THE EFFECTIVENESS OF CORRIDORS IN

FACILITATING CONNECTIVITY: ASSESSMENT OF A

MODEL SYSTEM FROM THE AUSTRALIAN WET TROPICS.

Kerrilee Horskins BSc. (Hons)

School of Natural Resource Sciences Queensland University of Technology Brisbane, Australia

This dissertation is submitted as a requirement of the Doctor of Philosophy Degree 2005

i

ABSTRACT

Wildlife corridors have become a widely adopted management strategy for the conservation of species in fragmented . Fragmentation reduces the size of patches and increases the isolation of the within them, potentially resulting in extinction due to stochastic processes. The provision of a corridor between habitat patches is believed to increase the level of connectivity through the integration of populations into a single demographic unit, thus increasing the probability of survival. This assumption remains largely untested due to both a lack of investigation, and design limitations in some of the few studies performed. Connectivity is often assumed to occur simply from the presence of individuals within the corridor.

Design criteria essential for the rigorous assessment of connectivity were identified and a landscape meeting these criteria selected. The vegetation within the corridor was found to be comparable in both structure and species composition to that of the patches that it connected. Two target species (Melomys cervinipes and Uromys caudimaculatus) were shown to occur along the corridor but not within the surrounding matrix. The combination of these factors indicated that the corridor was suitable for use as a model system and ensured that any subsequent results truly reflected the capacity of the corridor to function in the desired manner.

The structure was similar within the corridor and the connected patches for both species. Weights of individuals, sex ratios and the percentage of juveniles were consistent between the two system components, suggesting that the corridor contained breeding populations. Connectivity was therefore possible via generational for both species, while long distance movement events for U. caudimaculatus also indicated that direct movement between habitat patches may be possible for this larger species.

Despite all ecological parameters indicating that connectivity was likely, genetic markers (mtDNA and nDNA) revealed significant population differentiation between the connected patches for both species. Populations linked by the corridor and those in isolated habitats were found to show the same level of genetic differentiation. Sampling at a finer spatial scale within connected patches and a continuous control habitat showed that population differentiation was common for M. ii cervinipes. Given the continuity of suitable vegetation, and the presence of individuals of breeding age along the corridor system, this was attributed to social structuring. U. caudimaculatus populations also showed evidence of genetic differentiation within a connected patch and along the corridor, despite panmixia within the continuous habitat.

Having investigated a model system, the data from this study has implications for other studies and for landscape managers. Firstly, the advantages of using an integrated ecological and genetic approach have been demonstrated. While genetic data determined the level of connectivity, the ecological data provided an understanding of the processes operating within the system. Secondly, the level of scale at which wildlife corridor studies are conducted may need addressing. Most studies currently treat a fragmented landscape in a binary manner and consider the connected patch to be the finest “grain”. However, the processes responsible for the lack of connectivity were found to operate at the much finer within-patch scale. Finally, this study clearly indicated that not all wildlife corridors will provide connectivity between the connected populations and that connectivity cannot be inferred from the presence of individuals within the corridor. Given that social behaviour such as territorial defence and philopatry are common in many species, especially small , a lack of connectivity via a wildlife corridor may be more common than currently assumed. The successful use of wildlife corridors as a management strategy, and the accurate assessment of their effectiveness therefore requires careful consideration of not only structural attributes of the corridor, but also behavioural, demographic and genetic parameters of the target species. iii

TABLE OF CONTENTS

Abstract ...... i Table of Contents ...... iii List of Figures...... vi List of Tables ...... viii List of Appendices ...... x Glossary ...... xii Statement of Original Authorship ...... xiii Acknowledgements ...... xiv

1. PROJECT OVERVIEW ...... 1 1.1 The Problem ...... 1 1.2 Current conservation approaches...... 2 1.3 Selection of the appropriate approach...... 3 1.4 The current status of wildlife corridor assessment...... 5 1.5 Scope and structure of thesis ...... 6

2. WILDLIFE CORRIDORS AS A CONSERVATION STRATEGY ...... 8 2.1 What is a wildlife corridor?...... 8 2.2 Potential benefits of wildlife corridors ...... 10 2.3 Past assessment of wildlife corridor effectiveness ...... 14 2.3.1 Field-based ecological studies...... 15 2.3.2 Experimental studies...... 18 2.3.3 Genetic studies ...... 20 2.3.4 Computer modelling...... 22 2.3.5 Summary...... 23 2.4 Project aims ...... 24

3. THE MODEL SYSTEM: DESCRIPTION AND EVALUATION OF ITS SUITABILITY FOR

THE ASSESSMENT OF CONNECTIVITY...... 25 3.1 Introduction...... 25 3.2 The model system ...... 31 3.2.1 The study site...... 32 3.2.2 The study species...... 35 3.2.3 The conservation significance of the study region ...... 37 iv

3.2.4 Pleistocene refugia ...... 38 3.3 Methods ...... 39 3.3.1 Vegetation description ...... 39 3.3.2 Vegetation structure and composition...... 40 3.3.3 Trapping...... 42 3.3.4 Analyses ...... 45 3.4 Results...... 46 3.4.1 Qualitative overview of vegetation within the corridor system ...... 46 3.4.2 Resource availability between habitat types ...... 50 3.4.3 Differential use of the system components by small mammals ...... 52 3.4.4 Differential habitat use by M. cervinipes and U. caudimaculatus...... 55 3.5 Discussion ...... 57

4. POPULATION STRUCTURE AND MOVEMENT OF MELOMYS CERVINIPES AND

UROMYS CAUDIMACULATUS WITHIN THE CORRIDOR SYSTEM...... 61 4.1 Introduction...... 61 4.2 Methods ...... 64 4.2.1 Trapping regime...... 64 4.3 Results...... 65 4.3.1 Trap intensity ...... 65 4.3.2 Demographic parameters ...... 67 4.3.3 Recapture rates ...... 71 4.3.4 Movement ...... 74 4.4 Discussion ...... 75

5. DOES THE WILDLIFE CORRIDOR FACILITATE GENE FLOW? ...... 80 5.1 Introduction...... 80 5.2 Methods ...... 83 5.2.1 Sample collection...... 83 5.2.2 Mitochondrial DNA...... 84 5.2.3 Nuclear DNA...... 86 5.2.4 Statistical analyses ...... 87 5.3 Results...... 91 5.3.1 Tests for neutrality and linkage disequilibrium ...... 91 5.3.2 Haplotype distribution and diversity ...... 93 5.3.3 Allele distribution and heterozygosity...... 95 v

5.3.4 Population differentiation – landscape scale...... 97 5.3.5 Population differentiation – within patch scale...... 99 5.4 Discussion ...... 101 5.4.1 The effectiveness of corridors in maintaining ...... 101 5.4.2 Gene flow among M. cervinipes populations ...... 102 5.4.3 Gene flow among U. caudimaculatus populations...... 104 5.4.4 The effect of social structuring on wildlife corridor effectiveness ....106 5.4.5 The significance of appropriate sampling strategies...... 107 5.4.6 Summary...... 109

6. GENERAL DISCUSSION ...... 110 6.1 The benefits of an integrated approach ...... 110 6.2 The role of scale in wildlife corridor assessment ...... 112 6.2.1 Environmental scale...... 112 6.2.2 The scale of organism response...... 113 6.2.3 Observational scale ...... 114 6.3 The suitability of wildlife corridors for other small species ...... 115 6.4 Applications for management ...... 116 6.5 Conclusion...... 121

APPENDICES ...... 122

REFERENCES ...... 146

vi

LIST OF FIGURES

Figure 3.1 Diagrammatic representation of the essential design criteria necessary for the rigorous assessment of wildlife corridor effectiveness as identified from the literature...... 27

Figure 3.2 Stylized representation of the essential design criteria...... 27

Figure 3.3 Diagrammatic representation of additional design criteria identified as essential for the rigorous assessment of wildlife corridor effectiveness with respect to investigating connectivity...... 30

Figure 3.4 Location of the study system...... 34

Figure 3.5 The location of connected, corridor and matrix sites within the corridor system...... 35

Figure 3.6 Representation of the trapping grid at connected sites...... 44

Figure 3.7 Schematic map of vegetation types within the corridor system and the location of trapping sites along the corridor...... 48

Figure 3.8 Spatial representation of variation in vegetation structure between system components based on MDS...... 50

Figure 3.9 Species diversity per system component as measured by the Renyi index...... 51

Figure 3.10 Spatial representation of variation in resources between habitat types based on MDS...... 52

Figure 3.11 Captures of mammalian species expressed as the proportion of individuals per species trapped within each habitat component...... 54

Figure 3.12 Number of individuals (per 1000 trap nights) per site...... 56

Figure 4.1 Trap availability per night expressed as the percentage of traps remaining unutilized or unsprung for U. caudimaculatus and M. cervinipes...... 66

Figure 4.2 Percent of juveniles per trip per habitat type...... 70 vii

Figure 4.3 Mean weight (± s.e.) of individuals per sex and habitat type...... 71

Figure 5.1 Location of sampling sites for fine scale spatial studies within A) connected patches and B) continuous forest...... 84

Figure 6.1 Flowcharts outlining aspects of species behaviour and their potential affect on the ability of wildlife corridors to facilitate recolonization and gene flow...... 118

viii

LIST OF TABLES

Table 1.1 Examples of within Australia...... 2

Table 1.2 Current management strategies for the conservation of populations within fragmented habitats...... 4

Table 2.1 Examples of the different types of corridors used to address conservation issues...... 9

Table 3.1 Vegetation sampling allocation at each site for GC = ground cover; VC = vertical cover; PFC = percent foliage cover; Dens = stem density, BA = basal area; Spp. = species composition...... 42

Table 3.2 Trapping intensity shown as the number of trap nights (cage, elliot) per site per trip...... 44

Table 3.3 Variation in structural attributes of the vegetation within the corridor system...... 49

Table 3.4 Percent trap success for unique individuals per site for each trapping session...... 55

Table 4.1 The proportion of new individuals caught on the final night of trapping ...... 66

Table 4.2 Summary of demographic characteristics for the sampled population at each site per trapping session...... 68

Table 4.3 Reproductive status of mature females per trip...... 69

Table 4.4 The number of individuals marked per trip and the recapture rates on subsequent trips...... 72

Table 4.5 Recapture rates according to species, habitat, age and sex...... 73

Table 4.6 Average distances moved by individuals within the trapping grid of Connected R...... 75

Table 5.1 Tajima's D test for neutrality for the control region of the mtDNA...... 92 ix

Table 5.2 Linkage disequilibrium for each locus pair...... 93

Table 5.3 Haplotype number (H) and diversity (h) for each population...... 94

Table 5.4 Descriptive statistics of variation for nuclear markers for populations averaged across all loci...... 96

Table 5.5 Summary table of the pairwise differentiation of M. cervinipes populations at a landscape scale...... 98

Table 5.6 Summary table of the pairwise differentiation of U. caudimaculatus populations at a landscape scale...... 98

Table 5.7 Isolation by distance within the corridor system for M. cervinipes and U. caudimaculatus...... 99

Table 5.8 Pairwise differentiation of populations for M. cervinipes at a within patch scale...... 100

Table 5.9 Differences in the results of population structuring depending upon the location of sampling within Connected R...... 100

x

LIST OF APPENDICES

Appendix 1. Species identified as present within the corridor system by the collection of fruits and nuts...... 122

Appendix 2. Species abundances per habitat component of the corridor system ...... 123

Appendix 3. mtDNA procedures and protocols...... 124

Appendix 4. Nuclear DNA procedures and protocols...... 127

Appendix 5. mtDNA haplotype frequencies for M. cervinipes...... 130

Appendix 6. mtDNA haplotype frequencies for U. caudimaculatus...... 131

Appendix 7. mtDNA sequences: M. cervinipes...... 132

Appendix 8. mtDNA sequences: U. caudimaculatus...... 133

Appendix 9. The relative frequency of microsatellite alleles per population per locus for M. cervinipes...... 134

Appendix 10. The relative frequency of microsatellite alleles per population per locus for U. caudimaculatus ...... 137

Appendix 11. Hardy-Weinberg estimates for M. cervinipes ...... 139

Appendix 12. Hardy-Weinberg estimates for U. caudimaculatus...... 141

Appendix 13. Pairwise comparisons for all populations of M. cervinipes – mtDNA ...... 142

Appendix 14. Pairwise comparisons for all populations of U. caudimaculatus – mtDNA...... 143

xi

Appendix 15. Pairwise comparisons for all populations of M. cervinipes – nuclear DNA...... 144

Appendix 16. Pairwise comparisons for all populations of U. caudimaculatus - nuclear DNA ...... 145

xii

GLOSSARY

To ensure continuity in terminology, the following definitions will be used throughout this thesis: wildlife corridor – a permanent linear stretch of habitat, differing in structure and vegetation type to that of the matrix habitat, which links at least two habitat patches that would be otherwise isolated from each other but which have been connected in historical time connectedness - structural links between elements of the landscape; mappable structures (sensu Baudry and Merriam 1988). Also commonly referred to as . connectivity - a parameter of landscape function which measures the processes by which sub-populations of organisms are interconnected into a functional unit (sensu Merriam 1984). The degree of connectivity is often dependant upon the extent of connectedness between the sub-populations. matrix - the landscape element surrounding the habitat of primary interest. In highly disturbed regions the matrix will often be the dominant element within the landscape. corridor system – a collective term for the corridor, the connected patches and the surrounding matrix. xiii

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted for a degree or diploma at any other educational institution. To the best of my knowledge, this thesis contains no material previously published or written by another person except where due reference is made.

Kerrilee Horskins June 2005

xiv

ACKNOWLEDGEMENTS

This project has benefited from the generous help of many people. I would like to thank: The Cianelli, Ford, Laws, Kidd, Stocker, Schafer, Hapgood, Lorenson, Beattie and Wardell families for generously allowing fieldwork to be conducted on their property.

The numerous people who assisted with fieldwork, in particular David Elmouttie, Craig Streatfeild, Martine Adreaansen, Graeme Horskins and Adam Liedloff.

All members of the QUT population genetics lab, especially Natalie Baker, for advice on genetic procedures, Dave Hurwood for statistical advice and Vincent Chand and Craig Streatfeild for developing and optimizing the Uromys microsat primers respectively. Also, Martin Elphinstone (Southern Cross University) for providing the M. burtoni and U. hadrourus samples used as TGGE outgroups.

The staff at Queensland Parks and Wildlife Service (Lake Eacham and Townsville) for logistic support, especially Nigel Tucker for unselfishly sharing his wealth of knowledge and, along with David Elmouttie, identifying the fruits.

Adam Liedloff for assistance with Corel Draw and for proof reading this thesis.

My supervisors Dr John Wilson and Assoc. Prof. Peter Mather. Also Dr Ian Williamson for discussions on statistical matters.

My parents for their support, and in particular, for the interest they showed in my work. Also Glenn for help with computing and simply for caring.

Finally to Francine (Franki) Lacey: You are no longer with us but your words of encouragement have been an inspiration throughout.

Research was conducted with permission from the Queensland Environmental Protection Agency and QUT Animal Ethics. Funding was provided in the form of an Australian Post-Graduate Award, an Ethel Mary Read Research Grant from the Royal Zoological Society of New South Wales and a travel grant from the International Association of Landscape Ecology. - 1 -

1. PROJECT OVERVIEW

1.1 The Problem

Habitat fragmentation is a global issue that involves the conversion of a continuous stand of habitat into a landscape mosaic consisting of modified environments and residual stands of remnant vegetation (Wilcove et al. 1986, Kozakiewicz 1993, Murcia 1995). Common in most areas that have been settled, the practice occurs both in cities through , and in regional locations due to the conversion of natural habitat to agricultural land or the establishment of exotic forests. While the term habitat fragmentation in the strictest sense refers purely to the division of habitat without an accompanying loss in total habitat area (Fahrig 2003), it is commonly employed to describe large scale habitat clearing that drastically reduces the amount of natural habitat leaving only remnant patches (Burgman and Lindenmayer 1998, Bennett 1999, Pullin 2002). While acknowledging the differences in definition, the latter definition will be referred to herein in order to maintain continuity with the wildlife corridor literature.

Globally, habitat fragmentation varies with respect to the extent of clearing, the type of habitat cleared and the duration for which fragmentation has occurred. For example, the massive rainforest expanse of the Amazon Basin in South America is currently being fragmented (5 million km2 and 14% cleared as at 2001) (Laurance 2001), the forests in the small county of Warwickshire, England have been undergoing fragmentation since 400AD with very little native vegetation remaining by 1960 (Wilcove et al. 1986) and the heathlands of Dorset, England have been fragmented since the mid 1700s to the extent that the 15% of original habitat that remains (6000 ha) is split into 141 fragments (Webb 1997). A similar degree of variation exists within Australia with habitat fragmentation occurring throughout the country on a variety of spatial scales and in varying habitat types (Table 1.1), however the duration of fragmentation is considerably less than most instances throughout the world with human modification occurring predominantly in the last 100 years.

Chapter 1. Project overview - 2 -

The process of habitat fragmentation results in an overall loss of original habitat, a reduction in patch size and increased isolation of the patches (Bennett 1999, Pullin 2002). The cumulative effect of this process is the formation of smaller remnant habitats that are isolated both from each other and from any remaining continuous habitat. These remnants are often embedded in a matrix of modified habitat which differs greatly in structure, size, complexity and resource availability to the original habitat type. In the majority of cases, little to no thought is given to the spatial configuration of any vegetation that is to remain uncleared resulting in small blocks of remnant vegetation scattered throughout the modified area (Bennett 1990a).

Table 1.1 Examples of habitat fragmentation within Australia.

Natural Modified Extent of Location vegetation vegetation Total area clearing

Naringal, Victoria eucalypt forest pasture 20,000 ha ~90%1

Atherton and Evelyn pasture and Tablelands, Qld. rainforest agriculture ~79,000 ha ~96%2

Southern wheatbelt, 2 Western Australia eucalypt forest agriculture 140, 000km ~93%3 1 Bennett (1990b) 2 Winter et al. (1987) 3 Saunders et al. (1993)

Remnants that remain resource rich may have the capacity to sustain populations of flora and fauna which can resist extinction processes. Alternatively, population sizes may be restricted in remnants with limited resources and, in the absence of dispersal between remnants, they may be susceptible to extinction due to stochastic events, namely environmental, demographic and genetic stochasticity and catastrophes (Primack 1993).

1.2 Current conservation approaches

Current management strategies for fragmented landscapes include: • Increasing the size of the remnant patch: a strategy designed to provide additional resources which in turn should provide the opportunity for an increase Chapter 1. Project overview - 3 -

in population size.

• The translocation of individuals: the simulation of dispersal by the active transfer of individuals to a new habitat.

• The construction or retention of “stepping stone” habitats: small reserves of remnant habitat that provide limited refuge to organisms but assist in the transfer of individuals between larger habitat patches.

• The construction or retention of wildlife corridors: strips of habitat which connect isolated habitat patches and assist the transfer of organisms across the matrix habitat.

Despite the diversity of approach, all of these strategies attempt to reduce the potentially negative effects of habitat fragmentation by either providing additional resources to sustain populations or providing a means of transfer through the matrix so that the resources from a series of habitat patches can be exploited by a population. The construction of stepping stones or wildlife corridors to meet this latter objective not only enhances the physical connectedness of the landscape (i.e. physically observable or mappable landscape elements [sensu Baudry and Merriam 1988]) but also has the potential to increase the connectivity (i.e. the integration of populations into a functional unit [sensu Merriam 1984]) of the populations within the landscape (Bennett 1999).

1.3 Selection of the appropriate approach

The suitability of each conservation strategy to a particular scenario is dependent upon several factors including i) the characteristics of the target species such as their habitat requirements and dispersal ability, ii) the extent to which they can utilize the matrix habitat and iii) the spatial configuration of the landscape (Wiens 1997, Hobbs and Wilson 1998). It also depends on the desired degree of transfer of individuals between habitat patches and the duration for which human involvement can be maintained (Table 1.2).

Chapter 1. Project overview - 4 -

An increase in patch size is the most appropriate strategy to implement when the enhanced viability of a population within a remnant habitat is desired without an increase in the transfer of individuals between populations, or where no other patches contain individuals of the same species (as may occur with ). However, given the demographic and genetic benefits that the introduction of new individuals can confer to a population, an increase in dispersal rates is usually desired. Where the geographic distance between populations/patches prohibits the dispersal of individuals, the translocation of individuals is the only available strategy. This option is also appropriate when the population is expected to become self-sustaining following an initial translocation event, for example to replenish individuals of breeding age following the elimination of this subset of the population through (Simberloff and Cox 1987). However, as translocation does not alter the landscape in any way, any future dispersal events are reliant on human involvement.

Table 1.2 Current management strategies for the conservation of populations within fragmented habitats.

Suitable for: Management Potential increase Required duration of distance matrix strategy in dispersal levels human involvement limited spp. limited spp. Increase patch size nil Initial construction + + Translocation high Ongoing involvement + + Stepping stones high Initial construction + - Wildlife corridors high Initial construction + +

Alternatively, managers of fragmented landscapes often require a conservation strategy that involves initial human input but then continues in an unassisted manner. The options available in these instances are the implementation of either stepping stones or wildlife corridors, with the appropriateness of each strategy dependent upon the relative habitat use of the various landscape components by the target species (Bennett 1999). Those species which readily traverse the matrix habitat but are distance limited can benefit from the inclusion of stepping stone habitats which provide temporary refuge but do not alter the matrix habitat in any way (Bennett 1999). Conversely, those species that are intolerant of the matrix habitat and can not disperse between patches regardless of geographic distance will Chapter 1. Project overview - 5 - be best assisted by the construction of a wildlife corridor (Bennett 1999). While wildlife corridors can also assist distance-limited species, they are the only option that allows for the unassisted transfer of individuals of species that are unable to utilize the matrix habitat. In this case, successful transfer is achieved by reducing the hostility of inter-patch transition (Merriam 1984). Dispersal via wildlife corridors also has the advantage of allowing processes such as the selection and (non)acceptance of dispersing individuals to occur naturally relative to the enforced dispersal of particular individuals via translocation (Johnsingh et al. 1990).

Selection of the most appropriate management strategy therefore requires a detailed understanding of the target species and its interactions with the various components of the landscape mosaic (Wilson and Lindenmayer 1996b, Hobbs and Wilson 1998).

1.4 The current status of wildlife corridor assessment

The concept of wildlife corridors has become well entrenched within the disciplines of landscape ecology and and has been widely, although not unanimously, adopted by academics and land managers. Despite their apparent simplicity, the intrinsically simple concept of linking two otherwise isolated patches of habitat continues to confound the ecological community with respect to one aspect of its effectiveness. Whilst the use of corridors as additional habitat and the movement of animals along their length has been well established (eg. Bennett et al. 1994, Tewes 1994, Downes et al. 1997a,b, Sutcliffe and Thomas 1996), their effectiveness in increasing the connectivity between the populations within the connected patches is often assumed, yet is not well documented (Saunders and Hobbs 1991, Simberloff et al. 1992, Beier and Noss 1998, Bennett 1999). As connectivity requires integration of an individual into a recipient population, it cannot be estimated by simply assessing movement and habitat use within the wildlife corridor (Mech and Hallett 2001).

Relatively few studies have specifically addressed the enhancement of connectivity via wildlife corridors and of those that have, several have been confounded by the experimental design flaws highlighted by Beier and Noss (1998) along with additional limitations. These limitations reduce the power of any inferences that can Chapter 1. Project overview - 6 - be made on the degree of connectivity provided by wildlife corridors and include failure to investigate matrix use by the target species and the use of inappropriate landscapes with respect to spatial configurations. Many of the studies that have overcome these design limitations have been conducted in small-scale experimental plots (eg. Aars and Ims 1999, Coffman et al. 2001, Mabry and Barrett 2002) that offer little to the understanding of wildlife corridor effectiveness at the landscape scale (Beier and Noss 1998, Noss and Beier 2000, but see Haddad et al. 2000). Of those studies that have focussed on connectivity, very few have incorporated a genetic component into the project despite the facilitation of gene flow being stated as a major aim of wildlife corridors in most theoretical papers (eg. Harris and Scheck 1991, Simberloff et al. 1992, Rosenberg et al. 1998, Bolger et al. 2001). The few studies that have investigated gene flow at a landscape scale have also been hindered by design flaws that limit the validity of any conclusions drawn (Leung et al. 1993, Mech and Hallett 2001). Thus to date, substantial data on the effectiveness of wildlife corridors in facilitating connectivity from either an ecological or genetic viewpoint is minimal.

