’ homes have been removed from the map of the rural study area. In the following analyses, because the resource metric was distance to each resource, a negative coefficient indicates selection of that resource, whereas a positive coefficient indicates no selection. In all analyses, the R2c values were larger than the R2m values, indicating

Chapter Two: Companion cat spatial ecology that the inclusion of a random factor (CATID) within each model accounted for some of

the variation in the data (Appendix 8 - Tables 1-8).

2.3.6.2 Age and sex

Neither age nor sex influenced selection; however, model averaging indicated model fit was improved with the inclusion of sex and age.

Chapter Two: Companion cat spatial ecology

2.3.6.3 Second-order selection: Complete

An example of habitat maps of companion cat third-order selection can be found in Appendix 9 – Figure 1.

a. Rural-living cats

At the second-order of selection (i.e. placement of home range within the buffer, buffer), companion cats living outside the township selected Built features, followed by Cover, Sealed features and no evidence of selection for Grassland features. While the variable Wetland was significant, the relatively low effect size indicated companion cats showed no preference for this feature. The large R2m and R2c values (Figure 2.12; Appendix 8 - Table 1) indicate that these models explain a large amount of data variation; Model 1 was selected as providing the best representation of the available data (lowest AICc value and R2m and R2c values above 0.9, R2m = 0.959, R2c = 0.961; Appendix 8 - Table 1).

Figure 2.12: Second-order selection (Buffer): Complete, for 12 rural-living companion cats in the Te Anau Basin in the South Island, NZ. Coefficient plots of four models. a) Model 1, b) Model 2, c) Model 3, d) Model 4. Model information can be found in Appendix 8 - Table 1. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

Chapter Two: Companion cat spatial ecology

b. Urban-living cats

At the second-order of selection, urban-living companion cats also showed strong selection for Built features (homes, buildings and gardens), followed by Sealed features, then Cover. As with rural-living cats, urban cats also did not select Grassland features (Figure 2.13; Appendix 8 - Table 2). In the presence of Built features, the effect of Cover and Sealed features were severely diminished, indicating the preference by urban-living cats for Built features. The higher R2m and R2c values (R2m = 0.772, R2c = 0.780) and lower AICc values for Model 1 indicated Model 1 was a much better fit for the data compared to Model 2 (R2m = 0.163, R2c = 0.207; Appendix 8 - Table 2).

Figure 2.13: Second-order selection (Buffer): Complete, for 18 urban-living companion cats in the Te Anau Basin in the South Island of NZ. Coefficient plots of two models. a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 2. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

2.3.6.4 Third-order selection: Complete

An example of habitat maps of companion cat third-order selection can be found in Appendix 9 – Figure 2.

a. Rural-living cats

At the third-order of selection (i.e. use of habitat features within the home range, HR), rural-living cats selected strongly for Built over any other habitat feature. Cover and

Chapter Two: Companion cat spatial ecology

Sealed features were selected equally, as indicated by similar coefficients (Figure 2.14,

Appendix 8 - Table 3). In contrast to the second-order of selection, Grassland features were also selected for in addition to Cover and Sealed features (Figure 2.14, Appendix 8 - Table 3). Within the home range, Wetland habitats were not represented. R2m and R2c values indicated Model 1 (R2m = 0.658, R2c = 0.699) and Model 2 (R2m = 0.452, R2c = 0.491) both explain a moderate amount of variation in the data (Appendix 8 - Table 3). Model 1 is considered a better fit for the data, as denoted by the lower AICc value and higher R2m and R2c values (Appendix 8 - Table 3).

Figure 2.14: Third-order selection (HR): Complete for 14 rural-living companion cats in the Te Anau Basin in the South Island of NZ, a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 3. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

b. Urban-living cats

Similarly, within the home range, urban-living companion cats selected strongly for Built features, and for Sealed features, then Cover; they did not select for Grassland features (Figure 2.15; Appendix 8 - Table 4). Model 1 (R2m = 0.350, R2c = 0.363) explained a low to moderate amount of variation in the data, whereas Model 2 (R2m = 0.036, R2c = 0.044) explains very little of the variation in the data (Appendix 8 - Table 4). The lower AICc score of Model 1 indicates a better fit (Appendix 8 - Table 4).

Chapter Two: Companion cat spatial ecology

Figure 2.15: Third-order selection (HR): Complete for 11 urban-living companion cats in the Te Anau Basin in the South Island, NZ. Coefficient plots of two models. a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 4. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

2.3.6.5 Second-order selection: Outside

a. Rural-living cats

When outside the owners’ homes, at the second-order of selection (Buffer), rural-living companion cats selected predominantly for Cover, followed by Sealed features, then Built features (Figure 2.16, Appendix 8 - Table 5). Grassland features were not selected. Due to the relatively low effect size, there was no evidence that companion cats showed a preference for Wetland. Note: a negative coefficient indicated selection. The large R2m and R2c values (Appendix 8 - Table 5), indicated that these models explained a large amount of variation in the data. Model 1 is likely the best fit for the data with the lowest AICc value and high R2m and R2c values (R2m = 0.883, R2c = 0.890, Appendix 8 - Table 5).

Chapter Two: Companion cat spatial ecology

Figure 2.16: Second-order selection (Buffer): Outside for 12 rural-living companion cats in the Te Anau Basin in the South Island, NZ. Coefficient plots of four models. a) Model 1, b) Model 2 c) Model 3, d) Model 4. Model information can be found in Appendix 8 - Table 5. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

b. Urban-living cats

Even outside the owners’ homes, at the second-order of selection, urban-living companion cats selected strongly for Built features, followed by Sealed features, then Cover. Grassland features were not selected (Figure 2.17; Appendix 8 - Table 6). In the presence of Built features, however, the effect of Cover and Sealed features were greatly reduced, indicating the preference by urban-living cats for Built features. Model 1 (R2m = 0.738, R2c = 0.746) was a much better fit for the data compared to Model 2 (R2m = 0.164, R2c = 0.200) (as indicated by the much lower R2m and R2c and higher AICc values of Model 2; Appendix 8 - Table 6).

Chapter Two: Companion cat spatial ecology

Figure 2.17: Second-order selection (Buffer): Outside for 18 urban-living companion cats in the Te Anau Basin in the South Island of NZ. Coefficient plots of two models. a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 6. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

2.3.6.6 Third-order selection: Outside

a. Rural-living cats

When outside, at the third-order of selection (HR), the coefficients for Grassland features, Cover and Sealed features indicated no preference by companion cats. The strength of selection for Built features (i.e., buildings other than the owner’s home) was less than for the other habitat categories (Figure 2.18, Appendix 8 - Table 7). The lower AICc and higher R2m and R2c values indicated Model 2 (R2m =0.407, R2c = 0.452) was a better fit to the data and explained (albeit moderate to low) more variation in the data than Model 1 (R2m = 0.290, R2c = 0.348) (Appendix 8 - Table 7).

Chapter Two: Companion cat spatial ecology

Figure 2.18: Third-order selection (HR): Outside for 14 rural-living companion cats in the Te Anau Basin in the South Island, NZ. Coefficient plots of two models. a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 7. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

b. Urban-living cats

When outside, within the home range, urban-living companion cats selected strongly for Built features, but showed no preference for Sealed features, Cover, or Grassland features (Figure 2.19; Appendix 8 - Table 8). Model 1 (R2m = 0.299, R2c =0.312) explained a low-to-moderate amount of variation in the data, while Model 2 (R2m = 0.034, R2c = 0.040) explained very little of the variation in the data, indicating a very poor fit (Appendix 8 - Table 8).

Chapter Two: Companion cat spatial ecology

Figure 2.19: Third-order selection (HR): Outside for 11 urban-living companion cats in the Te Anau Basin in the South Island, NZ. Coefficient plots of two models. a) Model 1, b) Model 2. Model information can be found in Appendix 8 - Table 8. Note: negative coefficients indicate selection, while positive coefficients indicate no selection. Error bars are ± 2 standard errors; error bars appear absent where errors are small.

2.3.6.7 Time spent in habitats

For the following analyses, the proportion of random points placed in each available habitat feature approximated the relative proportions of actual area of each habitat category, as expected. Because not all cats had each habitat feature within the buffer of home range, Built is referred to as a combination of Urban and Building habitat features to increase the sample size. The following plots report Built as well as Urban and Building habitat categories to infer time in habitats at a finer scale. Each proportion is averaged across all cats for each group (rural n = 14, urban n = 18). The area of habitat available differs between cats as proportions were calculated without including Water habitat features, as water was not considered a habitat available for use. At the second-order of selection (within the buffer), rural-living companion cats spent approximately 40% of their time in Grassland features followed by Cover (~30%), Built (~18%), and Sealed features (~8%) (Figure 2.20a; Appendix 10 - Table 1). While available, rural-living companion cats did not spend any time in Wetland features (0%). In comparison to use, the most common available habitat feature within rural-living companion cat buffers was Grassland (~70%), followed by Cover (~20%), Sealed (~3%), Wetland (~2%), and lastly, Built (~2%) features (Figure 2.20a; Appendix 10 - Table 1). Urban-living companion cats spent approximately 80% of their time in Built features,

Chapter Two: Companion cat spatial ecology followed by Sealed (~ 10%), Grassland (~ 7%) and Cover (~ 2%) features (Figure 2.20b;

Appendix 10 - Table 2). Wetland was not available to urban-living companion cats (0%). The most common habitat feature available was Built (~ 45%), followed by Grassland (~ 30%), Sealed (~ 17%), and Cover (~ 7%) (Figure 2.20b; Appendix 10 - Table 2).

Chapter Two: Companion cat spatial ecology

Figure 2.20: A comparison of the average proportion area, available resource units and used GPS locations for the second- order (Buffer) selection: Complete, for (a) rural-living and (b) urban- living companion cats in the Te Anau Basin in the South Island, NZ. Available resource units were randomly generated throughout each cat’s buffer and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s buffer. Error bars are + 2 standard errors. Note: The Built category is a combination of the Urban and Building categories which was required to increase sample size.

Chapter Two: Companion cat spatial ecology

At the third-order of selection (within the 100% MCP; HR), rural-living companion cats

spent approximately 40% of their time in Grassland features followed by Cover (~ 30%), Built (~ 18%) and Sealed features (~ 8%) (Figure 2.21a; Appendix 10 - Table 3). Within the 100% MCP, rural-living companion cats did not spend any time in Wetland features as they were not available (0%). The most common habitat feature available within rural- living companion cat 100% MCPs was Grassland (~ 61%), followed by Cover (~ 30%), Sealed (~ 6%), and Built (~ 6%) (Figure 2.21a; Appendix 10 - Table 3). Urban-living companion cats spent approximately 80% of their time in Built features followed by Sealed (~ 10%), Grassland (~ 7%) and Cover (~ 2%) (Figure 2.21b; Appendix 10 - Table 4). The most common habitat feature available to urban-living companion cats within their 100% MCPs was Built (~ 60%) followed by Grassland (~ 30%), Sealed (~ 17%), and lastly, Cover (~ 5%) (Figure 2.21b; Appendix 10 - Table 4).

Chapter Two: Companion cat spatial ecology

Figure 2.21: A comparison of the average proportion area, available resource units and used GPS locations for the third-order (HR) selection: Complete for (a) rural-living and (b) urban-living companion cats in the Te Anau Basin in the South Island, NZ. Available resource units were randomly generated throughout each cat’s buffer and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s 100% MCP. Error bars are + 2 standard errors. Note: Built a combination of Urban and Building which was necessary to increase sample size.

Chapter Two: Companion cat spatial ecology

2.4 Discussion

2.4.1 Movements into conservation sensitive areas

Throughout this chapter, I have assessed companion cat spatial ecology in order to evaluate potential barriers to movement and possible impacts that companion cats living in the Te Anau Basin near Fiordland National Park (FNP) and the Kepler Mire conservation area (KMCA), might be having on biota living within. Encouragingly, even with unrestricted access (across bridges) into each of these areas and recorded movements of over a kilometre from the owner’s home, only one cat was observed using the KMCA and no cats were observed using the Kepler area of FNP. Exploitation of conservation-sensitive areas by companion cats living near protected areas in Otago, New Zealand (Metsers et al. 2010) and in New York State in North America (Kays & DeWan 2004) were also rare, even though access was unrestricted. In contrast, encroachments by free-ranging owned cats into Ojcow National Park in Poland were highly likely to negatively influence rare and protected species residing within this area (Wierzbowska et al. 2012). Companion cats might also pose a risk to prey species residing within habitats adjacent to formal National Park boundaries, as in the Booderee National Park in New South Wales, Australia (Meek 2003). Even if movements of companion cats into conservation-sensitive areas are rare, prey species themselves are not restricted to protected areas and cats might still contribute to declines through predation and sub- lethal effects (Dauphine & Cooper 2009, Bonnington et al. 2013). Companion cats are capable of travelling several kilometres from their home-base (Meek 2003, Metsers et al. 2010, this study), so it is possible that cats living within approximately two and a half kilometres of the Kepler area of FNP might regularly visit the area. Despite the large movements observed by companion cats in Te Anau, the positive selection for the type of habitat features found within the Kepler area of FNP (trees and scrub), and the proximity of the owners’ homes to the bridges (the closest home of a tracked cat was 1.6 km, while the closest home was 867 m from the control gates), forays in the direction of the control gates were rare. Companion cats might be discouraged from entering FNP due to water bodies (Waiau River, Lakes Te Anau and Manapouri) that restrict access to the bridges. These bodies of water largely disrupt the continuity of available habitat for cats and provide natural barriers to movement. In contrast, access into the KMCA is completely unrestricted as available habitat is completely contiguous. While potential physical barriers to movement might have

Chapter Two: Companion cat spatial ecology prevented companion cats entering FNP, there were no physical barriers into the KMCA,

and the movement of one cat into the KMCA was likely due to favourable habitat and the absence of any effective barriers. The extent to which companion cats penetrate into a conservation-sensitive area might also depend on the availability of the habitat within. While geographical barriers (e.g. rivers and lakes) can physically prevent movements, social barriers might also exist to limit companion cat movements (Liberg et al. 2000). The pattern of locations overlain on aerial photographs of the Te Anau Golf Course (see Appendix 11) suggest the golf course could be acting as a barrier to cat movements towards the control gates. Anecdotal evidence suggests there are unowned cats living on the grounds which might exclude companion cats by competing for space, food and preferred hunting and resting sites (See review: Ritchie & Johnson 2009). In North America and Australia, coyotes (Canis latrans) and red foxes (Vulpes vulpes), are suggested to competitively exclude (through exploitation and interference) feral cats and might limit feral cat movements into natural areas (Molsher 1999, Gehrt et al. 2013). In the absence of coyotes and foxes in New Zealand, feral cats in Te Anau might be acting to exclude companion cats from the area by aggressively defending their territories. The solitary nature of cats is widely reported within the literature (Leyhausen 1979, Calhoon & Haspel 1989, Natoli et al. 2001) (although cats can also exhibit group territoriality, Liberg 1980) and might act as a social barrier to movement. Even if social barriers prevent movement of companion cats across the Te Anau Golf Course into the Kepler area, unowned cats might still be crossing the control gates, entering the park and potentially breeding with resident feral cats and posing a risk to native wildlife. If this is the case, removal of unowned cats might lead to expansion by companion cats onto the golf course grounds, closer to FNP. If, however, there are no social barriers to companion cat movements, companion cats might avoid the golf course during the day, in response to human activity. Without further investigation of the cat populations inhabiting the golf course, this cannot be assessed.

2.4.2 Factors influencing companion cat movements

2.4.2.1 Overview

The estimated home range sizes of companion cats in the Te Anau Basin were generally larger than companion cats tracked previously (Barratt 1997, Meek 2003, Lilith et al.

Chapter Two: Companion cat spatial ecology

2008, Morgan et al. 2009, Horn et al. 2011, Coughlin & van Heezik 2014, Thomas et al.

2014, Wood et al. 2016), but comparable to those tracked in Otago, New Zealand (Metsers et al. 2010). But, 100% MCP and OREP home ranges of companion cats recorded here were much smaller than previously recorded 100% MCP and LoCoH home ranges to characterise feral and stray cat home ranges. This difference is possibly due to different types of ownership (i.e. owned versus un-owned) and the distribution of reliable food sources (Horn et al. 2011, Gehrt et al. 2013, Recio & Seddon 2013). Companion cats are generally not required to search for food as it is supplied by their owners, potentially resulting in reduced ranging behaviour. While a broad range of home range sizes were recorded in this study (100% MCP: 0.63 - 109.53 ha, OREP: 0.09 - 4.30 ha), I found sex, property type and time of day were the factors that influenced companion cat movements. Younger companion cats have previously been found to roam further and catch more prey (Morgan et al. 2009, van Heezik et al. 2010), but this was not observed in the Te Anau Basin. My sample of cats younger than 18 months was small (n = 2), which might have made any age related differences difficult to detect. It is possible, however, that companion cats might establish their home ranges at a relatively young age in the Te Anau Basin, with the result that cats of any age might equally encounter native wildlife.

2.4.2.2 Sex-related effects

I found that movements of male companion cats in the Te Anau Basin were much larger than female movements using both home range estimators and Euclidean distances, in contrast to previous studies where this difference was not statistically observed. Despite the lack of significance in previous studies, males tended to have larger average home range sizes than females (Barratt 1997, Meek 2003, Lilith et al. 2008, Morgan et al. 2009, Metsers et al. 2010, van Heezik et al. 2010, Horn et al. 2011, Thomas et al. 2014, Wood et al. 2016); however, this might have depended on the property type or home range estimator used (Metsers et al. 2010, Horn et al. 2011). Given the high variability of cat ranging behaviour, and the often small sample sizes of these studies, marked differences regarding sex-related ranging behaviour might be difficult to detect. Companion cats are often sterilized before or around the age of puberty (six – 14 weeks old) which has been suggested to contribute to the lack of sexually dimorphic

Chapter Two: Companion cat spatial ecology ranging behaviour (Kustritz 2002, Lilith et al. 2008). Fortunately, all cats volunteered for

use in my study had been sterilized, usually to prevent unwanted kittens and sex-related behaviours (e.g. spraying, howling; Pollard 1999). Sterilization prior to or around the age of maturation might act to reduce home range size, particularly of males, as movements of early sterilized male and female individuals appear to be more comparable, possibly due to the absence of sex-mediated dispersal (Lilith et al. 2008, Metsers et al. 2010, van Heezik et al. 2010). In contrast, free-ranging male cats sterilized as adults (post-puberty) on Catalina Island in California and in Chicago, exhibited larger ranges than females probably because these cats have already established their home range (Guttilla & Stapp 2010, Gehrt et al. 2013, respectively). Sexually dimorphic ranging behaviours have also been observed in sterilized, owned, free-ranging dogs (Canis familiaris) living on Bathurst Island, Australia (Sparkes et al. 2014). In the Te Anau Basin, companion cats might have established ranges prior to sterilization, resulting in the observed difference. Furthermore, cat age had very little effect on movements, indicating home range size might have been established at a young age. If sterilization were to have a negative effect on range size (notably male), the procedure should take place around puberty as sterilization of adults might not reduce the hormone-driven ranging behaviour of either sex (Guttilla & Stapp 2010). Additionally, expression of male sexual behaviours might extend post sterilization, given prior sexual experience (Rosenblatt & Aronson 1958), potentially causing males to continually seek females. Future studies might be able to incorporate age of sterilization into analyses in order to identify if this is a factor influencing ranging size between the sexes. In addition to hormonal-mediated differences between the sexes, body mass is also speculated to influence home range size (Haspel & Calhoon 1989, Germain et al. 2008). Larger animals are able to hold larger areas than smaller animals and have more energy to expend for foraging, hunting and exploration (Biro et al. 2004). As I found males to be heavier than females, it is possible this body mass difference between the sexes might have been influencing ranging behaviour. I did not investigate energy expenditure of cats, so conclusions drawn about potential energetic requirements of larger animals must be interpreted with some caution. Male cats might also be more confident than female cats and engage in riskier and explorative behaviours (e.g. crossing roads, encountering other cats, consuming substances away from the home and exploration of drain systems and house crawlspaces) (Loyd et al. 2013a), potentially ranging further. Bold temperaments might increase 72

Chapter Two: Companion cat spatial ecology transmission risk of certain diseases such as feline immune deficiency virus providing an

incentive for owners to keep their cats indoors (Natoli et al. 2005).

2.4.2.3 Property type As with other cats tracked in New Zealand and elsewhere, rural-living companion cats in the Te Anau Basin showed considerably larger home ranges and longer excursion distances from home than urban-living cats (Barratt 1997, Meek 2003, Biro et al. 2004, Kays & DeWan 2004, Metsers et al. 2010, van Heezik et al. 2010, Horn et al. 2011, Wierzbowska et al. 2012). While the spatial ecology of rural-living cats is often highly individual, movements might also be influenced by human modifications to the environment, such as the density and position of buildings within rural properties, traffic levels on nearby roads, as well as territorial pressure from conspecifics (Barratt 1997, Hall et al. 2000, Metsers et al. 2010). Rural properties are often large (> 2 ha) with sparsely spaced resources such as resting (e.g. buildings) and hunting (e.g. buildings and cover) sites which might encourage cats to travel further to fulfil biological and social requirements (Kays & DeWan 2004, Beyer et al. 2010). Furthermore, increased prey activity (e.g. birds, lizards, and rodents) around cover and buildings (Kays & DeWan 2004) might have encouraged movement of rural-living cats towards Cover and Built habitat features and promoted movement away from open grassy (Grassland) features (resource selection discussed further below). However, discrepancies between levels of care provided to rural-living cats in different studies make reliable comparisons between studies difficult. Increased levels of urbanisation (in conjunction with smaller property sizes) and a greater number of resting and hunting sites in urban areas has been found to reduce companion cat home range size and excursion distance from home (Lilith et al. 2008, Metsers et al. 2010). High levels of supplementary feeding and additional food sources in urban areas might also contribute to reduced range size (Boutin 1990, Koganezawa & Imaki 1999, Tennent & Downs 2008), as companion cats are often provided with a reasonably constant source of food which eliminates the need to search for it. Regardless of the owner’s property type (rural or urban) all cats in my study were fed at least once per day; with some owners providing food ad libitum, indicating that spatial distribution and home range size of companion cats are likely influenced by additional factors. Furthermore, in an environment with abundant food resources, companion cats might not

Chapter Two: Companion cat spatial ecology rely on predation for subsistence, but continue to hunt, indicating a disassociation

between hunger and the expression of individualised hunting behaviours (Dickman & Newsome 2014, Kitts-Morgan et al. 2015). Additional food sources (e.g. human refuse in urban areas) might also act as an attractant, leading to increased densities of cats in urban areas (Boutin 1990). Companion cat density, which was not incorporated into analyses in my study, and consequently the level of territoriality expressed by those cats is also likely to negatively influence home range size (Liberg et al. 2000, Davison et al. 2009). Companion cat densities in urban areas in New Zealand can be high: 252 cats/km2 in Christchurch (Morgan et al. 2009) and 223 cats/km2 in Dunedin (van Heezik et al. 2010), concentrating hunting into a small area even though smaller ranging sizes might decrease a cat’s individual chance of encountering prey (Thomas et al. 2014). A reduction in cat density might lead to increased ranging behaviours of any remaining cats (Thomas et al. 2014). Companion cat density estimates do not account for any unowned (e.g. stray or feral) cats that might also contribute to predation and sub-lethal effects.

2.4.2.4 Time of day

The 100% MCPs indicated companion cats in the Te Anau Basin appeared to roam further at night than during the day; however; this effect was not observed when calculating OREPs. This discrepancy might be a result of 100% MCP ranges capturing much larger range sizes due to the inclusion of outer occasional forays which can greatly increase the size of the range (Getz et al. 2007). Cats have a well-developed tapetum lucidum, contributing to their superior scotopic vision and enabling companion cats to roam during the night (Barratt 1997, Metsers et al. 2010, Thomas et al. 2014; but see: van Heezik et al. 2010). However, unlike feral cats (Alterio & Moller 1997), companion cats exhibit tendencies towards diurnal activity, potentially in response to their close human association (Driscoll et al. 2009a, 2009b). One cat in this study was confined indoors at night as the owner was concerned about nocturnal roaming behaviour and prey capture at night. Restricted outdoor access will prevent any nocturnal roaming, although there is the potential for unrestricted cats to travel further in response to lower cat density at night. Prey species might be equally

Chapter Two: Companion cat spatial ecology vulnerable to diurnal or nocturnal impacts of companion cats given diel roaming

behaviour of companion cats does not appear to be fixed to either day or night.

2.4.3 Resource selection

Resource selection for companion cats appeared to be enhanced at a coarse spatial scale (second-order selection - Buffer) rather than a fine spatial scale (third-order selection - HR), reflecting their generalist lifestyle and remarkable ability to adapt to many different ecosystems (Coman & Brunner 1972, Konecny 1987, Jongman 2007). At both levels of selection in analyses conducted using all filtered locations (complete analyses), regardless of landscape (rural or urban), companion cats were selecting strongly for Built habitat features as has been found previously (Barratt 1997, Metsers et al. 2010, Ferreira et al. 2011, Horn et al. 2011). Selection predominantly for Built features is not surprising, given that areas of high human activity are typically a reliable food source in the form of either owner-provided or human refuse, and shelter (Warner 1985, Metsers et al. 2010, Ferreira et al. 2011). Additionally, areas of human habitation (i.e. Built features) might support communities of prey species such as rodents (Glass et al. 2009, Kitts-Morgan et al. 2015). This aspect is postulated as one of the main attractants of wild cat ancestors (most likely F. silvestris lybica) to human settlements and consequently, a component in the development of cats as companion animals (Pollard 1999, Driscoll et al. 2009a). When the analysis was restricted to locations outside the owner’s home, I found urban-living cats selected for Built features; however, resolution was restricted to the property level, as I combined urban buildings and gardens into one habitat category (Built). In the UK and New Zealand, at the property level, urban-living companion cats spent a large amount of time within their owner’s property and selected preferentially for gardens over nearby green habitats and urban land use types (e.g. buildings, roads, etc.; van Heezik et al. 2010, Thomas et al. 2014). While I was unable to ascertain if urban- living companion cats were only ever inside their owner’s home other than being inside other buildings, it was clear that urban-living companion cats in Te Anau spent the majority of their time in Built features, and this likely comprised some time in urban gardens. Urban gardens do play an important ecological role by supporting biodiverse plant and animal communities (Parsons et al. 2006, van Heezik et al. 2013) and use of these environments by companion cats will lead to wildlife encounters.