1.5 Scope and structure of thesis

Given the proliferation of wildlife corridors being constructed worldwide (Kaiser 2001) and within Australia (Wet Tropics Management Authority 2003), and the lack of empirical evidence for their effectiveness in increasing connectivity, a well designed study using a combined ecological and genetic approach is required. Accordingly, this study aims to: • identify the experimental design limitations of previous studies from the wildlife corridor literature and develop an experimental protocol which unambiguously determines the effectiveness of wildlife corridors in providing connectivity (Chapters 2 and 3);

• apply the identified design criteria to a model landscape system (Chapter 3);

• investigate species/habitat interactions, population structure and population genetics in order to gain a comprehensive understanding of the factors influencing the effectiveness of the corridor system (Chapters 4,5 and 6):

Chapter 1. Project overview - 7 -

Each chapter within this thesis addresses a separate issue that builds on the findings of the previous chapter. Thus, the thesis presents a logical progression of the concepts necessary to fully investigate wildlife corridor effectiveness. Whilst each chapter expands on previous findings, each has sufficient background so that they can be read independently. The thesis concludes with a general discussion of the effectiveness of the model wildlife corridor system in providing connectivity and revisits key concepts in order to explain the processes driving the system. - 8 -

2. WILDLIFE CORRIDORS AS A CONSERVATION STRATEGY

2.1 What is a wildlife corridor?

Wildlife corridors are strips of habitat which serve to connect otherwise isolated habitat patches that were once connected in historical time (Saunders and Hobbs 1991). Having their origins in the island biogeography theory of the 1960’s (MacArthur and Wilson 1967) and originally referring to strips connecting terrestrial reserves (Wilson and Willis 1975), the definition has been expanded to include a variety of habitats that meet this physical configuration. These include clearings within forests (Haddad 1999a,b) or under powerlines (Hobbs 1992), temporary floodwaters that link usually free standing water bodies (Kirchner et al. 2003), creeks linking lakes and oceans (Puth and Wilson 2001), rivers linking tributaries (Fraser et al. 1999), highway underpasses (Simberloff et al. 1992, Bennett 1999) and drift macroalgae passing over a bare seabed matrix (Brooks and Bell 2001). Whilst all of these scenarios meet the criteria in the strictest sense, the most widely studied corridor type and those which are most commonly adopted for conservation purposes, are those habitat strips linking remnant vegetation patches within agricultural or plantation landscapes. It is this type of wildlife corridor that will be considered within this study. These corridors can exist as either “original” landscape entities that were retained at the time of habitat modification or as “constructed” corridors that have been subsequently created either as a means of habitat remediation or in the form of windbreaks or hedgerows (see Table 2.1 for examples). Regardless of the origins of a wildlife corridor, the potential benefits that it can confer to the connected populations are the same.

Wildlife corridors also play an important role in habitat networks that are prevalent in Europe and America. These networks consist of a collection of patches linked by a series of corridors and operate on a range of spatial scales including regional, national and European (Jongman 1995, Hobbs 2002). Habitat networks are designed on ecological theory incorporating the distribution and processes involved for particular species (Jongman 1995, Opdam 2002) and incorporate core, buffer and corridor zones (Noss 1992, Jongman 1995,). They provide the opportunity to increase connectivity across ecosystems (Vuilleumeir and Prélaz-Droux 2002) Chapter 2. Wildlife corridors as a conservation strategy - 9 - rather than just linking habitat patches within an ecosystem type as usually occurs with single corridors.

With their widespread adoption by land managers in recent years, corridors have become a “buzzword” in recent conservation literature and this has resulted in landscape configurations that do not meet the accepted definition also being called wildlife corridors. This is particularly the case with linear remnants of riparian or roadside vegetation which are not always true corridors as they do not connect isolated habitat patches. This distinction is often overlooked by researchers studying linear strips of natural or restored vegetation who term their study site a “wildlife corridor” despite the lack of linkage. Far from being a semantic argument, the distinction between linear remnants and wildlife corridors is crucial for wildlife conservation as linear remnants do not have the potential to function in the same manner as corridors (Beier and Noss 1998).

Table 2.1 Examples of the different types of corridors used to address conservation issues.

Location Length Type Status Target species Bogong High Plains, Mountain pygmy Australia1 60m underpass constructed possum

Atherton Tablelands, replanted rainforest Australia2 1.5km rainforest constructed specialists

Rajaji-Corbett N.P., degraded India3 6km forest natural

Lake Manyara N.P., regenerated various game Tanzania4 60km rangeland proposed species 1 Mansergh and Scotts (1989) 2 Tucker (2000) 3 Johnsingh et al. (1990) 4 Mwalyosi (1991)

It has also been suggested that the definition of a wildlife corridor should include a functional element and that only those linear strips that provide the means for organisms to move along them should be termed a wildlife corridor (Saunders and Hobbs 1991, Soulé 1991, Lidicker 1999). This scenario however is limited by the possibility that the corridor could experience differential use on a per species basis. In such circumstances, a landscape element would be defined as a wildlife corridor Chapter 2. Wildlife corridors as a conservation strategy - 10 - for one species but not for another, which would simply add to the confusion of terms which already exists within the landscape ecology literature. Further, the desired function of a corridor may be to increase connectivity rather than simply for habitat use. This presents the possibility that only after much research will it be known whether the landscape element can be called a wildlife corridor or not.

2.2 Potential benefits of wildlife corridors

Following the process of habitat fragmentation, many species exist as small populations within remnant patches. The extent to which these populations are adversely affected largely depends on the degree of isolation of the remnant which, in turn, depends on the differential use of each habitat component (i.e. remnant, matrix etc) within the landscape by individuals of the species (Wiens 1997). In the event that all components can be used equally, isolation will not result unless there is differential survival in dispersing between habitats. While this may occur for some habitat generalists, it is unlikely for habitat specialists given that in most heterogeneous landscapes each component has differing attributes that will likely lead to differential habitat use. Consequently, the isolation of a remnant due to habitat fragmentation will be a function of its geographic isolation from similar habitat and the degree to which the matrix presents a barrier to movement (Chesser 1983b, Saunders et al. 1991, Ricketts 2001). Isolation is therefore a species specific parameter and the sharpness of the remnant-matrix boundary can often provide an indication of the degree to which the matrix will be utilized (Lord and Norton 1990, Lidicker 1999).

It is in the absence of the connectivity provided by dispersal that small isolated populations face the biggest threat from habitat fragmentation as the size of the population is effectively limited to those animals that reside within the remnant (Lacy 1987, Lande and Barrowclough 1987, Lande 1988, Templeton et al. 1990). The persistence of such populations is threatened from any one, or combination of four random factors namely demographic, environmental and genetic stochasticity as well as catastrophes (Burkey 1995). Demographic stochasticity refers to random fluctuations in the population structure that can arise from a change in birth or death rates or the sex ratio of offspring. This process acts on the individuals that comprise the population (Gilpin and Soulé 1986). Conversely, environmental stochasticity Chapter 2. Wildlife corridors as a conservation strategy - 11 - acts upon the population as a whole and refers to the process of environmental change that can impact upon survivorship (Primack 1993). Extreme variations in environmental conditions can lead to catastrophes, which can have a greater impact upon population size, to the extent that they can lead to local extinction.

Isolation of small populations can also lead to potentially detrimental changes in population genetic structure such as the loss of alleles (Usher 1987, Sumner et al. 2004), a decrease in the level of heterozygosity (Wright 1978, Allendorf 1983, Lande 1988, Primack 1993) and an increased prevalence of deleterious recessive alleles in the homozygous state (Friend 1987, Amos and Harwood 1998). These effects can result in depression due to breeding between related individuals which can render populations more prone to extinction through a reduction in individual fitness (Mills and Smouse 1994). Further, the ability of a species to cope with random environmental changes is dependent upon the amount of heritable variation within the population (Frankel and Soulé 1981, Soulé and Simberloff 1986, Burgman and Lindenmayer 1998) and those populations that have experienced a reduction in genetic variation have limited ability to adapt (Franklin 1980, Sherwin and Moritz 2000). This can reduce the viability of the population and increase the potential for local extinction.

The impact that a catastrophe such as fire, , cyclones or disease outbreak, or stochastic process has on a population increases with decreasing population size. Hence an additive effect can occur whereby the effect of one process can make the population more vulnerable to the same or other processes (Gilpin and Soulé 1986). Much debate has occurred over the relative importance of the various stochastic processes (Tallmon et al. 2002) with Lande (1988) and Gaines et al. (1997) suggesting that environmental and demographic stochasticity are likely to play a more important role than genetic processes with regards to the viability of populations in fragmented habitats. This is based on the premise that populations can go extinct due to factors such as and climatic variation regardless of the genetic diversity within the population. However, regardless of their relative contribution, it is well recognized that each process can affect the vulnerability of a population and in the absence of appropriate management strategies, the process of local patch extinction can escalate to regional extinction (Saunders and Ingram 1987, Merriam 1991, Fahrig and Merriam 1994).

Chapter 2. Wildlife corridors as a conservation strategy - 12 -

Wildlife corridors offer a potential solution to the problem of isolation by altering the ecological isolation of the demographic and genetic subunits within the connected patches (Merriam 1991, Wilson and Lindenmayer 1996a). This in turn has the potential to increase the effective population size and increase population viability (Noss 1987). In the event that a breeding population is experiencing a skewed age or sex ratio due to demographic stochasticity, the incorporation of a new individual into the population via transit through the corridor may only slightly alter the size of the effective breeding population yet may greatly enhance the viability of that same population. Alternatively, the transfer of individuals between habitats via a corridor can supplement the number of animals in a declining population as per the “” of Brown and Kodric-Brown (1977). This can arrest population decline, considerably increase the effective population size, decrease the susceptibility of the population to demographic changes and ultimately increase the viability of the population (Burbrink et al. 1998).

The introduction of new genes into the population may also occur via the incorporation of relatively few individuals yet may considerably improve the genetic constitution of that population, especially when genetic diversity has been eroded due to prolonged isolation (Soulé 1991). An increase in gene flow can reduce inbreeding (Newman and Tallmon 2001) and enhance genetic diversity. This has the net effect of increasing population viability and reducing the chance of local extinction, however, such benefits are dependent upon the genetic constitution of the populations within the connected patches. A potential genetic benefit can only be derived from wildlife corridors if there is some degree of differentiation between the connected populations. Where the genetic structure of the populations in two patches is the same, no degree of dispersal or gene flow will result in a net increase in genetic diversity.

The facilitation of gene flow is not always viewed as a potential benefit of wildlife corridors with the potential homogenization of gene pools and the loss of adaptive traits an issue of concern (Noss 1987, Simberloff and Cox 1987). One must realize however, that wildlife corridors are only proposed to connect habitats that were once connected in historical time. Although not usually stated, this can be interpreted to represent those habitats connected prior to human alteration rather than those habitats having a previous geological connection (Bennett 1990a). Any specialized adaptations present within the populations therefore either arose i) after the Chapter 2. Wildlife corridors as a conservation strategy - 13 - alteration of the landscape and are therefore under the influence of human processes or ii) prior to the modification as a result of restricted gene flow due to natural processes. Neither of these situations are cause to consider wildlife corridors as an undesirable management option.

Another desired benefit of wildlife corridors is simply to provide additional habitat within the landscape (Wilson and Lindenmayer 1996a). Where the corridor provides only some of the resources needed for survival, it may assist the breeding population within a remnant habitat patch by increasing the area and resources available to that population (Lynch and Saunders 1991, Newmark 1993). However, should the corridor provide the full compliment of resources required by a species to complete its lifecycle, it may become a suitable breeding area in itself (Merriam and Saunders 1993). This not only increases the effective population size within the corridor system, but also provides a breeding link between the populations of the connected habitat patches (Bennett 1990a). Where the length of the corridor is beyond the dispersal capabilities of a single individual, as may be expected in a highly modified landscape, the arrival and incorporation of a new individual into the population within a connected remnant is dependent upon the existence of breeding populations within the corridor (Beier and Noss 1998, Lynch and Saunders 1991, Merriam 1991). For gene flow to occur in such scenarios, a series of matings, progressing along the corridor, is required so that the genes of the original dispersing individual become incorporated into the recipient population via its offspring or a member of subsequent generations (Burel 1989, Merriam 1991, Bennett 1999). An individual that enters and confers demographic or genetic benefits to a population within a remnant patch may therefore have been born within the corridor but have a genealogy originating from the alternate remnant population.

In the event that a population in a remnant patch does go extinct, the presence of a wildlife corridor can also assist in the recolonization of the remnant (Fahrig and Merriam 1985). In the absence of a corridor and any movement through the matrix, the patch would remain uninhabited by the target species, yet the provision of a suitable transit route offers the opportunity for recolonization which again increases the chance of both local and regional persistence of the species (Fahrig and Merriam 1994). While the recolonization of a vacant patch via a corridor requires the movement of a considerable number of individuals in order to establish a new breeding population, demographic and genetic benefits may result from the Chapter 2. Wildlife corridors as a conservation strategy - 14 - movement of a very small number of individuals through the corridor or from generational gene flow that negates the need for individual movement (Merriam 1991). Thus, a wildlife corridor may still provide a functional benefit to remnant populations in the event of little or no actual movement by individuals between patches via the corridor. This is despite the movement of individuals often being stated as the primary aim of wildlife corridors (Soulé 1991, Simberloff et al. 1992).

Wildlife corridors have the potential to ameliorate the negative effects of habitat fragmentation, however, these benefits can only be realized if individuals of the target species elect to use the corridor in the desired manner. Further, corridors can only provide potential benefits to populations in the remnants where the matrix either presents a complete barrier to movement or where limited dispersal may occur but a further increase would prove beneficial to the populations (Beier and Noss 1998). Regardless of whether a corridor is used or not, it does not have the potential to assist populations in environments where dispersal through the matrix is already sufficient to homogenize the demographic and genetic structure of the populations. In instances where matrix use is already considerable and homogenization occurs, the only benefit that a corridor can provide to the populations within connected remnants is that of additional habitat: a goal that would be best achieved by an alternative habitat configuration that is free from many of the potential problems associated with narrow linear strips eg. low interior to edge ratio (Hobbs and Saunders 1990, Simberloff et al. 1992).

2.3 Past assessment of wildlife corridor effectiveness

Despite the abundance of studies that have assessed some aspect of wildlife corridor effectiveness, very few have primarily investigated the ability of corridors to increase connectivity between the otherwise isolated populations (Várkonyi et al. 2003). The majority of studies have concentrated on alternative aspects of corridors such as their use as habitat by either the individual species (eg Lindenmayer et al. 1994, Beier 1995) or species assemblages (eg Burbrink et al. 1998, de Lima and Gascon 1999), transit into and along the corridor (eg Suckling 1984, Sutcliffe and Thomas 1996) and the effect of corridor presence on dispersal rates (La Polla and Barrett 1993, Danielson and Hubbard 2000). While these studies of corridor use can include aspects also included in studies on connectivity, they do not in Chapter 2. Wildlife corridors as a conservation strategy - 15 - themselves constitute an investigation into whether wildlife corridors successfully increase connectivity.

Those studies that have addressed connectivity issues have used a variety of approaches including field based ecological studies, genetic studies, experimental projects and computer modelling. Although the combination of these disciplines can provide a greater understanding of ecological processes including connectivity, relatively few studies have adopted an integrated approach. This is somewhat surprising given the high citation rates of the paper by Gilpin and Soulé (1986) which highlights the interaction between genetic and demographic processes (Young and Clark 2000).

Each method of assessment is reviewed below with specific emphasis placed on the design limitations of the approach where necessary. This style is not intended to diminish the importance of previous work, but is aimed at highlighting the experimental design flaws that are currently hindering our understanding of corridor effectiveness with respect to connectivity and to which special attention needs to be paid in the design of future studies, including the current study.

2.3.1 Field-based ecological studies The process of connectivity requires more than habitat use or the movement of animals from one patch to another. For dispersal along a corridor to be effective either demographically or genetically, an individual or its genes has to enter the breeding population and contribute to the next generation rather than merely residing within the connected patch. Given that incorporation into the breeding population is the end result of a complex series of ecological processes involving any of a number of parameters including philopatry, dispersal (including sex-biased), differential habitat use, spatial structuring and behavioural responses (eg territorial defence), the mere presence of individuals in a corridor, or movement into a patch does not necessarily translate into effective dispersal or gene flow (Whitlock and McCaughley 1999). Consequently, studies on aspects of corridor use such as habitat utilization should not be extrapolated to infer dispersal (eg Laurance 1990, Downes et al. 1997a,b) or its related benefits (eg gene flow) (Bennett 1990b). While these conclusions are possibly correct, they are not demonstrated by the data collected and thus, such studies should limit their conclusions to the use of the Chapter 2. Wildlife corridors as a conservation strategy - 16 - corridor by given species and at most should comment on the potential that corridors have to effectively increase connectivity.

Very few studies have been identified from the literature as using field based ecological techniques to rigorously address the issue of connectivity (with the suite of papers by Haddad and colleagues considered as one work). The study that best uses ecological techniques to demonstrate the ability of wildlife corridors to increase connectivity is that of Mansergh and Scotts (1989). Using pre- and post-corridor data, they showed that the social structure and survival rates of a previously fragmented mountain pygmy possum (Burramys parvus) population became consistent with those in an undisrupted habitat after the construction of a wildlife corridor.

Haddad and colleagues conducted research in a “natural” system (forest plantation) that had been experimentally manipulated to create an appropriate spatial configuration with corridors up to 380m in length (Haddad 1999a,b, Haddad and Baum 1999, Haddad 2000, Tewkesbury et al. 2002, Haddad et al. 2003). Among their results, Haddad and Baum (1999) showed population densities to be significantly greater in connected patches relative to isolates. This suggests that corridors may provide some protection against the stochastic events which can jeopardize the persistence of small populations in remnant patches. In these instances, wildlife corridors were found to have a positive effect on populations despite the detection of some matrix use by the target species (Haddad 1999a).

The remaining two studies illustrate the importance of tailoring the experimental design to each individual study with respect to the issue that is being investigated. Henderson et al. (1985) and Fahrig and Merriam (1985) both investigated the use of corridors, in the form of fencerows, by small mammals in an agricultural landscape. Fahrig and Merriam did not specifically investigate matrix use yet were able to provide valid comment on the role of corridors in facilitating recolonization of vacant patches as recolonization rates were greater in connected patches than in isolated remnants. However, the issue of whether the degree of movement through the matrix was sufficient to ensure population survival in the absence of corridors was not fully investigated. Alternatively, the lack of investigation into matrix use resulted in ambiguity in the data of Henderson et al. (1985). Recolonization of vacant patches was shown to occur with animals having to traverse small areas of matrix Chapter 2. Wildlife corridors as a conservation strategy - 17 - habitat to reach the patches. In the absence of detailed matrix trapping or the inclusion of isolated habitats within the design, the extent to which the species can utilize the matrix remains unknown, therefore precluding any conclusions on corridor effectiveness.

While the majority of the few studies investigating connectivity using a field-based ecological approach have been well designed, the potential exists for the data of future studies to be compromised by the design flaws highlighted by Beier and Noss (1998). These will be briefly summarized below, as recognition of these potential limitations are vital for the design of future studies, including the current study. Studies that have investigated aspects of corridor effectiveness other than connectivity will be used as examples.

Many of the studies addressing the issue of wildlife corridors have been undertaken within landscape configurations that do not meet the commonly accepted definition of a wildlife corridor. While linear remnants can provide additional habitat, they are unable to provide a means of transit between two independently breeding populations and are thus unable to provide the potentially advantageous demographic and genetic benefits of a corridor. In several instances, authors using these spatial configurations refer to the linear strips as corridors (Tewes 1994, Nicholls et al. 2001, Tigas et al. 2002) and draw inferences from their data on the conservation worth of corridors (Laurance and Laurance 1999, Lomolino and Perault 2000). This may be valid where the function under investigation is the provision of additional habitat or the tolerance of species to linear habitats with a high edge:interior ratio, however, any inferences beyond this scope are limited. Investigations of this nature would benefit from following the lead of Hill (1995) and Bolger et al. (2001) who investigated potential corridor use for a given species by determining the habitat use of linear remnants whilst outlining the theoretical differences between the two landscape types. It is important that wildlife managers are clear on the distinction between the terms and do not conclude that the construction of linear remnants can confer the same benefits to wildlife populations as can corridors. The potential benefits of corridors versus linear strips are so disparate that Beier and Noss (1998) excluded studies using linear remnants from their review on wildlife corridors.

Chapter 2. Wildlife corridors as a conservation strategy - 18 -

The use of inappropriate species is another area of concern with some studies reporting that wildlife corridors had no effect as the species were found to use the matrix habitat (eg. Bowne et al. 1999). These studies do not increase our understanding of the effect of wildlife corridors in any way and definitely cannot be used to infer that the corridors do not work. Of greater concern however, are those studies that suggest that wildlife corridors are effective conservation strategies while neglecting to investigate or report on the differential habitat use of the target species (eg. Johnsingh et al. 1990, Tewes 1994). Thus the reader does not know whether the stated effect is due to the presence of a corridor or the use of the matrix by the target species: the implications of which would be drastically different than the published interpretation of the data. Further, in some studies, matrix use has been shown to occur yet the presence of a wildlife corridor has still been reported as a positive attribute without any consideration as to whether the use of the matrix by individuals was sufficient to effectively reduce the isolation of the connected patches (eg. Wegner and Merriam 1979, Suckling 1984, Sutcliffe and Thomas 1996). As the connectivity studies of Fahrig and Merriam (1985) and Mansergh and Scotts (1989) demonstrated, investigation of the matrix is not essential if pre-corridor data are available or if experimental extinctions can be performed in both connected patches and isolates. However, most wildlife corridor studies are not afforded the luxury of these scenarios making matrix use an essential, yet often overlooked element of experimental design. This is despite the call for as much, if not more, emphasis to be placed on the matrix as on the fragments (Crome 1997).

2.3.2 Experimental studies Many of the problems associated with experimental design result from either i) the difficulty in finding a natural system that meets the necessary spatial criteria or ii) the ethical dilemma of simulating an extinction event in natural populations. One approach that has been commonly employed to address these issues is the use of small-scale experimental plots. While this approach eliminates any ambiguity from the data, the studies using this method have not added significantly to our understanding of wildlife corridor effectiveness on the landscape scale primarily due to the lack of realism associated with experiments conducted at a micro-scale. Such experiments are usually conducted at a spatial scale of 20-50m (La Polla and Barrett 1993, Boudjemadi et al. 1999, Coffman et al. 2001, Berggren et al. 2002) which can be up to several orders of magnitude less than that at which conservation Chapter 2. Wildlife corridors as a conservation strategy - 19 - strategies need to be implemented. Similarly, they use species with movement patterns relevant to these small spatial scales.

This small-scale approach has been criticized by many authors within the conservation biology literature (Murphy 1989, Beier and Noss 1998, Debinski and Holt 2000, Noss and Beier 2000, Hilty and Merenlender 2004) who have implored scientists to not only investigate topics relevant to conservation science, but to conduct the studies at a geographic scale that will provide meaningful information to those practitioners having to make real-world decisions on the conservation of natural resources. For instance, the comprehensive work of Aars, Ims and colleagues (Andreassen et al. 1998, Aars and Ims 1999, Aars et al. 1999, Ims and Andreassen 1999) was conducted in a system using corridors of approximately 50m x 0.5m linking habitat patches of ~20m x 37m and investigated the demographic consequences of movement, rates of transfer between habitats, interbreeding between individuals and spatio-social organization in voles in connected versus non- connected habitats. While the researchers made the valid point that the corridor length was twice the distance of the known male home range diameter and thus was an appropriate length for the species under investigation, it provides little information on the function of corridors which are required to extend for hundreds of metres or kilometres in order to connect remnants within highly fragmented systems. This study however, does provide a good example of the type of processes that need to be investigated in order to examine the issue of connectivity using a combined ecological/genetic approach.

Research conducted within experimental plots also often investigate processes within highly regulated environments which usually consist of only one animal species in a monoculture environment (eg. Boudjemadi et al. 1999, Davis-Born and Wolff 2000). Hence they lack many of the crucial ecological processes that can affect wildlife corridor effectiveness, such as inter-specific competition.

It should be noted that Haddad and colleagues have successfully used an experimental approach to show that corridors can increase connectivity. This work was described in the section on field based ecological studies (Section 2.3.1) as it was conducted using a large scale field-based design. This series of studies is an important contribution to the wildlife corridor literature not only because of their positive findings, but because they demonstrate that experimental studies can be Chapter 2. Wildlife corridors as a conservation strategy - 20 - implemented at a scale relevant to real conservation issues. Although the wildlife corridors only extended for a few hundred metres, the scale of the study was relevant to both the species under investigation and the forest management practices implemented within the region (Haddad 1999a).