Chapter Two: Companion cat spatial ecology

When outside the owner’s home, rural-living cats were selecting predominantly

for Cover features, which is seen throughout the Felidae (Edwards et al. 2002). Cover habitat supports their stalk-hunting strategy and is home to a variety of prey species (Hall et al. 2000, Edwards et al. 2002), enabling greater hunting success than in other habitat types, such as grassy areas. Cover is also thought to provide protection from predation (Edwards et al. 2002); however, the New Zealand environment is devoid of predators of cats, unlike in international localities (Edwards et al. 2002, Gehrt et al. 2013, Allen et al. 2014). While rural-living companion cats in the Te Anau Basin spent more of their time in Grassland habitat features than any other habitat type, Grassland habitat was largely more available than any other habitat feature, indicating use was linked to availability rather than preference. Indifferent use of Grassland features by rural-living companion cats might be a response to movement through grassy areas between Built and Cover habitat features where small mammal prey are abundant (Hall et al. 2000, Edwards et al. 2002). Rodents are often the most common prey item brought home, providing support for use of Cover habitat features where small mammalian prey might be found and caught (Fitzgerald & Turner 2000, Metsers et al. 2010, Wood et al. 2016). Indifferent use of grassland habitats, such as pasture, with a lack of cover and low species abundances might discourage cats from selecting grassland features (Warner 1985, Hall et al. 2000, Daniels et al. 2001, Metsers et al. 2010). Sealed features were also selected for at a coarse scale but not at a fine scale. Movement through the environment might be easier on roads; however, this might result in cats being more vulnerable to vehicle-related accidents. Even though companion cats have previously been recorded using Wetland habitat (Morgan et al. 2009), companion cats in the Te Anau Basin were ambivalent towards wetlands, perhaps due to the long distance these were located from home.

2.4.4 Risk posed to prey species

Over the short prey capture survey, companion cats in the Te Anau Basin were recorded capturing mammalian, avian, reptilian, and invertebrate prey species. These results are not unusual as many previous studies describe a similar range of fauna caught by companion cats (e.g. Barratt 1998, Metsers et al. 2010, van Heezik et al. 2010, Tschanz

Chapter Two: Companion cat spatial ecology et al. 2011, Krauze-Gryz et al. 2012, Loyd et al. 2013b, Loss et al. 2013, McDonald et

al. 2015), and this is characteristic of generalist opportunistic feeders (Sims et al. 2008). While no native prey species were identified (although chicks were unidentifiable to species level), exotic species are likely to be more abundant in urban areas (van Heezik et al. 2008), scrubland (Macleod et al. 2012a) and agricultural habitats (Macleod et al. 2012b) and low predation rates of native species might be a reflection of low prey numbers (Beckerman et al. 2007). Predation of reptiles by companion cats occurs elsewhere in New Zealand (van Heezik et al. 2010, Gaby 2014, although see: Metsers et al. 2010) and overseas (Australia: Dickman & Newsome 2014; North America: Loyd et al. 2013b) and might have a significant impact on these communities (Middlemiss 1995). Some individuals that have preferred prey types might travel further to find favoured prey (Barratt 1997, Meek 2003, Dickman & Newsome 2014). Sub-lethal effects might also contribute to prey productivity declines, where the presence of cats influences prey behaviour and reduces fecundity (Bonnington et al. 2013). However, prey survey results might not capture true predation rates due to a difference between captured, eaten or abandoned, versus returned prey (Loyd et al. 2013b). Additionally, owners might under- or over-report predation depending on the owner’s perceived acceptable cat predation rate (Baker et al. 2008, Loyd et al. 2013b, McDonald et al. 2015). Reconciling this difference might be difficult; however, direct observations and cat-borne video cameras offer some insight into actual prey capture rates. Observed hunting behaviours of 11 radio-tagged cats found that predation rates were greatly underestimated when using owner survey information and might be three times higher than results based solely on owner surveys (Kays & DeWan 2004). Additionally, video-borne studies of companion cats in North America reported less than a quarter of captures were returned to owners (Loyd et al. 2013b). The use of collar- mounted cameras and accelerometry have become useful to characterise companion cat fine scale behaviour associated with prey hunting (Wilson et al. 2012, Loyd et al. 2013b, Coughlin & van Heezik 2014, Gaby 2014)

2.4.5 Sources of Error

While GPS units provide an exceptional tool for investigating animal spatial ecology, measurement error including missing data (fix success rate - FSR) and inaccurate locations (location error - LE) can introduce bias into a study. These biases can influence 77

Chapter Two: Companion cat spatial ecology the conclusions drawn from animal tracking data and must be acknowledged in order to

produce reliable results.

2.4.5.1 Device size

Improvements in GPS device battery life have increased the length of time a device can be deployed, whilst maintaining or reducing the size and weight of the device (Dennis et al. 2010). This enables medium-to-small animals e.g. companion and feral cats (Metsers et al. 2010, Recio et al. 2010, Coughlin & van Heezik 2014) and brush-tail possums (Adams et al. 2014b), to be fitted with trackers that do not exceed 5% of the animal’s body mass (BM). This arbitrarily defined limit, however, might still cause discomfort and have negative impacts on a tracked animal’s behaviour, energy requirements, survivability and reproduction (Hawkins 2004, Wilson & McMahon 2006). Companion cats tracked in Dunedin, Otago exhibited reduced movements when wearing a heavier collar (~ 136 g) compared to a lighter collar (~ 36 g), indicating that devices might be required to be less than 3% or 2% of BM to prevent animals exhibiting altered spatial behavioural patterns (Coughlin & van Heezik 2014). Cats might, for example, exhibit more sedentary behaviours by spending more time resting, than they would otherwise. The weight added to the collar I used acted as a counter balance in order to reduce the obstruction of the signal by the cat’s body by keeping the GPS device dorsally positioned. This increased the acquisition rate of locations (fix success rate; FSR) of the device and the amount of data collected from an animal (Coughlin & van Heezik 2014). All collars used in my study were less than 3% BM of each cat, in order to maximise data collection while minimising behavioural alteration. Unfortunately, as there are no tracking devices that are small enough to have a completely negligible impact on an animal’s behaviour, efforts can only be made to minimise bias within a study.

2.4.5.2 Fix success rate (FSR) The data-logging capabilities of many GPS devices allow an animal to be tracked 24 hours a day with minimal human interaction disrupting normal behaviours. While data- logging features decrease human interactions with a tracked animal, interference from habitat configuration (e.g. open space versus canopy closure; sky availability) might obstruct or reflect signal transmission and prevent devices from acquiring successful locations (i.e. FSR, D’Eon et al. 2002, Jiang et al. 2008, van Heezik et al. 2010, Recio et 78

Chapter Two: Companion cat spatial ecology al. 2011, Adams et al. 2013). Locations could, therefore, be disproportionately biased

towards one habitat type regardless of animal behaviour (Frair et al. 2004). Buildings and vegetative cover can both interfere with FSR, leading to an underestimation of use of these habitat types (Recio et al. 2011, Adams et al. 2013, Coughlin 2014). Consequently, the use of Built and Cover features by Te Anau cats might have been underestimated. Missing data can bias resource use as the importance and types of habitat features used might be misidentified. Furthermore, home range size might also be underestimated. While I employed the use of a counterbalance to increase FSR (Coughlin & van Heezik 2014), current citizen science studies (Your Wild Life 2014, Cattracker 2015) have employed the use of a harness to keep the device dorsally positioned.

2.4.5.3 Accuracy/interference of GPS devices

In addition to fix success rate, location error can also limit the ability of the device to accurately identify an animal’s position in space. Several factors (sky availability, vegetation complexity, satellite geometry, e.g. horizontal dilution of precision - HDOP, and satellite number) might influence the accuracy of an acquired location (Recio et al. 2011, Adams et al. 2013). Location error of i-gotU devices have been recorded as ~ 30 m outside and up to ~ 60 m when inside (Coughlin & van Heezik 2014). These potentially inaccurate locations might lead to misidentification of resource use (Frair et al. 2004). For example, locations might be identified outside when the cat is actually inside. I opted to estimate resource selection by inferring the distance from each used and available location to the nearest of each habitat feature (thus implementing the distance approach) to accommodate for potentially inaccurate results and characterise the use of edge habitats (Conner et al. 2003). Alternate home range methods such as Brownian bridges can also be used to infer an animal’s space use (Horne et al. 2007). Brownian bridges create a utilization distribution of an animal’s space use from regularly collected data (e.g. GPS-derived data recorded at 15 minute intervals), whilst also incorporating uncertainty associated with each recorded location (Horne et al. 2007). Fortunately, lightweight GPS performance in highly urbanised (e.g. in Dunedin, Otago) and natural habitats are comparable, indicating locations are not disproportionally biased by interference from buildings compared to cover habitat types (Adams et al. 2013).

Chapter Two: Companion cat spatial ecology

2.4.5.4 Filtering of inaccurate locations

Identification and removal of erroneous locations prior to data analysis is important to improve the reliability of a tracking data-set and prevent inaccurate estimations of an animal’s spatial behaviour. Horizontal dilution of precision/satellite number can be used to filter data, whereby increased HDOP values relate to increased LE (Recio et al. 2011). Stationary tests can be used to select appropriate HDOP values that maximise the number of accurate locations collected (Recio et al. 2011). However, as the devices I used (i- gotU), do not record HDOP/satellite number information, tracking data were required to be screened based on alternative criteria; unrealistic speeds and turning angles that are more likely to characterise location error rather than normal animal behaviours (Hurford 2009, Augé et al. 2011, Recio & Seddon 2013, Adams et al. 2014b, Coughlin & van Heezik 2014). Inspection of my data and comparisons with previous studies were used to set speed and angle thresholds for removing locations; cats might not exceed speeds of ~ 0.2 m/s (Recio et al. 2010, Coughlin & van Heezik 2014) and turning angles of 160⁰, a value used previously to characterise brush-tail possum movements in a suburban environment (Adams et al. 2014b). However, high turning angles have been found to characterise foraging behaviours of large animals (e.g., in cattle, Bos taurus; de Weerd et al. 2015).

2.4.5 Limitations of study design

I tracked cats over November and December to coincide with the austral spring/summer when temperatures are warmer, there is less rainfall and there might be higher abundances of nestling and fledglings; consequently cats might exhibit increased movements (Germain et al. 2008). However, since over 60% of the cats experienced more than 50% rain days while wearing the collar, the movements I have reported might be underestimated as increased rainfall for the majority of study cats might have discouraged the majority of companion cats from going outside. Additionally, cats might have spent more time in and around buildings, contributing to the observed selection of Built features. To allow any naïve companion cats I tracked to become accustomed to the collar, they underwent an acclimation period of three days. Unfortunately, some cats experienced chafing and one cat had trapped its lower jaw in the collar, prompting early

Chapter Two: Companion cat spatial ecology removal of the collar. Potentially, the high rainfall level coupled with largely

inexperienced wearers might have contributed to rubbing of the collar. As a result, the cats might have altered their spatial behaviour due to discomfort, producing biased home range and resource selection results (Hawkins 2004). However, incremental area analyses of companion cats tracked for less than 12 days indicated fully revealed home ranges for all but one cat, Slinky, which was tracked for only 10 days. Short tracking periods might not completely capture an animal’s movements (van Heezik et al. 2010); however, the length of time a cat was tracked (period) had little influence on home range size. The sample size here was sufficiently large to draw reliable conclusions. Unequal numbers of male and females might have contributed to a Type 1 error (Tabachnick & Fidell 2007); however, I validated my results using bootstrapping to account for this difference in sample size (Burnham & Anderson 2010). Not all cats could be included in all resource selection analyses due to the absence of habitat categories within either the HR or buffer area. While this decreased the sample size available for some analyses, sample size was comparable to that in other studies of cat resource selection. I combined several habitat features in order to create a fairly broad habitat map. A more detailed habitat map might have resulted in the identification of more specific habitat selection; however, identification of used resources is limited by the accuracy of the i-gotU device, and given the measured location error, and classifying habitats at a finer scale would have been meaningless. The similarity between the relative sizes of the areas of each habitat type and the proportions of “available” points indicates available points appropriately characterised available habitat.

2.4.6 Management recommendations and future directions

Determining home ranges of companion cats living in close proximity to conservation- sensitive areas can inform owners as to the whereabouts of their cat when outside, and aid in developing management recommendations. Due to variable and often unpredictable cat home range size, the size of cat-exclusion buffer zones around conservation-sensitive areas might be difficult to set without first determining the movements of nearby companion cats. Previous authors have suggested using the maximum distance travelled of a tracked companion cat plus a 20% margin of error

Chapter Two: Companion cat spatial ecology

(Lilith et al. 2008, Metsers et al. 2010) to accommodate any further excursions. The size

of proposed cat-exclusion buffer zones around protected areas in urbanised areas has ranged from 360 m (Lilith et al. 2008) to 1.2 km (Metsers et al. 2010) based on local companion cat spatial ecology studies. In rural landscapes, buffer zones might need to be much larger (2.4 km; Metsers et al. 2010) to accommodate larger ranging habits of rural- living cats. In the Te Anau Basin, a companion cat-exclusion zone of at least 2.88 km would be required to be effective at preventing cats from entering both the FNP and KMCA. Additionally, spaces outside conservation-sensitive areas are likely to support significant biodiversity as prey species are not restricted to park boundaries. Therefore, even if companion cats are not crossing into conservation-sensitive areas, they are likely to have an impact on prey populations. Total confinement of cats indoors will restrict companion cat movements and can provide cat-specific benefits by reducing disease transmission (between conspecifics and between cats and avifauna), and vehicle-cat and conspecific interactions especially on busy roads (Jongman 2007, Loyd et al. 2013a). Cats kept indoors tend to live longer than outside cats (Warner 1985); however, total cat- exclusion or confinement in buffer zones around conservation-sensitive areas might not be widely accepted by the public (Grayson et al. 2002, Lilith et al. 2006, McDonald et al. 2015; although success has been reported in Australia: Baker 2001). There are also concerns regarding meso-predator release (e.g. contributions to increased predation or habitat degradation by invasive prey species following cat exclusion, Courchamp et al. 1999, Bergstrom et al. 2009, Dowding et al. 2009) that requires consideration and might require the development of alternate management strategies. Cat-free zones in new property developments or night curfews have been more acceptable to the Australian public (Buttriss 2001). Coordinated cat management and rodent control might be able to resolve the potential increase of rodents following cat exclusion. Even though rodents might be more harmful to native wildlife, rodent-only control might lead to prey switching of companion cats if small mammals are generally the main prey item. Collar-worn predation deterrent devices might be suitable alternatives to limit predation and maintain cat ownership near conservation-sensitive areas. Acoustic deterrents, such as bells and ultrasonic devices, worn on the collar have been found to reduce total prey capture by 53% in New Zealand (Gordon et al. 2010), ~ 50% across the UK (Nelson et al. 2005) and in England (Ruxton et al. 2002). While these results are encouraging, these studies were conducted from four to six weeks without the potential 82

Chapter Two: Companion cat spatial ecology for habituation to bell-wearing while hunting. Furthermore, several owners in Te Anau

remarked on the ability of their cat to bring home more prey when wearing a bell than they had done previously without. The owner(s) had then removed both the bell and collar. Morgan et al. (2002) also noted that urban bell–wearing companion cats did not bring home significantly different numbers of prey than urban non-bell-wearing cats. Bell-wearing cats might have been more active hunters regardless of bell-wearing, which might have been the case for the cats in my study. Pounce-protection devices, such as the CatBibTM, limit a cat’s movement when hunting and stopped 81% of cats capturing birds, 33% of cats catching herpetofauna and 45% from catching mammals in Western Australia (Calver et al. 2007). Pounce- protection devices might be highly beneficial in Australia with mammalian endemism as high as 87% (Chapman 2009). In the New Zealand context, prevention of mammal capture might not be the most appropriate option as all small terrestrial mammals are invasive. A device that limits a cat’s ability to catch rodents for example, might have disastrous flow-on effects for shared prey species. Visual deterrents, such as the BirdsBeSafe® collar, however, have reduced cat capture of avifauna and herpetofauna by 47%, but not of mammalian prey in Western Australia (Hall et al. 2015). In North America, captures of birds were reduced by 19 times (Willson et al. 2015). As these studies were conducted over a short period of time, as with the bells, long term studies should be conducted in order to determine the long term effects of wearing such devices and if the protection remains. The BirdsBeSafe® device exploits the existence of the well-developed colour receptors in the eyes of birds and herpetofauna alerting them to a cat’s presence (Hall et al. 2015). The Te Anau Basin cats could wear this device when outside; however, this would require compliance from owners. Regardless of the efficacy of a device, it’s use is limited by owner education regarding impacts of companion cats on wildlife. Additionally, owners must be willing to use such devices or comply with proposed management strategies (McDonald et al. 2015). Legislation surrounding cat ownership (e.g. restrictions on number owned as in Invercargill; Invercargill City Council 2013) may encourage owners to be more aware of their pet’s impact on wildlife and to practice responsible ownership. Mandatory registration and microchipping of cats (as with dogs, Department of Internal Affairs 2015) will connect owners with lost pets, and mays encourage owners to sterilize their cats to prevent unwanted kittens (which would require registration/microchipping), potentially reducing abandonment and prevent breeding 83

Chapter Two: Companion cat spatial ecology with stray populations (Natoli et al. 1999, 2006). Fortunately, all cats in my study were

sterilized; however, some had been allowed to breed prior to sterilization. While property confinement will restrict a cat’s roaming behaviour (all cats in my study moved outside their owner’s property), any wildlife using resources within properties might be at risk. As movement of companion cats in the Te Anau Basin into FNP is possibly limited, an assessment of the presence and density of unowned cats in the Te Anau Basin might be the next step to assess unowned cats on FNP, the KMCA and the surrounding landscape.

Chapter Three: Feral cat genetics

Chapter Three:

Identifying population genetic structure of feral cats (Felis catus) to aid management strategies in conservation-sensitive areas

The Ahuriri River, flanked on each side by mountains and beech forest, meanders down the Ahuriri Valley past matagouri (Discaria toumatou) stands and tussock grasslands.

Chapter Three: Feral cat genetics

3.1 Introduction

Biological invasions are considered one of the leading causes for significant biodiversity loss across the world (Allen & Lee 2006, Nentwig 2007). Invasive mammalian predators are thought to have contributed significantly to many declines and extinctions of native biodiversity on the mainland and islands worldwide (Nentwig 2007, Keitt et al. 2011). In New Zealand, invasive mammalian predators, such as feral cats (Felis catus), have led to the decline and extinction of many endemic fauna (Fitzgerald & Turner 2000, Allen & Lee 2006, Bonnaud et al. 2011). Eradication programmes to remove invasive mammalian predators aim to alleviate the detrimental impacts caused by invasive species and restore host ecosystems (Courchamp et al. 2003, Campbell et al. 2011). Eradication of invasive mammalian predators greatly benefits from an understanding of dispersal pathways that lead to naturally occurring population structure. Natural putative geographic barriers can aid in classifying invasive species into manageable isolated ‘eradication units’, where reinvasion is unlikely to occur (e.g. identification of presence of glaciers and ice that restrict rat, Rattus norvegicus, movement on South Georgia Island; Robertson & Gemmell 2004). The presence of natural geographic barriers, separating populations from each other, have been invaluable to many species conservation programmes by preventing reinvasion of invasive species (Moorhouse & Powlesland 1991, Towns 2011). Following on from the success of small island eradications where eradication units can be defined as whole islands, mainland and larger island projects have been undertaken where identification of eradication units might not be as straightforward. Molecular genetics have aided the identification of eradication units, population connectivity and the potential for reinvasion into managed areas (e.g. Robertson & Gemmell 2004, Hansen et al. 2007). Eradication efforts aimed at controlling a small portion of a larger population or a sink population are likely to be costly, ongoing and unsuccessful due to high levels of reinvasion (Hanski 1999, Bogich et al. 2008). In mainland conservation-sensitive areas, reinvasion of target invasive species is a common reoccurring problem even with extensive poisoning and trapping operations (e.g in the Tasman Valley, Rebergen & Woolmore 2015). Identification of geographic barriers that limit dispersal can provide information about target and help to determine if managed populations can be classified into isolated eradication units. Fortunately, molecular genetics can provide population genetic information to elucidate gene flow and genetic

Chapter Three: Feral cat genetics differentiation between populations, to identify if reinvasion can be reduced by exploiting natural putative barriers to invasive species dispersal. In New Zealand, feral cats have had devastating effects on endangered wildlife with current control efforts in place to reduce their impact (e.g. Rebergen & Woolmore 2015). Molecular genetic tools have been used overseas to inform feral cat control projects regarding feral cat dispersal in insular environments. On Hawai’i Island, for example, evidence for gene flow and migration between three sampled sites was found, indicating distance (> 60 km) and lava flows (putative geographic barriers) were not barriers to feral cat dispersal (Hansen et al. 2007). Without dispersal barriers, eradication units could not be specified and reinvasion into trapped areas was deemed highly likely, posing a threat to the native endemic fauna of the Hawai’ian archipelago. Reinvasion of feral cats into trapped areas is also of concern on Australian offshore islands, specifically, Dirk Hartog Island, a candidate island for native translocations following feral cat control (Koch et al. 2014). Multiple human-mediated invasion events of Dirk Hartog Island from different mainland feral cat source populations likely occurred, resulting in an observed increased genetic variation of the island-sampled sites (Kolbe et al. 2004, Koch et al 2014). In contrast, genetic differentiation between sampled sites on Grande Terre (a sub- Antarctic Island in the Kerguelen Archipelago) and mainland France, suggest existing effective barriers to movement (Say et al. 2003, Pontier et al. 2005). Barriers on Grande Terre might have only recently become effective after the past initial rapid colonisation process, hence providing evidence for potential eradication units and successful local eradication of feral cats (Pontier et al. 2005). The aforementioned studies highlight the dispersal ability of feral cats as well as the recovery potential, high fecundity and population connectivity of feral cats, which makes defining eradication units for cat management difficult. Feral cat management in New Zealand has been greatly aided by identifying their spatial ecology. Studies of feral cat home range and habitat use throughout New Zealand have identified very large home ranges of feral cats, often over 1000 ha (Gillies et al. 2007, Harper 2007, Recio et al. 2010, Recio & Seddon 2013, Cruz et al. 2014) with home range configuration likely determined by rabbit-occupied habitat (Recio & Seddon 2013, Cruz et al. 2014) or the availability of shelter in forested areas where preferred rat (Rattus sp.) prey abundance is ubiquitous across the landscape (Harper 2007). However, competitive exclusion by dominant feral cats from areas that support high abundances of preferred prey species has led sub-ordinates to occupy important habitat for endangered 87

Chapter Three: Feral cat genetics species, posing an even greater risk to endangered prey (Cruz et al. 2014). Additionally, feral cat movement is likely to be facilitated by man-made structures (i.e. tracks and roads) that reduce movement energetic costs through and between preferred habitats (Recio et al. 2015). Eradication units might be identified through discovery of feral cat reinvasion pathways and barriers to dispersal that restrict reinvasion into managed areas. As yet, large-scale dispersal patterns of feral cats in New Zealand has not been assessed using molecular genetic tools. Using measures of genetic diversity, gene flow, relatedness and migration, I explored the population genetic structure and potential dispersal pathways between feral cats sampled from three sites in the upper Waitaki Basin and at Macraes Flat in the South Island of New Zealand. The braided Tasman, Ohau and Ahuriri Rivers, situated in the upper Waitaki Basin, are conservation-sensitive areas frequented by various New Zealand endemic ground nesting birds (Keedwell et al. 2002a, Sanders & Maloney 2002). Intensive predator control programmes (Project River Recovery and Kaki Recovery Programme) managed by the Department of Conservation (DOC) and funded by hydro- electric power generation companies, Meridian and Genesis Energy (Rebergen & Woolmore 2015), are currently in operation in the Tasman Valley and at the upper Ohau River to target invasive mammalian predators including feral cats (Cleland et al. 2013). However, even though trapping is extensive, feral cat reinvasion occurs every year. Using molecular genetic tools, I aimed to identify the potential for feral cat populations to be managed as separate eradication units by determining isolation, estimated genetic differentiation and connectivity between different sites in the upper Waitaki Basin, using Macraes Flat as an outgroup. Genetic information can aid current management efforts by identifying the connectivity between feral cats sampled at different sites and define putative geographic barriers to movement.

Chapter Three: Feral cat genetics

3.2 Materials and methods

3.2.1 Study Sites

The Tasman and Ahuriri Valleys and the upper Ohau River (hereafter: Ohau River) are all situated within the upper Waitaki Basin near the township of Twizel (44.2500° S, 170.1000° E) (Figure 3.1). The Ohau River site is situated closest to the Twizel township (Euclidean distance ~ 7 km away across a bridge) compared to the Tasman Valley (Euclidean distance ~ 46 km) and Ahuriri Valley (Euclidean distance ~ 38 km) sites. The Ohau River and Tasman Valley sites are both conservation-sensitive areas that are currently managed by DOC for invasive predators and support high levels of endemic ground-nesting birds including the nationally critical Kaki (Black stilt; Himantopus novaezelandiae), nationally vulnerable banded dotterel (Charadrius bicinctus), wrybill (Anarhynchus frontalis) and the nationally endangered black-fronted tern (Chlidonias albostriatus) (Keedwell & Brown 2001). The Ahuriri Valley, also in the upper Waitaki Basin, is frequented by some of New Zealand’s ground-nesting birds (Sanders & Maloney 2002, per. obs.). Macraes Flat (45.2600° S, 170.2500° E), which lies ~ 140 km southeast of the upper Waitaki Basin, is another conservation-sensitive area close to human habituation (e.g. farms) that is extensively trapped and offers an outgroup from the upper Waitaki sites as is beyond the dispersal ability of feral cats (Figure 3.1). The Maraes Flat site provides important habitat for many endangered skink species, most notably, the Grand (Oligosoma grande) and the Otago skinks (O. otagense) requiring protection from mammalian predators (Norbury et al. 2006). In the Tasman Valley, seasonal snow and glacier melts from the surrounding mountains and moraines flow into the u-shaped valley, feeding the braided Tasman River (Kitson & Thiele 1910, Recio et al. 2010). The upper Ohau River, fed from Lake Ohau, lies south of the Tasman Valley (Figure 3.1). The Ohau riverbed is largely formed by sparsely vegetated expanses of gravel, creating ‘islands’ used by colonial nesting birds (Keedwell et al. 2002a, Rebergen & Woolmore 2015). The flows into the upper Ohau River have been greatly reduced (average flows of < 5 m3sec-1) due to water control for hydro-electric power generation (Rebergen & Woolmore 2015). The Ahuriri Valley lies to the south west of the Tasman Valley and Ohau River (Figure 3.1). The Ahuriri Valley floor is occupied by the braided Ahuriri River and shows typical glacial landforms with glacial gravel outwashes travelling downstream to form alluvial flats (Robertson et al. 1983). The river is set within tussock grasslands (Chionochloa rubra, Festucas sp., Poa 89

Chapter Three: Feral cat genetics sp.), surrounded by beech forest (Fuscospora sp.), matagouri (Discaria toumatou), swamp and sub alpine and alpine ecosystems (Robertson et al. 1983). The upper Ahuriri Valley lies within the Ahuriri Conservation Park, previously protected by predator control operations. However, logistic and time constraints led to the end of the operation in 2012. Macraes Flat to the south east of the upper Waitaki Basin (Figure 3.1) is characterised by exposed ridges and schist tors throughout tussock grassland hill country, providing habitat for some of New Zealand’s nationally endangered skink species (Norbury et al. 2006).

Chapterpopulation genetics Three: Feral cat

Figure 3.1: Sampling study sites of feral cats (Felis catus) in the upper Waitaki Basin and at Macraes Flat in the South Island of NZ. Triangles represent the sampling location and circles represent trap location. Tasman Valley, n = 60, Ahuriri Valley, n = 33, Ohau River, n = 70, Macraes Flat, n = 28.

Chapter Three: Feral cat population genetics

3.2.2 Feral cat trapping and sample collection

Feral cat samples were collected by DOC contractors and staff at the Ohau River, Tasman Valley and Macraes Flat sites during their normal predator control operations in 2014 and 2015 (Macraes Flat samples were only collected in 2015). Samples from DOC were collected under their standard operating procedures (SOP) for trapping feral cats. Within the Ahuriri Valley, I sampled feral cats from 15 April 2015 to 13 May 2015 during the austral autumn to coincide with predicted high feral cat numbers. Initially, I set up a ~ 8 km trap line with double spring Conibear traps (‘Twizel’ kill trap system to kill feral cats humanely; supplied by DOC) set in pairs at ~ 250 m intervals along Birchwood Road in the Ahuriri Conservation Area (Figure 3.2). Traps were preferentially set at least 10 m from the road and to maximise feral cat capture rate, traps were set under ‘cover’ habitats, generally stands of matagouri (Discaria toumatou) (S. Aitcheson pers. comm.; Figure 3.2). I opted to use a single trap rather than a double trap to increase the trapped area. I set each trap on 18 mm ply-wood board, and covered each trap with a Philproof double cat trap tunnel cover (supplied by DOC). I covered one end of each tunnel cover with wire mesh to direct feral cats into the Conibear trap while maintaining the view through the tunnel. On the advice of DOC feral cat trappers, I later extended the trap line south along Birchwood Road and into the Ben-Avon, Longslip and Ribbonwood areas to increase feral cat captures after obtaining landowner permission. This required traps to be moved from their original location. I baited each trap with rabbit (Oryctolagus cuniculus) meat. I cleared the traps every two to three days and rebaited them every five to six days, weather permitting. I recorded the GPS location of each trap on a Garmin handheld GPS. I also recorded date and trap number of the trap for each feral cat and any by-catch caught (Appendix 12). I took an ear biopsy from each feral cat and stored it in 90 - 100% ethanol at 4°C until DNA analysis. I also recorded sex and age class data for each captured feral cat if possible. I sexed each cat based upon the position and morphology of the anus and genital openings. Feral cats were also sexed by DOC staff (if possible) based on external morphological characteristics at the Tasman Valley, Ohau River and Macraes Flat sites (S. Aitcheson, S. Anderson, P. Liddy pers. comm.). I also aged each cat (< 1 year i.e. juvenile, > 1 year i.e. adult) based on tooth eruption, colour and morphology (Verstraete et al. 1996, Kressin 2009). Age class information was also collected for feral cats by

92 Chapter Three: Feral cat population genetics

DOC staff at the Tasman Valley and Ohau River sites. The Tasman Valley site assessed feral cat age based on dentition, while the Ohau River site aged feral cats based on size (Table 3.1). Therefore, feral cat ages reported for each site are approximate and very broad (Table 3.1). Most of the feral cat samples from the Ahuriri Valley were approximately less than one year old (< 1 year, Table 3.1). Most of the feral cat samples from the Tasman Valley and Ohau River sites were recorded as ‘adults’ (> 1 year) (Table 3.1). I removed the feral cats and any by-catch I captured from my trapping operation in the Ahuriri Valley site for disposal of by DOC. I visualised each study site and feral cat captures onto ortho-rectified aerial photographs of the upper Waitaki Basin and Macraes Flat (taken 2004-2010, NZTM2000 map projection, 0.5 m pixel resolution, 1:5,000 layout) (www.linz.govt.nz) in ArcGIS 10.1 (ESRI 2014).

Table 3.1: Age of feral cat samples from three sites in the upper Waitaki Basin in the South Island. Age data was not recorded at the Macraes Flat site.

Age (years) Site < 1 > 1

Tasman Valley 7 53 Ohau River 22 48 Ahuriri Valley 28 5

Total 47 106

Chapter Three: Feral cat population genetics

Figure 3.2: Location of a covered Conibear feral cat trap underneath a stand of matagouri (Discaria toumatou) in the Ahuriri Valley.

3.2.3 Multiplex genotyping

I extracted genomic DNA from tissue samples using a 5% Chelex protocol (Walsh et al. 1991). Each sample was genotyped using the “Meowplex”, an 11 panel microsatellite loci multiplex plus a sex identification marker specific to the Y-chromosome (Butler et al. 2002, Menotti-Raymond et al. 2007). I used a Qiagen Type-it Microsatellite PCR kit to perform multiple polymerase chain reactions (PCR) for multiplex amplification of the 11 microsatellite loci and sex marker (SRY). Forward primers for each locus were tagged with an M13 tag enabling each to be labelled with fluorescent dyes (Appendix 13 Schuelke 2000). Separate PCR reactions were performed for some loci to maximise DNA amplification (Appendix 13). Each 2 µL PCR reaction contained approximately 15 - 20 ng of dried template DNA, 1 µL of Qiagen Type-it microsatellite PCR mix, 2 pM of each end M13-labelled locus specific forward primer, 8 pM of each locus specific reverse-primer and 2 pM of the fluorescent dye specific to the reaction (FAM, VIC, PET, NED; Appendix 13). Thermal cycling consisted of an initial denaturation step at 95°C for 15 minutes, followed by eight cycles of 94°C for 30 seconds, touchdown 60°C for 90 seconds, 72°C for 60 seconds, then 25 cycles of 94°C for 30 seconds, 52°C for 90 seconds, 72°C for 60 seconds and a final extension step of 60°C for 30 minutes. After

PCR amplification, I added 25 µL of MilliQ H2O to the PCR product. I combined equal proportions of the diluted PCR reactions that were labelled with the same fluorescent dye (Appendix 13) before adding 0.2 µL GeneScanTMLIZ®500 size marker and 7.7 µL

Chapter Three: Feral cat population genetics

Hi-Di Formamide. Genotyping was performed on an Applied Biosystems 3730XL DNA Analyzer (Applied Biosystems Inc). I scored alleles visually using Geneious 6 (http://www.geneious.com, Kearse et al. 2012). I then used an automated binning program, Flexibin (Amos et al. 2007), to bin the scored peak sizes into allele sizes. This binning procedure uses an algorithm to bin alleles and reduce allele scoring errors (Amos et al. 2007). To check for genotyping errors, I re-ran 10% of the samples, chosen randomly, for each locus twice. To calculate the error rate per allele, I divided the number of mismatched alleles following binning by the total number of rerun (double-genotyped) alleles (Hoffman & Amos 2005).

3.2.4 Data analysis

I pooled samples collected over 2014 and 2015 from the Tasman Valley (2014: n = 27, 2015: n = 33) and Ohau River (2014: n = 31, 2015: n = 39) sites to increase sample size. Prior to data analysis, I removed locus F85 as MICROCHECKER (van Oosterhout et al. 2004) indicated the presence of null alleles as has been found previously (Koch et al. 2014). Across the 10 loci, the mean allele scoring error rate was low, with a mismatch score of 0.026 (± 0.069) (SD) per allele for 19 repeat genotyped samples. Of the 189 feral cat samples collected, I removed those that failed to amplify at three or more loci (n = 32) prior to further data analysis resulting in a total sample size of 157 individuals. Of these, 102 individuals were successfully genotyped at all 10 loci. I used GenAlEx 6.5 (Peakall & Smouse 2012) for data exploration, which included examining genotypic variation, heterozygosity and allelic richness at the 10 loci. Genotyping scores can be found in Appendix 14. I used GENEPOP 4.0.10 (Rousset 2008) to assess evidence of deviation from Hardy-Weinberg Equilibrium (HWE) and to test for the performance of loci as population genetic markers. I set the dememorization number, number of batches, and the number of iterations per batch all to 1,000. To correct for multiple tests, I applied a Bonferroni correction (P = 0.05/number of tests) to adjust the significance level and reduce Type 1 errors across multiple tests among loci (Rice 1989). I used FSTAT 2.9.3

(Goudet 2001) to estimate the FIS (inbreeding coefficient) and allelic richness (a measure of allelic diversity corrected for sample size). I used a one-way analysis of variance (ANOVA) and Tukey’s tests to assess differences in allelic richness among populations

Chapter Three: Feral cat population genetics using R 3.0.2 (R Core Team 2014). I estimated genetic relatedness of individuals using GenAlEx 6.5 (Peakall & Smouse 2012) by estimating the average pairwise relatedness among populations, with the permutation and bootstrap numbers both set to 999 (Queller & Goodnight 1989). Relatedness (r) values range between -1 and 1 such that approximate values of r indicate 0.5 for parent-offspring or full sibling relationships, 0.25 for half sibling and zero for unrelated relationships (Queller & Goodnight 1989). I used GenAlEx 6.5 to perform the following analyses within this paragraph unless otherwise stated. I performed a Mantel test (Mantel 1967, Peakall & Smouse 2012) to test for isolation by distance (IBD) between each site using Nei’s pairwise genetic relatedness and ln Euclidean geographic distance between each site to assess if dispersal was spatially limited. I estimated geographic distances between each site using ArcGIS 10.1 (ESRI 2014). To provide a visualisation of any genetic structure, I performed a Principal Coordinates Analysis (PCoA) (Peakall & Smouse 2012). I then used a multivariate framework to test for spatial autocorrelation to reduce stochasticity and increase the signal strength (compared to locus-by-locus or allele-by-allele analyses, Smouse & Peakall 1999). The autocorrelation coefficient, (r), computed via autocorrelation analyses provided a measure of the genetic similarity or distinctness between geographically separated pairs of individuals within specified distance classes. I used the point at which the correlogram crossed the x-axis to infer an estimate of the patch size, characterising the point of ‘non-random’ mating or restricted gene flow (Smouse & Peakall 1999). I used 20 distance classes each bounded at 10 km intervals. Positive correlations (above zero) are typically found for shorter distances and negative correlations for larger distance classes (below zero) to infer spatial genetic patterns. I set the permutation and bootstrap levels to 9,999. I then used ARLEQUIN 3.5.1.3 to calculate Slatkin’s lineralised FST to analyse the pairwise genetic differentiation between populations (Excoffier & Lischer 2010). I tested the significance of the FST values using 30,000 permutations. To infer estimates of migration, I implemented a Bayesian approach using BayesAss 3.0 (Wilson & Rannala 2003) to estimate recent migration rates of feral cats. I used a Bayesian Markov Chain Monte Carlo (MCMC) length of 3,000,000 iterations, a burn-in period of 200,000 iterations and various delta values for migration rates (m), allele frequencies (P) and inbreeding values (F) to achieve acceptable acceptance rates between 20% and 60%. I checked convergence using Tracer 1.5 (Rambaut & Drummond 2009). 96

Chapter Three: Feral cat population genetics

To estimate the number of population clusters (K) present in the samples of feral cats, I used two different Bayesian clustering methodologies. I used STRUCTURE 2.3.1 (Pritchard 2010) to implement a Bayesian MCMC approach to cluster individuals into the most appropriate population using the genetic information collected across multiple loci. I tested K for between two and four clusters, where a score of four indicates a distinct population between each site. I used an admixture (mixed ancestry model) and uncorrelated allele frequencies between populations to reduce algorithm instability and overestimation of K (Pritchard 2010). I set a LOCIPRIOR to use the sampling sites as prior information. I visually checked that the r (recombination rate) and alpha (the degree of admixture) parameters were below one to ensure convergence and that any structuring detected when using LOCPRIOR was not inaccurate. I ran the analysis with a burn-in of 100,000 iterations followed by 300,000 repetitions of MCMC chains for ten independent iterations. I ran the 10 loci dataset of population clusters to estimate K for all individuals together. The most likely value of K was selected using the “Evanno method” to calculate the mean-likelihood value of K based on the ad-hoc ΔK statistic using the web-based software, STRUCTURE HARVESTER (Evanno et al. 2005, Earl & vonHoldt 2012). Individuals were assigned a population cluster based on a threshold of q ≥ 0.8, which characterises individuals with high ancestry (Lecis et al. 2006, Bergl & Vigilant 2007, Nsubuga et al. 2010). I also used TESS 2.3.1 (Durand et al. 2009) to implement a spatially-explicit Bayesian clustering algorithm to determine population genetic structure (Chen et al. 2007). Unlike STRUCTURE, TESS includes the exact geographical co-ordinates for each individual as informed priors to create a spatially explicit model (Durand et al. 2009, Francois & Durand 2010). Therefore, TESS can help to distinguish between clinal versus clustering structure in continuous populations, which could be the case for feral cats from the Tasman Valley and Ohau River and Ahuriri Valley and Ohau River sites. I ran the TESS algorithm incorporating the spatial co-ordinates of individuals using the conditional autoregressive (CAR) Gaussian model of admixture with an interaction parameter of 0.6 (Chen et al. 2007). I ran the model for 50,000 iterations following a burn-in period of 10,000 iterations for Kmax = 2 through to Kmax = 4, with 100 replicates for each Kmax. I exported the 20% lowest likelihood runs for each Kmax to CLUMPP 1.1.2 to average individual membership coefficients to correct for discrepancies between runs (i.e. label-switching of population clusters; Jakobsson & Rosenberg 2007)). Individuals were assigned to a population based on the proportion of 97

Chapter Three: Feral cat population genetics

membership (q ≥ 0.8). Kmax was selected when the Q-matrix of posterior probabilities for individuals revealed no additional clusters and a plot of the deviance information criterion (DIC) against K was associated with stabilization of the DIC curve (i.e. when the curve reached plateau (Durand et al. 2009, Francois & Durand 2010). I then used DISTRUCT 1.1 (Rosenberg 2004) to visualise results from both STRUCTURE and TESS. I also conducted assignment tests to obtain further information pertaining to the distinctiveness of each population. I used GENECLASS 2 to self-classify individuals to their original site using partially-Bayesian methods (Rannala & Mountain 1997, Paetkau et al. 2004, Piry et al. 2004). I used the “leave one out option” and 10,000 simulations of individual genotypes (Cornuet et al. 1999).

Chapter Three: Feral cat population genetics

3.3 Results

3.3.1 Population structure of feral cats

Across all of the 10 polymorphic loci for all sites, a total of 93 alleles were scored, with an average of 7.6 alleles per locus, ranging from four to 13 alleles. The feral cat samples from the Ohau River site had the highest mean number of alleles (8.5), with the lowest at the Tasman Valley and Macraes Flat sites (both 7.1) (Table 3.2). Allelic richness was also lowest for the Tasman Valley site and highest for the Ohau River site; however, this difference was not statistically significant (Tukey HSD: P > 0.05). Expected (HE) and observed (HO) heterozygosity values were similar among sites (Table 3.2).

Inbreeding coefficient (FIS) values for each population were all low (Table 3.2). When testing for deviations from Hardy-Weinberg proportions (HWE) for the 10 loci, seven of the 40 tests showed significant values (P < 0.05), but this dropped to three after a standard Bonferroni correction.

Table 3.2: Population genetic information for feral cats caught at four different sites in the South Island of NZ and the genetic diversity that amplified at 10 polymorphic microsatellite loci (overall n = 157).

Population n A AC HO HE FIS

Tasman 7.1 6.4 0.68 0.74 45 0.081 Valley (2.8) (2.5) (0.18) (0.11)

Ohau 8.5 7.4 0.73 0.79 64 0.076 River (3.0) (2.3) (0.16) (0.08)

Ahuriri 7.5 7.0 0.70 0.74 27 0.054 Valley (2.7) (2.5) (0.18) (0.11)

Macraes 7.1 7.0 0.71 0.79 21 0.101 Flat (2.4) (2.3) (0.22) (0.12)

Overall 157 7.6 (0.7) 7.0 (0.4) 0.71(0.02) 0.77 (0.03)

n = sample size, A = mean number of alleles scored per locus (SD), AC = mean allelic richness per locus (SD), HO = observed heterozygosity, HE = expected heterozygosity, FIS = inbreeding coefficient.

Chapter Three: Feral cat population genetics

Average pairwise relatedness between the four sites showed that feral cats from the Ohau River site were the most unrelated, followed by Macraes Flat. The feral cats from the Tasman Valley and Ahuriri Valley sites were more related (Figure 3.3). Multiple captures from the same trap were removed to test for catching relatives, but no change in relatedness was observed (data not shown). When relatedness was assessed separately for each sex, relatedness was similar for both the Ohau River and Macraes Flat sites (i.e. males and females exhibited similar relatedness coefficients). In contrast, the Ahuriri Valley females were more related compared to male relatedness. At the Tasman Valley site, males were slightly more related to other males than females were to other females (Figure 3.4).

0.100 0.080 0.060 0.040 0.020 0.000 -0.020 -0.040 Tasman Valley Ohau River Ahuriri Valley Macraes Flat

Sampling Site Figure 3.3: Mean pairwise genetic relatedness for feral cats from four sites (Tasman Valley: n = 45, Ohau River: n = 64, Ahuriri Valley: n = 27, Macraes Flat: n = 21) in the South Island of NZ. Error bars are the 95% confidence intervals.

100

Chapter Three: Feral cat population genetics

Figure 3.4: Mean pairwise relatedness estimates for female (Total n = 95, Tasman Valley = 27, Ohau River = 34, Ahuriri Valley = 20, Macraes Flat = 14) and male (Total n = 62, Tasman Valley = 18, Ohau River = 30, Ahuriri Valley = 7, Macraes Flat = 7) feral cats from four sampling sites in the South Island, NZ. Values are based on molecular sex. Error bars are the 95% confidence interval.

The principal co-ordinates analysis (PCoA) did not show any particular clustering among the sites (Figure 3.5). In total, the first three principal axes explained 47.17% of the genetic diversity across the four populations (PC1: 8.72%, PC2: 16.10%, PC3: 22.35%).

Figure 3.5: Principal coordinates analysis (PCoA) showing genetic relationship of 157 feral cats collected from four different sites in the Upper Waitaki Basin (Tasman = Tasman Valley, Ohau = Ohau River, Ahuriri = Ahuriri Valley, Macraes = Macraes Flat). 101

Chapter Three: Feral cat population genetics

The Mantel test showed a non-significant trend towards isolation by distance 2 (IBD) (Figure 3.6, R = 0.52, P = 0.078). Based on pairwise FST values (Table 3.3), all sites showed low levels of significant genetic differentiation. The Ahuriri and

Tasman Valley sites were the most differentiated (pairwise FST = 0.051), followed closely by the Tasman Valley and Macraes Flat sites (pairwise FST = 0.049). The

Ohau River and Ahuriri Valley sites showed the least differentiation (pairwise FST = 0.010). Spatial autocorrelation did not show any overall significant spatial structure (P > 0.05); however, was significantly positively correlated at distances of 10 – 20 km and significantly negatively correlated for distances from 60 – 80 km (Figure 3.7). The correlogram crosses the x-axis at ~ 40 – 50 km, hence individuals separated by distances < 50 km are more genetically similar than those separated by > 50 km (Figure 3.7). This indicates non-random mating of individuals located spatially close together. There is an indication of non-significant genetic similarity at distance classes ~ 130 km (i.e. potentially between Ohau River and Macraes Flat).

0.250

0.200

0.150

0.100 R² = 0.5231 0.050

0.000 3.000 3.500 4.000 4.500 5.000 5.500

Ln (1 + geographic distance)

Figure 3.6: The relationship between Nei’s genetic distance and the natural log of geographic distance between pairs of feral cat populations in the South Island of NZ. N = 157.

102

Chapter Three: Feral cat population genetics

0.200

0.150

0.100

0.050

r 0.000 r -0.050 U/L -0.100

-0.150

-0.200

90 10 20 30 40 50 60 70 80

100 110 120 130 140 150 160 170 180 190 200

Distance Class (km)

Figure 3.7: Correlogram showing the spatial correlation (r) as a function of 10 km distance class sizes for 157 feral cats (Felis catus) in the South Island of NZ. Dashed red lines indicate the 95% CI about the null hypothesis of a random distribution of cats, and the error bars represent the 95% confidence intervals about the mean r calculated by bootstrapping.

Table 3.3: Pairwise Slatkin’s lineralised FST values among four sites of feral cats in the South Island of New Zealand. All values were significant (P < 0.01).

Macraes Site Tasman Valley Ohau River Ahuriri Valley Flat Tasman - Valley Ohau River 0.0220 - Ahuriri Valley 0.0505 0.0101 - Macraes Flat 0.0493 0.0181 0.0262 -

Based on the mean posterior probability for migration rate (m), the majority of individuals originated from the site they were sampled in (range m = 0.69 - 0.94, Table 3.4). There were low levels of migration detected from the Ohau River into the Ahuriri Valley site (posterior probability: 0.27), and from the Ohau River into the Tasman Valley site (posterior probability: 0.20). Both of these migration rates were asymmetrical, with

103

Chapter Three: Feral cat population genetics much smaller migration rates occurring into the Ohau River site (from Ahuriri Valley, m = 0.02, from Tasman Valley, m = 0.01, Table 3.4).

Table 3.4: Mean estimated migration rates (proportion of individuals) among four feral cat sites sampled in the South Island of NZ, calculated using BayesAss 3.0. Migrated into

Migrated from Tasman Ohau Ahuriri Macraes Valley River Valley Flat 0.7851 0.0135 0.0150 0.0157 Tasman Valley (0.0506) (0.0133) (0.0145) (0.0151)

0.1959 0.9360 0.2749 0.1594 Ohau River (0.0514) (0.0283) (0.0321) (0.0428) 0.0088 0.0234 0.6936 0.0217 Ahuriri Valley (0.0088) (0.0219) (0.0249) (0.0208) 0.0102 0.0271 0.0165 0.8031 Macraes Flat (0.0097) (0.0168) (0.0153) (0.0429)

Values are means of the posterior distribution of the migration rate per generation into each population with the standard deviation in brackets. Migration rates were estimated as the proportion of individuals in the column population that originated from the row population. Diagonal values are the proportion of individuals originating from the same population.

Based on ad-hoc statistics following the Bayesian clustering analyses performed in STRUCTURE (Pritchard 2010) and TESS (Durand et al. 2009), both indicated the presence of three clusters, although the assignment (Q) values differ between the two programs (Figure 3.8). The first cluster largely included Tasman Valley individuals, the second cluster largely included the Ohau River and Ahuriri Valley individuals and the third cluster comprises mostly Macraes Flat individuals. However, only 35% of the 157 feral cat individuals were allocated to one of the three clusters based on a threshold of q ≥ 0.80 in STRUCTURE (Table 3.5). None of the Macraes Flat individuals were assigned solely to one cluster. Interestingly, in STRUCTURE, where Kmax = 2, the Tasman Valley individuals were assigned to the first cluster, whereas in TESS, the Macraes Flat individuals were preferentially assigned to the first cluster compared to the other three sites (Figure 3.8).

104

Chapter Three: Feral cat population genetics

(a)

(b)

Figure 3.8: Posterior individual admixture coefficient (Q-matrices) estimates adjusted for label switching for four sites of feral cats from the South Island of NZ. Values were obtained from (a) STRUCTURE and (b) TESS for Kmax = 2 - 4 (n = 157). Each colour represents a different cluster with each column representing an individual. The vertical black bars separate each of the four sampling sites (Tasman Valley, Ohau River, Ahuriri Valley and Macraes Flat).

105

Chapter Three: Feral cat population genetics

Table 3.5: Number of individual feral cats with a cluster membership of q ≥ 0.8 from each of the four populations assigned to three clusters based on Bayesian clustering performed in STRUCTURE. Individuals with q ≤ 0.8 were considered unassigned. Bold values are the most likely cluster. Inferred Cluster Population n 1 2 3 Assigned Unassigned Tasman Valley 45 0.617 0.379 0.004 17 28 Ohau River 64 0.182 0.752 0.066 12 52 Ahuriri Valley 27 0.105 0.881 0.014 26 1 Macraes Flat 21 0.049 0.494 0.457 0 21

The assignment tests performed in GENECLASS 2 (Piry et al. 2004) correctly assigned 60.5% of individuals to their sampling location. Of the incorrectly assigned individuals, eight of them were not assigned to any of the sampled populations (c.f. Table 3.5), indicating these individuals might be migrants from a surrounding un- sampled population.

3.3.2 Demography of feral cats

Where visual data of sex (i.e. sex identification based on external morphological features) were available, there was a difference between the number of feral cats classed as either male or female when compared to the molecular data (presence of the genotyped male specific SRY gene, Table 3.6). In general, more females were sampled than males in all populations using both sexing techniques, although 60 of the 189 feral cats were unidentified when using external morphology (Table 3.6). This indicates a female-bias in the sampled populations if sampling was unbiased.

106

Chapter Three: Feral cat population genetics

Table 3.6: Sex of feral cat samples based on two different sexing-techniques (external morphology or presence of the SRY gene) from four sites within the South Island of NZ. Complete n = 189, samples used in analyses n = 157. External Morphology Molecular Evidence Site Male Female Unidentified Male Female Complete Tasman Valley 24 8 28 28 32 Ohau River 9 55 6 31 39 Ahuriri Valley 14 16 3 10 23 Macraes Flat 2 1 23 11 15 Total 49 80 60 80 109 Analysed Tasman Valley 15 6 24 18 27 Ohau River 8 50 6 30 34 Ahuriri Valley 10 15 2 7 20 Macraes Flat 4 2 15 7 14 Total 37 73 47 62 95

107

Chapter Three: Feral cat population genetics

3.4 Discussion

In this chapter, I examined the genetic population structure of feral cats from four different sites in the South Island of New Zealand. Genetic differentiation, clustering analyses, and migration tests suggest low levels of gene flow occurring between the sites. Movement is likely facilitated by human modifications to the environment (e.g. tracks and roads), and hastened by rough terrain (e.g. mountain ranges between valleys). However, contradictory results derived from different analyses made interpretation of the data equivocal.

3.4.1 Genetic diversity

Genetic diversity of the feral cats presented here (A= 7.1 – 8.5, HO = 0.71) was similar to that of feral cats on Dirk Hartog Island and mainland Australia (A = 6, HO = 0.7; Koch et al. 2014) and Hawai’i (A = 7.57 - 9.00, HO = 0.70; Hansen et al. 2007), but greater than reported genetic diversity of feral cats on Grande-Terre, a French sub-Antarctic island (A = 3.67 - 7.00, HO = 0.53; Pontier et al. 2005) and mainland France (A = 4.38

- 7.78, HO = 0.61, Say et al. 2003). Low genetic diversity on Grande Terre, which is expected on islands (Frankham 1997), likely arose due to a small founder population (Pontier et al. 2005). However, the large stray cat populations of urban France, which are characterised by low genetic diversity and high genetic differentiation are likely caused by busy high capacity roads creating a barrier to dispersal, preventing gene flow and homogenisation of allele frequencies (Say et al. 2003). Heavy traffic roads are often associated with high mortality rates as a result of collisions between vehicles and dispersing individuals (Holderegger & Di Giulio 2010). A reduction in dispersing individuals and gene exchange leads to functional isolation between colonies (Say et al. 2003, Holderegger & Di Giulio 2010). Additionally, due to the greater genetic diversity observed in my study, compared to those reported previously on islands, it is unlikely that the sampled feral cat populations were established with few founders, providing evidence for recent connectivity between sampled sites.