2.3.3 Genetic studies In contrast to the data obtained via mark-recapture techniques, the use of genetic markers enables an assessment of whether individuals do enter the breeding population (Slatkin 1994, Whitlock and McCaughley 1999). This type of study also offers the advantage of indicating the long-term cumulative effects of transfer between populations, making it more sensitive to long-distance or infrequent dispersal events which are likely to be missed by traditional trapping methods (Koenig et al. 1996, Perault and Lomolino 2000). They also provide an indication of the extent of gene flow via generational transfer without requiring detectable movement events. Despite the benefits that genetic methods of assessment offer the field of wildlife corridor research, a review of papers addressing the concept of wildlife corridor function reveals a large discrepancy between the theory of corridors and the empirical research. The vast majority of theoretical papers, along with many Introductions of empirical papers, mention the facilitation of gene flow between connected patches as a desired function of wildlife corridors. However this awareness of the issue has not translated into investigation of the topic, with very few papers investigating gene flow via corridors using genetic techniques (Leung et al. 1993, Aars and Ims 1999, Mech and Hallett 2001, Kirchener et al. 2003) and only two of these papers investigating gene flow through a landscape-scale terrestrial corridor (Leung et al. 1993, Mech and Hallett 2001). This lack of genetic research into wildlife corridor function is somewhat surprising given that population genetics has been used for many years to investigate the effects of habitat fragmentation on isolated populations with the results of these studies occasionally leading to the recommendation of wildlife corridors as an appropriate management strategy (eg. Campbell 1996, Simonsen et al. 1998).

Despite the potential that population genetics has to offer, little information on connectivity due to wildlife corridors can be obtained from the two landscape scale studies conducted to date. This is essentially due to confounding factors in the experimental design which preclude the interpretation of the primary effect of corridors, such as the use of inappropriate landscape configurations in which wildlife Chapter 2. Wildlife corridors as a conservation strategy - 21 - corridors do not provide the only potential means of gene transfer (Mech and Hallett 2001). In this instance, animals and/or genes can pass between the two locations designated as the end points of the corridor via an alternate, albeit less direct, route through the same vegetation type. Thus the sample sites would not be isolated in the absence of the corridor, limiting conclusions that can be drawn from this type of genetic study. Further, both Leung et al. (1993) and Mech and Hallett (2001) fail to implement appropriate control sites within either a continuous tract of habitat or in the form of totally isolated habitat patches. Similar to habitat fragmentation studies, (Taylor 1999, Williams et al. 2003) these controls are required as reference points against which the level of genetic differentiation between corridor connected populations can be compared.

Continuous habitats indicate the level of gene flow experienced in the absence of any human modification, while patches completely surrounded by the matrix habitat give insight into the effects of total isolation on gene flow and genetic diversity. These habitat types represent the extremes of the isolation/habitat fragmentation continuum and provide a basis for the interpretation of the effectiveness of wildlife corridors with respect to gene flow. Both studies attempted to include continuous and isolated control sites within the study design, but fell short of meeting these criteria and therefore, the extent to which their corridors facilitate gene flow remains unclear. The sampling locations designated by Mech and Hallett (2001) as “isolate” controls are simply locations within the continuous forest and the individuals sampled from these sites therefore represent a subset of the larger population from within the contiguous habitat. They are not isolated from similar habitat and the populations are therefore not subject to the potentially adverse affects arising from isolation and a disruption to dispersal. Mech and Hallett (2001) recognized this problem but drew comparisons between the “isolates” and corridor linked sites regardless of the limitations. Leung et al. (1993) selected an appropriate isolate site but compromised the study by pooling the two sampling locations within the continuous habitat. This resulted in samples from within small remnant habitats being compared to samples collected from a much larger geographic range within the continuous forest. Separation of the sites within the continuous habitat would have allowed for pairwise comparisons of connected sites relative to the natural variation exhibited over the same distance within control sites providing a more powerful interpretation of gene flow through the corridor. These examples highlight that although the adoption of a genetic approach has the potential to assess the Chapter 2. Wildlife corridors as a conservation strategy - 22 - extent of connectivity provided by corridors, additional considerations become important in the design of the study.

It must also be remembered that the use of molecular markers alone does not provide insight into the ecological functioning of individuals within a corridor system. This is especially pertinent when gene flow is not found to occur and an ecological explanation is required. For this reason, the use of molecular markers has been suggested as complimentary to ecological techniques (Lindenmayer 1994, Bierregaard et al. 1997). The integration of ecological and genetic techniques offers the greatest opportunities to not only assess the effectiveness of gene flow via corridors, but to also understand the processes involved. To date, no studies have been conducted that have rigorously applied this integrated approach in a field- based situation. Mech and Hallett (2001) solely presented genetic data whilst Leung et al. (1993) conducted only limited field work to compliment the genetic analysis. The most rigorous work using this combined approach has been undertaken by Aars, Ims and colleagues as outlined in the previous section on experimental studies.

Many authors are perhaps unaware of how design limitations can impact upon the conclusions drawn. Mech and Hallett (2001) and Hale et al. (2001) have recently been cited as providing evidence for the flow of genes through wildlife corridors (in Colgan et al. 2002, Hudgens and Haddad 2003, Mabry et al. 2003). While the design limitations of Mech and Hallett have been outlined, the study of Hale et al. (2001) was not mentioned as the authors do not deem the landscape configuration they investigated to be a corridor and did not call it as such in their paper. The revegetated area of land that they investigated did link two larger forests, however, it simply filled a linear strip that had been carved through the original forest. Hence the revegetated area had very little of its perimeter adjoining the matrix and due to the large proportion of perimeter adjoining continuous forest habitat, it was not subject to factors facing traditional corridor success such as territorial defence within the corridor. Despite this, the paper has been cited as demonstrating that gene flow can occur through corridors.

2.3.4 Computer modelling The final means of assessing wildlife corridor effectiveness involves the use of computer modelling. Models are commonly employed to assess the effects that Chapter 2. Wildlife corridors as a conservation strategy - 23 - corridors of different quality, quantity and spatial arrangement have upon a metapopulation when compared to unconnected habitats (Anderson and Danielson 1997, Jordan 2000) and have been written for specific species using published demographic parameters (Fahrig and Meriam 1985, Henein and Merriam 1990).

The ultimate test of the validity of a model is to compare the predicted results with those obtained through fieldwork. This validation process has been successfully undertaken for models written by Fahrig and Merriam (1985) and Haddad (1999b). Haddad (1999b) showed that consideration of movement behaviour by an individual at habitat boundaries could predict corridor use which was shown to lead to increased population density, while Fahrig and Merriam (1985) showed population growth rates to be significantly greater in connected patches when compared to isolates. These examples highlight the utility that models have in understanding the extent of connectivity provided by corridors and the potential that they have for assessing their potential in a given landscape scenario. Conversely, the study of Brooker et al. (1999) showed considerable discrepancies between the actual and simulated data for the movement of through corridors within an agricultural landscape.

Regardless of their predictive ability, the use of models to investigate the connectivity provided by wildlife corridors is currently limited by the lack of rigorous field data on the same issue. Thus, an increase in the number and quality of field studies investigating connectivity may also benefit the field of wildlife corridor research from a modelling perspective.

2.3.5 Summary It is unfortunate that design limitations have led to the current situation of numerous studies providing little information on the connectivity aspect of wildlife corridors and that many authors are perhaps unaware of how design limitations can impact upon the conclusions drawn. While many studies have been undertaken to assess some aspect of wildlife corridor effectiveness, and have shown corridors to be beneficial in aspects such as the provision of additional habitat, the vast majority investigate issues other than connectivity at a landscape scale. Of those studies that have addressed this issue, ecological evidence suggests that connectivity may occur however further studies, especially those providing complementary genetic data, are required. In light of this lack of data, acceptance of the connectivity benefits of Chapter 2. Wildlife corridors as a conservation strategy - 24 - wildlife corridors should be delayed until a conclusion can be based purely on rigorous research that shows unequivocally that they are beneficial rather than the positive interpretation of ambiguous data.

2.4 Project aims

The objective of this project was to assess the effectiveness of a wildlife corridor in providing connectivity within a model system. To achieve this, the following aims must be satisfied: • identify the design criteria necessary to unambiguously assess effectiveness;

• locate a model landscape system and identify candidate species that meet the necessary criteria;

• test the appropriateness of the selected landscape for use as a model system with respect to the identified essential design criteria.

• investigate the demographic parameters of the species within various components of the corridor system;

• monitor the movement of individuals within the corridor system and,

• assess the extent of gene flow through the corridor relative to unconnected and continuous habitats.

This study therefore aims to not only assess the degree of connectivity provided by the corridor, but also to investigate the ecological parameters that influence the function of the corridor system. In addition, this study aims to provide a template suitable for use by future studies in order to overcome the design limitations currently hindering this aspect of wildlife corridor research. - 25 -

3. THE MODEL SYSTEM: DESCRIPTION AND EVALUATION OF ITS

SUITABILITY FOR THE ASSESSMENT OF CONNECTIVITY

3.1 Introduction

Uncertainty exists over the ability of wildlife corridors to facilitate connectivity between populations within connected patches (Beier and Noss 1998). This is mainly due to previous studies focussing on aspects of corridor effectiveness other than the provision of connectivity and experimental design flaws in several of the studies that do investigate this issue. These limitations primarily concern the inability to partition the effect of the wildlife corridor from confounding factors. It is therefore essential that the spatial and biological characteristics of the study system be given special consideration. The following section focuses on essential design criteria and will emphasize parameters relevant to studies assessing connectivity through the combined use of ecological and genetic methods. While some experimental designs allow for the assessment of corridor function via alternate methods (eg. the comparison of population growth rates in connected and isolated patches (Fahrig and Merriam 1985)), these criteria, either in part or full, will be applicable to the vast majority of studies aiming to assess the extent of connectivity provided by wildlife corridors.

The provision of connectivity by wildlife corridors requires the movement of individuals and/or genes away from a habitat patch at one end of the corridor and their ultimate incorporation into the breeding population within the alternate connected habitat. This can occur via either the direct movement of organisms from one patch to another or via generational gene flow. Hence, to fully understand the process of connectivity, studies need to investigate whether gene flow occurs and the ecological processes underlying its occurrence (or otherwise). This requires two distinct lines of investigation: i) the assessment of movement patterns and demographic parameters and ii) gene flow. Several experimental design considerations for these investigations have been introduced previously (Section 2.3) and have been discussed within the literature (Nicholls and Margules 1991, Chapter 3. The model system - 26 -

Inglis and Underwood 1992, Beier and Noss 1998, Mech and Hallett 2001). These considerations include:

1. The study corridor should link at least two patches of habitat that would be otherwise isolated by an intervening matrix habitat (Saunders and Hobbs 1991, Soulé and Gilpin 1991, Beier and Noss 1998)(Figure 3.1a).

2. The matrix habitat should act as an isolating barrier and prevent movement of the target species between the connected patches. Failure to select a matrix intolerant species, or a corridor system embedded in a matrix intolerant to the target organism, provides opportunity for individuals to move between the connected patches via a means other than the corridor (Beier and Noss 1998)(Figure 3.1b). Consequently, even where species are known to avoid the matrix habitat type, trapping should be conducted within this component of the system so that data are available to confirm this important assumption (Cook et al. 2004).

Where a specific corridor is the unit of interest, the choice of a study organism is flexible and should be selected upon the basis of its relative habitat use. Where the unit of interest is a particular species, the study area should be determined by the known habitat preferences of the target organism. Selection of appropriate study sites is more problematic using this latter approach as the target species is likely to be endangered and/or have a restricted geographic distribution, thus limiting the number of corridor systems from which a suitable candidate can be selected.

3. Control sites are required within continuous habitat, and also in truly isolated habitat patches in order to indicate the extent to which isolation has affected gene flow within the landscape. These control sites should be of the same habitat type as the connected patches so that any variables such as resource availability and social behaviour that can affect the genetic structure of the populations remain constant (Taylor 1999, Mech and Hallett 2001).

These design considerations are presented in a stylized manner in Figure 3.2. Chapter 3. The model system - 27 -

a)

b)

Figure 3.1 Diagrammatic representation of the essential design criteria necessary for the rigorous assessment of wildlife corridor effectiveness as identified from the literature. The design criteria include the need to a) have the corridor linking two habitat patches, b) investigate matrix use by the target species in order to eliminate transfer via an alternate route.

Fragmented system Continuous (reference) system

distance x .

Figure 3.2 Stylized representation of the essential design criteria. = preferred habitat, = inhospitable matrix.

Chapter 3. The model system - 28 -

Additional criteria, not yet discussed in the literature, but also considered as essential for inclusion in connectivity studies, have been identified. These have primarily arisen from the consideration of genetic aspects of the project but include points relevant to non-genetic studies. The additional criteria involve i) biological properties considered essential for the selected wildlife corridor and ii) sampling strategies to ensure the collection of appropriate data once a system is deemed suitable for use. These criteria include:

1. Determining that the target species is able to utilize the connected patches and the corridor habitat. The effectiveness of a wildlife corridor in promoting connectivity between populations depends on many factors, including the behaviour of the target species. Before one can partition the success or failure of this aspect of corridor function to any species attributes, it must be known that the corridor itself does not provide a barrier to achieving the desired outcome. Where the desired function of a corridor is the provision of additional habitat, individuals must be shown to occur within the corridor in order for the corridor to be deemed as effective. The provision of connectivity however, can be achieved by infrequent dispersal events and a lack of habitat use over a short time frame does not indicate that the corridor is unlikely to provide this function. This criteria can be assessed either by detection of the target species within the habitat type and/or by inference from the assessment of vegetation characteristics along the corridor compared to that within patches where the species is known to reside.

The extent to which the physical factors and species composition of a corridor have been assessed in past studies varies considerably from the analysis of both form and composition (e.g. Bennett et al. 1994, Lindenmayer et al. 1994, Bentley and Catterall 1997, Downes et al. 1997a, Sieving et al. 2000), to the assessment of only: physical attributes of the vegetation (Bentley et al. 2000, Mönkkönen and Mutanen 2003), parameters of the corridor (such as width and length) (Lindenmayer et al. 1993) or the species present (Crome et al. 1994, Bolger et al. 2001). However, the majority of studies neglect to describe or assess the suitability of the vegetation. This could be due to the large number of studies that are performed on either a small scale or in monoculture which do not experience significant habitat variation.

Chapter 3. The model system - 29 -

2a. The location of traps within the connected patches is important with respect to mark-recapture success. Where a mark-recapture study is incorporated into the design, trapping location and intensity should maximize the chances of encountering individuals as they pass between components of the corridor system. This is required to increase both the number of animals uniquely identified and their chances of recapture. Given that many species may display social structuring or small home ranges relative to the size of the connected patch, these objectives can be best met by locating traps at the patch/corridor boundary. To date, little attention has been paid to the location of trap sites within patches. However, in the case of territorial species, individuals trapped and marked elsewhere within the patch may be denied access to the corridor by defensive con-specifics at the patch/corridor boundary or individuals of species with limited movement may never encounter the corridor (Figure 3.3a). Trapping within alternate areas of the connected patch could therefore lead to the interpretation that corridors are ineffective as no marked animals were found to enter the corridor and enter a new breeding population yet connectivity between the individuals residing at the respective patches/corridor boundaries may occur.

2b. If the effectiveness of wildlife corridors is to be determined, DNA samples need to be obtained from those animals most likely to be directly influenced from the acquisition of new genetic material via the corridor. This again most likely involves those animals residing at the corridor/patch boundary and while animals living throughout the patch may also benefit, the chance exists that territorial behaviour may exclude them from any effects of the corridor (Figure 3.3a).

3. Trapping is required at several locations within the corridor. This serves the dual purpose of i) increasing the chances of encountering uniquely marked individuals to obtain movement data in order to determine the likely mode of movement for the species and ii) establishing whether resident populations of the species are present within the corridor. This latter aspect is necessary if the length of the corridor is beyond the movement capabilities of the species and connectivity is therefore dependant upon generational connectivity (Figure 3.3b). Further, these data can be used retrospectively to explain any Chapter 3. The model system - 30 -

interruptions to gene flow through territorial defence by individuals residing within the corridor as suggested by Lidicker (1999) (Figure 3.3c).

a)

b)

c)

Figure 3.3 Diagrammatic representation of additional design criteria identified as essential for the rigorous assessment of wildlife corridor effectiveness with respect to investigating connectivity. The design criteria include the need to increase the chances of the study individuals accessing the corridor by locating traps at the patch/corridor boundary (a). Trapping within the corridor is also necessary in order to determine the likely mode of movement i.e. direct or generational (b) and to offer potential ecological explanations for any absence in gene flow despite the presence of the target species within the corridor (c).

4. To eliminate further confounding factors, the geographic distance between i) isolated sites and either one end of the corridor or each other and ii) the distance between sites within the continuous habitats, should ideally be equivalent to the length of the corridor. In landscapes where different matrix types are present, the matrix type within the corridor system should also be present between isolated sites. Given that the extent of dispersal between Chapter 3. The model system - 31 -

habitat patches depends on the degree of isolation of the habitat, which is in turn dependant on the nature of the intervening matrix, consistency in matrix types ensures that the effective distance and the geographic distance are equivalent. This allows any variation between sites to be attributed solely to the corridor.

While these requirements may seem numerous and difficult to obtain in natural landscapes, such a controlled approach is necessary to advance our currently limited knowledge on this aspect of wildlife corridor effectiveness.

In this study, many landscapes were assessed in light of these essential design criteria. The selected landscape satisfied those criteria that are assessable at an observational level such as spatial configuration, the uniformity of the matrix and the availability of suitable control sites and it was therefore investigated further for its suitability as a model system. This chapter describes the selected landscape and the investigation of its suitability for use with respect to the remaining essential design criteria i.e. those that cannot be assessed by observation alone. These parameters include the similarity in vegetation attributes within the corridor and the remnant patches and the use of the system components by the target species. Only after this final investigation can the landscape system be accepted as an appropriate model in which to investigate the effectiveness of the wildlife corridor in increasing connectivity.

3.2 The model system

The current study was undertaken using a single patch/corridor/patch configuration with appropriate control sites. As with all field studies, a balance exists between replication and the strength of data that can be obtained from each replicate system. Given that complex landscape systems cannot be truly replicated due to the large number of potentially confounding factors, it was decided to rigorously investigate one system. Hence, the selection of a suitable study system, and the testing of design assumptions within it, was an integral component of the project. This project therefore adopts a model system approach to the understanding of wildlife corridor function.

Chapter 3. The model system - 32 -

3.2.1 The study site The study system was located in the area surrounding the township of Malanda (17o 22’S, 145o 35’E) on the Atherton Tablelands of tropical northern Queensland. The Atherton Tablelands is a basaltic plateau approximately 50 km south-west of Cairns and was originally covered by complex notophyll and mesophyll rainforest (Tracey and Webb 1975). The area was extensively cleared in the early 1900’s for agriculture (Frawley 1983), creating a highly modified landscape mosaic consisting of agricultural land and remnant rainforest patches of various size. The study system is bordered to the east by a large tract of uncleared rainforest. Estimates of land-clearing indicate that by 1983, over 76,000 ha of rainforest had been removed from the Tablelands and clearing has continued to the present day (Winter et al. 1987).

Study sites consisted of both remnant patches of rainforest located on freehold land and sites within the unfragmented habitat of the World Heritage Area of Wooroonooran National Park (Figure 3.4). In order to meet the criteria outlined in Section 3.1, sites were categorized according to their degree of physical connectedness to other habitats of similar vegetation type. Site types included: i) connected sites consisting of remnant rainforest habitat surrounded by a pasture matrix but connected to another rainforest habitat by a corridor. ii) corridor sites located within the corridor that links the connected sites, iii) isolated sites of remnant rainforest habitat completely surrounded by a pasture matrix, iv) continuous sites within a large tract of rainforest habitat, v) matrix sites within the pasture surrounding the corridor/patch.

The Gwynne Creek corridor system, in which this study was conducted, consists of original riparian vegetation linking two rainforest remnants (Figure 3.4B) within an extensive matrix of short, heavily grazed cattle pasture (Figure 3.4C). The conversion of rainforest to pasture was completed prior to 1918 leaving the ~4.5km corridor as the only linkage between the remnant patches for at least 85 years (Mick Laws [landowner], pers. comm.). This corridor system was selected not only because of its spatial configuration and the nature of the surrounding matrix habitat, but also due to the period of time that the corridor has existed in its current form. Unlike recently constructed corridors within the region, the Gwynne Creek corridor is suitable for estimating long-term gene flow via the corridor. Chapter 3. The model system - 33 -

Five sites were selected within the corridor (Figure 3.4B, Figure 3.5) and while the spacing of these sites would ideally be equidistant along its length, site selection was restricted by landowner permission and accessibility. Matrix sites (Figure 3.5) were located in the pasture adjacent to corridor sites at Corridor KH and Corridor SK and also adjacent to the remnant/pasture boundary at Connected R and Connected M.

Despite the highly modified nature of the Atherton Tablelands very few rainforest remnants met the criteria necessary for inclusion as an “isolate” due to the nature of the surrounding matrix. The three sites that were chosen for use as isolates (Isolates B, L and W; Figure 3.4A) represent the only remnants on the Tablelands that are located within a suitable distance from the corridor and that are surrounded by a matrix similar to that found around the corridor system. These criteria are essential to ensure that geographic and effective distances between sites are equivalent. As the size of the isolated remnant patches differed in size from those linked by the corridor, the area from which animals were sampled was kept constant at all sites.

Chapter 3. The model system - 34 -

Figure 3.4 Location of the study system: A) regional map of the Atherton Tablelands (Australian Map Grid Zone 55, WGD’66) indicating "connected", "isolate" and "continuous" study sites. = sclerophyll forest, = rainforest, B) detailed map of the corridor system showing "corridor" sites, C) photograph of the corridor system highlighting the difference in the vegetation of the corridor and matrix. Chapter 3. The model system - 35 -

Connected R1

Corridor L (75m)

Corridor BR (20m)

Study site location within connected remnant Corridor K/H (60m) Study site location within corridor (number in parentheses denotes corridor width) Corridor K/S (87m) Study site location within the matrix Corridor S (79m) Stock crossing Connected M1 1 km

Figure 3.5 The location of connected, corridor and matrix sites within the corridor system.

Sites within the continuous forest were selected to match the conditions of the corridor as closely as possible. Pairs of sites were located along a creek-bed and within the same drainage basin. The selection of continuous sites represented the only instance in which a deviation from the ideal design outlined in Section 3.1 was required as steep terrain prevented the distance between pairs of sites (2.6km and 2.4km) being equal to the length of the corridor (~4.5km). Despite being less than the ideal distance, the distance between the sites within the continuous habitat is still well in excess of the reported movement distances for individuals of either species.

3.2.2 The study species

In previous studies, little emphasis has been placed on the criteria for species selection in investigating wildlife corridor effectiveness. Apart from matrix intolerance, the species should be present within all components of the system (connected patches, the corridor, isolates and continuous sites). While this is not necessary for connectivity to occur, it is required in a study assessing connectivity to Chapter 3. The model system - 36 - enable investigation of the ecological aspects of the populations within the corridor and how this may affect connectivity. Further, where population genetics is incorporated into the project design, the species should have a relatively short generation time. This allows many generations post-clearing so that any effect of fragmentation on gene flow within the modified system can be detected.

In the present study, two native rodent species fulfil these requirements: the fawn- footed melomys (Melomys cervinipes) and the giant white-tailed rat (Uromys caudimaculatus). These two species have very similar life history strategies yet differ in their movement and dispersal capabilities with respect to distance.

Melomys cervinipes is a small (35-150g, Watts and Aslin 1981) rodent from the family Muridae that is found only along the eastern seaboard of Australia. The species is restricted to closed forests in the northern part of its range, which includes the Atherton Tablelands, and movement from the closed forest into surrounding open pastures is rare (Watts and Aslin 1981, Strahan 1983). While the gestation period of M. cervinipes (38 days) is long relative to other small native rodents (Watts and Aslin 1981), the young develop rapidly and are known to reproduce within their first year (Watts and Kemper 1989). Litter sizes range from one to four and females have been recorded as producing many litters in a single year (Watts and Aslin 1981).

Little is known about the home range size of M. cervinipes, however the available data suggests that movement is minimal. Wood (1971) estimated home ranges of approximately 0.25ha while Smith (1985) found the average distance between captures of males to be greater than females at 71.5m. These data however, were obtained from the southern end of the species natural range, some 2000km from the present study site, and from within habitat types other than tropical rainforest.

Uromys caudimaculatus is a close relative of M. cervinipes (Watts and Kemper 1989) and is Australia’s largest rodent at 665-1000g (Watts and Aslin 1981). The species is restricted in distribution to the northern east coast of Australia and predominantly resides within tropical wet sclerophyll forests and rainforests although excursions into campgrounds and buildings have been recorded (Strahan 1983). Females become sexually mature at approximately 10 months, have a gestation period of 36 days and a litter size of one to three (Watts and Aslin 1981). Once Chapter 3. The model system - 37 - again little is known of home range size or dispersal potential, however the one published record of movement indicates overnight travel of 500m (Wellesley- Whitehouse 1981) suggesting that home range size and/or dispersal capabilities differ greatly from that of M. cervinipes.