The low FIS (0.05 – 0.10) values for each population provided little evidence for inbreeding at the Macraes Flat and Tasman Valley sites, with no evidence at the Ohau River and Ahuriri Valley sites. Levels of inbreeding documented in feral cats in other localities [(Grande-Terre FIS = 0.11 – 0.28, Pontier et al. 2005), mainland Europe (FIS = 108

Chapter Three: Feral cat population genetics

0.102, Pierpaoli et al. 2003; FIS = 0.019 – 0.190, Say et al. 2003), Hawai’i (f > 0.09, Hansen et al. 2007)] were generally higher or equal to those detected in this study and on Dirk Hartog Island and mainland Australia (FIS = 0.007 - 0.07, Koch et al. 2014). If inbreeding at the sampled sites (especially Macraes Flat and Tasman Valley) increase, populations could be subject to reduced reproduction, survival, and disease resistance (Frankham 2003), which might aid the long-term goal of suppressing feral cat populations. Long-term genetic monitoring might be able to identify if inbreeding and genetic differentiation increases due to trapping operations each year (Ehrich et al. 2009). When separated by sex, males at the Ahuriri Valley site were less related, indicating male-biased dispersal typical of mammalian species (e.g. in brush-tail possums, Trichosurus vulpecula, Adams et al. 2014a, and long-distance dispersal of feral cats on Hawai'i, Hansen et al. 2007). Interestingly, Tasman Valley males appeared to be more related to each other than females were to each other, suggesting female dispersal. Female-biased dispersal might arise due to inbreeding avoidance, local mate competition or resource competition between matriarchies, with sub-ordinate females dispersing (Devillard et al. 2003). While female dispersal is not uncommon in the Felidae (Spong and Creel 2001; Devillard et al. 2003), discrepancies between genetically-derived and morphologically-derived sex-identifications indicate sex- specific relatedness values reported for all sites must be interpreted with caution (discussed further below in section 3.4.3: Limitations of study design).

3.4.2 Population connectivity

There was evidence for gene flow, and hence connectivity between populations (Hansen et al. 2007, Stringham et al. 2012). The geographically closest sites (Ohau River and Ahuriri Valley) were also the closest genetically. The possibility of population connectivity occurring between these sites is supported by large home range and dispersal distances recorded for feral cats (Gillies et al. 2007, Hansen et al. 2007, Harper

2007, Robley et al. 2008, Recio et al. 2010, Cruz et al. 2014). The higher pairwise FST value between the Tasman Valley and Ahuriri Valley sites suggests a degree of genetic divergence, possibly due to geographic barriers (e.g. adjacent mountain ranges and valleys separating the populations) limiting dispersal between these sites. This is

109

Chapter Three: Feral cat population genetics supported further by high energetic costs associated with travelling across rough terrain preventing movement across valley systems (Recio et al. 2015). The high assignment rate of the Ahuriri Valley cats to a cluster might indicate movement across valleys (i.e. between steep, rough terrain) is limited. Assignment of individuals in the Ahuriri Valley and Ohau River sites largely to the same cluster is consistent with movement between them and the absence of geographic barriers that prevent dispersal. The level of unassigned individuals to the Macraes Flat and Ohau River sites and low relatedness between individuals suggest Macraes Flat and Ohau River might both be man-made sinks due to ongoing trapping operations, supporting high levels of feral cat immigration from the surrounding areas (King et al. 2000). Paradoxically, the Ohau River site appears to exhibit both source and sink characteristics as asymmetrical migration rates suggests low level dispersal out of Ohau River to other sampled areas indicative of a self-sustaining source population (Runge et al. 2006). High rabbit abundance at the Ohau River site (i.e. resource-rich habitat) might contribute to greater productivity and feral cat abundance, reproduction, and therefore, dispersal (Hansen et al. 2007, Ehrich et al. 2009, Cruz et al. 2013). However, all of the sites sampled support high rabbit abundance (pers. obs., Shaun Aitcheson, pers. comm.). At the same time, the Ohau River site exhibits low relatedness, indicative of a sink population that should be unable to support high population size required for producing migrants (Thomas & Kunin 1999). The Ohau River might in reality exhibit low productivity with competitive exclusion and movement of subordinate feral cats into empty territories created by predator control efforts, contributing towards immigration from a nearby, un-sampled feral cat population (Norbury et al. 1994, Spencer et al. 2015). Indirect supplementation from nearby stray and companion cat populations in Twizel could potentially account for this immigration. A similar scenario might occur for Macraes Flat, where relatedness was also low. It is possible that immigrant feral cats from nearby surrounding un- sampled sites are dispersing into the Macraes Flat area as a result of predator control efforts creating empty territories. I observed apparent immigration of feral cats from Ohau River into Macraes Flat, which is unlikely given the large distance (~ 135 km, beyond feral cat dispersal limits) and feral cat patch size (~ 50 km) between the two sites. While unexpected for feral cats, this observation might be evidence for leakage of unneutered companion cats from nearby human-occupied areas into the feral population. Several lines of evidence support the idea of ongoing recruitment of unneutered companion, abandoned companion, and stray cat populations into nearby feral 110

Chapter Three: Feral cat population genetics populations. The PCoA analysis showed there was considerable overlap of populations which is unexpected given the existence of an isolation by distance (IBD) pattern. In contrast to feral cats on Hawai’i Island, where distances of > 60 km do not prevent dispersal, the spatial autocorrelation analysis here confirms the isolation by distance pattern detected by the Mantel test and indicates a genetic patch size of ~ 40 - 50 km. Given the large movements of feral cats recorded from the Ohau River (maximum home range recorded: 6753 ha, Cruz et al. 2014), compared to the Tasman Valley (maximum home range recorded: 1606.8 ha, Recio et al. 2010), larger dispersal distances of Ohau River feral cats into the Ahuriri Valley and Tasman Valley are possible and a patch size of ~ 50 km is not unexpected. Ruiz-Garcia (1994, 1999) similarly describe an absence of spatial genetic structuring over relatively short distances between urban cat colonies in Europe. Ruiz-Garcia (1994, 1999) highlight the potential for high levels of gene flow between close-by colonies with populations close to panmixia and displaying homogenized allele frequencies. The low and non-significant, but positive, correlation between genetic similarity and distances > 120 km, might further indicate supplementation of feral cat colonies from surrounding stray and companion cat populations (e.g. Ohau and Macraes). It is possible, however, that selection is acting to homogenize allele frequency (Ruiz-Garcia 1999) between localities separated by large distances (i.e. > 100 km). It must be noted that spurious results for the higher distance class sizes in my study might have arisen from the low numbers of individuals in these distance classes (Smouse & Peakall 1999). Any distinctive companion or stray cat pelage markings from any recruitment might be lost over only a few generations in the wild, as the dominant tabby coat pattern masks all others (Pollard 1999, Sunquist & Sunquist 2002). Such tabby patterned coats (e.g. spots and stripes) are found throughout the mammalian class providing benefit through camouflage, predator evasion and social communication in the feral environment (Eizirik et al. 2010).

3.4.3 Limitations of study design

I found a discrepancy between the sex of feral cats using external morphological features and molecular techniques, indicating feral cat sex-identification might be difficult in the field, especially if the dead animals being identified are decomposing (pers. obs., Shaun Aitechson, pers. comm.). Captures of feral or stray cats are often male-biased (e.g. Brothers et al. 1985; Denny et al. 2002, Hansen et al. 2007, Recio & Seddon 2013, Cruz 111

Chapter Three: Feral cat population genetics et al. 2014), but can be female-biased (Wallace & Levy 2006), or show no difference (Bloomer & Bester 1991, Lal 2008). Based on molecular methods, the samples were female-biased, where detection of females was a result of the absence of the SRY PCR product (Butler et al. 2002). Such presence/absence tests can produce spurious results, as the absence of the SRY fragment is indistinguishable from amplification failure (due to technical difficulties) or biological variation, and should be treated with caution (Devlin et al. 2005, Robertson & Gemmell 2006). Interestingly, many more females were identified in the Ohau River site based on external morphological characteristics than were determined by molecular means. Misidentification of sex could result in a misrepresentative view of the sex ratio in the overall population. Male feral cats might not mature until approximately one year old, at which time testes growth-rate increases (H. Beattie, pers. comm., Jones & Coman 1982). This might give young males the external appearance of being female (i.e. absence of testes). Feral cat internal sexual features and a second independent sex test might be required to accurately identify the performance of the SRY presence/absence test for feral cats (Robertson & Gemmell 2006).

3.4.4 Management implications

This study was focused in conservation-sensitive areas in the South Island of NZ where feral cat management programmes are currently taking place to protect a number of vulnerable endemic ground nesting birds. Molecular genetic tools could be applied to other parts of New Zealand to define eradication units where feral cats pose a risk to native species (e.g. on Stewart Island, Harper 2004, Ewans 2014). Overall, there appears to be some connectivity between the sites sampled, despite the large distances between them. Unlike isolated stray cat colonies in France, where roads are likely barriers to dispersal, feral cat dispersal along valley systems, in the upper Waitaki Basin, is likely facilitated by roads that provide favourable habitat conducive to movement (Say et al. 2003, Recio et al. 2015). However, the detection of population connectivity and gene flow between long-distance geographically separate populations (e.g. between Macraes Flat and the upper Waitaki Basin separated by ~ 130 km) might have occurred due to recruitment from surrounding unneutered stray and companion cats in human-occupied areas, although this remains speculative. Due to potentially high

112

Chapter Three: Feral cat population genetics dispersal rates of feral cats and absence of current population separating into eradication units, management might benefit from a ‘whole basin’ approach, including education of and action by companion cat owners regarding feral cat management. Based on my findings, further predator control of the Ohau River site might reduce movement into adjoining areas, as is suggested based on migration rates out of the Ohau River site and feral cat dispersal ability of the cats at the Ohau River site (Cruz et al. 2014). Sustained trapping in the Tasman Valley could eventually lead to reduced population size, inbreeding and classification into an eradication unit if movements from the Ohau River (and potentially the surrounding areas) are reduced. A reduction in feral cat numbers is also likely to reduce encounter rates between the sexes, reducing pregnancy rate and fecundity (e.g. on Marion Island, Bester et al. 2002). Large fluctuations in abundance, as imposed by trapping operations, theoretically promote loss of genetic diversity, which could be detected via long-term genetic monitoring (Devillard et al. 2011). Removal of a staple prey item, i.e. rabbits, in conjunction with feral cat removal, might also contribute to feral cat declines and alleviate predation pressure on native birds (Keedwell et al. 2002b, Sanders & Maloney 2002) and reptiles (Middlemiss 1995). However, rabbit control must be carried out with caution as sudden reductions in staple prey items can cause short-term prey switching to native birds and lizards (Norbury et al. 2002). Continued genetic monitoring might be useful long-term, to assess the level of genetic diversity, connectivity and inbreeding to infer the success of management efforts (Devillard et al. 2011). However, my research suggests potential input from nearby non- feral cats might reduce the efficacy of feral cat management. Companion cat owner surveys regarding neuter rates of the nearby companion cats and a genetic assessment of nearby companion and stray cats to assess connectivity with feral cats might aid in determining companion and stray cat contribution towards feral cat populations (Bradshaw et al. 1999, Natoli et al. 2006). Further research into the role companion and stray cat populations’ play on feral cat populations is required to more accurately assess this relationship.

113

Chapter Four: Final Discussion

Chapter Four:

Final Discussion

Fiordland National Park across Lake Te Anau in the morning.

114 Chapter Four: Final Discussion

Throughout this thesis, I have assessed the home range and habitat selection of companion cats near conservation-sensitive areas in Fiordland and the population genetics of feral cats living in conservation-sensitive areas in Canterbury and Otago, New Zealand. I conducted this study in order to assess putative geographic barriers to movement at different spatial scales of two types of cats that differ in their association with humans. While the management strategies I proposed differed, the primary goal remained the same: to alleviate any detrimental effects of cat predation and presence on native wildlife and to reduce or remove their presence from conservation-sensitive areas. My research also highlights that the designation given to different types of cats can inherently influence what is considered acceptable management. Given the possibility of inter-breeding between unneutered companion and stray cats, and feral cats (as suggested by Chapter Three, Natoli et al. 2006 and Hansen et al. 2007), the long-term success of feral cat predator management programmes might be enhanced by regulations imposed on companion and stray cat populations.

4.1 Definition and perception of cats

Throughout the literature, the names and designations of cats are based upon their association with humans rather than reflecting any taxonomic distinctions, leading to inconsistencies between naming, management strategies and public awareness and action (Farnworth et al. 2011, 2014). Cats, specifically, Felis catus (rather than wild cat species Felis silvestris sp.) are described in different studies as pet, owned, unowned, house, feral house, indoor-outdoor, farm, domestic, feral domestic and free-roaming, in addition to the three terms used within this thesis (companion, stray and feral) (e.g. Middlemiss 1995, Barratt 1998, O’Hara 2007, Ramón et al. 2010, Loyd et al. 2013b, Loyd 2015). The designation of cats does have an influence on what publicly is considered acceptable management practice. A strong contrast to cat management is, for example, brush-tail possum management in New Zealand. Concerns regarding possum impacts (e.g. from an economical and conservation perspective) have prompted wide-scale control programmes with strong public support for existing and developing lethal (e.g. trapping, poisoning) and non-lethal control (e.g. sterilization) methods (King 2005, Potts 2009). Public concern for the welfare of species considered pests such as brush-tail possums, is much lower and considered less important than non-pest species (Farnworth et al. 2014). 115

Chapter Four: Final Discussion

Public attitudes towards cats (especially stray) are much more empathetic than for pests, and consequently non-lethal methods are considered more acceptable forms of control (Farnworth et al. 2011, 2014). Empathy with stray cats is observed throughout the world; for example, it is common practice that members of the public take care of stray cats in colonies by providing food, medicine and general care (Natoli et al. 1999, Devillard et al. 2003). In South Africa, survey participants at the conservancy-declared KwaZulu- Natal University Campus generally did not view introduced stray cats as an invasive species and supported the proposal for a university-funded, campus-wide feeding programme (Tennent et al. 2010). Most participants considered cat eradication unnecessary even though they understood that management is essential in an environment where removal of invasive species and regeneration of natives is a primary goal (Tennent et al. 2010). In New Zealand, a survey of the public revealed that there is a high level of support for the notion that companion cats killing wildlife is a problem in urban, peri- urban, rural and natural environments (van Heezik et al. 2014); however, 54% and 71% of survey respondents in Auckland and Dunedin, respectively, allow their cats free-reign outside. Fortunately, high sterilization rates (98% of respondents, van Heezik et al. 2014, and 100% of owners in Fiordland in this study), have a positive conservation outcome by minimising breeding and indirect population connectivity to stray and feral cats. However, only a few intact and abandoned individuals can compromise the successful operation of TNR programmes for managing stray cats (Castillo & Clarke 2003, Natoli et al. 2006).

4.2 What are acceptable management strategies?

Studies such as mine, which characterise movements to infer cat impact, are integral for maintaining and devising appropriate management strategies. These studies, and proposed management strategies, aim to reach a compromise between conservation and the human-cat relationship benefits. As well as providing companionship, cat ownership helps to enhance relaxation and relieve effects caused by high blood pressure, anxiety, depression, heart attacks and mood swings (Natoli et al. 1999, Bernstein 2007). However, cats are an invasive species and many conservation management protocols are in place to reduce the impact that invasive species have on New Zealand ecosystems (Saunders & Norton 2001, Gillies et al. 2003). Stray and feral cat management can include lethal 116

Chapter Four: Final Discussion

(e.g. poisoning, trapping, hunting) or non-lethal (e.g. trap-neuter-return, contraception, trap neuter re-home) (Farnworth et al. 2011) methods. Unlike stray and feral cats, management of companion cats is dependent upon compliance from owners. Management of stray or companion cats is considered a socially and politically sensitive issue, and management strategies employed must be considered alongside the social costs of proposed management strategies (Lohr et al. 2012). Management of companion cats may be met with opposition by the public, depending upon the proposed management strategy (Grayson et al. 2002, Lilith et al. 2006, Robertson 2008). Even though confining cats indoors at night is often considered unfavourable by owners, as it restricts the perceived important free-roaming behaviour of cats, it is more publicly acceptable than confining cats to properties or strictly indoors (van Heezik et al. 2014). Additionally, owners are often unaware of the risks (e.g. vehicle interactions, diseases, fighting) associated with cat free-roaming behaviour (Loyd et al. 2013a). Free-roaming prohibitions, exclusion zones (Lilith et al. 2006), and collar-worn predation deterrents (Ruxton et al. 2002, Calver et al. 2007, Hall et al. 2015, Willson et al. 2015) might help protect cats in the hazardous urban environment and reduce cat-wildlife interactions. Cats and are suggested to (over generations) adjust their behaviour away from prey hunting (Meek 2003). Companion cats living in Fiordland might need to be confined indoors due to wide-ranging movements and use of cover habitats, or be excluded from a ~ 3 km buffer zone around access points (Lilith et al. 2008, Metsers et al. 2010) to prevent movements into Fiordland National Park (FNP) or the Kepler Mire conservation area (KMCA). Additionally, use of predation deterrents, such as BirdsBeSafe®, CatBibTM or bells, can aid to reduce predation events on avian and herpetofauna prey (Ruxton et al. 2002, Calver et al. 2007, Gordon et al. 2010, Hall et al. 2015). These forms of non-lethal control can help to reduce cat-wildlife interactions. Feral cat management often involves trapping, hunting or poisoning to eradicate populations (Nogales et al. 2004). Trapping, the most publicly acceptable form of feral cat control is conducted in the upper Waitaki Basin, and at Macraes Flat (Farnworth et al. 2011, Cleland et al. 2013, Rebergen & Woolmore 2015). However, results from my genetic analyses suggest that these populations might benefit from tighter companion and stray cat control. Additionally, the Tasman Valley group might eventually be manageable as an eradication unit if movement from the Ohau River can be restricted. Cat owners are more opposed than non-owners to lethal control for stray or feral cats due to concerns for animal welfare (Farnworth et al. 2011). Additionally, close 117

Chapter Four: Final Discussion proximity of stray cats to urban areas and the interactions between humans and urban- living cats limits the use of harmful or emotionally upsetting types of control (Lohr et al. 2012). Even though the impacts of stray, feral or companion cats on native ecosystems might be similar and problematic (Metsers et al. 2010, van Heezik et al. 2010, van Heezik 2014, Loss et al. 2013, McDonald et al. 2015), the designation of each type of cat determines public perception of welfare and acceptance of management strategies (Farnworth et al. 2011). The ‘mesopredator release’ hypothesis suggests the removal of larger predators (apex-predators – such as cats) and therefore, predation pressure can lead to an explosion of smaller predators (mesopredators – such as rats), resulting in the rapid extinction of their shared prey (Courchamp et al. 1999, Woods et al. 2003, Fan et al. 2005). Cat management decisions must also consider controlling smaller mammals, specifically Rattus spp.. Cats might play an important role in limiting rodent population size which might have a more pronounced effect on wildlife populations (Courchamp et al. 1999, Woods et al. 2003).

4.3 Legislation

Currently, feral (NZ Conservation Act 1987, Biosecurity Act 1993) and stray (Animal Welfare Act 1999, O’Hara 2007) cat management is governed by different legislations. Farnworth et al. (2011) argue that feral and stray cats should be governed under the same legislation to prevent altered perception of what is acceptable for management of unowned invasive species. However, the different attitudes of owners and non-owners have implications for formulation of legislation (van Heezik et al. 2014), and any imposed regulations would rely on compliance from owners for effectiveness. Development and implementation of legislation pertaining to cat ownership, such as compulsory elective gonadectomy procedures, registration, micro-chipping and restrictions on the number of owned cats, aim to ensure owners become more responsible for their cat(s) behaviour (Meek 2003, Invercargill City Council 2013). However, many of these regulations are not compulsory. Compulsory companion cat registration, microchipping and sterilization might lead to reductions in the number of free-roaming, unwanted and abandoned cats to prevent impacts on wildlife (Jessup 2004, Robertson 2008). Additionally, identification of companion cats as such (i.e. via collars and microchips) might aid in stray cat management by distinguishing between unowned and 118

Chapter Four: Final Discussion owned cats during management operations. However, collaring and microchipping rates of companion cats are low (Harrod et al. 2015). Tighter regulations on ownership might enhance the success of unowned (stray and feral) cat management by restricting immigration and population growth as might be occurring in the South Island of New Zealand as potentially discovered via genetic analyses here (Natoli et al. 2006, Schmidt et al. 2009, Ramón et al. 2010). Long-term genetic monitoring of feral and stray cats, in conjunction with companion cat genetic analyses might help to elucidate cat relationships in New Zealand. Fortunately, natural geographic barriers (e.g. the Waiau River in FNP and mountain ranges in the upper Waitaki Basin) often exist that reduce cat movements; however, human-mediated dispersal (i.e. cats travelling along roads) are likely facilitators of movement in NZ (Recio et al. 2015), compromising feral cat predator control operations. Additionally, feral cat control operations and New Zealand native species might benefit from owner education into the free-roaming habits of their cats, the use of predation deterrents, sterilisation and the importance of identification measures to help to reduce cat-wildlife interactions (i.e. predation and predator presence). This study, and others like it, highlight the many different aspects of cats that need to be considered when developing and executing management decisions. The highly fluid nature of human-cat associations, and possible interactions between different types of cats, requires the exploration and application of a multidisciplinary approach to cat management in New Zealand. To aid protection of New Zealand's unique wildlife, especially in conservation-sensitive areas, cat management might require widespread cooperation from the public and conservation managers alike.

119

References

References:

Aarts G, MacKenzie M, McConnell B, Fedak M, Matthiopoulos J 2008. Estimating space-use and habitat preference from wildlife telemetry data. Ecography 31: 140– 160.

Adams AL, Dickinson KJM, Robertson BC, van Heezik Y 2013. An evaluation of the accuracy and performance of lightweight GPS collars in a suburban environment. PloS ONE 8: e68496. doi: 10.1371/journal.pone.0068496.

Adams AL, van Heezik Y, Dickinson KJM, Robertson BC 2014a. Identifying eradication units in an invasive mammalian pest species. Biological Invasions 16: 1481–1496.

Adams AL, Recio MR, Robertson BC, Dickinson KJM, van Heezik Y 2014b. Understanding home range behaviour and resource selection of invasive common brushtail possums (Trichosurus vulpecula) in urban environments. Biological Invasions 16: 1791–1804.

Aebischer N, Robertson P, Kenward R 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1315–1325.

Alderton D 1983. The cat: the most complete illustrated practical guide to cats and their world. MacDonald, London. 208 p.

Allen BL, Allen LR, Leung LKP 2014. Interactions between two naturalised invasive predators in Australia: are feral cats suppressed by dingoes? Biological Invasions 17: 761–776.

Allen R, Lee W 2006. Biological Invasions in New Zealand. Springer, New York. 465 p.

Allendorf FW, Luikart G, Aitken SN 2013. Conservation and the genetics of populations. Wiley-Blackwell, Chichester. 602 p.

Alterio N, Moller H 1997. Short communication daily activity of stoats (Mustela erminea), feral ferrets (Mustela furo) and feral house cats (Felis catus) in coastal grassland, Otago. New Zealand Journal of Ecology 21: 89–95.

Amos W, Hoffman JI, Frodsham A, Zhang L, Best S, Hill AVS 2007. Automated binning of microsatellite alleles: Problems and solutions. Molecular Ecology Notes 7: 10– 14.

Ancillotto L, Serangeli MT, Russo D 2013. Curiosity killed the bat: Domestic cats as bat predators. Mammalian Biology 78: 369–373.

Anderson D 1982. The home range: A new nonparametric estimation technique. Ecology 63: 103–112.

Arthur SM, Manly BFJ, Mcdonald LL, Garner GW, Manly BFJ 1996. Assessing habitat selection when availability changes. Ecological Society of America 77: 215–227.

120 References

Augé A, Chilvers B, Moore A, Davis L 2011. Foraging behaviour indicates marginal marine habitat for New Zealand sea lions: remnant versus recolonising populations. Marine Ecology Progress Series 432: 247–256.

Baasch DM, Tyre AJ, Millspaugh JJ, Hygnstrom SE, Vercauteren KC 2010. An evaluation of three statistical methods used to model resource selection. Ecological Modelling 221: 565–574.

Baker PJ, Bentley AJ, Ansell RJ, Harris S 2005. Impact of predation by domestic cats Felis catus in an urban area. Mammal Review 35: 302–312.

Baker PJ, Molony SE, Stone E, Cuthill IC, Harris S 2008. Cats about town: is predation by free-ranging pet cats Felis catus likely to affect urban bird populations? Ibis 150: 86–99.

Bar-Oz G, Weissbrod L, Tsahar E 2014. Cats in recent Chinese study on cat domestication are commensal, not domesticated. Proceedings of the National Academy of Sciences of the United States of America 111: E876.

Barratt DG 1997. Home range size, habitat utilisation and movement patterns of suburban and farm cats Felis catus. Ecography 20: 271–280.

Barratt DG 1998. Predation by house cats, Felis catus (L.), in Canberra, Australia. II. Factors affecting the amount of prey caught and estimates of the impact on wildlife. Wildlife Research 25: 475–487.

Barton K 2015. Package “ MuMIn .” 62 p.

Bateman PW, Fleming PA 2012. Big city life: carnivores in urban environments. In: Le Comber S ed. Journal of Zoology 287: 1–23.

Beckerman AP, Boots M, Gaston KJ 2007. Urban bird declines and the fear of cats. Animal Conservation 10: 320–325.

Bergl RA, Vigilant L 2007. Genetic analysis reveals population structure and recent migration within the highly fragmented range of the Cross River gorilla (Gorilla gorilla diehli). Molecular Ecology 16: 501–516.

Bergstrom DM, Lucieer A, Kiefer K, Wasley J, Belbin L, Pedersen TK, Chown SL 2009. Indirect effects of invasive species removal devastate World Heritage Island. Journal of Applied Ecology 46: 73–81.

Bernstein PL 2007. The human-cat relationship. In:The welfare of cats Rochlitz I ed. Springer. Pp. 47–89.

Bester MN, Bloomer JP, van Aarde RJ, Erasmus BH, van Rensburg PJJ, Skinner JD, Howell PG, Naude TW 2002. A review of the successful eradication of feral cats from sub-Antarctic Marion Island, Southern Indian Ocean. South African Journal of Wildlife Research 32: 65–73.

121

References

Beyer HL, Haydon DT, Morales JM, Frair JL, Hebblewhite M, Mitchell M, Matthiopoulos J 2010. The interpretation of habitat preference metrics under use- availability designs. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences 365: 2245–2254.

Biro Z, Szemethy L, Heltai M 2004. Home range sizes of wildcats (Felis silvestris) and feral domestic cats (Felis silvestris catus) in a hilly region of Hungary. Mammalian Biology 69: 302–310.

Blackburn TM, Cassey P, Duncan RP, Evans KL, Gaston KJ 2004. Avian extinction and mammalian introductions on oceanic islands. Science 305: 1955–1958.

Block M 2010. “Kitty Cam” reveals the secret life of roaming cats. http://www.npr.org/2012/08/10/158589981/kitty-cam-reveals-the-secret-life-of- roaming-cats. Accessed: 10/10/14.

Bloomer JP, Bester MN 1990. Diet of a declining feral cat Felis catus population on Marion Island. South African Journal of Wildlife Research 20: 1–4.

Bloomer JP, Bester MN 1991. Effects of hunting on population characteristics of feral cats on Marion Island. South African Journal of Wildlife Research 21: 97 – 102.

Boffa Miskell Limited 2006. Te-Anau Basin growth planning landscape capacity study. Wellington. 100 p.

Bogich TL, Liebhold AM, Shea K 2008. To sample or eradicate? A cost minimization model for monitoring and managing an invasive species. Journal of Applied Ecology 45: 1134–1142.

Bonnaud E, Bourgeois K, Vidal E, Kayser Y, Tranchant Y, Legrand J 2007. Feeding ecology of a feral cat population on a small Mediterranean Island. Journal of Mammalogy 88: 1074–1081.

Bonnaud E, Medina FM, Vidal E, Nogales M, Tershy B, Zavaleta E, Donlan CJ, Keitt B, Corre M, Horwath SV 2011. The diet of feral cats on islands: a review and a call for more studies. Biological Invasions 13: 581–603.

Bonnington C, Gaston KJ, Evans KL 2013. Fearing the feline: domestic cats reduce avian fecundity through trait-mediated indirect effects that increase nest predation by other species. Journal of Applied Ecology 50: 15–24.

Boutin S 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems, and the future. Canadian Journal of Zoology 68: 203–220.

Boyce MS, Vernier PR, Nielsen SE, Schmiegelow FK 2002. Evaluating resource selection functions. Ecological Modelling 157: 281–300.

Bradshaw JWS, Casey RA, Brown SL 1992. The behaviour of the domestic cat. CAB International, Wallingford. 219 p.

122

References

Bradshaw JW, Horsfield G, Allen J, Robinson I 1999. Feral cats: their role in the population dynamics of Felis catus. Applied Animal Behaviour Science 65: 273– 283.

Brimecombe I, Willans J, Wilson I 2014. Cats in the Kepler. Fiordland College, Te Anau. 2 p.

Burgman MA, Fox JC 2003. Bias in species range estimates from minimum convex polygons: implications for conservation and options for improved planning. Animal Conservation 6: 19–28.

Burnham KP, Anderson DR 2010. Model selection and multimodel interence: A practical information-theoretic approach. Springer, Verlag New York. 351 p.

Burt WH 1943. Territoriality and home range concepts as applied to mammals. Journal of Mammalogy 24: 346–352.

Buskirk SW, Millspaugh JJ 2006. Metrics for studies of sesource selection. Journal of Wildlife Management 70: 358–366.

Butler BJM, David VA, O’Brien SJ 2002. The MeowPlex: A new DNA test using tetranucleotide STR markers for the domestic cat. Profiles in DNA 5: 7–10.

Buttriss R 2001. No cat zone - City of Kingston. Proceedings of the 10th National Conference on Urban Animal Management in Australia. Australian Veterinary Committee: Melborne. Available at http:// www.uam.net.au/Publications.

Byrom AE 2002. Dispersal and survival of juvenile feral ferrets Mustela furo in New Zealand. Journal of Applied Ecology 39: 67–78.

Calhoon RE, Haspel C 1989. Urban cat populations compared by season, subhabitat and supplemental feeding. Journal of Animal Ecology 58: 321–328.

Calmet C, Pascal M, Samadi S 2001. Is it worth eradicating the invasive pest Rattus norvegicus from Molene archipelago?: Genetic structure as a decision-making tool. Biodiversity and Conservation 10: 911–928.

Calver M, Thomas S, Bradley S, McCutcheon H 2007. Reducing the rate of predation on wildlife by pet cats: The efficacy and practicability of collar-mounted pounce protectors. Biological Conservation 137: 341–348.

Campbell KJ, Harper G, Algar D, Hanson CC, Keitt BS, Robinson S 2011. Review of feral cat eradications on islands. In: Island invasives: erdication and management , Veitch CR, Clout MN, Towns DR ed. IUCN, Gland, Switzerland. Pp. 37–46.

Capowiez Y, Pierret A, Monestiez P, Belzunces L 2000. Evolution of burrow systems after the accidental introduction of a new earthworm species into a Swiss pre-alpine meadow. Biology and Fertility of Soils 31: 494–500.

Carter C 2013. Kids restoring the Kepler discover cats at park entrance. http://www.doc.govt.nz/news/media-releases/2013/kids-restoring-the-kepler- discover-cats-at-park-entrance/. Accessed 09/10/14.

123

References

Castillo D, Clarke AL 2003. Trap/neuter/release methods ineffectivein controlling domestic cat “colonies” on public lands. Natural Areas Journal 23: 247–253.

Cattracker 2015. Cat tracker nz. cattracker.nz. Accessed 09/10/14.

Chamberlain JS, Gibbs RA, Ranier JE, Nguyen PN, Caskey CT 1988. Deletion screening of the Duchenne muscular dystrophy locus via multiplex DNA amplification. Nucleic Acids Research 13: 3021–3030.

Chapman AD 2009. Numbers of living species in Australia and the world. Heritage 2nd: 84 p.

Chen C, Durand E, Forbes F, François O 2007. Bayesian clustering algorithms ascertaining spatial population structure: A new computer program and a comparison study. Molecular Ecology Notes 7: 747–756.

Churcher PB, Lawton JH 1987. Predation by domestic cats in an English village. Journal of Zoology 212: 439–455.

Ciarniello LM, Boyce MS, Seip DR, Heard DC 2007. Grizzly bear habitat selection is scale dependent. Ecological Applications 17: 1424–1440.

Cleland S, Aitcheson S, Currall G, Burke C, Winters J, Nelson D, Maloney R 2013. Predator control project report for kaki recovery programme Tasman Valley: March 2012 – February 2013. Twizel. 26 p.

Clout MN, Williams PA 2009. Invasive species management: A handbook of principles and techniques. Oxford University Press, Oxford. 330 p.

Clutton-Brock J 1987. A natural history of domesticated animals. Natural History Museum, London. 238 p.

Coman BJ, Brunner H 1972. Food habits of the feral house cat in Victoria. The Journal of Wildlife Management 36: 848–853.

Conner LM, Smith MD, Burger LW 2003. A comparision of distance based and classification based analyses of habitat use. Ecology 84: 526–531.

Cornuet JM, Piry S, Luikart G, Estoup A, Solignac M 1999. New methods employing multilocus genotypes to select or exclude populations as origins of individuals. Genetics 153: 1989–2000.

Coughlin CE 2014. Using technology to track animal behaviour: assessing instrumentation effects and the use of accelerometry to identify behaviours in domestic cats. Unpublished MSc Thesis. University of Otago. Otago, New Zealand.

Coughlin CE, van Heezik Y 2014. Weighed down by science: do collar-mounted devices affect domestic cat behaviour and movement? Wildlife Research 41: 606–614.

Council IC 2013. Invercargill City Council Bylaw 2013/2 – Keeping of Animals, Poultry and Bees.

124

References

Courchamp F, Chapuis J-L, Pascal M 2003. Mammal invaders on islands: impact, control and control impact. Biological Reviews 78: 347–383.

Courchamp F, Langlais M, Sugihara G 1999. Cats protecting birds: modelling the mesopredator release effect. Journal of Animal Ecology 68: 282–292.

Cromarty P, Scott DA 1995. A directory of wetlands in New Zealand. Department of Conservation, Wellington. 395 p.

Cruz J, Glen AS, Pech RP 2013. Modelling landscape-level numerical responses of predators to prey: The case of cats and rabbits. PLoS ONE 8: e73544. doi:10.1371/journal.pone.0073544.

Cruz J, Woolmore C, Latham MC, Latham ADM, Pech RP, Anderson DP 2014. Seasonal and individual variation in selection by feral cats for areas with widespread primary prey and localised alternative prey. Wildlife Research 41: 650–661.

D’Eon RG, Serrouya R, Smith G, Kochanny CO 2002. GPS radiotelemetry error and bias in mountainous terrain. Wildlife Society Bulletin 30: 430–439.

Daniel O, Kohli L, Schuler B, Zeyer J 1996. Surface cast production by the earthworm Aporrectodea nocturna in a pre-alpine meadow in Switzerland. Biology and Fertility of Soils 22: 171–178.

Daniels MJ, Beaumont MA, Johnson PJ, Balharry D, Macdonald DW, Barratt E 2001. Ecology and genetics of wild-living cats in the north-east of Scotland and the implications for the conservation of the wildcat. Journal of Applied Ecology 38: 146–161.

Dauphine N, Cooper RJ 2009. Impacts of free-ranging domestic cats (Felis catus) on birds in the united states: a review of recent research with conservation and management recommendations. Proceedings of the 4th International partners in flight Conference.

Davison J, Huck M, Delahay RJ, Roper TJ 2009. Restricted ranging behaviour in a high- density population of urban badgers. Journal of Zoology 277: 45–53.

De Barba M, Waits LP, Garton EO, Genovesi P, Randi E, Mustoni A, Groff C 2010. The power of genetic monitoring for studying demography, ecology and genetics of a reintroduced brown bear population. Molecular Ecology 19: 3938–51. de Weerd N, van Langevelde F, van Oeveren H, Nolet BA, Kölzsch A, Prins HHT, de Boer WF 2015. Deriving animal behaviour from high-frequency GPS: Tracking cows in open and forested habitat. PLoS ONE 10: e0129030. doi:10.1371/journal.pone.0129030.

Dennis TE, Chen WC, Shah SF, Walker MM, Laube P, Forer P 2010. Performance characteristics of small global-positioning-system tracking collars. Wildlife Biology in Practice 6: 14–31.

125

References

Denny E, Yakovlevich P, Eldridge MDB, Dickman C 2002. Social and genetic analysis of a popualtion of free-living cats (Felis catus L) exploiting a resource-rich habitat. Wildlife Research 29: 405–413.

Department of Conservation 2010. Fiordland National Park day walks. Te Anau. 32 p.

Department of Internal Affairs 2015. Dog Control. http://www.dia.govt.nz/Resource- material-Dog-Control-Index. Accessed 15/11/14.

Devillard S, Santin-Janin H, Say L, Pontier D 2011. Linking genetic diversity and temporal fluctuations in population abundance of the introduced feral cat (Felis silvestris catus) on the Kerguelen archipelago. Molecular Ecology 20: 5141–53.

Devillard S, Say L, Pontier D 2003. Dispersal pattern of domestic cats (Felis catus) in a promiscuous urban population: do females disperse or die? Journal of Animal Ecology 72: 203–211.

Devlin RH, Park L, Sakhrani D, Baker JD, Marshall AR, LaHood E, Kolesar SE, Mayo MR, Biagi CA, Uh M 2005. Variation of Y-chromosome DNA markers in Chinook salmon (Oncorhynchus tshawytscha) populations. Canadian Journal of Fisheries and Aquatic Sciences 62: 1386–1399.

Dickinson KJM, Chagué-Goff C, Mark AF, Cullen L 2002. Ecological processes and trophic status of two low-alpine patterned mires, south-central South Island, New Zealand. Austral Ecology 27: 369–384.

Dickman CR 1996. Overview of the impacts of feral cats on Australian native fauna. Canberra. 97 p.

Dickman CR, Newsome TM 2014. Individual hunting behaviour and prey specialisation in the house cat Felis catus: Implications for conservation and management. Applied Animal Behaviour Science. 173: 76–87.

Dowding JE, Murphy EC, Springer K, Peacock AJ, Krebs CJ 2009. Cats, rabbits, Myxoma virus, and vegetation on Macquarie Island: A comment on Bergstrom et al. (2009). Journal of Applied Ecology 46: 1129–1132.

Driscoll CA, Clutton-brock J, Kitchener AC, Brien SJO 2009a. Cat. Pp. 68–75.

Driscoll CA, Macdonald DW, O’Brien SJ 2009b. From wild animals to domestic pets, an evolutionary view of domestication. Proceedings of the National Academy of Sciences of the United States of America 106: 9971–9978.

Driscoll CA, Menotti-raymond M, Roca AL, Hupe K, Johnson E, Geffen E, Harley EH, Delibes M, Pontier D, Andrew C, Yamaguchi N, Brien SJO, Macdonald DW 2007. The Near Eastern origin of cat domestication. Science 317: 519–523.

Durand E, Chen C, Francois O 2009. Tess version 2.3 - Reference Manual August 2009

Earl DA, von Holdt BM 2012. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4: 359–361.

126

References

Edwards GP, Preu N DE, Crealy IV, Shakeshaft BJ 2002. Habitat selection by feral cats and dingoes in a semi-arid woodland environment in central Australia. Austral Ecology 27: 26–31.

Edwards A, Turner T 1999. The ultimate encyclopaedia of cats: cat breeds and cat care. Southwater, London. 256 p.

Ehrich D, Yoccoz NG, Ims RA 2009. Multi-annual density fluctuations and habitat size enhance genetic variability in two northern voles. Oikos 118: 1441–1452.

Eisert R 2011. Hypercarnivory and the brain: Protein requirements of cats reconsidered. Journal of Comparative Physiology B 181: 1–17.

Eizirik E, David VA, Buckley-Beason V, Roelke ME, Schäffer AA, Hannah SS, Narfström K, O’Brien SJ, Menotti-Raymond M 2010. Defining and mapping mammalian coat pattern genes: Multiple genomic regions implicated in domestic cat stripes and spots. Genetics 184: 267–275.

ESRI 2014. ArcGIS. Environmental Systems Resource Institute. Redlands, California.

Evanno G, Regnaut S, Goudet J 2005. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology 14: 2611–2620.

Ewans R 2014. Predator Free Rakiura Halfmoon Bay Project — methods for predator removal. Invercargill. 56 p.

Excoffier L, Lischer HEL 2010. Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: 564–567.

Fan M, Kuang Y, Feng Z 2005. Cats protecting birds revisited. Bulletin of Mathematical Biology 67: 1081–106.

Farnworth MJ, Campbell J, Adams NJ 2011. What’s in a name? Perceptions of stray and feral cat welfare and control in Aotearoa, New Zealand. Journal of Applied Animal Welfare Science 14: 59–74.

Farnworth MJ, Watson H, Adams NJ 2014. Understanding attitudes toward the control of nonnative wild and feral mammals: similarities and differences in the opinions of the general public, animal protectionists, and conservationists in New Zealand (Aotearoa). Journal of Applied Animal Welfare Science 17: 1–17.

Ferreira JP, Leitão I, Santos-Reis M, Revilla E 2011. Human-related factors regulate the spatial ecology of domestic cats in sensitive areas for conservation. PloS ONE 6: e25970. doi:10.1371/journal.pone.0025970.

Fieberg J, Matthiopoulos J, Hebblewhite M, Boyce MS, Frair JL 2010. Correlation and studies of habitat selection: problem, red herring or opportunity? Philosophical transactions of the Royal Society of London. Series B, Biological sciences 365: 2233–2244.

127

References

Fiordland Conservation Trust 2013. The Fiordland Conservation Trust. http://www.fiordlandconservationtrust.org.nz/. Accessed: 09/10/14.

Fitzgerald B 1990. House cat. In: Handbook of New Zealand Mammals, King C ed. Oxford University Press, Auckland. Pp. 330–348.

Fitzgerald AM, Karl BJ 1979. Foods of feral house cats (Felis catus L.) in forest of the Orongorongo Valley, Wellington. New Zealand Journal of Zoology 6: 107–126.

Fitzgerald BM, Turner DC 2000. Hunting behavior of domestic cats and their impact on prey populations. In: Turner DC, In: The domestic cat: The biology of its behaviour, Bateson P ed. Cambridge University Press, Cambridge. Pp. 151–175.

Fitzgerald BM, Veitch CR 1985. The cats of Herekopare Island, New Zealand; their history, ecology and affects on birdlife. New Zealand Journal of Zoology 12: 319– 330.

Fitzgerald BM, Veitch CR, Bag P, Hutt L 1991. The diet of feral cats (Felis catus) on Raoul Island, Kermadec group. New Zealand Journal of Ecology 67: 123–129.

Frair JL, Nielsen SE, Merrill EH, Lele SR, Boyce MS, Munro RHM, Stenhouse GB, Beyer HL 2004. Removing GPS collar bias in habitat selection studies. Journal of Applied Ecology 41: 201–212.

Francois O, Durand E 2010. Spatially explicit Bayesian clustering models in population genetics. Molecular Ecology Resources 10: 773–784.

Frankel O 1970. Genetic conservation in perspective. In: Genetic resources in plants – their exploration and conservation, Frankel O, Bennett E ed. Bell and Bain Ltd, Glasgow. Pp. 469–489.

Frankham R 1997. Do island populations have less genetic variation than mainland populations? Heredity 78: 311–327.

Frankham R 2003. Genetics and conservation biology. Comptes Rendus Biologies 326: 22–29.

Frankham R, Ballou JD, Briscoe DA 2002. Introduction to conservation genetics. Cambridge University Press, Cambridge. 617 p.

Gaby MJ 2014. A report on free-roaming domestic cat activities in urban areas: What do owned free-ranging domestic cats get up to? Wellington. 15 p.

Galbreath R, Brown D 2004. The tale of the lighthouse-keeper’s cat : Discovery and extinction of the Stephens Island wren (Traversia lyalli). Notornis 51: 193–200.

Gehrt SD, Wilson EC, Brown JL, Anchor C 2013. Population ecology of free-roaming cats and interference competition by coyotes in urban parks. PloS ONE 8: e75718. doi:10.1371/journal.pone.0075718.

128

References

Germain E, Benhamou S, Poulle M-L 2008. Spatio-temporal sharing between the European wildcat, the domestic cat and their hybrids. Journal of Zoology 276: 195– 203.

Getz WM, Fortmann-Roe S, Cross PC, Lyons AJ, Ryan SJ, Wilmers CC 2007. LoCoH: nonparameteric kernel methods for constructing home ranges and utilization distributions. PloS ONE 2: e207. doi:10.1371/journal.pone.0000207.

Getz WM, Wilmers CC 2004. A local nearest-neighbor convex-hull construction of home ranges and utilization distributions. Ecography 27: 489–505.

Gillies C 2001. Advances in New Zealand mammalogy 1990–2000: House cat. Journal of the Royal Society of New Zealand 31: 205–218.

Gillies C, Clout M 2003. The prey of domestic cats (Felis catus) in two suburbs of Auckland City, New Zealand. Journal of Zoology 259: 309–315.

Gillies CA, Graham PJ, Clout MN 2007. Home ranges of introduced mammalian carnivores at Trounson Kauri Park, Northland, New Zealand. New Zealand Journal of Zoology 34: 317–333.

Gillies CS, Hebblewhite M, Nielsen SE, Krawchuk MA., Aldridge CL, Frair JL, Saher DJ, Stevens CE, Jerde CL 2006. Application of random effects to the study of resource selection by animals. Journal of Animal Ecology 75: 887–898.

Gillies CA, Leach MR, Coad NB, Theobald SW, Campbell J, Herbert T, Graham PJ, Pierce RJ 2003. Six years of intensive pest mammal control at Trounson Kauri Park, a Department of Conservation “mainland island”, June 1996—July 2002. New Zealand Journal of Zoology 30: 399–420.

Ginsburg L, Delibruas G, Minaut-Gout A, Valladas H, Zivie A 1991. On the Egyptian origin of the domestic cat. Bulletin of the National Natural History Museum 13: 107–113.

Glass GE, Gardner-Santana LC, Holt RD, Chen J, Shields TM, Roy M, Schachterle S, Klein SL 2009. Trophic Garnishes: Cat–rat interactions in an urban environment. PLoS ONE 4: e5794. doi:10.1371/journal.pone.0005794.

Goldstein D, Schlotterer C 1999. Microsatellites – Evolution and applications. Oxford University Press, Oxford. 352 p.

Gordon JK, Matthaei C, van Heezik Y 2010. Belled collars reduce catch of domestic cats in New Zealand by half. Wildlife Research 37: 372–378.

Goudet J 2001. FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3). Available from: http://www2.unil.ch/popgen/softwares/fstat.htm

Govenment NZ 2013. 2013 Census QuickStats about a place: Te Anau. www.stats.govt.nz. Accessed 11/10/14.

129

References

Grayson J, Calver M, Styles I 2002. Attitudes of suburban Western Australians to proposed cat control legislation. Australian Veterinary Journal 80: 536–43.

Grueber CE, Nakagawa S, Laws RJ, Jamieson IG 2011. Multimodel inference in ecology and evolution: Challenges and solutions. Journal of Evolutionary Biology 24: 699– 711.

Gurevitch J, Padilla DK 2004. Are invasive species a major cause of extinctions? Trends in Ecology & Evolution 19: 470–474.

Guttilla DA, Stapp P 2010. Effects of sterilization on movements of feral cats at a wildland – urban interface. Journal of Mammalogy 91: 482–489.

Hall CM, Fontaine JB, Bryant KA, Calver MC 2015. Assessing the effectiveness of the Birdsbesafe® anti-predation collar cover in reducing predation on wildlife by pet cats in Western Australia. Applied Animal Behaviour Science doi.org/10.1016/j.applanim.2015.01.004 0.

Hall LS, Kasparian MA, Van Vuren D, Kelt DA 2000. Spatial organization and habitat use of feral cats (Felis catus L.) in Mediterranean California. Mammalia 64: 19–28.

Hampton JO, Spencer PBS, Alpers DL, Twigg LE, Woolnough AP, Doust J, Higgs T, Pluske J 2004. Molecular techniques, wildlife management and the importance of genetic population structure and dispersal: a case study with feral pigs. Journal of Applied Ecology 41: 735–743.

Hansen H, Hess SC, Cole D, Banko PC 2007. Using population genetic tools to develop a control strategy for feral cats (Felis catus) in Hawai’i. Wildlife Research 34: 587- 596.

Hanski I 1999. Habitat connectivity, habitat continuit, and metapopulations in dynamic landscapes. Oikos 87: 209–219.

Harper GA 2004. Feral cats on Stewart Island/Rakiura. DOC Internal Report, Wellington. 35 p.

Harper GA 2007. Habitat selection of feral cats (Felis catus) on a temperate, forested island. Austral Ecology 32: 305–314.

Harrod M, Keown AJ, Farnworth MJ 2015. Use and perception of collars for companion cats in New Zealand. New Zealand Veterinary Journal 169: 1–4.

Hartl DL, Clark AG 2007. Principles of population genetics. Sunderland, Sinauer Associates. 652 p.

Haspel C, Calhoon E 1989. Home ranges of free-ranging cats (Felis catus) in Brooklyn, New York. Canadian Journal of Zoology 67: 178-181.

Hawkins P 2004. Bio-logging and animal welfare: practical refinements. Memiors of National Institute of Polar Research 58: 58–68.

130

References

Hebblewhite M, Merrill E 2008. Modelling wildlife-human relationships for social species with mixed-effects resource selection models. Journal of Applied Ecology 45: 834–844.

Hedrick PW 1995. The Florida gene flow and genetic restoration: Panther as a case study. Conservation Biology 9: 996–1007.

Hedrick P 2005. Genetics of populations. Jones and Bartlett Publishers, Massachusetts. 675 p.

Heffner RS, Heffner HE 1985. Hearing range of the domestic cat. Hearing Research 19: 85–88.

Hemson G, Johnson P, South A, Kenward R, Ripley R, Macdonald D 2005. Are kernels the mustard? Data from global positioning system (GPS) collars suggests problems for kernel home-range analyses with least-squares cross-validation. Journal of Animal Ecology 74: 455–463.

Hoffman JI, Amos W 2005. Microsatellite genotyping errors: Detection approaches, common sources and consequences for paternal exclusion. Molecular Ecology 14: 599–612.

Holderegger R, Di Giulio M 2010. The genetic effects of roads: A review of empirical evidence. Basic and Applied Ecology 11: 522–531.

Horn JA, Mateus-Pinilla N, Warner RE, Heske EJ 2011. Home range, habitat use, and activity patterns of free-roaming domestic cats. The Journal of Wildlife Management 75: 1177–1185.

Horne JS, Garton EO, Krone SM, Lewis JS 2007. Analyzing animal movements using Brownian bridges. Ecology 88: 2354–2363.

Hosmer DW, Lemeshow S, Sturdivant RX 2013. Applied logistic regression. Wiley and Sons, New York. 397 p.

Hu Y, Hu S, Wang W, Wu X, Marshall FB, Chen X, Hou L, Wang C 2014. Earliest evidence for commensal processes of cat domestication. Proceedings of the National Academy of Sciences of the United States of America 111: 116–120.

Hubbard AL, Mcorist S, Jones TW, Boid R, Scott R, Easterbeet N 1992. Is survival of European wildcats Felis silvestris in Britain threatened by interbreeding with domestic cats? Biological Conservation 61: 203–208.

Hubert P, Sirguey P, Poulle M 2015. Performance and accuracy of lightweight and low- cost GPS data loggers according to antenna positions, fix intervals, habitats and animal movements. PLoS ONE 10: e0129271. doi: 10.1371/journal.pone.0129271.

Hurford A 2009. GPS measurement error gives rise to spurious 180° turning angles and strong directional biases in animal movement data. PLoS ONE 4: e5632. doi:10.1371/journal.pone.0005632.

131

References

Jakobsson M, Rosenberg NA 2007. CLUMPP: A cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23: 1801–1806.

Jamieson IG, Ludwig K 2012. Rat-wise robins quickly lose fear of rats when introduced to a rat-free island. Elsevier Ltd. Animal Behaviour 84: 225–229.

Jarne P, Lagoda PJ 1996. Microsatellites, from molecules to populations and back. Trends in Ecology & Evolution 11: 424–429.

Jessup DA 2004. The welfare of feral cats and wildlife. Journal of the American Veterinary Medical Association 225: 1377–1383.

Jiang Z, Sugita M, Kitahara M, Takatsuki S, Goto T, Yoshida Y 2008. Effects of habitat feature, antenna position, movement, and fix interval on GPS radio collar performance in Mount Fuji, central Japan. Ecological Research 23: 581–588.

Johnson D 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65–71.

Jones E, Coman B 1982. Ecology of the feral cat, Felis catus (L.) in South-Eastern Australia II. Reproduction. Wildlife Research 9: 111–119.

Jones C, Norbury G 2006. Habitat use as a predictor of nest raiding by individual hedgehogs Erinaceus europaeus in New Zealand. Pacific Conservation Biology 12: 180–188.

Jongman EC 2007. Adaptation of domestic cats to confinement. Journal of Veterinary Behavior: Clinical Applications and Research 2: 193–196.

Karl BJ, Best HA 1982. Feral cats on Stewart Island; their foods, and their effects on kakapo. New Zealand Journal of Zoology 9: 287–293.

Kays RW, DeWan A 2004. Ecological impact of inside/outside house cats around a suburban nature preserve. Animal Conservation 7: 273–283.

Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Meintjes P, Drummond A 2012. Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28: 1647–1649.

Keedwell RJ, Brown KP 2001. Relative abundance of mammalian predators in the upper Waitaki Basin, South Island, New Zealand. New Zealand Journal of Zoology 28: 31–38.

Keedwell RJ, Maloney RF, Murray DP 2002a. Predator control for protecting kaki (Himantopus novaezelandiae)—lessons from 20 years of management. Biological Conservation 105: 369–374.

Keedwell RJ, Sanders MD, Alley M, Twentyman C 2002b. Causes of mortality of Black- fronted terns Sterna albostriata on the Ohau River, South Island, New Zealand. Pacific Conservation Biology 8: 170–176.