In addition, both M. cervinipes and U. caudimaculatus are ideal candidate species as wildlife corridors have been postulated as a management strategy suitable for their conservation within the Atherton Tablelands (Campbell 1996) as well as for M. cervinipes in other areas of Australia (Cox et al. 2003). Further, in other independent studies, M. cervinipes is one of the species being used to assess the effectiveness of constructed corridors on the Atherton Tablelands (N. Tucker [Director, Centre for Tropical Restoration, Queensland National Parks and Wildlife]).

Much debate exists over the choice of species to be used in studies of conservation issues including wildlife corridors. Laurance and Laurance (1999) called for future research on wildlife corridors to focus on species vulnerable to habitat fragmentation that are most likely to benefit from corridors while Doak and Mills (1994) highlighted several of the problems involved with working on endangered species. In this study, the selected species are locally abundant yet have been shown to be vulnerable to fragmentation. The challenge to researchers who take this approach lies in formulating general concepts that are applicable to a wide suite of species including those that are endangered.

3.2.3 The conservation significance of the study region The wet-tropics rainforests, of which the Atherton Tablelands are a part, have the richest fauna in Australia (Rainforest Conservation Society of Queensland, 1986). Not only does the area contain a high proportion of the total of Australia (eg 62% of all butterfly species, 60% of all bats, 30% of all marsupials and 30% of all frogs) (Rainforest Conservation Society of Queensland, 1986), the region is also renowned for its high degree of endemism. For example, 54 species of vertebrates alone are known to occur only within the wet-tropics and the area has the highest number of endemic rainforest mammals of any region in Australia (Winter 1991). Of this larger region, the Atherton Uplands has the highest species richness of vertebrates in the wet tropics (Williams et al. 1996). Due to the high levels of species richness and endemism, and the extensive level of habitat modification which threatens these same values, the Atherton Tablelands is an area of high Chapter 3. The model system - 38 - conservation significance. In an attempt to conserve the natural values of the area, several wildlife corridors have been constructed in the region by government authorities and public organizations within the last decade. A further 41 corridors have been identified as priority linkages by the Wet Tropics Management Authority for the greater Wet Tropics area (Wet Tropics Management Authority 2003). The study site for this project was selected purely on the appropriateness of its spatial configuration, however, its location within an area of conservation significance adds an applied aspect to the project in addition to the primary aim of a theoretical consideration of ecological processes within a model corridor system.

3.2.4 Pleistocene refugia As a result of their studies on the genetic structure of populations from the wet Tropics of northern Queensland, Cunningham and Moritz (1998), warned about the potentially confounding effect of historical events on the interpretation of current gene flow patterns. Climatic change during the Pleistocene resulted in contraction of the rainforest of the wet-tropics to small refugia (Webb and Tracey 1981, Winter 1997). The now continuous rainforest that defines the eastern and southern border of the Atherton Tablelands is thought to have been disjunct throughout this period of contraction with Webb and Tracey (1981) suggesting that refuges were located around Mt Bartle Frere to the east and the Cardwell Range to the south. This historical isolation was found to have exerted a much greater effect on the genetic structure of some populations of lizards from sites on the Tablelands than had human alteration of the landscape (Cunningham and Moritz 1998). Populations within the eastern range varied considerably from those within the central and southern Tablelands with the central/southern regions sharing only one of 16 haplotypes. As a result, they cautioned against misinterpretation of results due to this potentially confounding factor. A component of the present study examines the genetic structure of animals within sites from the same region studied by Cunningham and Moritz but historical isolation is not considered to be an issue of concern for two reasons. Firstly, the results of Campbell (1996), in a study similar in scope to that of Cunningham and Moritz, do not indicate regional variation to the same extent for either M. cervinipes or U. caudimaculatus. In his study, 38% of mtDNA haplotypes from M. cervinipes were shared between the central/southern and eastern regions. While this figure dropped to 19% for U. caudimaculatus, this

could be attributable to the largely disparate sample sizes (ncentral/southern = 27; neastern = 214) and the sampling of most central/southern individuals from populations within Chapter 3. The model system - 39 - small rainforest remnants (24 of 27) which rapidly lose the rarer haplotypes. Secondly, all inferences regarding gene flow in the present study will be based on paired site comparisons with both sites of each pair located within the same region of the Tablelands.

3.3 Methods

Although the selection of appropriate sites has been emphasized as a necessary criteria for the assessment of corridor effectiveness, it was not deemed necessary to investigate the vegetation attributes at all sites. The history of land use on the Atherton Tablelands is well documented (Winter et al. 1987) and thus, the small isolated patches of rainforest that remain are known to be remnants of the rainforest that once covered the entire region. Such vegetation is still present in Wooroonooran National Park, in which the continuous control sites are located. The vegetation within the isolate and continuous sites was therefore not assessed as, in the absence of human modification, the vegetation is expected to represent the original rainforest type. Similarly, the remnant sites connected by the corridor were also expected to represent the original state, however the vegetation in these connected patches was assessed to provide a reference point against which the vegetation of the corridor could be compared. Hence, the isolate and continuous sites are only used as controls for the comparison of population genetic structure within this study and are not mentioned further until Chapter 5.

3.3.1 Vegetation description A general description of vegetation community type along the corridor was obtained by walking the length of the corridor. Vegetation was categorized on the basis of both species composition and structural attributes with all vegetation in the corridor being assigned to one of the following classes: Category 1: Rainforest type 5A, 5B or 1A (Tracey 1982) - representative of the pre-fragmentation vegetation. Sparse lower-strata, minimal to moderate mid-stratum, dense canopy. Category 2: Predominantly rainforest (as in Category 1) with the presence of some introduced or early successional species eg lantana (Lantana camara), wild tobacco (Solanum mauritianum) and stinging trees Chapter 3. The model system - 40 -

(Dendrocnide moroides). No differences in structure from Category 1 were apparent. Category 3: Lantana dominated providing a dense low/mid-strata with a sparse overstorey of Category 1 trees. Category 4: Revegetated creekbank. Acacia spp dominated overstorey with a bare/grassy understorey. Rainforest trees scattered throughout the overstorey. Category 5: Assorted shrubs, long reeds and trees (not rainforest or lantana dominated) Category 6: with scattered trees and long reeds Category 7: Creekbank consisting of long native grass backing onto bracken (Pteridium esculentum), lantana and some established rainforest trees. Very dense lower-stratum but mid-stratum absent and sparse canopy.

3.3.2 Vegetation structure and composition Structural and floristic attributes of the riparian vegetation were assessed at five sites along the corridor and five random sites along the creek within each of the connected patches.

At each corridor site, remnant vegetation on each side of the creek was divided into five parallel strata based on distance from the creek (0-5m, 5-10m, 10-15m, 15-20m and >20m). Sampling was stratified in this manner to account for any changes in vegetation attributable to factors associated with distance from the creek (eg slope or soil moisture content).

Ground, vertical and canopy cover were determined at five random locations within each available belt transect (50m x stratum width). To determine ground cover, a digital photo of the ground was taken from a height of 1.5m. The image was overlaid with a grid of 100 points and the percent ground cover calculated by scoring the number of grid intersects covered with vegetation. Vertical cover at ground level was determined by photographing a vertically placed 1m2 board upon which 100 grid intersects were marked. Photographs were taken through vegetation from a constant distance so that the board filled the image and the percent cover determined as the number of grid intersects covered by vegetation. Percent foliage cover (PFC) was estimated as ([1- visible sky]*100) from an image taken with a fish- Chapter 3. The model system - 41 - eye lens. The visible sky value was obtained using the canopy analysis software HemiView 2.1 (1999).

The diversity and abundance of species present within the corridor were estimated at one random location within each available 50m x strata width belt transect. All fruit and present in the leaf litter and the O horizon of the soil within a 0.5m x 0.5m quadrat were collected during August 2002. This period of the year coincides with the period of peak fleshy fruit abundance and sampling was undertaken during this time to maximise the number of species represented (Elmouttie pers. comm.). Data was recorded as the total number of fruits per quadrat for each species.

The point-centre quarter technique (Krebs 1989) and diameter at breast height (DBH) were used to determine stem density (as per Pollard 1971) and basal area in three height classes (1-4m, 4-10m and >10m) at five random locations within each available 50m x ½ corridor width belt transect. Absolute stem density was calculated using the same data but considering the nearest tree (>1m) in each quadrant regardless of height category. Absolute basal area was calculated as absolute stem density x mean basal area of all height classes.

Due to the varying width of the corridor, the number of strata available at each site varied. At two sites it was not possible to sample both sides of the creek due to lack of landowner permission. The number of samples resulting from this design for each vegetation attribute is shown in Table 3.1.

Within each connected patch, five random sites were selected along the creekbank and six (50m long) parallel strata were established (0-5m, 5-10m, 10-15m, 15-20m, 20-30m and 30-50m). The addition of an extra strata over the five designated in the corridor was to ensure that the “interior” of the patch was included in the sampling design. The approaches used to determine vegetation attributes were the same as used within the corridor with ground, vertical and canopy cover determined at five random locations within each strata, sampling of vegetation composition determined one random location within each stratum and stem density and basal area determined at five random locations within strata 1-3 and 4-6.

Chapter 3. The model system - 42 -

Table 3.1 Vegetation sampling allocation at each site for GC = ground cover; VC = vertical cover; PFC = percent foliage cover; Dens = stem density, BA = basal area; Spp. = species composition.

# strata per bank Number of sampling locations for : East West GC VC PFC Dens BA Spp. Corridor L 5 - 25252510105 Corridor BR 3 - 15 15 15 5 5 3 Corridor KH 5 5 50 50 50 20 20 10 Corridor KS 5 2 35 35 35 10 10 7 Corridor S 5 5 505050202010 Connected R1-R5* - 6 30 30 30 10 10 6 Connected M1-M5* - 6 30 30 30 10 10 6 * allocation for each site.

3.3.3 Trapping Small mammal trapping was undertaken every 3 months from February 2002 until March 2003 (five trapping sessions) in: i) the two connected rainforest remnants at the point of intersection with the corridor, ii) five sites within the corridor (with the exception of Corridor KS and KH which were not trapped in February 2002) iii) four sites within the matrix (all trapping sites are shown in Figure 3.5).

All traps were baited with linseed oil soaked cardboard and checked for captures the following morning. Upon capture of an individual M. cervinipes or U. caudimaculatus, the trap location and species were recorded and a uniquely numbered microchip (Compliance No. ISO 11784, Veterinary Marketing Network) was inserted into the nape of the neck of the animal. Upon recapture, the animal was scanned (Pocket Reader EX, Destron Fearing) and the microchip number recorded. The trap location and species of all individuals from other species were also recorded.

The two species of Melomys encountered during trapping (M. cervinipes and M. burtoni) show remarkable morphological similarities with the only discernible differences being the arrangement of the molars in the skull (Keith 1970) and the number of roots on the upper molars (Knox 1978). Due to these similarities, tissue samples were taken from all individuals of M. cervinipes and Melomys burtoni Chapter 3. The model system - 43 - encountered within the matrix and corridor sites for identification using genetic techniques (see Appendix 3 for details of methods). Samples were not collected from all Melomys within connected patches as M. burtoni is known to inhabit only grass dominated habitats and previous analysis had shown all samples from within the rainforest remnants to be M. cervinipes. Although corridor sites are also predominantly closed forest, samples were collected from all Melomys individuals due to the close proximity of the sampling sites to open pastures.

In the connected patches, a permanent trapping grid was established at the patch/corridor boundary. Both trapping grids consisted of a cage trap (20cm x 20cm x 56cm; Mascot Wire Works) and an elliot trap (30cm x 9cm x 10cm; Elliot Scientific) placed at 25m x 25m grid intervals with an additional elliot trap placed between each cage/elliot pair in the east-west direction (Figure 3.6). Trapping continued for eight consecutive nights.

Trapping at sites within the corridor and matrix was undertaken for eight and six nights respectively. A permanent 190m transect was established at each of five sites within the corridor (Corridor L, BR, KH, KS and S; Figure 3.5) with one elliot trap and one cage trap being placed at each of 20 points located 10m apart. Traplines within the matrix (Matrix R, KH, KS and M; Figure 3.5) were located 10m from the corridor/pasture border during the Feb 2002 trapping session and 50m from the boundary on subsequent trips. Temporary electric fences were erected around the trap-lines within the matrix to exclude cattle when necessary. Trapping intensity throughout the study is shown in Table 3.2.

Chapter 3. The model system - 44 -

12.5 m 25 m

25 m

Figure 3.6 Representation of the trapping grid at connected sites. O indicates the location of a cage and elliot trap; + indicates an elliot trap only.

Table 3.2 Trapping intensity shown as the number of trap nights (cage, elliot) per site per trip.

February May August November February Total Combined total Connected sites Connected R 351, 648 312, 576 312, 576 312, 576 312, 576 1599, 2952 Connected M 312, 576 312, 576 312, 576 312, 576 312, 576 1560, 2880 3159, 5832 8991 Corridor sites Corridor L 160, 160 160, 160 160, 160 160, 160 160, 160 800, 800 Corridor BR 160, 160 160, 160 160, 160 160, 160 160, 160 800, 800 Corridor KH - 160, 160 160, 160 160, 160 160, 160 640, 640 Corridor KS - 160, 160 160, 160 160, 160 160, 160 640, 640 Corridor S 160, 160 160, 160 160, 160 160, 160 160, 160 800, 800 3680, 3680 7360 Matrix sites Matrix R 120*, 120* 120, 120 120, 120 120, 120 120, 120 600, 600 Matrix KH 120*, 120* 120, 120 120, 120 120, 120 120, 120 600, 600 Matrix KS 120*, 120* 120, 120 120, 120 120, 120 120, 120 600, 600 Matrix M 120*, 120* 120, 120 120, 120 120, 120 100, 100 580, 580 2380, 2380 4760 21111 * indicates that traps were located only 10m into the matrix

Chapter 3. The model system - 45 -

3.3.4 Analyses The similarity of corridor sites to the connected patches vegetation was initially assessed by multivariate analyses. Data from the two remnants (Connected R and Connected M) were not pooled as these patches, although once part of a contiguous rainforest, have been described as different rainforest types (Tracey 1982) which may lead to variation between the sites. For this reason, Connected M and Connected R are considered as distinct system components within this chapter rather than being classed together as in the remainder of this thesis.

Seven variables describing the structural attributes of the vegetation were included in the analyses: %ground cover, %vertical cover 0-1m, %foliage cover (PFC), absolute density (individuals per/100m2), absolute basal area (cm/m2), % density (1- 4m) and % density (>10m). Absolute density was expressed as /100m2 rather than the customary /m2 to ensure that all variables were of a similar magnitude in order to prevent any scaling bias towards particular variables (see max/min values in Table 3.3). Further, to ensure the independence of variables, the % relative density of the 4-10m height strata was not included in the analyses as its value is a function of the 1-4m and >10m strata. Data from the outermost strata of corridor sites was omitted from PFC analyses as the image from the fish eye lens included a large proportion of the adjacent pasture, thus severely underestimating the canopy cover of the corridor. Analysis of vegetation structure consisted of: 1. Multidimensional scaling (MDS) to provide a visual representation of the variation between sites based on vegetation attributes. A similarity matrix was generated using the Simplified Morisita’s index of similarity (Horn 1966) as this similarity index is considered the most robust for the type of data generated by this study (Wolda 1981, Krebs 1989).

2. K-means cluster analysis with an a priori level of three groups (Connected R, Connected M and corridor – the major components of the corridor system) and results overlaid on the spatial arrangement of the MDS.

3. Discriminate function analysis (DFA) to determine which variables contributed to the variation depicted by MDS and K-means cluster analysis. All variables expressed as percentages were converted to proportions and arcsine transformed. All three DFA methods (standard, forward and backward) were performed with concurring results. The data presented here uses the forward Chapter 3. The model system - 46 -

step-wise model with a tolerance value of 0.01, an F to enter of 1 and an F to remove of 0. Each variable, including those in the resulting model, was tested for significance between groups by one-way ANOVA. For all analyses on vegetation data, site was considered as the experimental unit with analyses performed using the mean values for each site (n = 5 for each system component).

The variation in resources throughout the corridor system was analysed in a similar way to the vegetation data in that it was described with MDS (using Morisita’s co- efficient of similarity) and K-means clustering. Species diversity was assessed using Renyi’s diversity index to overcome the inconsistencies produced by indices such as Shannon-Wiener and Simpson (Southwood and Henderson 2000). This index is calculated as:

s log pβ ∑ i i=1 Hβ = 1− β

where pi = the proportional abundance of the ith species, log is to the base of choice (in this case, e) and β is the order or scale parameter (β >0, ≠1).

The structural attributes and species composition of the matrix habitat were not included in any analyses as all sites consisted entirely of exotic pasture species. All trapping data were standardized for differential trap effort between the four system components (Connected R, Connected M, corridor and matrix). Mus domesticus were considered capable of being caught in elliot traps only because individuals are able to pass between the mesh grid of the cage traps. Alternatively, elliot traps were not suitable for trapping Perameles nasuta ( 500-900g), Hydromys chrysogaster ( 620-1200g), Uromys caudimaculatus ( 500-900g) and Trichosurus vulpecular (1.5–4kg) due to larger body size (Menkhorst and Knight 2001). Both trap types were considered suitable for trapping all other species.

3.4 Results

3.4.1 Qualitative overview of vegetation within the corridor system

While not as undisturbed as the connected patches, the vast majority of the corridor was found to consist of category 2 vegetation (see Section 3.3.1, Figure 3.7), which Chapter 3. The model system - 47 - consisted primarily of rainforest species. The majority of the remainder of the corridor was covered by lantana. In most instances where this vegetation type was found, Category 2 habitat was located on the opposite side of the creek. Given that rodents can readily traverse fallen logs and swim where necessary, it is reasonable to assume that the corridor consists of a vegetation type that provides a high degree of aerial cover and thus should provide a suitable habitat for residency and/or transit of rainforest rodents (Figure 3.7).

A large degree of variation was found for most vegetation attributes between the system components (Table 3.3). This was especially prevalent for the variables associated with ground layer vegetation (% ground cover and vertical structure 0- 1m) and absolute basal area. Discriminate function analysis identified three variables capable of distinguishing between the three system components (Wilks λ = 0.34, F = 2.4, p = 0.06). These variables were arcsin PFC (partial λ = 0.65, p = 0.12), arcsin vertical structure 0-1m (partial λ = 0.64, p = 0.10) and total basal area (partial λ = 0.81, p = 0.34). When tested using one-way ANOVA’s, only arcsin PFC showed any significant differences (Table 3.3), indicating that only one variable contributed significantly to the differentiation between the groups. Tukeys post-hoc tests showed PFC in the corridor to differ from that of Connected M but not Connected R. Yet, as stated above, a mean PFC of 93% within the corridor still represents considerable canopy cover.

Chapter 3. The model system - 48 -

Figure 3.7 Schematic map of vegetation types within the corridor system and the location of trapping sites along the corridor. See Section 3.3.1 for more detailed descriptions of vegetation categories.

Chapter 3. The model system - 49 -

Table 3.3 Variation in structural attributes of the vegetation within the corridor system. Mean (± s.e.) per site with maximum and minimum values in parentheses.

Connected M Connected R Corridor F, p* Ground cover (%) 17.2 ± 2.1 21.8 ± 1.9 31.5 ± 6.4 2.91, 0.09 (11.3 - 24.1) (17.7 - 28.5) (11.0 - 44.8) Vertical structure 0-1m (%) 14.6 ± 2.9 25.5 ± 3.0 27.9 ± 5.5 3.39, 0.07 (9.1 - 25.2) (16.3 - 24.8) (12.0 - 44.0) P.F.C. (%) 98.0 ± 0.3 97.5 ± 0.3 91.5 ± 2.8 4.14, 0.04 (97.3 - 99.1) (96.5 - 98.5) (82.5 - 97.3) % density 1-4m 57.3 ± 3.5 63.6 ± 3.4 58.5 ± 6.0 0.55, 0.59 (50.0 - 69.0) (52.4 - 70.8) (42.9 - 76.8) % density >10m 16.9 ± 2.0 19.6 ± 1.3 19.2 ± 3.9 0.30, 0.75 (13.0 - 24.5) (15.8 - 23.1) (9.2 - 30.8) Absolute density *100 (No./100m2) 43.8 ± 3.5 36.4 ± 2.2 31.4 ± 10.3 0.95, 0.41 (31.2 - 51.0) (32.1 - 38.4) (6.3 - 62.7) Absolute basal area (cm/m2) 32.2 ± 10.0 51.4 ± 13.7 46.0 ± 11.9 0.69, 0.52 (12.5 - 70.2) (20.6 - 92.7) (10.9 - 84.1) * n=5 for each system component.

While the spatial representation of sites in the MDS and K-means cluster initially suggested that some sites within the corridor component of the system may possess a slightly different vegetation structure (Figure 3.8), discriminate function analysis failed to detect significant differentiation between the system components. This suggests that while the corridor displays a greater degree of variation in vegetation structure than Connected R and M, the corridor component as a whole does not differ to any significant extent from the remnant habitat patches that it connects.

Chapter 3. The model system - 50 -

Figure 3.8 Spatial representation of variation in vegetation structure between system components based on MDS. Broken lines represent groups identified by K-means clustering. r1-5 = Connected R 1-5, m1-5 = Connected M 1-5, c1 = Corridor L, c2 = Corridor BR, c3 = Corridor KH, c4 = Corridor KS, c5 = Corridor S.

3.4.2 Resource availability between habitat types

Fifty-five species of were identified from the sampling locations within the corridor system (Appendix 1), all of which were known rainforest species (Hyland et al. 1999, Williams 1999). Twenty-two species were found within Connected R, 33 from Connected M and 30 from within the corridor, with 21 species (38%) found in more than one habitat component and nine species present in all three system components (16%). Neither the total number of seeds/fruit per site nor the total

species richness per site varied between the system components (No. fruit: F(2,12) =

1.32, p = 0.30, Species richness: F(2,12) = 0.16, p = 0.86).

One species (Aleurites moluccana - candlenut) was present in much greater numbers within two sites at Connected R than any other species/site combination (Appendix 2). Although an important component of the vegetation at Connected R, this species was removed from the dataset used for the Renyi Diversity index and the multivariate analyses on the premise that the high abundance of this one Chapter 3. The model system - 51 - species could potentially influence the between site comparison of the remaining 54 species.

Interpretation of Renyi’s index of diversity is based on the visualization of the curves for each site, or in this instance, system component. As the relative magnitude of the index is dependant upon the magnitude of the scale parameter (Southwood and Henderson 2000), system components represented by curves which do not intersect other curves can be described as significantly more/less diverse from each other. Figure 3.9 indicates that no substantial differences can be discerned between the system components with the corridor having similar species diversity to the vegetation within both connected patches, which themselves were comparable when the suite of species minus A. moluccana was considered.

3.5

3.0

Renyi index Renyi 2.5

2.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Scale of magnitude

Figure 3.9 Species diversity per system component as measured by the Renyi index. ● = Connected R, ■ = Connected M, ▲ = Corridor.

Multidimensional scaling revealed a large spread of sites with no distinct clusters apparent (Figure 3.10). The only noticeable feature of the spatial distribution of the sites is the apparent differences in resources at Corridor S (C5) which partitions separately from all other sites. This is primarily due to the dense stand of lantana Chapter 3. The model system - 52 - present at this site (Figure 3.7) which prevents the growth of native rainforest species in the ground and mid-strata and the low density of rainforest trees in the upper strata. The apparent lack of distinct differences in composition between the habitat components was also suggested by K-means clustering (n = 3) which clustered 13 of the 15 sites in the one cluster (Cluster 1: Connected M2, Cluster 2: Corridor 5, Cluster 3: Connected R1-5, Connected M1 & M3-5, Corridor 1-4) (Figure 3.10). These results concur with those of the vegetation structure and suggest that the three system components do not differ greatly from each other.

1.4

1.2 m5 m2 1.0

0.8 m1 0.6 r5 c2 0.4 c1 c5

2 0.2

n r2

o

i s

n 0.0

e r1 m

i -0.2 m4 D

-0.4 c3 -0.6 r4 c4

-0.8 m3

-1.0 r3 -1.2

-1.4 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Dimension 1

Figure 3.10 Spatial representation of variation in resources between habitat types based on MDS. Broken lines represent groups identified by K-means clustering. r1-5 = Connected R 1-5, m1-5 = Connected M 1-5, c1 = Corridor L, c2 = Corridor BR, c3 = Corridor KH, c4 = Corridor KS, c5 = Corridor S.

3.4.3 Differential use of the system components by small mammals

Along with the three system components considered in the vegetation analyses, trapping was also conducted within the matrix. The number of small mammals species trapped and the suite of species present, was similar within the matrix habitat to within the corridor and connected patches (matrix = 8 species, Connected Chapter 3. The model system - 53 -

M = 8, Connected R = 8, corridor = 10), (Figure 3.11). However, as expected, the relative abundance of individuals differed considerably between the forested and matrix habitats (Figure 3.11). In addition to those mammal species shown in Figure 3.11, individuals of the black rat (Rattus rattus: Connected M and corridor), the swamp rat (Rattus lutreolus: corridor) and the northern brown bandicoot (Isoodon macrourus: Connected R and matrix) were also captured but in very low numbers.