132

References

Keitt B, Campbell K, Saunders A, Clout M, Wang Y, Heinz R, Newton K, Tershy B 2011. The global islands invasive vertebrate eradication database: A tool to improve and facilitate restoration of island ecosystems. In: Island invasives eradication and management, Veitch CR, Clout M, Towns DR. IUCN, Gland. Pp. 74–77.

Keller LF, Waller DM 2002. Inbreeding effects in wild populations. Trends in Ecology and Evolution 17: 19–23.

Kenward RE, Casey NM, Walls SS, South AB 2014. Ranges 9 : For the analysis of tracking and location data. Wareham, UK. Online manual.

Kids Restore the Kepler 2013. Kids restore the kepler. www.kidsrestorethekepler.co.nz. Accessed 08/10/14.

King C 2005. The handbook of New Zealand Mammals. Oxford University Press, Oxford. 610 p.

King TM, Williams M, Lambert DM 2000. Dams, ducks and DNA: identifying the effects of a hydro-electric scheme on New Zealand’s endangered blue duck. Conservation Genetics 1: 103–113.

Kitson AE, Thiele O 1910. The geography of the Upper Waitaki Basin, New Zealand. The Geographical Journal 36: 537–551.

Kitts-Morgan SE, Caires KC, Bohannon LA, Parsons EI 2015. Free-ranging farm cats: home range size and predation on a livestock unit in Northwest Georgia. PLoS ONE 10: e0120513. doi:10.1371/journal.pone.0120513.

Koch K, Algar D, Schwenk K 2014. Population structure and management of invasive cats on an Australian Island. The Journal of Wildlife Management 78: 968–975.

Koch K, Algar D, Searle JB, Pfenninger M, Schwenk K 2015. A voyage to Terra Australis: human-mediated dispersal of cats. BMC Evolutionary Biology 15. doi: 10.1186/s12862-015-0542-7.

Koganezawa M, Imaki H 1999. The effects of food sources on Japanese monkey home range size and location, and population dynamics. Primates 40: 177–185.

Koike F, Clout M, Kawamichi M, De Poorter M, K I 2006. Assessment and control of biological invasion risks. Gland. 216 p.

Kok OB, Nel JAJ 2004. Convergence and divergence in prey of sympatric canids and felids: opportunism or phylogenetic constraint? Biological Journal of the Linnean Society 83: 527–538.

Kolbe JJ, Glor RE, Rodríguez Schettino L, Lara AC, Larson A, Losos JB 2004. Genetic variation increases during biological invasion by a Cuban lizard. Nature 431: 177– 181.

Konecny MJ 1987. Home range and activity patterns of feral house cats in the Galapagos Islands. Oikos 50: 17-23.

133

References

Krauze-Gryz D, Gryz J, Goszczyński J 2012. Predation by domestic cats in rural areas of central Poland: An assessment based on two methods. Journal of Zoology 288: 260–266.

Kressin D 2009. Oral examination of cats and dogs. Compendium Continued Vetinary Education 31: 72–85.

Kurushima JD, Ikram S, Knudsen J, Bleiberg E, Grahn RA, Lyons LA 2012. Cats of the pharaohs: Genetic comparison of Egyptian cat mummies to their feline contemporaries. Journal of Archaeological Science 39: 3217–3223.

Kustritz MVR 2002. Early spay-neuter: clinical considerations. Clinical Techniques in Small Animal Practice 17: 124–128.

Kustritz MVR 2007. Determining the optimal age for gonadectomy of dogs and cats. Veterinary Medicine Today 231: 1665–1675.

Lal A 2008. A preliminary evaluation of mammalian predator trapping efficacy at Macraes Flat, Otago. Unpublished postgraduate WILM report, University of Otago, Otago, New Zealand. 48 p.

Langham NPE 1990. The diet of feral cats (Felis catus L.) on Hawke’s Bay farmland, New Zealand. New Zealand Journal of Zoology 17: 243–255.

Langham N, Porter R 1991. Feral Cats (Felis Catus L.) On New Zealand Farmland. I. Home Range. Wilflife Research 18: 741-760.

Lecis R, Pierpaoli M, Bir ZS, Szemethy L, Ragni B, Vercillo F, Randi E 2006. Bayesian analyses of admixture in wild and domestic cats (Felis silvestris) using linked microsatellite loci. Molecular Ecology 15: 119–131.

Leyhausen P 1979. Cat behavior : the predatory and social behavior of domestic and wild cats. Garland Press, New York. 340 p.

Liberg O 1980. Spacing patterns in a population of rural free roaming domestic cats. Oikos 35: 336–349.

Liberg O, Sandell M, Pontier D, Natoli E 2000. Density, spatial organisation and repordutive tactics in the domestic cat and other felids. In: The domestic cat: Biology of its behaviour, Turner DC, Bateson PPG ed. Cambridge University Press, Cambridge. 222 p.

Lichti NI, Swihart RK 2011. Estimating utilization distributions with kernel versus local convex hull methods. The Journal of Wildlife Management 75: 413–422.

Lilith M, Calver M, Garkaklis M 2008. Roaming habits of pet cats on the suburban fringe in Perth, Western Australia: what size buffer zone is needed to protect wildlife in reserve? In: Too close for comfort: contentious issues in human-wilflife encounters, Lunney D, Munn A, Meikle W ed. Royal Zoological Society of New South Wales, Mosman. Pp. 65–72.

134

References

Lilith M, Calver M, Styles I, Garkaklis M 2006. Protecting wildlife from predation by owned domestic cats: Application of a precautionary approach to the acceptability of proposed cat regulations. Austral Ecology 31: 176–189.

Linseele V, van Neer W, Hendrickx S 2007. Evidence for early cat taming in Egypt. Journal of Archaeological Science 34: 2081–2090.

Linseele V, van Neer W, Hendrickx S 2008. Early cat taming in Egypt: a correction. Journal of Archaeological Science 35: 2672–2673.

Loeschcke V, Tomiuk J, Jain SK 1994. Conservation genetics. Springer, New York. 436 p.

Lohr CA, Cox LJ, Lepczyk CA 2012. Costs and benefits of trap-neuter-release and euthanasia for removal of urban cats in Oahu, Hawai’i. The Journal of the Society for Conservation Biology 27: 64–73.

Loss SR, Will T, Marra PP 2013. The impact of free-ranging domestic cats on wildlife of the United States. Nature Communications 4: 1 -7.

Lowe S, Browne M, Boudjelas S, De Poorter M 2000. 100 of the world’s worst invasive alien species. A selection from the Global Invasive Species Database. The Invasive Species Specialist Group (ISSG) a specialist group of the Species Survival Commission (SSC) of the World Conservation Union (IUCN): 12 p.

Loyd KAT 2015. Sociopolitical, ecological and behavioural aspects of free-roaming cats. Unpublished PhD Thesis. University of Georgia, Georgia. Greece.

Loyd KAT, Hernandez SM, Abernathy KJ, Shock BC, Marshall GJ 2013a. Risk behaviours exhibited by free-roaming cats in a suburban US town. The Veterinary Record 173. doi: 10.1136/vr.101222.

Loyd KAT, Hernandez SM, Carroll JP, Abernathy KJ, Marshall GJ 2013b. Quantifying free-roaming domestic cat predation using animal-borne video cameras. Biological Conservation 160: 183–189.

Mackay J 2011. Companion Animals in New Zealand. New Zealand Compaion Council Inc. Wellington. 62 p.

Macleod C, Blackwell G, Weller F, Moller H 2012a. Advances in tools for bird populations: Designing a bird monitoring scheme for New Zealand’s agricultural sectors. New Zealand Journal of Ecology 36: 1–12.

Macleod CJ, Greene TC, Mackenzie DI, Allen RB 2012b. Monitoring widespread and common bird species on New Zealand’s conservation lands: a pilot study. New Zealand Journal of Ecology 36: 1–12.

Manly B, McDonald LL, Thomas DL, McDonald TL, Erickson WP 2002. Resource Selection by Animals. Kluwer Academic Publishers, London. 221 p.

Mantel N 1967. The detection of disease clustering and a generalized regression approach. Nature 214: 637–637.

135

References

Markoulatos P, Siafakas N, Moncany M 2002. Multiplex polymerase chain reaction: A practical approach. Journal of Clinical Laboratory Analysis 16: 47–51.

Marsh J 2013. Kids Restore the Kepler. Restoration SoultioNZ, Te Anau. 33p.

McDonald JL, Maclean M, Evans MR, Hodgson DJ 2015. Reconciling actual and perceived rates of predation by domestic cats. Ecology and Evolution 5: 2745–2753.

McLintock A 1966. An encyclopedia of New Zealand. New Zealand Government, Wellington. Available from: http://www.teara.govt.nz/en/1966.

McLoughlin PD, Case RL, Gau RJ, Cluff HD, Mulders R, Messier F 2002. Hierarchical habitat selection by barren-ground grizzly bears in the central Canadian Arctic. Oecologia 132: 102–108.

Meek PD 2003. Home range of house cats Felis catus living within a National Park. Australian Mammalogy 25: 51–60.

Menotti-Raymond M, David VA, Brien SJO 2007. STR based forensic analysis of felid samples from domestic and exotic cats. In: Nonhuman DNA Typing: Theory and casework applications, Coyle HM. CRC Press, New York. Pp. 71–94.

Menotti-Raymond MA, David VA, Wachter LL, Butler JM, O’Brien SJ 2005. An STR forensic typing system for genetic individualization of domestic cat (Felis catus) samples. Journal of Forensic Sciences 50: 1061–70.

Menotti-Raymond M, David VA, Weir BS, O’Brien SJ 2012. A population genetic database of cat breeds developed in coordination with a domestic cat STR multiplex. Journal of Forensic Sciences 57: 596–601.

Metsers L 2008. Movement behaviour and habitat use of domestic cats in relation to threatened native lizard habitat. Unpublisehd MSc Thesis. University of Otago. Otago, New Zealand.

Metsers EM, Seddon PJ, van Heezik YM 2010. Cat-exclusion zones in rural and urban- fringe landscapes: how large would they have to be? Wildlife Research 37: 47-56.

Middlemiss A 1995. Predation of lizards by feral house cats (Felis catus) and ferrets (Mustela furo) in the tussock grassland of Otago. Unpublisehd MSc Thesis. University of Otago. Otago, New Zealand.

Mohr C 1947. Table of equivalent populations of North American small mammals. Midland Naturalist 37: 223–249.

Molsher RL 1999. The ecology of feral cats, Felis catus, in open forest in New South Wales: interactions with food resources and foxes. Unpublished PhD Thesis. Univerity of Sydney, Australia.

Molsher R, Dickman C, Newsome A, Müller W 2005. Home ranges of feral cats (Felis catus) in central-western New South Wales, Australia. Wildlife Research 32: 587– 595.

136

References

Moorhouse RJ, Powlesland RG 1991. Aspects of the ecology of Kakapo Strigops habroptilus liberated on Little Barrier Island (Hauturu), New Zealand. Biological Conservation 56: 349–365.

Morgan SAA, Hansen CMA, Ross JGA, Hickling GJB, Ogilvie SCA 2009. Urban cat (Felis catus) movement and predation activity associated with a wetland reserve in New Zealand. 36: 574–580.

Mulongoy KJ, Webbe J, Ferreira M, Mittermeier C 2006. The wealth of islands, A global call for conservation. Montreal. 24 p.

Nakagawa S, Freckleton RP 2011. Model averaging, missing data and multiple imputation: A case study for behavioural ecology. Behavioral Ecology and Sociobiology 65: 103–116.

Nakagawa S, Schielzeth H 2013. A general and simple method for obtaining R 2 from generalized linear mixed-effects models. Methods in Ecology and Evolution 4: 133– 142.

Natoli E, Baggio A, Pontier D 2001. Male and female agonistic and affiliative relationships in a social group of farm cats (Felis catus L.). Behavioural processes 53: 137–143.

Natoli E, Ferrari M, Bolletti E, Pontier D 1999. Relationships between cat lovers and feral cats in Rome. Anthrozoos 12: 16–23.

Natoli E, Maragliano L, Cariola G, Faini A, Bonanni R, Cafazzo S, Fantini C 2006. Management of feral domestic cats in the urban environment of Rome (Italy). Preventive Veterinary Medicine 77: 180–185.

Natoli E, Say L, Cafazzo S, Bonanni R, Schmid M, Pontier D 2005. Bold attitude makes male urban feral domestic cats more vulnerable to Feline Immunodeficiency Virus. Neuroscience and Biobehavioral Reviews 29: 151–157.

Nelson SH, Evans AD, Bradbury RB 2005. The efficacy of collar-mounted devices in reducing the rate of predation of wildlife by domestic cats. Applied Animal Behaviour Science 94: 273–285.

Nentwig W 2007. Biological Invasions. Springer, Berlin. 441 p.

Neter J, Kutner MH, Wasserman W, Nachtsheim CJ 1996. Applied linear statistical models. McGraw-Hill Publishers, New York. 1408 p.

New Zealand Government 2007. Fiordland National Park Management Plan. Invercargill. 366 p.

Nielsen R, Slatkin M 2013. An introduction to population genetics: theory and applications. Sinauer Associates, Sunderland. 287 p.

Nislow KH, Hudy M, Letcher BH, Smith EP 2011. Variation in local abundance and species richness of stream fishes in relation to dispersal barriers: Implications for management and conservation. Freshwater Biology 56: 2135–2144.

137

References

Nogales M, Mart A, Tershy BR, Donlan CJ, Veitch D 2004. A review of feral cat eradication on islands. Conservation Biology 18: 310–319.

Norbury GL, C. ND, Oliver AJ 1994. Facultative behavior in unpredictable environments - Mobility of red kangaroos in arid Western-Australia. Journal of Animal Ecology 63: 410–418.

Norbury G, Heyward R, Parkes J 2002. Short-term ecological effects of rabbithaemorrhagic disease in the short-tussock grasslands of the South Island, New Zealand. Wildlife Research 29: 599–604.

Norbury G, Reardon JT, McKinlay B 2006. Grand and Otago skink recovery plan 2006- 2016. DOC Threatened Species Recovery Plan. 28 p.

Nowell K, Jackson P 1996. European wildcat, Felis silvestris group Schreber, 1775. In: Nowell K, Jackson P ed. Wild cats: status survey and conservation action plan. IUCN, Gland. Pp. 110-113.

Nsubuga AM, Holzman J, Chemnick LG, Ryder OA 2010. The cryptic genetic structure of the North American captive gorilla population. Conservation Genetics 11: 161– 172.

O’Donnell CFJ, Christie J, Corben C, Sedgeley JA, Simpson W 1999. Rediscovery of short-tailed bats (Mystacina sp) in Fordland, New Zealand: Preliminary observations of taxonomy, echolocation calls, population size, home range, and habitat use. New Zealand Journal of Ecology 23: 21–30.

O’Hara P 2007. Animal welfare (companion cats) code of welfare 2007. National Animal Welfare Advisory Committee Ministry of Agriculture and Fisheries. Wellington. 41p.

Paetkau D, Slade R, Burden M, Estoup A 2004. Genetic assignment methods for the direct, real-time estimation of migration rate: A simulation-based exploration of accuracy and power. Molecular Ecology 13: 55–65.

Parsons H, Major RE, French K 2006. Species interactions and habitat associations of birds inhabiting urban areas of Sydney, Australia. Austral Ecology 31: 217–227.

Peakall R, Smouse PE 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research--an update. Bioinformatics 28: 2537–2539.

Pierpaoli M, Biro ZS, Herrmann M, Hupe K, Fernandes M, Ragni B, Szemethy L, Randi E 2003. Genetic distinction of wildcat (Felis silvestris) populations in Europe, and hybridization with domestic cats in Hungary. Molecular Ecology 12: 2585–2598.

Pinheiro J, Douglas B, Saikat D, Deepayan S, Siem H, Bert V 2016. Package “ nlme .” 336 p.

Piry S, Alapetite A, Cornuet JM, Paetkau D, Baudouin L, Estoup A 2004. GENECLASS2: A software for genetic assignment and first-generation migrant detection. Journal of Heredity 95: 536–539.

138

References

Plantinga EA, Bosch G, Hendriks WH 2011. Estimation of the dietary nutrient profile of free-roaming feral cats: possible implications for nutrition of domestic cats. The British Journal of Nutrition 106: 35–48.

Pollard M 1999. The encyclopedia of the cat. Dempsey Parr, Bath. 384 p.

Pontier D, Say L, Devillard S, Bonhomme F 2005. Genetic structure of the feral cat (Felis catus L.) introduced 50years ago to a sub-Antarctic island. Polar Biology 28: 268– 275.

Potts A 2009. Kiwis against possums: A critical analysis of anti-possum rhetoric in Aotearoa, New Zealand. Society and Animals 17: 1–20.

Pritchard JK 2010. Documentation for STRUCTURE software: Version 2.3. Pp 321-326.

Queller DC, Goodnight KF 1989. Estimating relatedness using genetic markers. Society for the Study of Evolution 43: 258–275.

Queller DC, Strassmann JE, Hughes CR 1993. Microsatellites and kinship. Trends in Ecology & Evolution 8: 285–288.

R Core Team 2014. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. Available from: http:// R-project.org.

Rambaut A, Drummond A 2009. Tracer v1.5. Available from: http://tree.bio.ed.ac.uk/software/tracer.

Ramón ME, Slater MR, Ward MP 2010. Companion animal knowledge, attachment and pet cat care and their associations with household demographics for residents of a rural Texas town. Preventive Veterinary Medicine 94: 251–63.

Randi E, Ragni B 1991. Genetic variability and biochemical systematics of domestic and wild cat populations (Felis catus). Journal of Mammalogy 72: 79–88.

Rannala B, Mountain JL 1997. Detecting immigration by using multilocus genotypes. Proceedings of the National Academy of Sciences of the United States of America 94: 9197–9201.

Rebergen AL, Woolmore C 2015. Project River Recovery Strategic Plan 2012-2019. Twizel. 24 p.

Recio MR, Mathieu R, Denys P, Sirguey P, Seddon PJ 2011. Lightweight GPS-tags, one giant leap for wildlife tracking? An assessment approach. PloS ONE 6: e28225. doi:10.1371/journal.pone.0028225.

Recio MR, Mathieu R, Maloney R, Seddon PJ 2010. First results of feral cats (Felis catus) monitored with GPS collars in New Zealand. 34: 288-296.

Recio MR, Seddon PJ 2013. Understanding determinants of home range behaviour of feral cats as introduced apex predators in insular ecosystems: a spatial approach. Behavioral Ecology and Sociobiology 67: 1971–1981.

139

References

Recio MR, Seddon PJ, Moore AB 2015. Niche and movement models identify corridors of introduced feral cats infringing ecologically sensitive areas in New Zealand. Biological Conservation 192: 48–56.

Rice WR 1989. Analyzing tables of statistical tests. Ecology 43: 223–225.

Ritchie EG, Johnson CN 2009. Predator interactions, mesopredator release and biodiversity conservation. Ecology Letters 12: 982–998.

Robertson SA 2008. A review of feral cat control. Journal of Feline Medicine and Surgery 10: 366–375.

Robertson H, Dowding J, Elliott G, Hitchmough R, Miskelly C, O’Donnell C, Powlesland R, Sagar P, Scofield R, Taylor G 2013. Conservation status of New Zealand birds. Wellington. 26 p.

Robertson BC, Gemmell NJ 2004. Defining eradication units to control invasive pests. Journal of Applied Ecology 41: 1042–1048.

Robertson BC, Gemmell NJ 2006. PCR-based sexing in conservation biology: Wrong answers from an accurate methodology? Conservation Genetics 7: 267–271.

Robertson CJR, O’Donnell CFJ, Overmars FB 1983. Habitat requirements of wetland birds in the Ahuriri River catchment New Zealand. Department of Internal Affairs, Wellington. 455 p.

Robinson R, Cox HW 1970. Reproductive performance in a cat colony over a 10-year period. Laboratory Animals 4: 99–112.

Robley A, Ramsey D, Woodford L, Lindeman M, Johnston M, Forsyth D 2008. Evaluation of detection methods and sampling designs used to determine the abundance of feral cats. Technical Report Series No. 181. 28 p.

Rosenberg NA 2004. DISTRUCT: A program for the graphical display of population structure. Molecular Ecology Notes 4: 137–138.

Rosenblatt JS, Aronson LR 1958. The decline of sexual behavior in male cats after castration with special reference to the role of prior sexual experience. Behaviour 12: 285–338.

Rothwell T 2004. Evidence for taming of cats. Science 305: 1714–1715.

Rousset F 2008. GENEPOP’007: A complete re-implementation of the GENEPOP software for Windows and Linux. Molecular Ecology Resources 8: 103–106.

Ruiz-Garcia M 1994. Genetic structure of the Marseilles cat population: is there really a strong founder effect? Genetics Selection Evolution 26: 317–331.

Ruiz-Garcia M 1999. Genetic structure of different cat populations in Europe and South America at a microgeographic level: importance of the choice of an adequate sampling level in the accuracy of population genetics interpretations. Genetics and Molecular Biology 22: 493–505.

140

References

Runge JP, Runge MC, Nichols JD 2006. The role of local populations within a landscape context: Defining and classifying sources and sinks. The American Naturalist 167: 925–938.

Ruxton G, Thomas S, Wright J 2002. Bells reduce predation of wildlife by domestic cats (Felis catus). Journal of Zoology, London 256: 81–83.

Sanders MD, Maloney RF 2002. Causes of mortality at nests of ground-nesting birds in the Upper Waitaki Basin, South Island, New Zealand: a 5-year video study. Biological Conservation 106: 225–236.

Saunders A, Norton D. 2001. Ecological restoration at Mainland Islands in New Zealand. Biological Conservation 99: 109–119.

Say L, Bonhomme F, Desmarais E, Pontier D 2003. Microspatial genetic heterogeneity and gene flow in stray cats (Felis catus L.): a comparison of coat colour and microsatellite loci. Molecular Ecology 12: 1669–1674.

Schmidt PM, Swannack TM, Lopez RR, Slater MR 2009. Evaluation of euthanasia and trap–neuter–return (TNR) programs in managing free-roaming cat populations. Wildlife Research 36: 117.

Schuelke M 2000. An economic method for the fluorescent labeling of PCR fragments A poor man’s approach to genotyping for research and high-throughput diagnostics. Nature Biotechnology 18: 1–2.

Seaman DE, Powell RA 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecological Society of America 77: 2075–2085.

Selkoe KA, Toonen RJ 2006. Microsatellites for ecologists: a practical guide to using and evaluating microsatellite markers. Ecology Letters 9: 615–29.

Seppala M, Koutaniemi L 1985. Formation of a string and pool topography as expressed by morphology, stratigraphy and current processes on a mire in Kuusamo, Finland. Boreas 14: 287–309.

Serpell S 1988. The domestication of the cat. Cambridge University Press, Cambridge. Pp. 151–158.

Shea K, Chesson P 2002. Community ecology theory as a framework for biological invasions. Trends in Ecology and Evolution 17: 170–176.

Simonoff J 1996. Smoothing methods in statistics. Springer-Verlag, New York. 338 p.

Sims V, Evans KL, Newson SE, Tratalos JA, Gaston KJ 2008. Avian assemblage structure and domestic cat densities in urban environments. Diversity and Distributions 14: 387–399.

Smouse PE, Peakall R 1999. Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure. Heredity 82: 561–573.

141

References

Sparkes J, Kortner G, Ballard G, Fleming PJS, Brown WY 2014. Effects of sex and reproductive state on interactions between free-roaming domestic dogs. PLoS ONE 9: e116053. doi:10.1371/journal.pone.0116053.

Spencer PBS, Yurchenko AA, David VA, Scott R, Koepfli K, Driscoll C, Brien SJO, Menotti-raymond M 2015. The population origins and expansion of feral cats in Australia. Journal of Heredity 107: 104–114.

Spitzen-van der Sluijs A, Spitzen J, Houston D, Stumpel AHP 2009. Skink predation by hedgehogs at Macraes Flat, Otago, New Zealand. New Zealand Journal of Ecology 33: 205–207.

Steiniger S, Hunter AJS 2013. A scaled line-based kernel density estimator for the retrieval of utilization distributions and home ranges from GPS movement tracks. Ecological Informatics 13: 1–8.

Stringham SA, Mulroy EE, Xing J, Record D, Guernsey MW, Aldenhoven JT, Osborne EJ, Shapiro MD 2012. Divergence, convergence, and the ancestry of feral populations in the domestic rock pigeon. Current Biology 22: 302–308.

Sunnucks P 2000. Efficient genetic markers for population biology. Trends in Ecology & Evolution 15: 199–203.

Sunquist M, Sunquist F 2002. Wild cats of the world. University of Chicago Press, Chicago. 452 p.

Tabachnick BG, Fidell LS 2007. Using multivariate statistics. Pearson Education, London. 980 p.

Tennent J, Downs CT 2008. Abundance and home ranges of feral cats in an urban conservancy where there is supplemental feeding: a case study from South Africa. Zoological Society of Southern Africa 43: 218–229.

Tennent JK, Downs CT, Wald DM, Watson HK 2010. Public perceptions of feral cats within an urban conservancy on a campus of the University of KwaZulu-Natal. Zoological Society of Southern Africa 1: 16–26.

Thomas RL, Baker PJ, Fellowes MDE 2014. Ranging characteristics of the domestic cat (Felis catus) in an urban environment. Urban Ecosystems doi: 10.1007/s11252-014- 0360-5.

Thomas CD, Kunin WE 1999. The spatial structure of populations. Journal of Animal Ecology 68: 647–657.

Towns DR 2011. Eradications of vertebrate pests from islands around New Zealand: what have we delivered and what have we learned? In: Island invasives: eradication and management, Veitch CR, Clout MN, Towns DR. IUCN, Auckland. Pp. 364– 371.

142

References

Townsend CR, Winterbourn MJ 1992. Assessment of the environmental risk posed by an exotic fish: the proposed introduction of channel catfish (Ictalurus punctatus) to New Zealand. Conservation Biology 6: 273–282.

Tschanz B, Hegglin D, Gloor S, Bontadina F 2011. Hunters and non-hunters: Skewed predation rate by domestic cats in a rural village. European Journal of Wildlife Research 57: 597–602.

Tsutsui T, Stabenfeldt GH 1992. Biology of ovarian cycles, pregnancy and pseudopregnancy in the domestic cat. Journal of Reproduction and Fertility 47: 29– 35. van Aarde R 1980. The diet and feeding behaviour of feral cats, Felis catus at Marion Island. South African Journal of Wildlife Research 10: 123–128. van Heezik Y, Adams N, Hight S 2014. Comunity attitudes towards pet cats. Report to Respondents. Dunedin. 17 p. van Heezik Y, Freeman C, Porter S, Dickinson KJM 2013. Garden size, householder knowledge, and socio-economic status influence plant and bird diversity at the scale of individual gardens. Ecosystems 16: 1442–1454. van Heezik Y, Lei P, Maloney R, Sancha E 2005. Captive breeding for reintroduction: Influence of management practices and biological factors on survival of captive kaki (black stilt). Zoo Biology 24: 459–474. van Heezik Y, Smyth A, Adams A, Gordon J 2010. Do domestic cats impose an unsustainable harvest on urban bird populations? Biological Conservation 143: 121–130. van Heezik Y, Smyth A, Mathieu R 2008. Diversity of native and exotic birds across an urban gradient in a New Zealand city. Landscape and Urban Planning 87: 223–232.