Captures within the matrix were predominantly of the introduced house mouse (Mus domesticus) or of native species that are known habitat generalists such as the long-nosed bandicoot (Perameles nasuta) and the canefield rat (Rattus sordidus). The suite of species captured predominantly within the connected patches and corridor was similar and consisted primarily of rainforest or woodland specialists such as M. cervinipes, U. caudimaculatus, bush rats (Rattus leucopus), water rats (Hydromys chrysogaster), and the coppery brushtail possum (Trichosurus vulpecula). Four individual M. cervinipes were trapped within the matrix on trip 1 when transects were located 10m from the corridor (ntrap nights = 960) (Figure 3.11). However, as expected from their known habitat preferences, neither U. caudimaculatus nor M. cervinipes were trapped within the matrix when traplines were located 50m from the corridor-matrix boundary. Captures of Melomys at traplines 50m in the matrix were restricted to the grassland melomys (M. burtoni).

Chapter 3. The model system - 54 -

1.0 a) Connected R H.c.: Hydromys chrysogaster (15) M.c.: Melomys cervinipes* (508)

0.8 U.c.: Uromys caudimaculatus (137) T.v.: Trichosurus vulpecula (77) R.l.: Rattus leucopus (559)

0.6 P.n. : Perameles nasuta (85 ) M.b.: Melomys burtoni (43) R.s.: Rattus sordidus (30)

0.4 M.d.: Mus domesticus (229)

0.2

t

n

e

n

o p b) Connected M

m

o

c

t

a

t

i

b

a

h

h

c

a

e

n

i

h

t

i

w

d

e

p

p

a

r

t

1.0

)

s

e

i c) corridor

c

e 0.8

p

s

r

e

p

( 0.6

s

l

a

u

d i 0.4

v

i

d

n

i

f

o 0.2

n

o

i

t

r

o

p

o

r P d) matrix

Species

Figure 3.11 Captures of mammalian species expressed as the proportion of individuals per species trapped within each habitat component when corrected for trap effort. Numbers in parentheses denote the total number of captures. * all captures in matrix at 10m transect. Chapter 3. The model system - 55 -

3.4.4 Differential habitat use by M. cervinipes and U. caudimaculatus Trap success at all sites within the corridor and connected patches was very low, with a maximum value at a corridor or connected site of 6.88% and 4.81% for M. cervinipes and U. caudimaculatus respectively (Table 3.4). These values represent the capture of unique individuals for each trip and do not include any recaptures in a given trapping session. It does however, allow for the inclusion of the same individual in more than one trip. Trap success for M. cervinipes did not vary across

month or site (random block ANOVA, site: F(6,22) = 2.37, p = 0.07, month: F(4,22) = 1.69, p = 0.19) however significant variation was detected between sites for U. caudimaculatus (random block ANOVA, site: F(6,22) = 7.45, p = <0.001, month: F(4,22) = 0.63, p = 0.65). Connected R and Corridor L were found to both have significantly higher trap success than Corridor KS, S and M.

Table 3.4 Percent trap success for unique individuals per site for each trapping session. Numbers in parentheses indicate the number of individuals caught and the total number of trap nights.

February May August November February Total M. cervinipes Connected R1 2.20 (22,999) 2.02 (18,888) 2.25 (20,888) 0.90 (8,888) 1.58 (14,888) 1.43 (65,4551) Corridor L 0.93 (3,320) 0.63 (2,320) 1.56 (5,320) 1.25 (4,320) 0.63 (2,320) 0.68 (11,1600) Corridor BR 1.88 (6,320) 1.88 (6,320) 0.00 (0,320) 1.56 (5,320) 1.88 (6,320) 1.44 (23,1600) Corridor KH - 6.88 (22,320) 4.69 (15,320) 2.19 (7,320) 0.31 (1,320) 2.97 (38,1280) Corridor KS - 5.63 (18,320) 4.38 (14,320) 1.56 (5,320) 0.31 (1,320) 2.66 (34,1280) Corridor S 0.31 (1,320) 1.56 (5,320) 0.94 (3,320) 0.94 (3,320) 1.25 (4,320) 0.75 (12,1600) Connected M1 0.56 (5,888) 0.79 (7,888) 1.13 (10,320) 0.56 (5,320) 2.36 (21,888) 1.04 (46,4440) U. caudimaculatus Connected R1 2.28 (8,351) 3.85 (12,312) 1.28 (4,312) 1.60 (5,312) 4.81 (15,312) 2.19 (35,1599) Corridor L 3.75 (6,160) 3.13 (5,160) 3.75 (6,160) 2.50 (4,160) 1.88 (3,160) 2.63 (21,800) Corridor BR 1.88 (3,160) 1.88 (3,160) 1.88 (3,160) 1.25 (2,160) 0.00 (0,160) 1.25 (10,800) Corridor KH - 1.88 (3,160) 0.63 (1,160) 1.25 (2,160) 1.25 (2,160) 0.94 (6,640) Corridor KS - 0.63 (1,160) 0.63 (1,160) 0.00 (0,160) 1.25 (2,160) 0.47 (3,640) Corridor S 0.63 (1,160) 0.63 (1,160) 0.00 (0,160) 0.63 (1,160) 0.00 (0,160) 0.38 (3,800) Connected M1 0.64 (2,312) 0.32 (1,312) 0.96 (3,312) 0.64 (2,312) 1.28 (4,312) 0.64 (10,1560)

When the total number of unique individuals trapped per system component (Connected R, Connected M and corridor) was considered, neither M. cervinipes nor U. caudimaculatus were spread uniformly across the corridor system (M. cervinipes: 2 2 χ (2) = 6.47,p = 0.039, U. caudimaculatus: χ (2) = 15.7,p < 0.001). For M. cervinipes, this significance was due to an over-representation within the corridor and an under-representation within Connected M, with Connected R being Chapter 3. The model system - 56 - equivalent to both other components. Conversely the observed and expected numbers of U. caudimaculatus within the corridor were comparable and the significant difference was due to greater than expected numbers in Connected R and fewer individuals within Connected M. However, when the two connected sites were pooled and compared to pooled data from the corridor, the number of individuals of both species did not differ between the habitat types (M. cervinipes: 2 2 χ (1) = 3.69, p = 0.06, U. caudimaculatus: χ (1) = 1.56, p = 0.22).

Both species also showed disparate use of sites within the corridor (M. cervinipes: 2 2 χ (4) = 40.38, p < 0.001, U. caudimaculatus: χ (4) = 20.01, p = 0.005) (Figure 3.12). No relationship between the number of individuals of each species per site was 2 apparent (r (1,5) = 0.11, p = 0.47) when corrected for trap success suggesting that the presence of one species had little impact on the numbers of the other species at a given site. Hence, although not evenly distributed, the three forested components of the corridor system are all capable of supporting populations of both M. cervinipes and U. caudimaculatus.

35

30

25

20

15

10

5 Number of individuals per 1000 trap nights

0 Conn R Corr L Corr BR Corr KH Corr KS Corr S Conn M Site

Figure 3.12 Number of individuals (per 1000 trap nights) per site [M. cervinipes (■) and U. caudimaculatus (□)]. Chapter 3. The model system - 57 -

A trend was apparent in the number of U. caudimaculatus individuals caught at corridor sites, with numbers being proportional to the proximity of the site to Connected R. As expected, given the relative area covered by traps, Connected R had a greater number of individuals than any corridor site (Table 3.4). However, when expressed as trap success, the highest value was obtained at Corridor L, suggesting a greater density of animals (Table 3.4, Figure 3.12). Trap success and the total number of individuals both decreased in the remaining corridor sites as distance from Connected R increased. While the relationship between the geographic distance from Connected R and the total number of individuals per site is significant (n = 5, r2 = 0.85, p = 0.03), the limited number of sites within the corridor did not allow the potential causes for this pattern to be explored.

3.5 Discussion

The comparative assessment of physical structure and species composition between the corridor and the connected patches both indicated that the vegetation within the corridor was representative of that within the connected patches. Sites from Connected R, Connected M and the corridor all showed considerable variation suggesting that while the vegetation within the corridor may be more heterogeneous due to factors such as and the intrusion of exotic species, the connected patches also display a degree of heterogeneity. The only significant difference detected in either structure or composition was for percent foliage cover with the corridor having significantly less cover than Connected M. Given that the corridor still offered an average of 93% canopy cover, this factor was not expected to affect rodent use of the corridor.

Comparison of vegetation composition, based on the presence of fruit and nut resources also indicated a continuity of vegetation type along the corridor. Corridor S stood out as the only significantly different site with a low and non-diverse resource base primarily due to the infestation of lantana on one of the creek-banks and the low density of rainforest trees. This site, and Corridor BR, presented the only sites where differences in vegetation type could present a barrier to movement. Corridor BR is the narrowest section of the corridor, contains the lowest resource levels and consists largely of replanted vegetation to supplement the pre-existing vegetation. If these sites were found to be capable of supporting rodents, either as Chapter 3. The model system - 58 - residents or by the provision of adequate resources to allow transit, then the remaining vegetation results suggest that the habitat within the corridor should be suitable for use by rodents and that the corridor should be suitable for investigation as part of the model system.

Interestingly, of the five trapping sites within the corridor, Corridor BR contained the second highest number of both M. cervinipes and U. caudimaculatus, clearly indicating that the change in vegetation type was not a deterrent to either species. Conversely, Corridor S was quite depauperate in both species. However, their presence, albeit in low numbers, indicates that the site does not prevent a break in the continuity of suitable habitat along the corridor. Admittedly, trapping could not be conducted along the entire length of the corridor. However, the vast majority of the corridor consists of a vegetation type similar to that of at least one of the trapping sites that has been shown to be suitable for use by the target species and the length of the corridor is thus considered suitable for use by the target species.

In assessing the suitability of a corridor for use in a model system, the differential use of the system components is as crucial as the demonstrated use of the corridor. As is expected when habitats display vastly different vegetation attributes, the relative abundances of the suite of animals found within matrix and corridor/patch habitats differed greatly with only one generalist species, the long-nosed bandicoot (Perameles nasuta) showing little preference for either habitat type. Those animals located predominantly within the rainforest vegetation comprised native species with known affinities for rainforest or heavily wooded habitats (M. cervinipes, U. caudimaculatus, Hydromys chrosogaster and Trichosurus vulpecula). The near absence of these species within the matrix, suggests that the pasture habitat presents a barrier to movement and that all could potentially benefit from the inclusion of wildlife corridors within the landscape configuration. This is especially pertinent for U. caudimaculatus with no individuals captured within the matrix and M. cervinipes only being captured within the matrix when traps were at a distance of 10m into the pasture.

The presence of M. burtoni (Grassland melomys) within the corridor initially presented an unexpected potential barrier to the continuity of suitable habitat. While M. burtoni was expected to occur within the corridor system, it was assumed that the pastures would be the primary habitat for the species given its preference for Chapter 3. The model system - 59 - grassland habitats. It has been reported that the differential habitat use of the two Melomys species is due to competitive exclusion (Watts and Aslin 1981). However, the presence of M. cervinipes at Corridor BR showed that the two species can live sympatrically and that the presence of M. burtoni within the corridor should not hinder the progression of M. cervinipes individuals or genes along the corridor.

Some individuals of M. cervinipes were located within the matrix at a distance of 10m from the corridor/matrix boundary suggesting that this species is not totally averse to venturing into areas lacking aerial cover. While they may have been enticed into the open by the presence of bait, it demonstrates that given the motivation, gaps of an equivalent distance within the corridor in the form of stock crossings and a minor road (less than 10m wide) should not inhibit movement of individuals. Similarly Goosem (2001) found that where rainforest abutted the road verges, both U. caudimaculatus and M. cervinipes frequently crossed up to 20m in width.

At the commencement of this study, U. caudimaculatus was selected as a candidate species due to its known preference for rainforest habitat. A recent publication (Vernes 2003) has shown that the species also resides within Allocasuarina and wet sclerophyll forests on the Atherton Tablelands, however both of these habitat types offer high levels of structure and aerial cover and do not suggest that the species can tolerate open habitats. The species is therefore still considered to be vulnerable to the isolating effects of extreme habitat fragmentation and a suitable candidate species for the investigation of wildlife corridor effectiveness over the spatial scales under consideration in this study.

In order to be appropriate for use as a model system, a wildlife corridor must fulfil several criteria. The Gwynne Creek corridor meets those criteria identified from the literature, namely the appropriate spatial configuration and the differential use of the corridor and the matrix by the target species. It also satisfies the additional criteria identified as necessary including the continuity of suitable vegetation along the corridor length, the ability of the target species to utilize the corridor and the connected patches and the location of appropriate control sites within a uniform matrix. The Gwynne Creek corridor system is therefore considered suitable for use as a model system in which to investigate the effectiveness of a corridor to facilitate connectivity. This assessment will involve the incorporation of several of the design Chapter 3. The model system - 60 - criteria identified during this study including the location of trapping sites at the corridor/patch boundary and trapping at several locations within the corridor. - 61 -

4. POPULATION STRUCTURE AND MOVEMENT OF MELOMYS

CERVINIPES AND UROMYS CAUDIMACULATUS WITHIN THE

CORRIDOR SYSTEM.

4.1 Introduction

It is through disruption to the process of dispersal that fragmentation can lead to the isolation of populations within remnants. Dispersal, or the movement of an individual away from its natal home range (Stenseth and Lidicker 1992) can enhance the fitness of an animal by changing the habitat in which it lives or the individuals with which it associates (Cockburn 1992). Dispersal can also benefit both the source and recipient populations by altering sex ratios, age structures, social dynamics and the genetic constitution of populations (Stenseth and Lidicker (1992). The dispersal process consists of three phases: emigration, travel and immigration (Lidicker and Stenseth 1992), all of which can be affected negatively by encountering a hostile habitat component as is likely during dispersal within a highly fragmented landscape. Emigration, the initial phase of dispersal, involves the movement of individuals away from the home territory. When an animal is motivated to disperse but is prevented from doing so by inadequate resources within the surrounding matrix habitat, dispersal may become frustrated (Lidicker and Stenseth 1992, Adler and Levins 1994), resulting in increased population density and pressure on social structures and resource availability in the natal habitat (Cockburn 1992). Should emigration occur, the dispersing animal often faces greater risks within an unsuitable habitat, such as increased predation and decreased food and/or shelter resources. This reduces the chances of successful transfer to a destination population. Finally, the dispersing individual, having reached an alternate habitat, must be integrated into the recipient population for dispersal to affect the recipient population. For successful dispersal to be accomplished, the individual must enter the breeding population and produce viable offspring (Barton 1992). Just as social factors can be the cause of involuntary emigration, they can also impede immigration and integration (Brandt 1992). Intolerance by conspecifics and non-acceptance into the recipient population may result if the individual has moved to a population experiencing shortages of space or resources as may occur in the remnant patches created by habitat fragmentation. Chapter 4. Population structure and movement within the corridor - 62 -

Wildlife corridors have the potential to assist the process of dispersal by reducing the chances of encounter with a hostile habitat. For this potential to be realized, it is necessary that individuals likely to contribute to subsequent generations elect to utilize the corridor. Soulé and Gilpin (1991) suggest that in general, individuals entering a corridor will represent a non-random subset of those individuals within the connected patches. The individuals most likely to enter the corridor will be “breeders” searching for suitable sites for reproduction, dispersing juveniles and surplus individuals unable to find a suitable territory within the patch or aged individuals who have been displaced from their usual territory. Where individuals entering the corridor consist of breeders or juveniles, connectivity between patches has the potential to occur. However if corridor individuals consist only of displaced aged animals or surplus individuals who may be socially excluded, connectivity is highly unlikely and the corridor may confer no benefits to the populations of the connected patches despite individuals of the species residing within or moving along the wildlife corridor.

The issue of sex biased use of the corridor is particularly important where the length of the linkage is beyond the expected movement capabilities of individual animals. In these instances, the movement of animals and/or the flow of genes between the patches is reliant on generational movement along the corridor whereby the offspring of individuals (or generations subsequent), rather than the original dispersing individuals themselves immigrate into the destination patch. Soulé and Gilpin (1991) further suggested that individuals within the corridor may have skewed sex ratios, especially in those species that exhibit sex-biased dispersal as occurs in many mammals. Where ratios are highly skewed, the corridor may be ineffective as breeding would be unlikely to occur. In addition, successful recolonization of vacant patches following a local extinction requires dispersal of both sexes along the corridor length.

While several studies note the sex of animals located within corridors (eg. Bennett 1990), the differential movement between patches by the sexes (eg. La Polla and Barrett 1993, Aars and Ims 1999, Mabry and Barrett 2002) or the sex ratios of populations in experimentally connected versus isolated patches (eg. Ims and Andreassen 1999, Davis-Born and Wolff 2000), very little attention has been paid to the age or sex ratios of individuals within corridors relative to those within the remnant and how this may impact upon the likely effectiveness of the corridor under Chapter 4. Population structure and movement within the corridor - 63 - investigation (but see Downes et al. 1997a,b). Similarly, while most studies report the length of the study corridor, few relate this to the expected movement distances of individuals and thus it is not known whether direct or generational movement is the most likely means of connectivity.

The distinction between direct and generational movement between the connected patches also has implications for the resource base required within the corridor. Where generational movement is required to obtain connectivity, the potential effectiveness of the corridor is dependant upon the organism being able to complete its entire life cycle within the corridor. The corridor must therefore provide the resources required for all stages of the life cycle (Csuti 1991, Saunders and Hobbs 1991, Harrison 1992) and these resources may be considerably different to those required for temporary shelter during direct dispersal along corridors (Beier and Loe 1992). The increased resource requirements necessary to enable generational gene flow within long corridors make them more difficult to construct. However, they are necessary in landscapes that have experienced extensive clearing of the natural vegetation or where the target organism has limited movement capabilities relative to the scale of patchiness within the landscape (Bennett 1999). Corridors that are beyond the expected movement scale of individuals but fail to provide adequate resources for residency and breeding may in turn act as “sinks” and be detrimental to the populations within the connected patches (Soulé and Gilpin 1991).

Debate exists within the literature over whether linkages that contain permanent residents should be termed wildlife corridors. While Bennett (1999) cites generational gene flow, which requires animals to reside within the corridor, as the most effective means of transferring genes, and Merriam and Saunders (1993) state that if possible, wildlife corridors would have all the features of a breeding habitat, Lidicker (1999) diametrically opposes these viewpoints. Instead, he proposes that corridors be considered solely as places that facilitate the movement of animals between local populations and suggests that any strips of habitat that support breeding populations be termed “linear habitat patches”. This opinion is based on the potential for resident individuals to disrupt dispersal, and while Lidicker considers only the movement of individuals and not the transfer of genetic material, his concept of a wildlife corridor precludes the possibility of gene flow via generational means. Adoption of his restrictive definition would mean that “wildlife corridors” could never exist in those situations where the distance between patches Chapter 4. Population structure and movement within the corridor - 64 - is beyond the movement capabilities of the individual. This viewpoint once again raises the possibility of the one landscape configuration being termed a wildlife corridor only on a species specific basis depending on movement capabilities of the species. Further, it presents the scenario that those “corridors” that best meet the resource requirements of a species, and are therefore most likely to facilitate movement and/or gene flow between the connected patches (Bennett 1999), are also the spatial configurations that would be deemed inappropriate to be termed “wildlife corridors” as they are most likely to contain resident individuals. Here I take the view that if a physical structure meets the definition of a corridor based on spatial configuration, it is deemed to be a wildlife corridor regardless of the residency status of the individuals contained within.

This chapter investigates ecological aspects of the target species within the model system. Whilst these data assess corridor use and do not provide evidence for any functional connectivity between the populations within the connected patches, they can provide an indication as to whether interaction between the populations may be expected to occur. In particular, the following will be investigated: i) demographic parameters of individuals within the corridor relative to the connected patches, ii) recapture rates for individuals of both species, iii) the likely mode of movement (direct versus generational) based upon the degree of movement by individuals within the system These data will be integrated to assess the potential that the corridor has for increasing connectivity between the populations within the connected remnants.

4.2 Methods

4.2.1 Trapping regime

Mark-recapture trapping was undertaken every three months from February 2002 until March 2003 (see Section 3.3.3). Upon initial capture of a Melomys cervinipes or Uromys caudimaculatus, a microchip was inserted into the nape of the animal and the following data were recorded: trap location, weight, sex and sexual condition of the individual (recorded as mature female [perforate], reproducing female [extended nipples or other visible signs of pregnancy], immature female [imperforate], mature male [scrotal testes] or immature male [abdominal testes]). Chapter 4. Population structure and movement within the corridor - 65 -

These data, and the microchip number were recorded for the first capture of each individual for each trip with only trap location and the microchip number recorded for subsequent captures within any given trip. Only ten of 220 M. cervinipes individuals and one of 86 U. caudimaculatus were found to have lost chips upon subsequent recapture (identifiable from the ear biopsy performed upon initial capture to obtain tissue for genetic analysis; see Chapter 5). These animals were re-chipped and were treated as a new individual in the movement database as it was not possible to determine the original identification of each animal. Re-chipping occurred on trips two and three which allowed sufficient time for re-capture of these few individuals on subsequent trips.

4.3 Results

4.3.1 Trap intensity Trapping resulted in the capture of 220 individuals of M. cervinipes on 508 occasions and 137 captures of U. caudimaculatus consisting of 86 individuals. Trap availability was not a limiting factor in the number of rodents trapped with the majority of traps remaining empty and unsprung on a high proportion of trapping nights (Figure 4.1). Both elliot and cage traps were considered as a potential trapping source for M. cervinipes. At least 78% of traps remained available for potential use by M. cervinipes on 80% of the nights trapping was conducted and trap availability was below 50% on only 1.1% of nights. Due to size constraints, only cage traps were suitable for capture of the larger species (U. caudimaculatus). Similar figures were obtained for this larger species with 74% of traps available on 80% of nights and trap availability falling below 50% on only 1.5% of trapping nights.

For the majority of trapping sessions at each site, no new individuals were caught on the last night of trapping for either species (M. cervinipes 58%; U. caudimaculatus 76%) (Table 4.1). This proportion increased greatly upon inclusion of those trapping sessions when only one or less new individuals were trapped on the final night at any site (M. cervinipes 85%; U. caudimaculatus 100%). This, along with high trap availability, indicates a high probability of capture for those individuals present within the trapping grid over the eight night trapping period.

Chapter 4. Population structure and movement within the corridor - 66 -

160

140

120

100

80

60 Number of observations 40

20

0 1-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100

% trap availability

Figure 4.1 Trap availability per night expressed as the percentage of traps remaining unutilized or unsprung for M. cervinipes („) and U. caudimaculatus ( ).

Table 4.1 The proportion of new individuals caught on the final night of trapping

Month Site February May August November February M. cervinipes Connected R 0.08 (2,24) 0.13 (3,24) 0.05 (1,21) 0 0 Connected M 0.20 (1,20) 0.25 (2,8) 0.09 (1,11) 0.20 (1,5) 0 Corridor L 0 0 0 0 0 Corridor BR 0 0.14 (1,7) - 0.20 (1,5) 0 Corridor KH * 0 0 0.29 (2,7) 0 Corridor SK * 0 0.07 (1,15) 0.20 (1,5) 0 Corridor S 0 0.20 (1,5) 0 0 0 U. caudimaculatus Connected R 0 0.08 (1,12) 0 0 0.07 (1,15) Connected M 0 0 0 0 0.25 (1,4) Corridor L 0 0.20 (1,5) 0 0.25 (1,4) 0.33 (1,33) Corridor BR 0.33 (1,3) 0 0 0 - Corridor KH * 0 0 0 0 Corridor SK * 0 0 - 0 Corridor S 0 1.0 (1,1) - 0 - Figures in italics denote proportions representing only one individual; numbers in parentheses denote the number of new individuals trapped on the final night of trapping and the total number caught per site per trip; * indicates that trapping was not undertaken; - denotes that no individuals were caught.

Chapter 4. Population structure and movement within the corridor - 67 -

4.3.2 Demographic parameters Although the number of individuals per site showed some variation within the corridor system (Chapter 3), the constitution of the populations was relatively consistent. Analysis of sex ratios (expressed as the proportion of males in the population [arcsin transformed]) per site for each trapping session showed no

significant differences for M. cervinipes (mean ± s.e.: 0.55 ± 0.05) (site: F(6, 21) =

1.96 p = 0.12, month: F(4, 21) = 1.28, p = 0.31) and minimal variation for U.

caudimaculatus (0.45 ± 0.07) (site: F(6, 18) = 2.72, p = 0.046, month: F(4, 18) = 0.56, p = 0.69) (Table 4.2). The only difference between sites was detected between Corridor S and Corridor KH. A maximum of one individual was trapped at Corridor S per trip, thus forcing the proportion of males to be either 0% or 100%, hence, little emphasis can be placed on this significant difference.

Sex ratios were also consistent between the corridor and connected patches when sites were pooled across the entire study. Males were most abundant for M. cervinipes but were in relatively equal proportions within each system component 2 (ncorridor = 69♂, 46♀; nconnected patch = 62♂, 47♀: χ (1) = 0.15, p = 0.69). Conversely, both sexes of U. caudimaculatus were trapped in approximately equal numbers 2 within each system component (ncorridor = 21♂, 20♀; nconnected patch = 20♂, 25♀: χ (1) = 0.19, p = 0.67).

Individuals in reproductive condition (pregnant or lactating) were found in both the corridor and the connected patches for both species. Breeding in M. cervinipes occurred predominantly from November through to March (Table 4.3). No distinct breeding period was detected for U. caudimaculatus as very low numbers of pregnant females were trapped throughout the year. Chapter 4. Population structure and movement within the corridor - 68 -

Table 4.2 Summary of demographic characteristics for the sampled population at each site per trapping session. (Trap nights per session: M. cervinipes: connected sites = 576, corridor sites = 160; U. caudimaculatus: connected sites = 888, corridor sites = 320).

February May AugustNovember February MF MF MF MF MF mean ± s.e. M. cervinipes Connected R1 adult 7 4 8 4 11 4 4 4 4 4 Male:Female 0.71 ± 0.08 juvenile 9 1 2 3 1 4 0 0 4 2 Adult:Juvenile 0.29 ± 0.08 Corridor L adult 1 1 1 1 5 0 2 2 2 0 M:F 0.73 ± 0.13 juvenile 1 0 0 0 0 0 0 0 0 0 A:J 0.93 ± 0.07 Corridor BR adult 1 2 4 0 0 0 3 2 2 0 M:F 0.73 ± 0.06 juvenile 3 0 1 1 0 0 0 0 3 1 A:J 0.63 ± 0.14 Corridor KH adult - - 11 2 9 0 1 4 1 0 M:F 0.65 ± 0.15 juvenile - - 4 5 0 5 1 1 0 0 A:J 0.73 ± 0.09 Corridor KS adult - - 6 5 8 4 2 1 0 1 M:F 0.38 ± 0.13 juvenile - - 4 3 0 2 0 2 0 0 A:J 0.77 ± 0.10 Corridor S adult 0 1 1 1 1 0 0 2 1 2 M:F 0.35 ± 0.10 juvenile 0 0 2 1 0 2 1 0 1 0 A:J 0.63 ± 0.12 Connected M1 adult 3 2 1 3 4 4 1 3 5 10 M:F 0.37 ± 0.09 juvenile 0 0 0 3 0 2 0 1 6 0 A:J. 0.78 ± 0.07 M:F (overall) 0.55 ± 0.05 A:J (overall) 0.74 ± 0.04 U. caudimaculatus Connected R1 adult 1 3 4 3 2 1 4 1 3 6 M:F 0.53 ± 0.08 juvenile 2 2 1 4 0 0 0 0 3 3 A:J 0.75 ± 0.10 Corridor L adult 1 2 1 1 3 1 2 1 0 2 M:F 0.42 ± 0.12 juvenile 1 2 2 1 1 1 0 1 0 0 A:J 0.66 ± 0.10 Corridor BR adult 0 1 1 1 1 0 1 1 0 0 M:F 0.38 ± 0.14 juvenile 0 2 0 1 1 1 0 0 0 0 A:J 0.58 ± 0.16 Corridor KH adult - - 2 0 1 0 2 0 1 0 M:F 0.92 ± 0.08 juvenile - - 0 1 0 0 0 0 1 0 A:J 0.79 ± 0.13 Corridor KS adult - - 0 1 0 0 0 0 1 1 M:F 0.17 ± 0.17 juvenile - - 0 0 0 1 0 0 0 0 A:J 0.67 ± 0.33 Corridor S adult 0 0 0 1 0 0 0 1 0 0 M:F 0.00 ± 0.00 juvenile 0 1 0 0 0 0 0 0 0 0 A:J 0.67 ± 0.33 Connected M1 adult 2 0 0 1 0 2 1 0 2 0 M:F 0.50 ± 0.22 juvenile 0 0 0 0 0 1 1 0 0 2 A:J 0.80 ± 0.12 M:F (overall) 0.45 ± 0.07 A:J (overall) 0.71 ± 0.06

Chapter 4. Population structure and movement within the corridor - 69 -

Table 4.3 Reproductive status of mature females per trip.

Corridor sites Connected patches No. mature % No. mature % Month females reproductive females reproductive M. cervinipes February 4 75 6 83 May90714 August 4080 November 11 91 7 100 February 6 50 14 64 U. caudimaculatus February 3 0 3 67 May40425 August 2030 November 3 33 1 0 February 3 0 6 50

As expected, given the similar breeding seasons for M. cervinipes, and the varying but very low numbers of reproductive females for U. caudimaculatus, the proportion of juveniles per trip was found not to differ between system components for either species (M. cervinipes: tpaired (4) = 2.19, p = 0.09; U. caudimaculatus: tpaired (4) = 0.85, p = 0.40) (Figure 4.2). Juveniles were present in all corridor sites except Corridor S (Table 4.2) suggesting that breeding is occurring within the corridor for at least M. cervinipes as the distance from remnant patches is likely to be beyond the dispersal capabilities of the species. Further, even if dispersal distances could be obtained, it would be unlikely that individuals would remain as juveniles and at such low weights for the duration of dispersal if a connected patch was acting as the source for juveniles trapped within the corridor (see minimum weights Figure 4.3). Insufficient data due to low numbers of individuals prevented analysis on a per site basis.

Chapter 4. Population structure and movement within the corridor - 70 -

100 a) M. cervinipes

80

60

40

20

s

e

l 0 i February May August November February n

ve

u

j

100

% b) U. caudimaculatus

80

60

40

20

0 February May August November February

Trip

Figure 4.2 Percent of juveniles per trip per habitat type (corridor , connected patch „).

The mean weight of individuals did not differ between connected patches and the corridor for either sex of both species (Figure 4.3) (U. caudimaculatus: male, F(1, 38) =

0.012, p = 0.91; female, F(1, 44) = 0.003, p = 0.96)( M. cervinipes: male, F(1, 129) =

1.71, p = 0.19; female, F(1, 91) = 0.60, p = 0.44) suggesting that the corridor provides adequate resources and/or that animals within the corridor do not represent a cohort consisting fully of juvenile, subordinate or unhealthy individuals. These results are Chapter 4. Population structure and movement within the corridor - 71 - based on the individual being the unit of replication such that data points represent the mean weight for those individuals caught on more than one trip. When combined with the data on sex ratios, these data suggest that individuals within the corridor are representative of the structure of the populations within the connected patches and the corridor is not being used solely by displaced old animals or by dispersing young in a sex-biased manner.

600

500

144 166 400 762 828 178 176 (20) (20) 716 998 (21) (26) 300

200 20 20 30 16 84 86 98 78 Mean weight per individual (gm) 100 (46) (47) (69) (62)

0 UF -C UM -C MF -C MM -C UF -P UM -P MF -P MM -P Species, sex and habitat

Figure 4.3 Mean weight (± s.e.) of individuals per sex and habitat type. Numbers immediately under or above bars represent the minimum/maximum weight recorded. Figures in parentheses denote sample size. Labels on x-axis: U/M = species, F/M = sex, P/C = habitat (connected patch/corridor) e.g. UF-C = U. caudimaculatus (female) from the corridor.

4.3.3 Recapture rates

A total of 232 individuals of M. cervinipes and 85 U. caudimaculatus were individually marked with approximately 50% and 70% of markings occurring by the completion of trips two and three respectively (Table 4.4). These figures do not correspond to the total number of individuals caught of each species as reported in Section 4.3.1. A small number of individuals were chipped twice due to loss of the Chapter 4. Population structure and movement within the corridor - 72 - original chip (but were only scored as one individual in total counts) while some individuals were ear clipped, sexed, weighed and counted as present but were not chipped due to their poor condition.

The marking of a high proportion of individuals in the early trips ensured that the majority of animals had the potential to be re-trapped and their movements noted over a period of six to nine months. Despite this, the recapture rate of individuals of both species on subsequent trips was low with fewer than 6% of the marked M. cervinipes and 13% of the U. caudimaculatus recaptured on each of the final two trips (Table 4.4).

Table 4.4 The number of individuals marked per trip and the recapture rates on subsequent trips. Numbers in parentheses represents the cumulative total for the study.

# recaught on subsequent trips % of total Trip # marked marked 1 2 3 4 5 M. cervinipes February 37 (37) 15.9 - 11 7 2 0 May 77 (114) 49.1 - - 14 4 3 August 47 (161) 69.4 - - - 3 3 November 28 (189) 81.5 - - - - 1 February 43 (232) 100.0 - - - - - % recapture 29.8 18.4 5.6 3.7

U. caudimaculatus February 20 (20) 23.5 - 0 1 1 3 May 27 (47) 55.3 - - 2 4 2 August 15 (62) 72.9 - - - 1 1 November 9 (71) 83.5 - - - - 5 February 12 (85) 100.0 - - - - -

% recapture 0 6.4 9.7 12.7

Of those individuals with the potential for recapture on more than one trip (i.e. marked on trips 1-4), 39 M. cervinipes (21%) and 19 U. caudimaculatus (27%) were re-trapped on a subsequent trip.

Chapter 4. Population structure and movement within the corridor - 73 -

Both species showed similar recapture rates of marked individuals on subsequent trips (Table 4.5). The habitat component in which capture occurred and the sex of the animal were also found to have no influence on the success rate of recapturing marked individuals for either species. The only factor found to influence recapture rates was the maturity of U. caudimaculatus upon initial capture with a much lower proportion of juveniles (11.5%) being trapped again compared to adults (38.7%) (Table 4.5). Differential recapture rates for male and female juveniles were not able to be determined statistically due to low values in some cells. The data however, suggests no obvious disparity between the sexes, suggesting that the low recapture rates of juveniles is similar for both sexes (Table 4.5).

Table 4.5 Recapture rates according to species, habitat, age and sex.

* no analysis was performed.

Chapter 4. Population structure and movement within the corridor - 74 -

The recapture of marked individuals within a trip occurred more frequently for M. cervinipes than U. caudimaculatus. Thirty-seven percent of M. cervinipes individuals were trapped on more than one night of any trip with the mean number of captures being 1.6 and a maximum of seven. Conversely, only 25% of all marked U. caudimaculatus individuals were trapped more than once on a trip with a mean capture rate of 1.3 and a maximum of three.

4.3.4 Movement No M. cervinipes were recorded moving between sites while three inter-site movements from two individuals were recorded for U. caudimaculatus. One mature female moved from Corridor S to Corridor KS between trips 4 and 5 while a mature male was recorded in Connected M in trip 1, and both Corridor KS and Connected M in trip 5. These movement events equate to travel of approximately 390m for the female and a total of 1.9km for the male, 940m of which occurred within a two week period.

The distance moved by individuals between recaptures both within and between trips was assessed within Connected R using the individual as the unit of replication. Multiple data from the same individual was averaged and included only once in analyses. Individuals of U. caudimaculatus moved significantly greater distances between trips than did M. cervinipes (F(1,24) = 4.39, p = 0.047) (Table 4.6). No difference in distance moved between trips was present between sexes for either species (U. caudimaculatus: F(1,6) = 0.03, p = 0.87; M. cervinipes: F(1,16) = 2.3, p = 0.15). Due to low recapture rates, movement distances within trips could only be determined for M. cervinipes (n = 1 for U. caudimaculatus). Average distances obtained between successive captures within a trip ranged from 0m to 70.7m with a

mean of 24.1 ± 2.6 (Table 4.6) and did not differ between the sexes (F(1,29) = 2.62, p = 0.12). Comparative assessment of within and between trip movement was not possible for individuals from Connected M due to the low number of individuals recaptured. Only three individuals of M. cervinipes and two U. caudimaculatus were trapped on more than one trip while the number of individuals re-trapped within a trip was six and four respectively.

Chapter 4. Population structure and movement within the corridor - 75 -

Table 4.6 Average distances moved by individuals within the trapping grid of Connected R.

n mean (s.e.) Between trips M. cervinipes (all) 18 49.4 ± 4.7 male 11 54.9 ± 5.7 female 7 40.8 ± 7.3 U. caudimaculatus (all) 8 70.5 ± 11.0 male 4 72.5 ± 15.9 female 4 68.5 ± 17.6 Within trips M. cervinipes (all) 31 24.1 ± 2.6 male 17 27.8 ± 3.9 female 14 19.7 ± 2.8

4.4 Discussion

All aspects of population structure indicated that the corridor is likely to facilitate connectivity for both species. With i) age and sex ratios comparable between the corridor and the connected patches, ii) juveniles and pregnant females present within the corridor and iii) individuals within the corridor having similar weights to those within the remnants, the corridor appears to provide all of the resources necessary for the residency of self-perpetuating populations and/or the transit of individuals in breeding condition. Given this, no potential barriers to connectivity were identified.

The most likely mode of dispersal between the connected patches was found to differ for each species. With recorded movements by an individual over a distance representing over 40% of the corridor length, direct dispersal is the likely mode of movement between the connected patches for U. caudimaculatus. This is supported by the higher recapture rates for adults than juveniles suggesting that juveniles leave the natal territory and have the opportunity to disperse to the alternate connected remnant. Further, both sexes covered comparable distances within Connected R, had similar recapture rates within the corridor system and demonstrated the ability to move between trapping sites, suggesting that both males and females may be capable of direct dispersal. Additional recordings of an individual travelling one kilometre in a period of two days (Goosem 2001) suggests Chapter 4. Population structure and movement within the corridor - 76 - that these long distance movement events fall within the normal movement range of the species, further supporting the likelihood of direct dispersal.

Alternatively, generational movement was found to be the most probable means of connectivity for M. cervinipes. The detection of long-distance movement events has been shown to be problematic (Koenig et al. 1996, and Whitlock and McCaughley 1999) and direct movement cannot be eliminated as a possible means of dispersal solely from the lack of any inter-site movement by M. cervinipes. However, intra-site distances within Connected R were shorter than recorded for U. caudimaculatus, and recapture rates and movement distances were similar between age classes and sexes. This suggests that despite the prevalence of male-biased dispersal in small mammals, males were as likely to remain within the trapping area between trips as females. Similarly, juveniles of both sexes were as likely to be recaptured as adults suggesting limited dispersal away from the natal patch by juveniles. These results concur with the literature which also suggests minimal movement by individuals of the species relative to the length of the corridor (Wood 1971, Smith 1985)

Both species demonstrated a low percentage of recapture for individuals on trips subsequent to their initial capture indicating either a high mortality rate or movement away from the area of original encounter. Given the apparently healthy state of individuals upon initial capture, and the known lifespan of the species, mortality is unlikely to be responsible. It is therefore suggested that individuals of both species had movement capabilities enabling them to move away from the trapping grid, but that this distance was greater for U. caudimaculatus than for M. cervinipes.

The differing mode of movement most likely for each species has important implications for the necessary demographic constitution of the species within the wildlife corridor. The process of generational connectivity places the greatest demands on the corridor as it necessitates the presence of breeding populations resident within the corridor. This in turn requires the corridor to provide all of the resources necessary for completion of the life cycle. M. cervinipes within the corridor displayed similar average weights and sex ratios to those within the connected patches indicating that the populations within both habitats are functioning in a similar manner. This is further supported by the synchronized breeding seasons and similar proportions of juveniles within the two habitats. These combined data indicate that breeding populations reside within the corridor which Chapter 4. Population structure and movement within the corridor - 77 - thus suggests that the corridor provides adequate resources and that requirements necessary for generational connectivity were met.

Weight and sex ratios of animals both within the corridor and connected patches were very similar which made for easy interpretation of the data. Similarly, Bennett et al. (1994) found individuals of all sex and age categories to be resident within fencerows that served as corridors between woodlots. Conversely, Downes et al. (1997a,b) found individuals of native rodents to be significantly lighter and to consist of a disproportionately higher number of males and juveniles within the corridor relative to the connected patches. The reduced weight of individuals may result from the provision of inadequate resources by the corridor or from the increased proportion of juvenile animals. This suggests that the corridor is predominantly used by dispersing individuals in a sex biased manner and that completion of the lifecycle and thus breeding within the corridor is unlikely to occur. However, the implications of these data are harder to interpret because the potential for generational connectivity requires the presence of breeding individuals within the corridor rather than the presence of demographically similar populations. Therefore although the individuals within the corridor consist predominantly of males and juveniles, generational connectivity may still be possible if some breeding females are resident along the corridor. This example highlights how ecological data can play a complimentary role to genetic studies. Should gene flow be shown not to occur between the patches investigated by Downes et al. (1997a,b), the demographic data can provide a retrospective explanation for the lack of connectivity despite the presence of the species along the corridor length.

In contrast to the specific stipulations necessary for generational connectivity, a species capable of direct movement only requires the corridor to supply those resources needed during travel within the corridor (Merriam and Saunders 1993) and is not reliant upon the presence of any con-specifics within the corridor. That individuals of U. caudimaculatus were found to reside and breed within the corridor indicates that the necessary resources can be acquired. The sex ratios and weights of corridor trapped animals were again consistent with those within connected patches. Although direct dispersal has been suggested as the most likely mode of movement, the presence of breeding populations within the corridor provides a second possible means of connectivity via generational gene flow. However, it must Chapter 4. Population structure and movement within the corridor - 78 - also be acknowledged that the presence of corridor resident animals also introduces a potential barrier to connectivity through territorial exclusion.

One anomaly in the data collected for U. caudimaculatus is the discrepancy between the proportion of juveniles within the corridor populations and the low numbers of females encountered in reproductive condition. Two possible explanations can account for this data. First is the time lag between pregnancy and the time at which juveniles are independent and can be trapped i.e. the juveniles trapped within the first two trips may have resulted from pregnancies that occurred prior to the commencement of trapping. Secondly, reproductively active females may have resided within the corridor but in an area outside of the trapping grids with juveniles from these females encountering the trapping grid during either exploratory behaviour or during dispersal away from the natal territory. Regardless of the explanation, it is believed that the juveniles were born within the corridor as it is unlikely that the juvenile weights within the corridor would be as low as observed if those individuals had covered the distance required for dispersal from a connected patch.

Both species in this study were found to have breeding populations resident within the corridor indicating that the corridor provides all of the resources necessary for survival and reproduction. When combined with the data from Chapter 3, which shows that the entire corridor is suitable for use by both species, it can be inferred that the wildlife corridor has the potential to increase connectivity between the populations within the connected patches albeit by different modes. If only ecological techniques were employed to assess the extent of connectivity provided by the corridor as per previous studies (eg. Downes et al. 1997a,b, Bennett 1990b), a favourable conclusion would most likely be drawn based on the results obtained.

It has been suggested however, that connectivity requires more than just corridors (Merriam 1995). This statement can be expanded further to state that connectivity requires more than just movement or residency, as effective dispersal between connected patches depends on the integration of new individuals into a recipient population. Given that trapping is unable to determine the reproductive success of individuals, especially males, a large number of movement events need to be detected before any comment can be made about the likelihood of connectivity occurring. Should trapping show movement between populations to be a frequent Chapter 4. Population structure and movement within the corridor - 79 - event, it could be suggested with confidence that the probability of some dispersing individuals becoming integrated into the new population is high. The data from this current study demonstrates the difficulty in obtaining the large number of movement events necessary for such a conclusion. The high proportion of traps available each night and the low number of new individuals encountered on the last night of trapping suggests that those animals residing within the trapping grid had a high chance of capture during each trapping period. However, with over 6,800 trap nights conducted in the connected patches and the corridor for U. caudimaculatus, only three long distance movement events were detected despite it being demonstrated that individuals were capable of moving the distance between trapping locations. It must be therefore considered unlikely that sufficient data will be obtained from trapping studies, conducted within wildlife corridors at the scale of kilometres, to allow comment on the degree of connectivity provided by the corridor. Thus, in light of Merriam’s (1995) suggestion, alternative methods, such as the use of molecular markers, are required in order to fully address the issue of connectivity. - 80 -

5. DOES THE WILDLIFE CORRIDOR FACILITATE GENE FLOW?

5.1 Introduction

Realized gene flow necessitates that DNA is passed to the subsequent generation through mating and that offspring survive to reproduce (Slatkin 1987, Whitlock and McCaughley 1999, Prugnolle and de Meuss 2002). Gene flow among populations can occur through regular dispersal of juveniles, occasional long-distance dispersal events or as a stepwise progression via several generations that requires minimal dispersal by individuals (Bennett 1999). Further, gene flow may occur as a result of periodic dispersal episodes rather than as a constant event occurring every generation (Slatkin 1987). As the evolutionary importance of gene flow on populations is a cumulative effect over many generations (Whitlock and McCaughley 1999, Brooks 2003), each of these methods may contribute to gene flow among populations in spite of the large variations in the frequency and distance of dispersal by individuals. Direct methods for assessing gene flow (i.e. trapping) are highly unlikely to detect long distance or periodic dispersal events (Koenig et al. 1996, Whitlock and McCaughley 1999), or to allow inferences suggesting that generational gene flow has occurred. Estimates of gene flow using these techniques are therefore likely to be underestimates. Conversely, overestimates of gene flow from direct methods of assessment can occur when habitat use by individuals does not result in dispersal or when dispersal does not translate into successful breeding (Slatkin 1994, Whitlock and McCaughley 1999). Complex interactions between habitat heterogeneity and social behaviours are one potential reason why differences in habitat use can fail to translate into effective gene flow.

Populations of most species are not uniformly distributed in space (Amarasekare 1994) and commonly consist of segregated breeding groups (Sugg et al. 1996). Such groups can form from disruptions to dispersal due to geographic distance, physical barriers or behavioural factors (Chesser 1983b), all of which can be influenced by the heterogeneity of the habitat. Simple examples of how habitat heterogeneity may affect dispersal include geographic distance acting as a barrier between suitable habitats within a heterogeneous landscape, or when a suitable area is surrounded by inhospitable habitat, thus limiting individuals to their natal patch. More subtle however, is the role that habitat heterogeneity may Chapter 5. Gene flow - 81 - play in the spatial distribution of social groups. The formation of these groups occurs when control is exerted over which individuals of a species are included within a particular social group (Hughes 1998, Storz 1999). Regulation can occur in the form of social exclusion resulting in the failure of an individual to become incorporated into a new breeding unit, the philopatry of individuals born into the group or in more extreme cases, the exclusion of individuals from a territory via active defence. Social structuring in this manner can occur in environments that are relatively homogenous (e.g. house mice within barns, Selander 1970), but more commonly it can be attributed to the heterogeneous distribution of resources.

The formation and maintenance of social groups can provide individuals with the benefits of unified defence or the seizure and protection of important resources (Chesser 1998). These attributes are of most benefit to individuals living in heterogeneous environments where resources are located in a patchy manner or may be in short supply. When the abundance or spatial distribution of resources varies over time, nomadic social groups may develop (Chesser 1998) or regulation of immigration and/or dispersal may be relaxed, leading to a change in the social structuring of the populations (Brandt 1992). Variation of this type has been detected in Gunnison’s prairie dog (Cynomys gunnisoni) which displays a monogamous mating system when resources are evenly distributed but becomes polygynous and polyandrous when spatial heterogeneity of resources increases (Travis et al. 1995). The behaviour of deer mice (Peromyscus maniculatus) also changes from being tolerant of immigrants to actively defending territories when resource levels decline (Metzgar 1979, 1980). Experimental manipulation of food availability has been shown to alter social systems in many species with the level of immigration increasing as food supplies become more abundant (see Table 5.4 in Brandt 1992 for review).

Social groups are especially common in mammals with 83% of known mammalian mating systems exhibiting some degree of female philopatry (Dobson 1982; in Chesser 1998). This results in replacement of breeding females by their female offspring, and polygynous mating which samples a non-random sample of the gene- pool of adult males (Anderson 1970 [in Lidicker 1995], Chesser 1998). The combination of these factors, along with social exclusion, can contribute to the maintenance of a closed social group (Chesser 1998, Storz 1999).