Van Neer W, Linseele V, Friedman R, De Cupere B 2014. More evidence for cat taming at the Predynastic elite cemetery of Hierakonpolis (Upper Egypt). Journal of Archaeological Science 45: 103–111. van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P 2004. Micro-Checker: Software for Identifying and Correcting Genotyping Errors in Microsatellite Data. Molecular Ecology Notes 4: 535–538.

Verstraete FJ, van Aarde RJ, Nieuwoudt BA, Mauer E, Kass PH 1996. The dental pathology of feral cats on Marion Island, part II: periodontitis, external odontoclastic resorption lesions and mandibular thickening. Journal of Comparative Pathology 115: 283–297.

Vigne J, Guilaine J, Debue K, Haye L 2004. Early taming of the cat in cyprus. Science 304: 259.

Wallace JL, Levy JK 2006. Population characteristics of feral cats admitted to seven trap- neuter-return programs in the United States. Journal of Feline Medicine and Surgery 8: 279–284.

143

References

Walsh P, Metzger DA, Higuchi R 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques 10: 506 – 513.

Warner RE 1985. Demography and movements of free-ranging domestic cats in rural Illinois. The Journal of Wildlife Management 49: 340–346.

White GC, Garrott RA 1990. Analysis of wildlife radio-tracking data. Academic Press, San Deigo. 383 p.

Whitlock MC, Ingvarsson PK, Hatfield T 2000. Local drift load and the heterosis of interconnected populations. Heredity 84 : 452–457.

Whitlock MC, Schluter D 2009. The analysis of biological data. Roberts and Company Publishers, Greewood Village, Colorado. 700 p.

Wierzbowska IA, Olko J, Hędrzak M, Crooks KR 2012. Free-ranging domestic cats reduce the effective protected area of a Polish national park. Mammalian Biology 77: 204–210.

Willson SK, Okunlola IA, Novak JA 2015. Birds be safe: Can a novel cat collar reduce avian mortality by domestic cats (Felis catus)? Global Ecology and Conservation 3: 359–366.

Wilson RR, Gilbert-Norton L, Gese EM 2012. Beyond use versus availability: behaviour-explicit resource selection. Wildlife Biology 18: 424–430.

Wilson RP, McMahon CR 2006. Measuring devices on wild animals: What consititutes acceptable practice? Ecological Society of America 4: 147–154.

Wilson GA, Rannala B 2003. Bayesian inference of recent migration rates using multilocus genotypes. Genetics 163: 1177–1191.

Wood V, Seddon PJ, Beaven B, van Heezik Y 2016. Movement and diet of domestic cats on Stewart Island/Rakiura, New Zealand. New Zealand Ecological Society. 40: 1- 5.

Woods M, Mcdonald RA, Phen STE, Ris HAR 2003. Predation of wildlife by domestic cats Felis catus in Great Britain. 33: 174–188.

Woolmore CB, Anderson SJ, Garside R 2012. Project River Recovery Annual Report 2012/01. Department of Conservation, Twizel. Pp. 21.

YourWildLife 2014. Cat tracker. cats.yourwildlfie.org. Accessed: 18/10/14.

Zane L, Bargelloni L, Patarnello T 2002. Strategies for microsatellite isolation: a review. Molecular ecology 11: 1–16.

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM, Ebooks Corporation. 2009. Mixed effects models and extensions in ecology with R. Statistics for Biology and Health. 579 p.

144

Appendix 1: “What does your cat get up to?” flier (cat photo courtesy of Cayley Coughlin).

Appendices

145

Appendices

Appendix 2: Study information for participants.

146

Appendices

Appendix 3: Survey used to collect information on study companion cats.

147

Appendix 4: Prey record sheet. Appendices

148

Appendices

Appendix 5: Community involvement.

When I initially arrived in Te Anau, I presented my research proposal to the local schools to increase awareness of the project and companion cat movements (Table 1). In addition, I was able to bring students with me to collar cats (following owner and parent permission, Table 1). I also was invited to the Enviroschools Hui on the 23 October 2014 which gave me the opportunity to see the amazing conservation work carried out by the local schools and to talk to students from around Southland about my research. I also conducted several follow-up presentations detailing my preliminary findings (Table 1). Below are some photos taken during my presentations and outings with school children.

I borrowed a stuffed feral cat from DoC for my Showing the children at Fiordland presentations (photo courtesy of Claire Kindergarten how the collar is worn by the Shaw). cats (photo courtesy of Claire Shaw).

Children at the Southern Stars Early Learning Students and teachers at Mararoa Primary Centre accompanied by the stuffed feral cat. School.

Tom Kelly Clarke with McMertrie, Lily. William Hamilton and Blake DeBettencor with Junior.

Jessica Willans, Millie Iona Walker Brimecombe and and Angus Isabella Wilson Holmes- with Jerry Crombie with Houdini.

149

Appendices

Table 1: Table detailing presentations and student field trips given in Te Anau. Field trips involved either GPS collaring a cat or changing the battery over.

Event School/Venue Date Who? Proposal presentation Fiordland Kindergarten 7/11/2014 3 - 4 year olds Proposal Te Anau Primary 14/11/2014 Junior and senior presentation School syndicates Proposal Fiordland College 18/11/2014 Interested students presentation Iona Brimecombe, Student ‘field trip’ Fiordland College 19/11/2014 Jessica Willans, Isabella Wilson Blake Te Anau Primary DeBettencor, Tom Student ‘field trip’ 19/11/2014 School McMurtrie, William Hamilton Proposal Southern Stars Early 20/11/2014 1 - 2 year olds presentation learning Centre Proposal Mararoa Primary 21/11/2014 5 - 10 year olds Presentation School Student ‘field trip’ Fiordland College 21/11/2014 Kelly Clarke Angus Holmes- Student ‘field trip’ Fiordland College 4/12/2014 Crombie, Millie Walker Preliminary results Fiordland Kindergarten 16/05/2015 3 - 4 year olds presentation Preliminary Mararoa Primary results 16/05/2015 5 - 10 year olds School presentation Preliminary Department of Members of the results 16/05/2015 Conservation public presentation Preliminary results Fiordland College 17/05/2015 Interested students presentation

150

Appendices

Appendix 6:

Examples of Incremental Area Analyses that correspond to home ranges that were and were not fully revealed.

Alvin (fully revealed) Dexter (fully revealed)

Merlot (fully revealed) Munchkin (not fully revealed)

Debbie (not fully revealed) Houdini (not fully revealed)

151

Appendices Appendix 7: Data tables Table 1: Details of each companion cat tracked in the Te Anau Basin including breed, age, approximate body mass (BM), sex – male (M) and female (F), property type, bell or collar wearer, period of tracking, date tracking started (start date) and number of rain days (RD) over the tracking period acquired from the owner-filled survey. Bell and collar indicate whether the cat wore a bell or collar before starting the study. Table continued over page. “-“ indicates no information supplied. DSH stands for domestic short hair. Age Approx. Property Period Name Breed Sex Bell Collar Start Date No. RD (years/months) BM (kg) type (days)

Alvin a, 2 Mixed 7/6 7 M Rural No No 13 6/11/2014 10

Austin Burmilla 6 6.3 M Rural No No 13 4/12/2014 9

Bear Burmese 2 4.8 M Rural No No 13 4/11/2014 9 Benjie DSH 1 4.5 M Urban Yes Yes 13 5/12/2014 9

c Charlie - 6 5.1 M Urban No No 14 1/12/2014 4 Debbie d - 3 4.3 F Rural Yes Yes 13 3/12/2014 4

Dexter 2 Ragdoll 2 5.1 M Urban Yes Yes 13 2/11/2014 9 Dipstick d - 3 5.8 M Rural Yes Yes 13 3/12/2014 4

Dora e Mancoon 8 3.4 F Urban No No 14 19/11/2014 9 Felix a, 2 Russian Black 3/6 8 M Rural No Yes 13 6/11/2014 10

Houdini - 8 6.8 F Urban No No 13 4/12/2014 5

Jerry - 14 7.2 M Rural No No 13 19/11/2014 9

Junior e - 7 6.7 M Urban No No 13 19/11/2014 9

Karma f - 14 4 F Rural No No 13 17/11/2014 11

Kelvin ½ Burmese 4 6.9 M Rural No No 13 2/12/2014 4

Lily1, 2, g - 1/9 5 F Urban No Yes 13 16/11/2014 11

Lola - 5/4 4.6 F Urban No No 11 8/12/2014 6

152

Appendices Age Approx. Property Period Name Breed Sex Bell Collar Start Date No. RD (years/months) BM (kg) Type (days) Lulu - 4/3 5.9 F Urban No No 13 5/12/2014 6

Max c - 4 6.1 M Urban No No 14 1/12/2014 4

Merlot Tabby 4 5.1 F Rural No No 13 5/11/2014 9

Mia Burman 10 3.4 F Rural No No 13 5/11/2014 9

Misty - 6/7 4.9 M Urban No No 13 2/11/2014 9

Mittens - 3 4.1 F Urban No No 11 4/11/2014 7

Munchkin f - 13 4.7 M Rural No No 13 17/11/2014 11

Paddy - 3 4.7 M Urban No No 13 3/11/2014 9

Rocky DSH 5/1 5.9 M Rural No No 10 17/11/2014 9

Rosie b Burmese 2 3.8 F Rural No No 13 4/11/2014 9

Slinky - 16 5.2 M Urban No No 11 20/11/2014 8

Smokey - 1 5 M Urban No No 13 2/12/2014 4

Socks - 14 or 15 4.1 M Urban No No 13 18/11/2014 10

Spider 2 DSH Burmese 5/2 5 M Urban Yes Yes 11 2/11/2014 7

Tom1, 3, g - 1/9 5 M Urban No Yes - - -

Wolfie 2 - 3 5 M Urban No No 14 5/12/2014 7

a: Alvin and Felix live in the same household b: Bear and Rosie live in the same household c: Charlie and Max live in the same household d: Debbie and Dipstick live in the same household e: Dora and Junior live in the same household f: Karma and Munchkin live in the same household g: Lily and Tom from same household 1: Owner supplied collar.

2: Multiple cats in household that were not tracked 3: Not tracked, lost collar

153

Appendices Table 2: Further details about each companion cat tracked in the Te Anau Basin detailing the number of cats in household (no. cats), number of times cat is fed per day (No. fed), frequency cat brings home prey, number of nights the cat spends outside, number of nights the cats has disappeared at a time (No. nights disappeared), fed

canned food and meat (wet) or dry food (dry). Table continued over page.

Name No. cats No. fed Prey frequency Nights out No. nights disappeared Wet Dry Alvin a,2 3 2 Daily 6 3 – 7 Yes Yes

Austin 1 2 Monthly, >1/yr 7 0 Yes Yes Bear b 2 2 <1/yr 7 0 Yes Yes Benjie 1 2 Monthly 7 0 Yes Yes

Charlie c 2 ad. lib. Never 7 3 No Yes

Debbie d 2 ad. lib. Weekly 7 0 No Yes

Dexter 2 2 3 Never 7 1 No Yes

Dipstick d 2 ad. lib. Weekly 7 0 No Yes

Dora e 2 1 <1/yr 4 2 No Yes

Felix a, 2 3 2 Daily 3 2 Yes Yes

Houdini 1 2 >1/yr 0 0 Yes Yes

Jerry 1 ad. lib. <1/yr 7 If owners go away Yes Yes e

Junior 2 1 Never 3-4 7 No Yes Karma f 2 2 Fortnightly/monthly 3 0 Yes Yes

Daily for weeks at Kelvin 1 2 7 0 Yes Yes a time

Lilyg,1, 2 2 2 Monthly 7 0 Yes Yes

154

Appendices Name No. cats No. fed Prey Frequency Nights out No. nights disappeared Wet Dry Lola 1 2-3 <1/yr 4 2 Yes Yes

Lulu 1 3 >1/yr 1-2 0 Yes Yes

c Max 2 ad. lib. Never 7 2 No Yes Merlot 1 2 Weekly 7 2 No Yes

Mia 1 ad. lib. <1/yr 7 Sometimes Yes Yes

Misty 1 2 <1/yr 7 0 Yes Yes

Mittens 1 3 Weekly 3 0 Yes Yes

f Yes Yes Munchkin 2 2 Fortnightly/Monthly 7 Sometimes (injured) Paddy 1 3-4 >1/yr 7 0 Yes Yes

Rocky 1 3 Weekly/Fortnightly 7 0 No Yes

Rosie b 2 2 Weekly 7 0 No Yes

Slinky 1 2 Monthly 7 0 No Yes Yes Yes Smokey 1 6 <1/yr 2 0 Socks 1 2 <1/yr 7 0 Yes Yes

Spider 2 2 ad. lib. Monthly 3 0 Yes No

Tom g, 3 2 2 Monthly 7 0 Yes Yes

Wolfie 2 4 2 <1/yr 3 0 Yes Yes a: Alvin and Felix from same household b: Bear and Rosie from same household

c: Charlie and Max from same household d: Debbie and Dipstick from same household e: Dora and Junior from same household f: Karma and Munchkin from same household

g: Tom and Lily same household 1: Owner supplied collar

155 2: Multiple cats in household were not tracked 3: Not tracked, lost collar

Appendices Table 3: Home range size (OREP, 95% MCP, 100% MCP), maximum (max.) and average (av.) Euclidean distance (ED) moved from home and diurnal (D) 100% MCP and OREP and nocturnal (N) 100% MCP and OREP home range sizes of companion cats in the Te Anau Basin in relation to sex and property type. Mean values are presented (±2 standard errors) in bold for total and female (F) and male (M) groups separately. Each

value is reported to two decimal places. Table continued over page.

Property 95% MCP 100% OREP Max. ED Av. ED D 100% N 100% D OREP N OREP Name Sex type (ha) MCP (ha) (ha) (m) (m) MCP (ha) MCP (ha) (ha) (ha)

Debbie F Rural 1.25 2.89 0.31 130.19 27.88 2.09 1.45 0.3 0.1

Karma F Rural 0.46 2.94 0.17 208.35 18.81 1.36 2.59 0.13 0.14

Merlot F Rural 55.68 61.16 1.69 756.40 160.34 5.36 61.16 0.32 2.39

Mia F Rural 0.75 2.33 0.21 148.85 22.27 1.11 2.32 0.13 0.12

Rosie F Rural 3.01 6.89 0.44 292.89 36.65 6.89 1.28 0.4 0.15

Dora F Urban 0.54 1.68 0.28 117.86 28.54 1.25 1.62 0.18 0.18

Houdini F Urban 0.57 2.20 0.24 143.92 21.78 1.76 0.97 0.24 0.1 Lily F Urban 5.6 7.72 0.63 293.27 52.91 4.6 7.16 0.43 0.46

Lola F Urban 0.69 2.84 0.22 194.85 23.43 1.28 2.18 0.15 0.19 Lulu F Urban 0.36 0.63 0.09 71.40 16.39 0.57 0.33 0.08 0.04

Mittens F Urban 2.4 5.81 0.64 242.00 47.15 3.56 4.43 0.59 0.35 Mean 6.48 8.83 0.45 236.36 41.47 2.71 7.77 0.27 0.38

(n = 11) ± 4.94 ± 5.28 ± 0.14 ± 56.26 ± 12.39 ± 0.31 ± 2.68 ± 0.02 ± 0.10 Alvin M Rural 28.58 70.22 3.59 1322.85 214.66 32.94 69.07 1.26 3.91

Austin M Rural 2.66 18.52 0.62 494.70 42.91 2.73 18.29 0.3 0.52

Bear M Rural 1 3.18 0.26 155.09 26.61 2.9 1.11 0.27 0.13

Dipstick M Rural 4.16 12.47 0.77 343.41 49.59 6.94 11.99 0.53 0.5

Felix M Rural 76.65 84.90 2.67 2445.41 653.53 75.77 68.89 1.7 3.87

156

Property 95% MCP 100% OREP Max. ED Av. ED D 100% N 100% D OREP N OREP Appendices Name Sex type (ha) MCP (ha) (ha) (m) (m) MCP (ha) MCP (ha) (ha) (ha) Jerry M Rural 36.77 48.10 2.53 764.07 163.90 5.55 48.1 0.23 2.47

Kelvin M Rural 66.37 109.53 4.30 985.87 324.61 40.59 105.07 0.86 6.46

Munchkin M Rural 4.36 25.90 0.60 639.42 40.54 15.83 10.96 0.55 0.35 Rocky M Rural 1.57 6.53 0.40 389.93 39.04 2.25 5.12 0.27 0.31

Benjie M Urban 5.08 14.65 1.02 406.37 46.84 7.55 13.4 0.46 1.22 Charlie M Urban 1.01 8.49 0.34 342.54 56.61 1.66 8.49 0.22 0.38

Dexter M Urban 2.43 7.31 0.60 275.02 47.73 7.31 1.53 0.62 0.17 Junior M Urban 1.38 13.91 0.59 611.94 45.16 2.24 11.71 0.23 0.55 Max M Urban 4.08 7.62 1.18 309.14 113.31 7.16 5.3 0.98 0.52

Misty M Urban 2.04 10.53 0.56 323.54 41.77 1.32 10.53 0.22 0.48

Paddy M Urban 0.96 4.18 0.31 228.50 29.44 1.51 3.65 0.22 0.19

Slinky M Urban 1.58 4.22 0.25 198.19 27.08 1.68 4.2 0.19 0.44

Smokey M Urban 1.29 7.03 0.47 286.88 36.85 2.02 6.88 0.28 0.51

Socks M Urban 1.77 3.12 0.34 202.47 36.20 2.75 2.79 0.29 0.23

Spider M Urban 0.67 2.23 0.22 157.38 23.25 0.82 2.09 0.11 0.21

Wolfie M Urban 1.11 2.10 0.33 151.15 41.66 1.63 1.54 0.26 0.16

Mean 11.69 22.13 1.05 525.42 100.06 10.63 19.56 0.48 1.12

(n = 21) ± 4.80 ± 6.58 ± 0.26 ± 115.79 ± 32.15 ± 7.95 ± 6.19 ± 0.18 ± 0.73

Overall 9.90 17.56 0.84 426.06 79.92 7.91 15.51 0.41 0.87

Mean ± 3.56 ± 9.53 ± 0.36 ± 81.46 ± 21.90 ± 5.42 ± 8.58 ± 0.13 ± 0.50

157

Appendices

Appendix 8: Resource Selection model outputs

Table 1: Second-order resource selection: Complete, for rural-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 10037 0.959 0.961 Intercept 2.640 0.202 13.08 Built -22.341 0.475 -47.05 Cover -18.328 0.938 -19.54 Wetland -0.796 0.047 -16.99 Age -0.096 0.124 -0.78 Sex (Males) 0.382 0.247 1.55 Model 2 10597 0.961 0.963 Intercept 2.240 0.180 12.42 Built -25.992 0.468 -55.58 Grassland 15.626 1.890 8.27 Wetland -0.728 0.046 -15.94 Age -0.105 0.110 -0.95 Sex (Males) 0.342 0.219 1.57 Model 3 14249 0.895 0.901 Intercept 2.089 0.212 9.83 Sealed -17.835 0.440 -40.55 Cover -28.093 0.828 -33.93 Wetland -0.899 0.037 -24.14 Age -0.183 0.133 -1.37 Sex (Males) 0.291 0.266 1.09 Model 4 16629 0.891 0.893 Intercept 0.979 0.106 9.24 Sealed -22.254 0.460 -48.42 Grassland 7.837 0.906 8.65 Wetland -0.512 0.033 -15.42 Age -0.096 0.062 -1.56 Sex (Males) 0.177 0.123 1.44

158

Appendices

Table 2: Second-order resource selection: Complete, for urban-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 30089 0.772 0.780 Intercept -0.596 0.152 -3.92 Built -82.449 2.391 -34.48 Sealed 1.582 0.868 1.82 Cover -1.870 0.200 -9.37 Age 0.033 0.021 1.53 Sex (Males) 0.051 0.184 0.28 Model 2 31718 0.163 0.207 Intercept -1.357 0.176 -7.71 Grassland 14.663 0.340 43.15 Sealed -7.993 0.690 -11.58 Cover -4.973 0.243 -20.47 Age 0.043 0.025 1.72 Sex (Males) -0.013 0.215 -0.06

Table 3: Third-order resource selection: Complete, for rural-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 23510 0.658 0.699 Intercept 0.209 0.303 0.69 Built -22.428 0.453 -49.47 Grassland -18.104 0.988 -18.33 Cover -10.724 0.689 -15.56 Age 0.043 0.042 1.01 Sex (Males) 0.208 0.376 0.55 Model 2 26259 0.452 0.491 Intercept -0.324 0.225 -1.44 Grassland -12.766 0.900 -14.19 Sealed -13.906 0.420 -33.10 Cover -15.710 0.638 -24.61 Age 0.044 0.031 1.41 Sex (Males) 0.175 0.280 0.62

159

Appendices

Table 4: Third-order resource selection: Complete, for urban-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 20671 0.350 0.363 Intercept -0.879 0.123 -7.137 Built -59.070 2.413 -24.476 Sealed -2.169 1.067 -2.032 Cover -3.435 0.379 -9.063 Age 0.007 0.021 0.364 Sex (Males) 0.248 0.161 1.542 Model 2 21670 0.036 0.044 Intercept -1.375 0.082 -16.777 Grassland 8.154 0.668 12.200 Sealed -9.384 0.960 -9.774 Cover -4.018 0.510 -7.872 Age 0.007 0.013 0.504 Sex (Males) 0.143 0.104 1.373

160

Appendices

Table 5: Second-order resource selection: Outside, for rural-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 12158 0.883 0.890 Intercept 2.284 0.219 10.42 Built -11.525 0.264 -43.60 Cover -24.334 0.859 -28.31 Wetland -0.778 0.039 -19.91 Age -0.036 0.030 -1.21 Sex (Males) 0.173 0.273 0.63 Model 2 13627 0.883 0.888 Intercept 1.668 0.174 9.60 Built -14.500 0.277 -52.32 Grassland 9.095 1.164 7.81 Wetland -0.609 0.037 -16.68 Age -0.027 0.023 -1.17 Sex (Males) 0.102 0.213 0.48 Model 3 13496 0.881 0.887 Intercept 1.799 0.204 8.81 Cover -27.095 0.831 -32.61 Sealed -16.294 0.425 -38.36 Wetland -0.847 0.037 -22.78 Age -0.035 0.028 -1.27 Sex (Males) 0.300 0.255 1.18 Model 4 15694 0.874 0.876 Intercept 0.714 0.107 6.67 Sealed -20.261 0.446 -45.39 Grassland 6.629 0.903 7.34 Wetland -0.476 0.033 -14.55 Age -0.018 0.014 -1.29 Sex (Males) 0.200 0.124 1.61

161

Appendices

Table 6: Second-order resource selection: Outside, for urban-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 28565 0.738 0.746 Intercept -0.741 0.142 -5.21 Built -74.102 2.207 -33.57 Cover -1.653 0.203 -8.13 Sealed 0.840 0.888 0.95 Age 0.027 0.020 1.33 Sex (Males) 0.071 0.172 0.42 Model 2 29750 0.164 0.200 Intercept -1.472 0.161 -9.12 Grassland 14.481 0.349 41.51 Cover -4.955 0.253 -19.61 Sealed -8.087 0.719 -11.25 Age 0.036 0.023 1.56 Sex (Males) 0.015 0.196 0.08

Table 7: Third-order resource selection: Outside, for rural-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 25434 0.290 0.348 Intercept -0.663 0.243 -2.728 Cover -17.927 0.635 -28.225 Grassland -15.070 0.909 -16.58 Built -3.360 0.170 -19.771 Age 0.023 0.034 0.689 Sex (Males) 0.186 0.302 0.618 Model 2 24247 0.407 0.452 Intercept -0.553 0.235 -2.351 Cover -15.404 0.644 -23.928 Grassland -12.444 0.923 -13.482 Sealed -12.159 0.405 -29.993 Age 0.045 0.033 1.388 Sex (Males) 0.214 0.292 0.733

162

Appendices

Table 8: Third-order resource selection: Outside, for urban-living companion cats in the Te Anau Basin. Table of models displaying effect size (β), standard error (S.E.), z-value, Akaike’s Information Criterion corrected for small sample sizes (AICc), marginal (R2m) and conditional (R2c) R squared values. Values are reported to three decimal places where applicable.

β S.E. z-value AICc R2m R2c Model 1 19481 0.299 0.312 Intercept -1.010 0.118 -8.546 Built -52.237 2.311 -22.606 Sealed -2.282 1.091 -2.092 Cover -3.250 0.390 -8.342 Age 0.011 0.020 0.546 Sex (Males) 0.262 0.154 1.704 Model 2 20290 0.034 0.040 Intercept -1.465 0.073 -20.181 Sealed -9.227 0.975 -9.46 Grassland 7.771 0.687 11.308 Cover -3.978 0.525 -7.576 Age 0.007 0.012 0.64 Sex (Males) 0.156 0.090 1.731

163

Appendix 9: Habitat maps Appendices

Figure 1: Examples of 100% MCP home ranges overlain on habitat buffers for two

rural and 18 urban-living companion cats in the Te Anau Basin. Note: for analyses,

Building and Urban features were combined into the Built feature.

164

Appendices

Figure 2: Examples of filtered GPS locations within the home range overlain on the habitat map for two rural- and 18 urban-living companion cats in the Te Anau Basin. Note: Different coloured locations represent different cats, for analyses, Building and Urban features were combined into the Built feature.

165

Appendices

Appendix 10: Proportions for time spent in habitats

Table 1: Average proportions for each habitat feature at the second-order of selection for rural-living companion cats in the Te Anau Basin. Available resource units were randomly generated throughout each cat’s buffer and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s buffer. Built is a combination of Urban and Building. All values reported to three decimal places where applicable. ‘Grass’ refers to Grassland features.

Average Proportion Grass Sealed Cover Wetland Built Urban Building Area 0.712 0.026 0.221 0.024 0.017 0.014 0.004 Available 0.737 0.025 0.206 0.025 0.017 0.014 0.003 Used 0.444 0.084 0.295 0 0.177 0.024 0.153

Table 2: Average proportions for each habitat feature at the second-order of selection for urban-living companion cats in the Te Anau Basin. Available resource units were randomly generated throughout each cat’s buffer and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s buffer. Built is a combination of Urban and Building. All values reported to three decimal places. ‘Grass’ refers to Grassland features.

Average Proportion Grass Sealed Cover Wetland Built Urban Building Area 0.311 0.169 0.066 0 0.454 0.422 0.031 Available 0.309 0.172 0.067 0 0.452 0.421 0.031 Used 0.071 0.101 0.016 0 0.812 0.795 0.017

Table 3: Average proportions for each habitat feature at the third-order of selection for rural-living cats in the Te Anau Basin. Available resource units were randomly generated throughout each cat’s 100% MCP and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s 100% MCP. Built is a combination of Urban and Building. All values reported to three decimal places. ‘Grass’ refers to Grassland features.