Chapter 5. Gene flow - 82 -

While not an inevitable outcome (Lidicker and Patton 1987, Storz 1999), temporal stability of social groups can lead to the formation of genetically differentiated subpopulations. A synthesis of data on the relationship between social and genetic structuring in mammals (Storz 1999) and rodents in particular (Lidicker and Patton 1987) found examples of species where social barriers to gene flow had large impacts on the genetic structure of the population (e.g. black-tailed prairie dogs; Chesser 1983a, red howler monkeys; Pope 1992), however no clear relationship was found between either social cohesion or taxa to explain the variability in genetic structuring (as measured by FST) among species. Some species with a low level of social cohesion were found to exhibit genetic structures similar to taxa displaying tight within-group social behaviour and conversely, the level of genetic differentiation within groups was sometimes found to exceed that displayed between geographically divided sub-populations (e.g. black-tailed prairie dogs; Chesser 1983a, California voles; Bowen 1982). Further, Fuller et al. (1997) reported extreme variation in genetic structuring in different habitats within one species (the European rabbit, Oryctolagus cuniculus) with the variation attributed to social structure, environmental conditions and population stability (see also Kurt 1991 cited in Wang and Schreiber 2001). Thus, although socially-mediated gene flow is known to occur, its existence is neither obvious nor predictive (Lidicker and Patton 1987).

Although well established within the literature as a means of detecting gene flow, direct methods (i.e. trapping) are actually indicators of dispersal or habitat use and their use as an indicator of gene flow is potentially constrained by an inability to detect all dispersal events or to determine whether those dispersal events detected actually translate into gene flow due to the above factors (Slatkin 1994). Extrapolation of these parameters to that of effective gene flow is therefore potentially erroneous in both directions (Lidicker and Patton 1987, Whitlock and McCaughley 1999).

By contrast, indirect methods of estimation (i.e. the use of genetic markers) assess the cumulative contribution to the recipient population of only effective dispersal events while not discriminating between the spatial and temporal manner through which gene flow occurred (Slatkin 1994, Neigel 2002). Thus, this approach is more sensitive to long distance or infrequent dispersal events that are likely to be missed using mark-recapture methods (Slatkin 1994, Whitlock and McCaughley 1999). While the population genetic approach appears ideal for use in studies of wildlife Chapter 5. Gene flow - 83 - corridors, it has the disadvantage of being only suitable for scenarios where modification of the habitat in the form of land clearing or corridor construction has existed for sufficient time to allow the genetic attributes of a population to have been affected by the habitat change. How long one must wait after modification of the landscape before it is appropriate to use indirect methods is not constant and will depend upon the individual life-history traits of the target organism.

Given the large number of factors that may disrupt gene flow, one must question the validity of drawing inferences about the effectiveness of gene flow via corridors based solely on direct measures. This is especially pertinent where mammals are the study organism and/or where the corridor is narrow which may accentuate the social interactions that have the potential to disrupt gene flow. Indirect measures of gene flow were therefore used to investigate whether gene flow occurs via the corridor.

5.2 Methods

5.2.1 Sample collection Sample sites for the landscape scale study were selected according to the rationale outlined in Section 3.1. This involved collecting samples from within connected patches, including at the patch/corridor boundary (Connected R1 and M1) (Figure 5.1a), within isolated remnant habitats (Isolates B, W and L), from within the corridor (Corridor sites L, BR, K/H, K/S and S) and from within the continuous habitat (Continuous 1, 2, 3 and 4). Samples from within Connected R, M and the corridor sites were collected during the trapping trips undertaken to collect demographic and movement data while samples from the remaining sites were collected on an ad-hoc basis throughout the duration of the field work. Both species were trapped using bait of either linseed oil soaked cardboard or a mixture of rolled oats, peanut butter and honey. Ear biopsies were taken for DNA samples from individuals and samples were stored in 70% ethanol.

To investigate the degree of differentiation between populations of M. cervinipes on a fine (within-patch) geographic scale, additional samples were taken from within Connected R (R2 & R3) and Connected M (M2 - M5) and were analysed along with the original samples from Connected R1 and M1. Additional sites were also Chapter 5. Gene flow - 84 - sampled within the continuous forest with sites 1A and 2A located 300m away from the original trapping sites of Continuous 1 and 2. This distance was selected as it is approximates the distance between sites within Connected R and Connected M. Population structuring of U. caudimaculatus was not studied over smaller distances within the continuous forest, or within Connected M, as individuals were observed to move among these sites (pers. obs.). Thus only structuring between connected R1 and R3 was studied for this species as it provided the greatest distance available for a “within site” comparison. In total, 554 M. cervinipes and 240 U. caudimaculatus samples were collected during the study.

Figure 5.1 Location of sampling sites for fine scale spatial studies within A) connected patches and B) continuous forest. Circles indicate sites also used for landscape scale gene flow analyses; diamonds represent sampling locations used solely for fine-scale analyses.

5.2.2 Mitochondrial DNA

Mitochondrial DNA (mtDNA) is a small double-stranded, circular, haploid molecule which is almost exclusively maternally inherited in mammals (Randi 2000, Ballard and Whitlock 2004). The uniparental mode of inheritance reduces the effective

population size (Ne) to ¼ that of nuclear genes, making mtDNA sensitive to genetic Chapter 5. Gene flow - 85 - drift in small and isolated populations (Birky et al. 1989, Harrison 1989, Moritz 1994, Randi 2000). Maternal inheritance also makes mtDNA an appropriate marker to detect demic structure where it may be present due to female philopatry (Moritz et al. 1987).

After extraction of genomic DNA from the tissue (see Appendix 3 for all mtDNA protocols) polymerase chain reaction (PCR) was used to amplify the displacement loop region (D-loop) of the mitochondrial DNA. In contrast to other regions of the mtDNA, this region lacks structural genes (Moritz et al. 1987). This implies that any mutations that accumulate are essentially neutral and do not confer any selective advantage or disadvantage on the individual. Where a locus that confers a selective advantage is used, populations may show the same genetic structure even in the absence of gene flow (Slatkin 1994). As a consequence of neutrality, the D-loop region is the most rapidly evolving part of the mtDNA genome (Saccone et al. 1987) with changes accumulating rapidly, both within and among species (Moritz et al. 1987). The rapid accumulation of mutations makes it ideal for studies that investigate the extent of contemporary gene flow. Amplification was performed using the light strand primer MT15996L (5’CTCCACCATCAGCACCCAAAGC3’, M.S. Elphinstone, B.A.Williams and P.R. Baverstock, unpublished) located in the tRNA proline gene and the heavy strand primer MT 16498H (5’CCTGAAGTAGGAACCAGATG3’, Meyer et al. 1990) located within a highly conserved region of the mammalian control region, respectively. This resulted in amplification of a 414 base pair fragment located in the highly variable tRNA end of the control region from which a 367 and 372 base pair fragment was analyzed for all M. cervinipes and U. caudimaculatus analyses, respectively.

Amplified DNA was screened for variability using temperature gradient gel electrophoresis (TGGE) (Rosenbaum and Reissner 1987). The electrophoresis of double-stranded DNA through a temperature gradient results in differential denaturation due to variation within the base-pair composition of the fragment and interactions across the molecule (Lessa and Applebaum 1993). As denatured DNA has a forked configuration which slows its rate of migration through the gel (Campbell et al. 1995), haplotypes with the lowest thermal stability denature first and migrate the least. Thus, detection of DNA variants by TGGE occurs due to differences in nucleotide sequence rather than by fragment length as per classical gel electrophoresis (Lessa and Applebaum 1993). Chapter 5. Gene flow - 86 -

While TGGE has a high resolving power, it is still possible for two distinct haplotypes (unique mtDNA sequences) to have melting points that are too similar to enable accurate discrimination on a gel. The addition of heteroduplex analysis to the TGGE protocol provides greater discriminatory power to the procedure. This technique involves denaturing two similar but non-identical DNA fragments and re-annealing them to enable the formation of homoduplexes of each fragment type, along with heteroduplexes which consist of “hybrids” of the two original DNA fragments (Lessa and Applebaum 1993, Elphinstone and Baverstock 1997). Heteroduplex molecules containing base-pair mismatches (non-Watson-Crick bonding) in their double strands have reduced electrophoretic mobility relative to homoduplexes and can display reduced thermal stability (Campbell et al. 1995) allowing increased resolution among haplotypes. Campbell et al. (1995) advanced this procedure further by introducing a different, yet closely related species as the reference to which samples were heteroduplexed. This procedure, known as Outgroup Heteroduplex Analysis, when used in conjunction with TGGE, was shown to have greater discriminatory power than TGGE alone or TGGE with ingroup heteroduplex analysis (Campbell et al. 1995). The technique was able to discriminate among haplotypes differing by a single base-pair and this approach was used in the current study. The use of morphologically similar congeners as the outgroup (Melomys burtoni and Uromys hadrourus) also enabled verification of the identification of each sample as distinct heteroduplex bands were not present when outgroups were annealed to samples of the same species from within the same geographic region. At least one individual of each unique haplotype was sequenced in both the forward and reverse direction. This was performed to verify the uniqueness of the haplotypes identified via TGGE.

5.2.3 Nuclear DNA

Microsatellites are variable length nucleotide repeats (e.g. ATATAT; CCTCCTCCT) which are located widely and randomly throughout the nuclear genome of almost all organisms (Queller et al. 1993). With the exception of some trinucleotide repeats which have been linked to diseases in humans (Queller et al. 1993), microsatellites are selectively neutral, are highly polymorphic, are inherited in a mendelian fashion and commonly have a high mutation rate (Jarne and Lagoda 1996). Polymorphisms result generally from slippage during replication leading to the formation of new alleles which differ in size (either addition or subtraction) by one tandem repeat (Ellegren 2000). Alleles are also susceptible to the processes of drift and migration Chapter 5. Gene flow - 87 -

(Chambers and MacEvoy 2000) making them suitable markers for population genetics.

Unlike mtDNA, nuclear DNA is diploid by virtue of being bi-parentally inherited (Sunnucks 2000). This allows for the screening of not only allele frequency (the nuclear equivalent of haplotype frequency), but also the extent of heterozygosity within populations. Heterozygosity is expressed when an individual carries two different alleles at a locus and is thus able to pass one of two allele types to any offspring. Heterozygosity can be important in maintaining genetic diversity and ultimately the fitness of a population (Gilpin and Soulé 1986, Allendorf and Leary 1986), yet it is readily eroded by factors such as small population size and inbreeding associated with population/habitat isolation (McCauley 1993).

After extraction, genomic tissue was amplified (see Appendix 4 for nuclear protocols and primer information) and microsatellite allelic variation screened using gel electrophoresis. This process separates alleles on the basis of their electrophoretic potential which is determined by the fragment size and represents variation in the number of nucleotide repeats. Five loci were screened for M. cervinipes and six for U. caudimaculatus. Gels were run on a Gel-Scan 2000 (Corbett Research) with a size marker (T350, Applied BioSystems) and a reference individual run on every gel to ensure consistent scoring across gels. Digital images from Gel-Scan 2000 were scored using One-DScan (Scanalytics Inc.) with all output manually cross-checked.

Mitochondrial DNA and nuclear DNA are suitable for use as independent indicators of gene flow as they are unlinked and patterns of variation between the markers may not be concordant (Harrison 1989).

5.2.4 Statistical analyses

Mitochondrial DNA The DNA sequences of all haplotypes were double checked for assignment to the

target species using the BLAST program (NR vertebrate database in GenBank) on the web interface of the Australian National Genomic Information Service (ANGIS: www.angis.org.au). This database contains many known sequences of M. cervinipes and U. caudimaculatus including samples from geographic regions where the conspecifics do not occur. mtDNA sequences were then aligned and checked

for uniqueness in the CLUSTALW program (Thompson et al. 1994) of ANGIS. Chapter 5. Gene flow - 88 -

The following analyses were performed for the mtDNA data using ARLEQUIN VER 2.0 (Schneider et al. 2000): 1. Tajima’s (1989) test for selective neutrality: a test appropriate for short DNA sequences. This was performed for each site separately and also for all samples combined.

2. Estimates of haplotype diversity (Nei 1987).

3. Pairwise comparison of population differentiation via FST (Weir and Cockerham 1984).

4. Estimation of pairwise differentiation using Raymond and Rousset’s (1995) analogue to Fisher’s exact test. p-values were determined using a Monte Carlo Markov chain with 1,000 iterations.

5. Isolation by distance for sites within the corridor system was determined through the regression of ln-geographic distance and pairwise estimates of

FST/(1- FST) (Rousset 1997). Mantel tests (Mantel 1967) were performed to test the statistical association between the two matrices using the program

MANTEL VER 2.0 (Liedloff 1999) (1000 iterations).

Where data can be partitioned into different categories or levels, the common

analytical approach is via AMOVA (Analysis of Molecular Variation) (Excoffier et al.

1992). Similar to ANOVA for ecological data, AMOVA partitions variation in the data according to pre-determined hierarchical levels and compares the variation within

and among groups. If AMOVA were applied to the data obtained in this study, sites would have to be grouped according to their degree of structural connectedness i.e. isolates, connected sites and continuous sites, with variation within these groupings compared to that among groups. This form of analysis is inappropriate for this study as isolate sites were selected on the basis of their geographic distance from one of

the connected sites. Analysis via AMOVA would calculate the within group variation between isolate sites, yet this is invalid as the geographic distance between isolate sites was not standardized and thus could be confounded by any isolation-by- distance effect. Therefore, pairwise analyses of populations were performed on haplotype frequency data using FST and exact tests to circumvent this potential problem. Chapter 5. Gene flow - 89 -

While analysis of mtDNA sequence variation is often the preferred method, Aars et al. (1998) noted that the timescale over which mutations accumulate is anomalous to the timescale relevant to the investigation of population effects of anthropogenic habitat fragmentation. Hence, Aars et al. (1998) proposed that haplotype frequency data should be used in the analysis of studies investigating habitat fragmentation as these data reflect changes resulting from contemporary processes such as . Using this approach, haplotypes were simply scored as identical or different and the extent of any sequence variation between haplotypes was not considered.

To account for the possibility of territoriality at the remnant/corridor boundary, population differentiation analyses involving Connected R and Connected M were conducted in two ways: i) using all samples from within the connected patches and ii) using only those samples collected at the corridor/patch boundary (Connected R1 and M1). This approach was undertaken for both mtDNA and nuclear DNA data.

Nuclear DNA The various alleles for each locus were identified according to the length of the fragment and were identified alphabetically with allele “A” corresponding to the shortest fragment. The following analyses were then conducted for the microsatellite data for both species: 1. An exact test of linkage disequilibrium for all pairs of loci using an extension of Fisher’s exact test (Slatkin 1994) with p-values determined by a Monte Carlo Markov chain with 1,000 iterations.

2. Tests for departures from Hardy-Weinberg equilibrium using Guo and Thompson’s (1992) extension of Fisher’s exact test.

3. Calculation of FIS (Weir and Cockerham 1984) which provides a measure of inbreeding within a population by measuring the correlation of genes within individuals.

4. Calculation of the average M ratio per population, where M is the ratio of the total number of alleles (k) to the range in allele size (r) (Garza and Williamson 2001). Variation in M-ratios among habitat types were tested

using a one-way ANOVA.

Chapter 5. Gene flow - 90 -

5. The significance of differences in allelic frequency between populations was tested using a log-likelihood (G) based exact test (Goudet et al. 1996), the principle of which is analogous to Fisher’s exact test. This test gives equal

weighting to all loci and is more powerful than other exact FST -estimator tests especially in the event of uneven sample sizes (Balloux and Lugon- Moulin 2002). Significance values were determined using a Monte Carlo Markov chain with 1,000 iterations.

6. Estimation of pairwise differentiation was also assessed through the

calculation of FST (Wright 1978). While the assumptions of the analogous

RST (Slatkin 1985) estimator are thought to be more accurate for

microsatellites than the infinite-allele model on which FST is based (Balloux

and Lugon-Moulin 2002), FST’s are still commonly reported and are thought to be more suitable for smaller, recently diverged populations (Gaggiotti et al. 1999, Balloux and Lugon-Moulin 2002). This is due primarily to the high

variance of RST estimates. FST values indicate the proportion of genetic variation that occurs among populations relative to that found within.

7. Isolation by distance for sites within the corridor system was determined through the regression of geographic distance (natural log) and pairwise

estimates of FST/(1- FST) (Rousset 1997). Mantel tests (Mantel 1967) were performed to test the statistical association between the two matrices using

the program MANTEL VER 2.0 (Liedloff 1999) (1000 iterations).

Linkage disequilibrium and Hardy-Weinberg parameters were tested using ARLEQUIN

VER 2.0 (Schneider et al. 2000) while FIS and FST values and exact tests of population differentiation were calculated using GENEPOP ON THE WEB VER 3.1C (available at http://biomed.curtin.edu.au/genepop/, Raymond and Rousset 1995).

Where multiple comparisons were made, α values were adjusted according to the sequential Bonferroni correction technique (Rice 1989). This method reduces the probability of Type-1 error, yet it is not as conservative as the traditional Bonferroni correction which calculates ά as α/n. The sequential technique initially calculates ά in the same manner but then computes subsequent ά values according to α/n-1, α/n-2, α/n-3, α/n-k, where k = the total number of comparisons, such that in the final comparison ά = α. In the instance of global comparisons using discrete data for Chapter 5. Gene flow - 91 - each test, and where p-values can be deemed significant by chance alone (e.g. Hardy-Weinberg exact tests), the initial ά value was calculated as α/n where n = the total number of comparisons. Alternatively, when multiple tests used discrete data sets, yet addressed a null hypothesis common to only a subset of the data, n was calculated as the number of tests involved in testing the common null hypothesis (Rice 1989). For instance, n was calculated as the number of populations involved in testing a particular loci pair during tests for linkage disequilibrium. Further, due to the large number of non-relevant pairwise comparisons in the half matrices when all sites were represented (eg Appendices 13-16), alpha correction was only performed for those subsets of the matrix that were specifically analyzed rather than according to the total number of comparisons in the half-matrix. In these instances, α was adjusted according to the number of comparisons for which a specific data set was involved.

5.3 Results

5.3.1 Tests for neutrality and linkage disequilibrium Following sequential Bonferroni α correction (Rice 1989), no populations of either species displayed a significant deviation from neutrality for the mtDNA marker (Table 5.1). This result was supported when all samples were analyzed together, confirming that as expected, the DNA fragment was not under the influence of selection and was suitable for use as a neutral indicator of gene flow.

Chapter 5. Gene flow - 92 -

Table 5.1 Tajima's D test for neutrality for the control region of the mtDNA.

M. cervinipes U. caudimaculatus Population Tajima's D P Tajima's D p Connected R1 0.41 0.70 0.15 0.74 Connected R2 0.79 0.81 - - Connected R3 0.07 0.58 -1.60 0.02 Connected M1 1.74 0.97 -0.86 0.21 Connected M2 0.20 0.62 - - Connected M3 0.31 0.70 - - Connected M4 1.05 0.88 - - Connected M5 0.21 0.65 - - Corridor L 0.13 0.59 0.00 1.00 Corridor BR 1.32 0.93 0.69 0.77 Corridor K/H 1.25 0.96 - - Corridor K/S -0.40 0.35 - - Corridor S -1.34 0.10 - - Isolate B 1.31 0.92 -1.76 0.02 Isolate L 0.99 0.88 - - Isolate W 1.34 0.91 0.51 0.75 Continuous 1 -0.60 0.31 2.18 0.99 Continuous 1A -1.53 0.05 - - Continuous 2 -1.42 0.06 1.99 0.98 Continuous 2A -0.17 0.45 - - Continuous 3 -1.17 0.14 - - Continuous 4 -0.75 0.23 -0.32 0.43 All sites combined -1.09 0.14 -0.18 0.46 - denotes location either not sampled or sample size not sufficient for inclusion.

The five micosatellite loci used for M. cervinipes were also shown to be suitable for use as independent markers. Although some linkage disequilibrium was observed among loci (Table 5.2), no consistent patterns were evident in either the locus pairs showing linkage or the populations from which the significant results were obtained and so all loci were retained. In contrast, two of the six loci analysed for U. caudimaculatus were found to be highly and significantly linked in all populations (p<0.0001) (UVC19 and UVC245). Locus UVC19 was therefore eliminated from all further analyses (Table 5.2). While the number of significant results in the remaining comparisons was higher than expected, there again was no pattern in the pairs of loci linked with the maximum number or significantly linked pairs being 4 out of 11 populations. If pairs of loci were truly linked, significant results would be expected in the majority of populations as was detected for the UVC19/UVC245 pair.

Chapter 5. Gene flow - 93 -

Table 5.2 Linkage disequilibrium for each locus pair.

+ denotes significance after sequential Bonferroni correction of α.

5.3.2 Haplotype distribution and diversity Analysis of the control region of M. cervinipes revealed 34 unique haplotypes in 554 individuals (Appendix 5) with no single haplotype dominating any population. Fewer haplotypes and lower haplotype diversity, (analogous to heterozygosity for diploid markers), were found for populations within the isolated and connected sites

compared with continuous sites (H: F(2, 14) = 8.27, p = 0.004; h: F(2, 14) = 19.5, p < 0.001). This was in spite of populations from continuous sites having smaller sample sizes ([mean ± s.e.] isolates: 36.7 ± 6.2, connected sites 25.1 ± 1.3, continuous sites 22.5 ± 6.0). No differences were apparent in the number of haplotypes present or diversity of haplotypes within isolate and connected sites. Chapter 5. Gene flow - 94 -

Fourteen haplotypes were detected among the 240 U. caudimaculatus individuals sampled (Appendix 6). Populations from connected and isolate sites were depauperate in genetic variation compared with those from continuous sites (F(2, 5) = 11.91, p = 0.01) with only three haplotypes detected within the largest patch (Connected R). All populations from isolated or connected sites were dominated by a single haplotype, “A”, which represented an average of 83% of all samples in each population (Appendix 6). Populations from continuous sites also displayed higher haplotypic diversity when compared with populations from other site types (F(2, 5) = 40.16, p = 0.001). This was the case in spite of possessing smaller sample sizes ([mean ± s.e.] isolates: 29.5 ± 0.5, connected sites 24.0 ± 6.6, continuous sites 23.3 ± 3.5).

Table 5.3 Haplotype number (H) and diversity (± s.e.) (h) for each population.

M. cervinipes U. caudimaculatus Population n H h nH h Connected R1 28 70.73±0.07 32 2 0.27 ± 0.09 Connected R2 18 70.71±0.06- -- Connected R3 26 12 0.85 ± 0.04 11 2 0.18 ± 0.14 Connected M1 26 70.72±0.06 29 4 0.30 ± 0.10 Connected M2 25 90.77±0.06- -- Connected M3 30 11 0.82 ± 0.05 - -- Connected M4 22 90.75±0.07- -- Connected M5 26 10 0.83 ± 0.05 - -- Corridor L 11 50.80±0.08 19 1 0.00 ± 0.00 Corridor BR 23 50.74±0.04 9 3 0.72 ± 0.10 Corridor K/H 32 80.82±0.04 7 1 0.00 ± 0.00 Corridor K/S 26 60.79±0.05 11 - Corridor S 12 50.67±0.14 3 1 0.00 ± 0.00 Isolate B 49 80.72±0.04 29 4 0.54 ± 0.05 Isolate L 31 50.74±0.04- -- Isolate W 30 70.75±0.05 30 2 0.30 ± 0.09 Continuous 1 28 15 0.94 ± 0.02 30 10 0.87 ± 0.03 Continuous 1A 24 12 0.84 ± 0.05 - -- Continuous 2 32 17 0.93 ± 0.03 22 7 0.87 ± 0.04 Continuous 2A 18 10 0.87 ± 0.05 - -- Continuous 3 11 90.97±0.05- -- Continuous 4 22 15 0.97 ± 0.02 18 7 0.85 ± 0.06

- denotes location either not sampled or sample size not sufficient for inclusion.

Chapter 5. Gene flow - 95 -

5.3.3 Allele distribution and heterozygosity All microsatellite loci were found to be highly polymorphic with the total number of alleles per locus ranging from 13 to 25 for M. cervinipes and 15 to 19 for U. caudimaculatus (Appendices 9 and 10). A maximum of 16 alleles per locus were detected per population for both species. The alleles present within populations from the connected, corridor and isolate sites generally consisted of a subset of those present in the continuous forest with 83 of the 100 alleles (83%) across all loci from M. cervinipes and 78 of the 84 alleles (93%) from U. caudimaculatus present within the continuous habitat.

Several M. cervinipes samples failed to amplify at the Mc2E, Mc2O and Mc2P microsatellite loci (Appendix 9) and any individuals that amplified at only a single locus were excluded from the data set. Failure to amplify is often attributed to mutations at the priming site, resulting in null alleles, which can in turn result in a homozygote excess and a departure from Hardy-Weinberg equilibrium. In this instance however, few departures from Hardy-Weinberg were observed after Bonferroni correction (Appendix 11) with no apparent conformity in the patterns of the observed deviations.

The majority of U. caudimaculatus samples amplified at all loci (Appendix 10) and relatively few departures from Hardy-Weinberg were detected by exact tests (Appendix 12). Most departures were attributable to one population (Continuous 1)

however, the observed disparity between Ho and HE values were very small for this population (Table 5.4, Appendix 12), and some loci exhibited a heterozygote excess. These results were therefore attributed to sampling effects and the population was included in subsequent analyses.