Average Proportion Grass Sealed Cover Wetland Built Urban Building Area 0.612 0.060 0.269 0 0.058 0.045 0.013 Available 0.610 0.059 0.273 0 0.058 0.045 0.013 Used 0.444 0.084 0.295 0 0.187 0.024 0.153

166

Appendices

Table 4: Average proportions for each habitat feature at the third-order of selection by urban-living companion cats in the Te Anau Basin. Available resource units were randomly generated throughout each cat’s 100% MCP and used resource units were the filtered GPS locations obtained from each cat. Area was calculated by finding the proportion of each habitat feature (excluding water) within each cat’s 100% MCP. Built is a combination of Urban and Building. All values reported to three. ‘Grass’ refers to Grassland features.

Average Proportion Grass Sealed Cover Wetland Built Urban Building Area 0.188 0.168 0.050 0 0.594 0.576 0.018 Available 0.187 0.169 0.051 0 0.594 0.578 0.016 Used 0.071 0.101 0.016 0 0.812 0.795 0.017

167

Appendix 11: Golf course

Appendices GPS locations of rural-living companion cats tracked in the Te Anau Basin in relation to the Te Anau golf course.

168

Appendices

Appendix 12: Trap captures

Table 1: Frequency different mammal species caught in the Ahuriri Valley in the South Island of New Zealand. Species Frequency

Ferret (Mustela furo) 49 Cat (Felis catus) 33 Hedgehog (Erinaceus europaeus) 17 Stoat (Mustela nivalis) 3 Weasel (Mustela ermine) 2 Total 104

169

Appendices

Appendix 13: Microsatellite loci Table 1: Details of the 11 polymorphic microsatellite loci and a sex specific marker (SRY) (Butler et al. 2002; Menotti-Raymond et al. 2005) used to determine feral cat population structure from the South Island of NZ. Loci tagged with the same dye label were run in a single PCR reaction (except for those marked with an *, which were run singularly) and then pooled for genotyping. Labels in brackets used previously within the literature (e.g. Butler et al. 2002).

Locus Primer pairs (5’  3’) Dye label Repeat motif

Compound FCA733 F: GATCCATCAATAGGTAAATGGATAAAGAAG 6-FAM [ATAG] (C08) R: ATGTGGCTGAGTAATATTCCACTGTCTCTC [ATAC] Compound FCA723 F: TGAAGGCTAAGGCACGATAGATAGTC 6-FAM [AAGG] (B04) R: GTGTCTTCCACCCAGGTGTCCTGCTTC [AAAG] FCA731 F: ATCCATCTGTCCATCCATCTATT 6-FAM Simple [ATCC] (G11) R: GGTCAGCATCTCCACTTGAGG

F: GTGTCTTGATCGGTAGGTAGGTAGATATAG Simple FCA441* VIC R :ATATGGCATAAGCCTTGAAGCAAA [ATAG] Compound FCA736* F: CCGAGCTCTGTTCTGGGTATGAA VIC [ACAT] (D09) R: TGTCTTTCTAGTTGGTCGGTCTGTCTATCTG [ATAG] Compound F: TGTGCTGGGTATGAAGCCTACTG F124* VIC [AGGA] R: GTGTCTTCCATGCCCATAAAGGCTCTGA [AGAA] FCA742 F: AAATTTCAATGTCTTGACAACGCATAAG NED Simple [CTTT] (C09) R: GCCAGGAACACCATGTTGGGCTA Complex F: TAAATCTGGTCCTCACGTTTTC F85* NED [TTTC] R: GCCTGAAAATGTATCCATCACTTCAGAT [TCTC]

FCA740 F: CCAAGGAGCTCTGTGATGCAAA NED Simple [TATC] (D06) R: GTTCCCACAGGTAAACATCAACCAA

F: GAGGAGCTTACTTAAGAGCATGCGTTC Complex FCA749* R:GTGTCTTAAACCTATATTCGGATTGTGCCTG PET [AGAT] (C12) CT [ACAT]

F: CCTATGTTGGGAGTAGAGATCACCT F53* PET Simple [AAGA] R: GTGTCTTGAGTGGCTGTGGCATTTCC

F: TGCGAACTTTGCACGGAGAG SRY* PET Gender ID R:GCGTTCATGGGTCGTTTGAG

170

Appendix 14: Genotyping scores Appendices Table 1: Genotyping scores of feral cat (Felis catus) samples used within Chapter Three. Note: Genotypes for each cat have been split into two tables. ID, site and sex (based on molecular analysis) information is repeated in each table, “0” values indicate missing data due to genotyping error. Colours correspond to fluorescent dye colour. N = 157. ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 A001 Ahuriri F 215 231 293 293 371 391 150 150 194 206 A002 Ahuriri F 203 203 293 309 391 391 146 154 198 198 A003 Ahuriri M 203 203 285 313 375 391 150 150 194 194 A004 Ahuriri F 203 219 285 293 399 399 150 154 0 0 A005 Ahuriri F 215 231 293 301 391 399 0 0 206 206 A006 Ahuriri F 215 219 293 301 0 0 150 154 194 206 A007 Ahuriri F 219 231 281 293 391 399 150 150 202 202 A009 Ahuriri F 219 219 309 309 399 399 150 154 194 194 A010 Ahuriri F 207 219 293 301 391 391 146 150 206 206

A011 Ahuriri F 231 235 285 293 391 391 146 150 202 202 A012 Ahuriri F 223 231 309 317 391 391 150 150 206 206 A013 Ahuriri F 215 235 317 321 399 399 150 150 202 206 A014 Ahuriri M 215 231 285 301 391 399 146 150 202 202 A015 Ahuriri M 215 215 281 293 371 399 150 154 194 206 A017 Ahuriri M 203 203 293 293 391 399 150 150 202 202 A018 Ahuriri M 203 219 277 281 391 391 146 150 0 0 A019 Ahuriri M 211 219 281 309 379 391 146 154 0 0 A020 Ahuriri F 215 215 293 309 391 399 150 150 198 206 A022 Ahuriri F 215 215 297 317 391 391 146 150 194 202 A025 Ahuriri F 203 203 317 317 399 399 146 154 206 206 A026 Ahuriri F 207 219 289 293 391 391 146 150 206 206 A027 Ahuriri F 203 215 297 317 391 399 146 154 202 214

171

Appendices ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 A028 Ahuriri F 211 215 281 297 371 391 150 150 202 202

A029 Ahuriri F 203 215 293 309 391 399 146 150 206 206 A030 Ahuriri F 231 235 293 317 391 391 150 154 194 194

A032 Ahuriri F 203 231 0 0 0 0 146 150 202 202

A033 Ahuriri M 219 231 289 293 399 419 150 154 198 206 M001 Macraes F 203 211 289 321 391 391 150 150 0 0 M002 Macraes M 203 211 281 313 391 399 150 150 202 202 M003 Macraes M 215 247 293 297 391 419 146 150 194 194

M005 Macraes F 203 207 281 285 371 391 0 0 198 198 M006 Macraes F 163 203 321 321 371 419 146 146 206 210 M008 Macraes F 203 235 281 309 391 399 150 150 194 202

M009 Macraes M 211 219 281 293 391 391 146 150 206 206 M011 Macraes F 163 215 289 297 391 419 150 154 210 210

M012 Macraes F 215 219 309 309 391 399 150 150 206 206 M013 Macraes F 211 247 293 313 0 0 150 150 206 210 M014 Macraes M 215 219 281 297 375 399 150 150 202 206

M016 Macraes F 163 219 305 309 371 419 146 150 210 210 M017 Macraes F 163 231 281 289 391 399 146 154 194 194 M018 Macraes F 211 215 289 289 395 399 150 150 206 206 M019 Macraes M 215 235 313 321 391 391 146 150 194 194

M020 Macraes F 203 219 281 305 391 391 146 150 206 206 M021 Macraes F 163 215 289 297 391 419 150 154 210 210 M023 Macraes F 203 235 297 309 0 0 146 150 194 210

M024 Macraes F 235 247 281 321 371 419 146 150 198 198 M025 Macraes M 215 219 281 309 371 399 0 0 206 206

M026 Macraes M 203 203 297 313 371 419 150 150 0 0 O001 Ohau M 211 211 289 313 399 419 150 154 198 206

172

Appendices ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 O002 Ohau M 215 223 281 297 391 399 150 158 206 206

O004 Ohau F 203 203 293 297 391 391 150 154 210 210

O005 Ohau F 219 223 293 293 399 399 146 154 206 206

O006 Ohau F 219 247 289 293 391 399 146 150 194 202 O007 Ohau F 215 219 285 285 391 391 150 150 202 210 O008 Ohau M 203 247 277 281 0 0 150 150 210 210 O009 Ohau M 223 231 277 309 395 399 146 150 194 194 O010 Ohau M 203 219 285 309 391 399 146 150 194 202

O011 Ohau F 247 247 301 317 391 399 150 154 194 194 O012 Ohau M 203 219 281 281 391 399 146 150 198 202 O013 Ohau F 203 219 281 301 399 399 146 150 198 198

O014 Ohau M 203 219 293 317 399 399 150 154 202 202 O015 Ohau F 203 223 293 305 371 391 150 150 202 202

O016 Ohau M 235 247 313 317 399 399 150 150 206 206 O017 Ohau F 219 247 301 309 391 399 150 150 194 206

O018 Ohau F 219 219 293 317 379 399 146 150 206 206 O019 Ohau F 219 219 281 289 391 399 150 150 194 194 O020 Ohau F 203 219 281 301 391 399 146 150 198 206 O022 Ohau M 211 215 289 297 399 399 150 150 202 202 O023 Ohau F 215 219 293 309 399 399 146 146 194 206

O024 Ohau M 203 215 289 293 399 419 150 154 206 214 O025 Ohau F 203 215 293 293 399 399 150 154 198 206 O026 Ohau F 219 247 293 297 371 371 146 150 210 210

O027 Ohau F 203 211 285 297 391 399 154 158 194 194 O028 Ohau F 203 215 289 297 371 399 146 150 206 206

O029 Ohau M 203 219 277 297 391 395 150 150 206 210 O030 Ohau F 231 247 285 309 391 391 146 150 194 194

173

Appendices ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 O031 Ohau M 203 247 281 309 399 399 154 154 206 206

O032 Ohau M 219 223 281 281 391 391 150 158 206 206

O033 Ohau F 231 247 309 309 391 391 150 158 206 206

O034 Ohau M 219 235 289 301 0 0 150 158 206 206 O037 Ohau M 219 247 309 309 391 391 146 150 194 210 O038 Ohau F 203 215 305 321 391 399 150 150 194 206 O039 Ohau F 223 223 281 317 391 399 150 158 202 202 O040 Ohau M 215 247 277 309 371 391 146 150 194 194

O041 Ohau F 203 223 293 317 391 391 146 150 206 214 O042 Ohau M 219 247 305 309 391 399 150 154 206 206 O043 Ohau F 203 219 277 277 391 419 158 158 202 202

O045 Ohau F 211 223 281 281 391 391 150 154 206 206 O046 Ohau F 219 247 281 293 371 419 154 154 0 0

O047 Ohau M 215 231 281 309 391 399 146 150 194 194 O048 Ohau M 219 235 285 293 391 399 146 158 206 206

O049 Ohau F 215 235 289 293 391 399 150 150 206 206 O050 Ohau F 219 219 277 301 399 419 0 0 198 198 O051 Ohau M 219 223 289 293 399 419 154 158 194 194 O052 Ohau M 231 247 293 309 391 391 150 154 194 206 O053 Ohau M 211 211 297 313 371 371 150 150 206 206

O054 Ohau M 203 203 281 281 391 399 146 146 206 214 O055 Ohau M 211 219 301 317 391 391 150 154 198 198 O056 Ohau F 211 215 297 309 391 419 0 0 194 214

O057 Ohau F 231 247 293 301 391 391 0 0 0 0 O058 Ohau F 207 219 289 293 391 399 146 150 194 194

O059 Ohau M 203 211 277 285 371 399 150 154 202 202 O060 Ohau M 211 235 281 309 399 419 146 158 194 198

174

Appendices ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 O061 Ohau F 215 247 297 309 399 419 146 154 194 194

O062 Ohau F 215 247 277 321 391 399 146 150 194 198

O063 Ohau M 203 215 285 293 0 0 154 154 202 202

O064 Ohau F 219 219 277 293 395 399 146 154 202 202 O066 Ohau M 159 219 289 297 399 399 150 150 0 0 O067 Ohau M 207 215 281 309 371 391 150 150 206 206 O068 Ohau M 219 247 289 297 399 399 154 158 198 198 O069 Ohau F 219 223 309 317 391 399 146 146 202 206

O070 Ohau F 215 215 293 309 399 399 146 158 202 206 T001 Tasman M 207 211 309 309 399 399 150 150 198 198 T002 Tasman M 215 219 277 289 371 399 150 154 206 206

T004 Tasman F 219 223 277 289 391 399 150 154 194 206 T006 Tasman F 203 219 281 289 371 391 150 158 194 194

T100 Tasman M 211 219 277 305 371 399 150 154 194 194 T101 Tasman F 203 247 309 313 371 391 146 146 194 206

T102 Tasman F 247 247 289 293 391 399 150 154 206 206 T105 Tasman F 203 203 277 317 371 399 150 158 202 206 T107 Tasman M 207 219 297 313 399 399 154 158 202 202 T108 Tasman M 203 219 293 309 399 399 150 150 206 206 T109 Tasman M 207 219 289 289 391 399 150 154 206 206

T111 Tasman F 219 219 289 321 399 399 150 154 194 194 T112 Tasman M 219 219 289 297 399 399 150 154 194 194 T113 Tasman F 219 219 281 321 391 399 154 154 194 194

T114 Tasman M 207 219 289 293 399 399 0 0 194 194 T115 Tasman M 219 247 293 293 391 399 146 150 206 206

T117 Tasman F 211 223 0 0 0 0 150 150 194 202 TG001 Tasman F 211 219 297 321 399 399 150 154 194 194

175

Appendices ID Site Sex FCA733 FCA723 FCA731 FCA441 FCA736 TG002 Tasman F 215 219 285 313 391 399 154 154 194 194

TG003 Tasman F 219 219 289 317 371 399 146 154 206 206 TG004 Tasman F 207 219 293 293 375 375 154 154 194 194

TG005 Tasman M 207 219 289 293 0 0 154 154 194 194

TG006 Tasman F 219 219 289 301 375 391 150 154 194 194 TG007 Tasman F 203 211 297 297 391 391 150 154 194 202

TG008 Tasman F 211 219 309 309 399 399 146 150 206 206 TG009 Tasman M 211 219 309 317 391 399 150 150 194 194

TG010 Tasman M 219 219 289 309 0 0 154 158 194 202 TG011 Tasman F 203 207 281 289 391 391 150 150 194 206 TG012 Tasman M 211 215 285 293 391 399 150 150 206 206 TG013 Tasman F 207 219 309 313 399 399 150 154 202 206 TG014 Tasman M 211 219 309 309 399 399 146 158 202 202

TG015 Tasman M 203 211 281 309 391 399 154 158 206 206 TL003 Tasman F 219 219 293 293 391 399 150 150 206 206 TL006 Tasman F 219 231 285 309 371 399 150 150 202 202

TL007 Tasman F 215 219 277 281 399 399 150 150 0 0 TL009 Tasman M 219 219 293 309 399 399 150 150 206 206

TL010 Tasman M 219 219 277 285 399 399 146 146 194 206 TL011 Tasman F 211 215 309 317 391 399 150 150 194 206

TL012 Tasman F 207 231 281 285 399 399 146 158 194 194 TL015 Tasman M 203 219 281 301 399 399 146 150 194 206 TL016 Tasman F 203 211 301 309 391 399 150 150 194 194 TR001 Tasman F 203 247 281 309 391 399 146 154 194 194 TR002 Tasman F 203 223 281 289 371 391 150 154 206 206

TR017 Tasman F 203 231 309 309 371 391 146 146 206 206 TR018 Tasman F 215 223 281 289 371 371 150 154 206 210

176

Table 2: Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749

A001 Ahuriri F 291 291 183 187 343 355 160 196 386 386 A002 Ahuriri F 311 319 179 199 343 343 164 180 386 390 A003 Ahuriri M 0 0 175 179 335 343 168 172 390 390 A004 Ahuriri F 307 311 175 187 335 343 172 172 386 394 A005 Ahuriri F 291 303 175 187 335 339 0 0 378 398

A006 Ahuriri F 291 307 183 191 343 355 160 160 386 390 A007 Ahuriri F 303 315 175 191 343 343 160 168 378 386 A009 Ahuriri F 0 0 175 175 343 355 156 184 378 390 A010 Ahuriri F 303 315 175 175 339 343 148 164 386 386 A011 Ahuriri F 303 307 175 191 335 343 160 160 378 386 A012 Ahuriri F 307 323 187 195 343 343 156 168 386 386

A013 Ahuriri F 311 327 175 195 343 355 160 164 386 390 A014 Ahuriri M 291 307 175 187 335 339 160 180 386 398

A015 Ahuriri M 291 323 179 183 343 343 160 168 386 386 A017 Ahuriri M 307 319 179 195 343 355 168 180 386 394 A018 Ahuriri M 323 323 175 187 343 355 164 168 382 390

A019 Ahuriri M 311 315 175 187 355 355 164 184 0 0 A020 Ahuriri F 299 327 0 0 0 0 160 172 386 386

A022 Ahuriri F 303 323 175 191 343 355 172 176 386 390 A025 Ahuriri F 291 323 167 175 343 343 148 196 386 390 A026 Ahuriri F 303 315 175 179 339 355 164 192 386 470

A027 Ahuriri F 315 319 159 175 335 355 168 176 386 402 A028 Ahuriri F 307 307 175 195 343 351 0 0 386 390

A029 Ahuriri F 311 319 175 179 343 343 180 192 390 390 A030 Ahuriri F 299 307 175 187 355 355 156 168 386 386 A032 Ahuriri F 303 311 175 191 343 355 168 176 386 390 A033 Ahuriri M 291 315 175 175 343 343 160 160 390 390

177

Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749 M001 Macraes F 319 319 167 195 343 347 160 176 386 390

M002 Macraes M 291 291 167 199 0 0 164 172 386 386

M003 Macraes M 291 327 187 187 339 343 164 168 378 386 M005 Macraes F 295 311 175 179 0 0 160 164 390 390

M006 Macraes F 327 327 167 195 343 347 168 172 0 0 M008 Macraes F 307 319 167 179 335 347 156 180 386 386 M009 Macraes M 311 315 175 179 335 355 164 184 362 390 M011 Macraes F 315 323 167 199 343 347 148 164 386 386

M012 Macraes F 303 315 175 175 343 355 160 184 362 378 M013 Macraes F 311 315 167 179 343 355 160 164 362 394 M014 Macraes M 307 311 167 187 343 343 160 168 390 390

M016 Macraes F 307 315 175 175 339 347 164 172 0 0 M017 Macraes F 303 327 175 175 343 347 156 156 386 394

M018 Macraes F 303 303 167 195 339 347 164 168 386 386 M019 Macraes M 295 307 179 195 339 343 164 168 390 394 M020 Macraes F 311 315 179 187 335 339 184 192 362 394

M021 Macraes F 315 323 167 199 343 347 148 164 386 386 M023 Macraes F 303 307 195 195 343 343 152 184 382 394 M024 Macraes F 291 319 179 195 347 351 168 172 386 386 M025 Macraes M 303 311 179 191 339 355 164 168 386 386

M026 Macraes M 0 0 187 187 343 343 172 172 386 386 O001 Ohau M 291 311 175 195 335 343 160 184 386 390 O002 Ohau M 315 331 171 171 339 343 160 160 386 386

O004 Ohau F 303 323 167 175 335 339 168 176 378 394 O005 Ohau F 307 335 179 199 343 343 164 164 386 390

O006 Ohau F 311 311 187 195 335 355 148 156 386 386 O007 Ohau F 331 331 159 199 335 347 164 172 378 378

178

Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749 O008 Ohau M 323 323 191 199 0 0 164 184 378 386

O009 Ohau M 291 311 175 191 335 359 160 164 386 390 O010 Ohau M 303 323 179 179 347 355 172 192 378 474

O011 Ohau F 307 327 171 175 347 355 160 172 378 378

O012 Ohau M 311 315 175 179 335 359 168 176 382 386 O013 Ohau F 307 311 179 179 347 359 172 176 382 390 O014 Ohau M 307 323 187 199 343 347 0 0 390 474 O015 Ohau F 311 311 175 179 335 355 156 160 390 398

O016 Ohau M 307 315 175 175 343 347 164 168 378 390 O017 Ohau F 315 327 171 175 347 355 160 172 386 386 O018 Ohau F 307 307 167 187 335 339 176 192 386 390

O019 Ohau F 311 323 175 187 343 343 168 172 378 386 O020 Ohau F 307 315 167 179 335 347 176 180 378 390

O022 Ohau M 311 327 175 183 343 343 164 172 390 390 O023 Ohau F 323 331 187 187 335 343 156 160 378 390 O024 Ohau M 307 311 175 179 343 355 164 180 386 386

O025 Ohau F 291 331 175 195 355 355 160 164 378 386 O026 Ohau F 307 311 175 199 343 355 160 168 390 398 O027 Ohau F 303 311 167 175 343 343 148 164 378 390 O028 Ohau F 307 315 187 195 343 355 164 176 382 394

O029 Ohau M 311 311 187 199 355 355 168 184 386 390 O030 Ohau F 303 323 175 187 343 343 160 164 386 394 O031 Ohau M 0 0 175 179 343 347 0 0 386 386

O032 Ohau M 311 315 167 175 355 355 160 180 0 0 O033 Ohau F 303 311 199 199 343 343 160 164 386 398

O034 Ohau M 303 307 175 175 335 355 148 192 0 0 O037 Ohau M 307 319 159 175 343 343 160 172 386 390

179

Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749 O038 Ohau F 0 0 171 187 343 343 152 176 386 394

O039 Ohau F 303 307 167 179 335 355 156 168 378 390 O040 Ohau M 311 319 175 187 343 359 160 184 378 390

O041 Ohau F 303 331 167 175 335 335 172 180 378 390

O042 Ohau M 311 315 175 199 343 355 164 176 386 398 O043 Ohau F 307 311 175 179 0 0 180 180 378 386 O045 Ohau F 0 0 179 199 0 0 148 148 378 390 O046 Ohau F 311 323 175 195 343 359 160 164 378 390

O047 Ohau M 311 323 175 187 343 359 184 184 378 394 O048 Ohau M 303 323 175 199 343 347 164 168 386 390 O049 Ohau F 307 307 175 183 343 355 148 164 386 386

O050 Ohau F 311 311 175 175 355 355 168 176 390 398 O051 Ohau M 307 307 175 199 343 343 160 168 386 390

O052 Ohau M 307 311 167 175 343 347 156 168 378 386 O053 Ohau M 0 0 167 175 0 0 160 160 390 390 O054 Ohau M 315 331 175 179 335 355 156 168 378 386

O055 Ohau M 307 327 175 203 355 359 160 164 378 386 O056 Ohau F 307 311 187 191 343 355 156 164 386 390 O057 Ohau F 311 311 171 179 343 343 164 164 386 386 O058 Ohau F 295 331 191 195 343 355 148 176 386 390

O059 Ohau M 303 311 175 199 339 355 160 164 378 386 O060 Ohau M 307 311 179 179 355 355 160 204 386 390 O061 Ohau F 307 311 175 187 339 343 168 172 386 390

O062 Ohau F 327 327 179 191 343 355 164 180 386 386 O063 Ohau M 303 307 179 179 343 359 148 164 390 394

O064 Ohau F 307 311 175 199 343 347 168 176 386 390 O066 Ohau M 311 311 175 191 343 355 160 172 378 386

180

Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749 O067 Ohau M 303 319 0 0 0 0 148 192 390 390

O068 Ohau M 311 311 175 175 343 355 160 168 386 390 O069 Ohau F 303 331 175 191 339 343 148 172 386 390

O070 Ohau F 323 327 187 187 335 343 160 172 378 390

T001 Tasman M 303 319 175 179 335 355 164 176 378 386 T002 Tasman M 319 323 159 167 343 347 172 184 386 386 T004 Tasman F 303 323 159 167 347 347 148 168 378 378 T006 Tasman F 303 303 187 195 343 343 168 184 0 0

T100 Tasman M 303 307 179 187 343 343 0 0 378 386 T101 Tasman F 303 303 175 195 335 343 148 164 378 386 T102 Tasman F 291 323 175 187 343 347 164 184 386 390

T105 Tasman F 311 315 0 0 0 0 156 160 386 394 T107 Tasman M 307 323 187 187 335 335 172 196 378 378

T108 Tasman M 319 319 175 175 347 355 160 168 386 390 T109 Tasman M 0 0 175 187 343 343 148 164 378 378 T111 Tasman F 311 315 187 195 335 335 172 180 378 386

T112 Tasman M 303 315 187 191 335 335 160 164 378 394 T113 Tasman F 303 315 0 0 0 0 168 180 378 386 T114 Tasman M 311 323 179 191 343 343 164 176 390 398 T115 Tasman M 311 311 167 191 335 343 172 176 378 378

T117 Tasman F 307 311 175 175 335 343 164 180 378 378 TG001 Tasman F 303 323 187 195 335 343 160 168 378 386 TG002 Tasman F 303 323 175 191 343 347 148 164 0 0

TG003 Tasman F 291 303 187 191 343 347 172 176 378 386 TG004 Tasman F 303 335 191 195 343 347 172 176 390 398

TG005 Tasman M 303 335 175 187 343 347 148 172 378 398

181

Appendices ID Site Sex F124 FCA742 FCA740 F53 FCA749 TG006 Tasman F 303 323 191 191 343 347 164 180 386 402

TG007 Tasman F 315 323 175 191 343 343 168 176 386 398 TG008 Tasman F 303 319 175 191 335 347 168 196 378 378 TG009 Tasman M 303 319 0 0 0 0 168 196 378 378 TG010 Tasman M 291 303 175 203 335 347 168 196 378 378

TG011 Tasman F 303 307 159 191 343 347 176 184 378 378 TG012 Tasman M 307 311 175 187 335 335 0 0 378 390 TG013 Tasman F 307 319 175 191 335 347 148 176 378 386 TG014 Tasman M 307 319 0 0 0 0 168 196 378 378 TG015 Tasman M 307 323 179 187 343 343 164 172 386 390

TL003 Tasman F 307 307 179 179 335 343 160 168 386 390 TL006 Tasman F 307 307 175 187 343 343 0 0 378 386 TL007 Tasman F 0 0 179 179 335 343 160 164 390 390

TL009 Tasman M 311 315 179 187 343 347 164 176 386 386 TL010 Tasman M 311 311 175 179 343 347 168 168 378 382 TL011 Tasman F 303 311 175 175 343 347 164 176 386 386 TL012 Tasman F 291 323 179 199 335 343 164 172 0 0

TL015 Tasman M 311 311 175 179 343 347 164 168 0 0 TL016 Tasman F 307 311 179 179 343 347 160 164 378 386 TR001 Tasman F 303 303 159 191 343 347 164 184 378 386 TR002 Tasman F 303 307 159 191 343 347 0 0 378 378 TR017 Tasman F 303 323 175 179 0 0 160 160 386 386

TR018 Tasman F 311 323 0 0 0 0 160 160 378 386

182