FIS measures the extent of inbreeding within a population and can range from -1 to 1, where -1 represents complete outbreeding or 100% heterozygotes and 1 equates to complete inbreeding with 100% homozygotes. The majority of observed FIS values were very close to zero, with the largest deviation in the direction of inbreeding being 0.14 and 0.16 for M. cervinipes and U. caudimaculatus respectively. These data confirm that inbreeding was not a problem in either species (Table 5.4). This was further supported by the comparable M-ratios (when Chapter 5. Gene flow - 96 -

averaged across loci per population) within habitat types (M. cervinipes: F(3, 18) =

2.89, p = 0.064, U. caudimaculatus: F(3, 10) = 0.353, p = 0.788).

Table 5.4 Descriptive statistics of variation for nuclear markers for populations averaged across all loci. N = sample size, A = mean number of alleles, Ho = mean observed heterozygosity, He = mean expected heterozygosity, FIS = fixation index, M = M ratio.

Site N A Ho He F IS M ± s.e.

M. cervinipes Connected R1 28 9.6 0.77 0.83 0.07 0.42 ± 0.03 R2 18 8.6 0.85 0.86 0.01 0.42 ± 0.04 R3 26 10.4 0.86 0.86 0.00 0.45 ± 0.03 M1 26 10.4 0.67 0.79 0.14 0.39 ± 0.04 M2 25 9.4 0.85 0.84 -0.02 0.48 ± 0.06 M3 30 11.2 0.77 0.85 0.09 0.46 ± 0.02 M4 22 10.0 0.88 0.87 -0.03 0.43 ± 0.03 M5 26 11.2 0.79 0.84 0.06 0.42 ± 0.06 Corridor L 11 7.2 0.85 0.84 -0.02 0.40 ± 0.04 BR 23 8.8 0.90 0.84 -0.08 0.41 ± 0.03 KH 32 10.2 0.74 0.83 0.11 0.39 ± 0.06 KS 26 10.2 0.85 0.85 -0.02 0.40 ± 0.03 S 12 7.6 0.84 0.85 0.00 0.30 ± 0.01 Isolate B 49 9.2 0.77 0.81 0.06 0.41 ± 0.04 L 31 8.8 0.84 0.83 -0.01 0.41 ± 0.04 W 30 10.0 0.85 0.84 -0.02 0.39 ±0.02 Continuous 1 28 12.4 0.76 0.90 0.14 0.37 ± 0.03 1A 24 11.6 0.90 0.87 -0.04 0.37 ± 0.03 2 32 11.4 0.85 0.88 0.03 0.47 ± 0.03 2A 18 10.0 0.93 0.87 -0.08 0.40 ± 0.03 3 11 9.0 0.88 0.90 0.01 0.35 ± 0.05 4 22 12.0 0.91 0.86 -0.05 0.38 ± 0.03 U. caudimaculatus Connected R1 32 7.2 0.79 0.78 0.00 0.46 ± 0.04 R3 11 5.4 0.71 0.76 0.04 0.39 ± 0.06 M1 29 6.8 0.71 0.76 0.01 0.37 ± 0.03 Corridor L 19 7.0 0.71 0.73 0.01 0.35 ± 0.03 BR 9 4.2 0.78 0.72 -0.11 0.39 ± 0.05 KH 7 3.4 0.66 0.60 -0.12 0.51 ± 0.13 KS 1 2.0 - - - 0.67 ± 0.14 S 3 3.2 0.80 0.76 -0.12 0.41 ± 0.10 Isolate B 29 8.2 0.79 0.75 -0.05 0.34 ± 0.04 W 30 6.6 0.82 0.74 -0.11 0.47 ± 0.06 Continuous 1 30 13.8 0.90 0.90 -0.01 0.46 ± 0.02 2 22 12.6 0.76 0.90 0.16 0.41 ± 0.01 4 18 10.0 0.80 0.84 0.05 0.36 ± 0.03

- denotes insufficient data for parameter estimation.

Chapter 5. Gene flow - 97 -

Comparison of the five individuals represented by the anomalous mtDNA haplotypes revealed no evidence of nuclear divergence. While these individuals did display some unique microsatellite alleles, all individuals were heterozygotes and possessed at least one allele also present in non-divergent conspecifics. Given this, along with the low incidence of the individuals from the second clade and the wide spatial segregation of the individuals across the continuous forest (represented in four populations), individuals with divergent mtDNA haplotypes were retained within the dataset.

The mean number of microsatellite alleles per population across all loci was significantly higher in populations from within the continuous habitat relative to either the connected or isolate patches for U. caudimaculatus (F(2, 5) = 12.68, p = 0.01). No

difference in allelic richness was detected for M. cervinipes (F(2,14) = 3.15, p = 0.07). While expected heterozygosity levels were relatively high in all populations (Table 5.4), levels within the continuous forest were greater than those within either connected or isolated habitats for both species when averaged across loci per population (M. cervinipes: F(2, 14) = 8.26, p = 0.004, U. caudimaculatus: F(2, 5) = 26.2, p = 0.002).

5.3.4 Population differentiation – landscape scale Significant population differentiation was detected among M. cervinipes populations connected by the corridor, as well as those isolated from each other by the matrix for both mtDNA and nDNA markers (Table 5.5). The same results were obtained regardless of whether samples from the entire connected patch were used or only those samples obtained from the patch/corridor boundary. Disparate results were obtained from the markers for comparisons between populations within the continuous control with microsatellite results suggesting restricted or no gene flow and mtDNA showing no population differentiation.

MtDNA and microsatellite marker results both indicated significant differences in the structure of populations from within Connected R and each of the two isolated sites for U. caudimaculatus (Table 5.6). However, results from the two markers were not congruent for the remaining pairwise comparisons. mtDNA suggested similar population composition between i) the connected patches and ii) both pairs of populations within the continuous forest while nDNA indicated significant differentiation for all but one of the continuous pairs. Chapter 5. Gene flow - 98 -

Table 5.5 Summary table of the pairwise differentiation of M. cervinipes populations at a landscape scale.

mtDNA nDNA

Sites Exact p F ST Exact p F ST Connected Connected R1 / Connected M1 <0.001* 0.20 <0.001* 0.05 Connected R (all) / Connected M (all) <0.001* 0.09 <0.001* 0.03 Continuous Continuous 1 / Continuous 2 0.475 <0.01 0.020* 0.01 Continuous 3 / Continuous 4 0.765 <0.01 0.042* 0.02 Isolates Connected R1 / Isolate B <0.001* 0.20 <0.001* 0.05 Connected R1 / Isolate L <0.001* 0.23 <0.001* 0.08 Connected R1 / Isolate W <0.001* 0.18 <0.001* 0.05 Connected R (all) / Isolate B <0.001* 0.15 <0.001* 0.05 Connected R (all) / Isolate L <0.001* 0.13 <0.001* 0.07 Connected R (all) / Isolate W <0.001* 0.13 <0.001* 0.05

* denotes significant p-values after sequential Bonferroni correction.

Table 5.6 Summary table of the pairwise differentiation of U. caudimaculatus populations at a landscape scale.

mtDNA nDNA

Sites Exact p F ST Exact p F ST Connected Connected R1/ Connected M1 0.069 0.13 <0.001* 0.05 Connected R (all) / Connected M1 0.165 <0.01 <0.001* 0.05 Continuous Continuous 1 / Continuous 2 0.858 <0.01 0.307 0.01 Continuous 1 / Continuous 4 0.371 0.01 <0.001* 0.04 Isolates Connected R1 / Isolate B <0.001* 0.19 <0.001* 0.05 Connected R1 / Isolate W 0.003* 0.06 <0.001* 0.05 Connected R (all) / Isolate B <0.001* 0.20 <0.001* 0.04 Connected R (all) / Isolate W 0.003* 0.06 <0.001* 0.05

* denotes significant p-values after sequential Bonferroni correction.

Chapter 5. Gene flow - 99 -

At a slightly finer level of scale, significant differences were detected between populations within the corridor system, including adjacent populations, for both species (Appendices 13-16). Differences were more pronounced for M. cervinipes than for U. caudimaculatus and for nDNA than mtDNA. No isolation-by-distance effect was apparent for either species (Table 5.7).

Table 5.7 Isolation by distance within the corridor system for M. cervinipes and U. caudimaculatus.

G r2 p M. cervinipes mtDNA 1.25 0.14 >0.05 nDNA 0.79 0.05 >0.05 U. caudimaculatus mtDNA -0.65 0.03 >0.05 nDNA -0.12 0.01 >0.05

5.3.5 Population differentiation – within patch scale Considerable population differentiation was apparent over a scale of 300m within both connected patches and the continuous forest for M. cervinipes (Table 5.8). These differences were evident for both genetic markers, however, nDNA results were generally more discriminating. Only a single within-patch comparison was possible for U. caudimaculatus with nDNA (p = 0.003) but not mtDNA (p = 0.09) showing genetic structure.

Chapter 5. Gene flow - 100 -

Table 5.8 Pairwise differentiation of populations for M. cervinipes at a within patch scale. ** = p <0.01, *** = p <0.001.

mtDNA nDNA Sites p F ST p F ST Within Connected R R1 / R2 *** 0.14 *** 0.02 R1 / R3 *** 0.12 *** 0.02 R2 / R3 *** 0.15 ** 0.01 Within Connected M M1 / M2 ** 0.04 *** 0.02 M1 / M3 0.21 0.01 ** 0.02 M1 / M4 0.96 <0.01 *** 0.04 M1 / M5 0.33 <0.01 *** 0.02 M2 / M3 ** 0.03 *** 0.01 M2 / M4 ** 0.04 0.015 0.01 M2 / M5 *** 0.05 ** 0.01 M3 / M4 0.74 <0.01 ** 0.02 M3 / M5 0.63 <0.01 ** 0.01 M4 / M5 0.82 <0.01 *** 0.02 Within Continuous 1 / 1A 0.06 0.03 0.153 0.01 2 / 2A *** 0.06 *** 0.03

The location of sampling within Connected R influenced the results for pairwise tests between the patch and corridor sites for M. cervinipes. Despite significant population structuring among i) populations from all sites within Connected R, and ii) Connected R1 and all corridor sites, Connected R2 and R3 were shown to be similar to some corridor sites, primarily for the mtDNA marker (Table 5.9).

Table 5.9 Differences in the results of population structuring depending upon the location of sampling within Connected R.

Connected R1 Connected R2 Connected R3 mtDNA nDNA mtDNA nDNA mtDNA nDNA Corridor L****0.120.11 Corridor BR**0.13*** Corridor KH****** Corridor KS****** Corridor S****0.07*

* indicates significant population differentiation.

Chapter 5. Gene flow - 101 -

5.4 Discussion

Genetic theory predicts that isolation of habitat patches as a result of habitat fragmentation will reduce effective population size and genetic diversity within the population (Amos and Harwood 1998, Lindenmayer and Peakall 2000), will increase inbreeding levels (Frankham and Ralls 1998, Soulé and Mills 1998) and may eventually compromise the adaptive potential and fitness of a population (McCaughley 1993, Saccheri et al. 1998). These disruptions arise from a reduction of, or more likely, a cessation to dispersal which leads to decreased levels of gene flow (Burgman and Lindenmayer 1998). Conservation biology theory predicts that construction or retention of a wildlife corridor between two otherwise isolated habitat patches may increase or maintain the levels of inter-patch dispersal respectively (Kozakiewicz 1993, Primack 1993, MacMahon and Holl 2001). This can translate into gene flow between the populations within the habitats, which can in turn lead to elevated levels of genetic diversity, and a general homogenization of their genetic composition. This has the potential to ameliorate any potentially negative effects of isolation (Harris and Scheck 1991, Merriam 1991, Bolen and Robinson 1995).

5.4.1 The effectiveness of corridors in maintaining genetic diversity Allelic and haplotypic richness and genetic diversity levels were generally higher in continuous habitats compared with isolated patches. This indicates that populations in remnant habitats i) have been affected by isolation, most probably due to the combination of large fluctuations in population size and a lack of dispersal which has led to genetic drift and ii) could benefit from management strategies designed to increase gene flow. These results are in accord with those of Campbell (1996) who showed significant reductions in haplotype diversity for U. caudimaculatus within habitat fragments on the Atherton Tablelands relative to populations in the continuous forest and similar, yet non-significant trends for M. cervinipes. Differences in the extent to which the allelic/haplotype richness was reduced is probably a function of the relative individual body size of the two species. Larger bodied individuals generally have greater resource requirements, leading to fewer animals being accommodated within a habitat patch (Wood 1971, Wellesley- Whitehouse 1981) and smaller populations being more susceptible to genetic drift (Campbell 1996).

Chapter 5. Gene flow - 102 -

The presence of a wildlife corridor between two remnant habitats did not apparently provide any significant benefit to the populations within the connected patches with respect to haplotype/allele richness or diversity for either species. Genetic diversity/richness were comparable in remnant patches regardless of the presence of a corridor. Campbell (1996) postulated that the greater conservation of haplotypic richness in M. cervinipes relative to U. caudimaculatus in remnant patches may be due to more effective use of wildlife corridors by the former species. The present data however, do not support this hypothesis but add weight to the argument that relative population size is the primary factor that influences haplotype richness.

On the basis of his mtDNA analysis and the allozyme data of Leung et al. (1993), Campbell (1996) also suggested that populations of M. cervinipes within remnant patches have not experienced severe reductions in effective population sizes, also known as a bottleneck effect. All measures of inbreeding based on microsatellites from this study support this notion. Thus, while populations have undergone a reduction in haplotypic richness, most likely due to genetic drift, effective population sizes have remained large enough to prevent significant inbreeding and any associated potentially detrimental effects. This outcome was again independent of any corridor effect and applied to both species.

5.4.2 Gene flow among M. cervinipes populations The wildlife corridor failed to provide any apparent connectivity between the populations in connected patches in the form of gene flow with mtDNA and nuclear markers producing highly congruent results for M. cervinipes. Both markers indicated a highly significant degree of population differentiation, and hence limited, if any, gene flow among populations confined to remnant patches regardless of whether or not they were linked by a corridor. Populations within remnant patches connected by the corridor were as dissimilar to each other as those in patches totally surrounded by pasture matrix.

As isolation-by-distance effects were not present within the corridor for either species, and the corridor has been shown to be suitable for residency and/or transit, social factors (termed “behavioural distance” by Chesser 1983b) are the likely cause of the observed population differentiation. That fine scale geographic structuring of populations was also present within continuous habitat therefore suggests that this scenario represents the natural state for M. cervinipes. This interpretation of the Chapter 5. Gene flow - 103 - data contradicts the suggestion by Campbell (1996) who suggested that wildlife corridors which can support resident populations may be critical for the conservation of species that show fine-scale structuring.

The formation of social groups that are unreceptive to new individuals is likely to be a response to the naturally heterogeneous distribution of food and shelter resources within the available habitat (Chesser 1998, Hughes 1998). Given the dynamic nature of ecosystems, and the temporal variation in food resources used by small mammals within rainforest habitats of the Atherton Tablelands (Elmouttie and Streatfeild unpublished data), social groups may be temporary arrangements which break down in times of resource abundance, as has been shown to occur in other mammal species (Metzgar 1979, 1980, Brandt 1992). Such relaxation of social systems could lead to the integration of family groups and consequently homogenize gene pools that had been previously segregated. However, in the present study, a relatively high number of mtDNA haplotypes and nDNA alleles were observed in only one of the two connected patches or in the corridor but not in either of the connected patches. This suggests that discrete social structuring and significant differentiation among populations is representative of the common state rather than a limited “snapshot” taken at a time of resource limitation and consequently high social structuring.

Discordant results were obtained for pairwise comparisons within the continuous habitat with nDNA but not mtDNA indicating population differentiation between the populations separated by the greatest distance (Continuous 1 & 2, 3 & 4). However, all but one other pairwise comparison between populations along the creekbank indicated significant structuring using at least one marker. This includes the pairwise comparisons that were performed (Appendices 13 and 15), but were not mentioned in the results as they have no direct comparison within the corridor system (eg Continuous 1A & 2A, 2 & 1A). Thus, the non-significant results are thought to be attributed to sampling rather than gene flow between the populations. This is further supported by the fact that 48% (Continuous 1 & 2) and 67% (Continuous 3 & 4) of the total number of haplotypes within each pair of populations occur in only one of the populations.

Thus, although the social assemblage of M. cervinipes within the wildlife corridor resembles that within the continuous forest, the wildlife corridor appears to confer Chapter 5. Gene flow - 104 - little benefit. This is apparently due to the large differences in the corridor length and the spatial scale at which the species operates. With behaviourally mediated social organization occurring over spatial scales of hundreds of metres, any corridor that extends for kilometers is unlikely to assist gene flow among populations. Given the extent of past rainforest clearing on the Atherton Tablelands, distances in the order of kilometers are common between remaining habitat remnants.

5.4.3 Gene flow among U. caudimaculatus populations Only one of the two pairwise comparisons among populations from within the continuous forest showed genetic structuring with the discrepancy likely to be due to physical parameters. The pair of populations located along the same creek (Continuous 1 & Continuous 2), as per the configuration of the corridor system, showed no population differentiation. However, the alternate pair of populations (Continuous 1 & Continuous 4), located in different drainage basins and separated by a steep embankment which may act as a barrier to dispersal, showed significant structuring. While the inclusion of samples from Continuous 4 provided additional information with respect to the diversity of haplotypes and alleles within the continuous habitat, this comparison highlights the importance of replicating the conditions of the corridor system within the continuous control. Lindenmayer and Peakall (2000) also reported genetic structuring between populations of native Australian rodents located on parallel creekbeds. It was suggested that these results may indicate inequality between the geographic and effective distance between the populations (Lindenmayer and Peakall 2000).

If only the truly replicated comparison from this present study is considered, i.e. that with both pairs of populations along the same creek and with no apparent physical or geographical barriers, then both mtDNA and nDNA indicate that U. caudimaculatus does not show structuring over the distance of kilometers within the continuous forest. Given this result, considerable gene flow would also be expected between the connected patches should the corridor function as desired.

This expectation was not met, with significant population differentiation occurring between both connected and unconnected populations. Exact tests based on GST have been described as very sensitive to small differences in allelic frequencies and have a propensity to overemphasize the significance of population differentiation. (Balloux and Lugon-Moulin 2002). However, the p-values for all comparisons Chapter 5. Gene flow - 105 - between pairs of populations from within connected or isolated remnant sites were highly significant (<0.001), clearly suggesting very restricted or no gene flow between the populations. The distribution of haplotypes and microsatellite alleles along the corridor provides additional evidence for the apparent lack of gene flow among the connected patches. Only two sites within the corridor produced sufficient U. caudimaculatus individuals for meaningful pairwise comparisons with the population from one of these sites containing several alleles that were not found in populations from either of the connected patches. Both populations also differed significantly from each other and from either of the connected populations. Even though a smaller proportion of the total population was sampled within the connected patches than at each corridor site, the sample sizes from the connected sites were considerably larger. Hence, if gene flow was occurring between the patches, it would be expected that alleles present within the corridor would be also detected within the patches, and that the corridor populations would represent the genetic structure within the connected patches. Given that population differentiation is also found where the habitat of Connected R narrows considerably, it is possible that narrow linear habitats influence the extent of gene flow by enhancing antagonisitic interactions between dispersing and resident individuals.

In contrast to the results for M. cervinipes, results for U. caudimaculatus are discordant between mitochondrial and nuclear markers among the connected populations. mtDNA results suggest that the wildlife corridor facilitates gene flow while those in remnant habitats isolated by the matrix show marked differentiation. These results initially appear to provide evidence for the effectiveness of wildlife corridors with respect to providing gene flow, however it is highly likely that the results are confounded by the severe reduction in mtDNA diversity. With a maximum of only four haplotypes per population, very little variation in haplotype frequency will provide significant differences in genetic structure. Thus, significant differences were observed in only some of the pairwise comparisons in spite of all of the populations being dominated by the same mtDNA haplotype (A), and all of the populations to which Connected R was compared possessing a unique haplotype. While discordant results between mtDNA and nDNA markers are not unusual, it is usually the mtDNA that indicates structuring to a greater extent due to the effects of maternal inheritance and female philopatry (Moritz et al. 1987). For these reasons, greater emphasis has been placed on the results obtained from the nDNA data. Chapter 5. Gene flow - 106 -

5.4.4 The effect of social structuring on wildlife corridor effectiveness The organization of a species into social groups can restrict the extent to which a wildlife corridor can assist a species in several ways. Most obviously, direct gene flow is unlikely to occur if dispersing individuals are unable to traverse the length of the corridor to an alternate connected patch due to resistance from corridor-resident conspecifics. However, even if movement occurs, the social exclusion of an individual from the breeding unit of a new population will restrict gene flow in either a direct or generational manner.

The connection between social exclusion and/or territorial defence in linear habitats and gene flow has been identified within the wildlife corridor literature with Lidicker (1999) and Soulé and Gilpin (1991) recognizing that the rate and directionality of movement by individuals may be affected by the presence of conspecifics within the corridor. In support, Aars et al. (1998) found bank vole populations to show highly significant differentiation for mtDNA along a linear riparian zone while gene flow for the same species within a non-linear habitat was less restricted. Scant attention however, has been given to the possibility that wildlife corridors may be of little value to species which exhibit well developed social structures that operate on a spatial scale much smaller than that of the corridor length (but see Amarasekare [1994] who, in a study of the ecology of the banner-tailed kangaroo rat, suggested that corridors of kilometers in length were unlikely to benefit the species due to the small scale at which structuring occurs).

The results of this study also indicate that the propensity of a species to show social structuring within a linear habitat cannot necessarily be predicted from their social organization within the continuous control site. It may be expected that species showing social organization within the continuous habitat will show similar behaviour within the connected patches and the corridor, as was demonstrated by M. cervinipes. However, the significant differentiation of U. caudimaculatus populations along the corridor and within Connected R at a point where the patch narrows considerably, suggests that species which can exhibit panmixia within a large habitat tract may also show social structuring within linear habitats.

The presence of discrete social groups also alters the traditional view that if a patch becomes vacant through extinction, recolonization will occur through the corridor by individuals from the alternate patch. In the event that a social group goes extinct Chapter 5. Gene flow - 107 - within a connected patch, it is highly likely that the vacant area will be recolonized by another social group from within that patch. Further, with social groups relatively independent of each other within a patch, as indicated by the lack of gene flow, it is unlikely that demographic or genetic stochasticity will lead to the simultaneous extinction of several social groups. Thus, for an entire patch to become vacant it is probable that the most likely cause would be environmental stochasticity or a catastrophe. Under these circumstances recolonization may not be possible as individuals within the corridor and an alternate connected patch are also likely to be affected (Earn et al. 2000). In the event that only one connected patch does become vacant, recolonization will most probably result from range expansion of individuals resident within the corridor. In this instance, the corridor would play a vital role in the recolonization process even though recolonizing individuals may not originate from the alternate patch.

5.4.5 The significance of appropriate sampling strategies

The data from both species within the corridor system highlights the importance of sampling at the patch/corridor boundary. Where species show fine-scale population structuring relative to the size of the patch, it is unlikely that individuals living elsewhere in the patch will have access to the corridor. This situation can arise from the corridor being outside the home range of the individuals such that the corridor is not encountered, or through intolerance of individuals from an alternate social group at the corridor boundary. Thus, sampling strategies that sample areas other than the corridor/patch border may not accurately interpret the effectiveness of the corridor as those individuals with the highest chance of benefiting from the corridor may not be included.

The importance of selecting appropriate control sites to provide a better understanding of the biological processes that operate within a corridor system was also reinforced. If populations within a contiguous habitat are assumed to exist in a panmictic state, then significant genetic structuring among populations connected by a corridor would be interpreted as indicating a lack of gene flow due to some inherent property of the corridor. Such population differentiation between connected populations was found for M. cervinipes in this study, however, equivalent results were found within the continuous control. This indicates that the differentiation within the corridor was most likely due to social factors in the target organisms and Chapter 5. Gene flow - 108 - thus it is the behaviour of the target species, rather than any physical characteristics of the corridor, that appears to be the critical factor limiting gene flow. U. caudimaculatus meanwhile showed population differentiation in linear remnants that was not detected in the equivalent configuration within the continuous habitat. This suggests that the configuration of the habitat may influence the behaviour of a population. While the end result remains the same for both species, i.e. that gene flow does not occur between populations connected by the corridor, the reason that each species fails to benefit from the corridor appears to be entirely different. Thus in the absence of comparable data from continuous control sites, the correct overall conclusion may be drawn but the wrong mechanism could be inferred. This would demonstrate a lack of understanding about the corridor system and could have serious ramifications for the perception of wildlife corridors as a conservation strategy.

The inclusion of isolate control sites was also vital in interpreting the erosion of genetic diversity within the fragmented landscape as they allowed the relative influence of the corridor to be interpreted. Diversity and haplotype/allele richness levels were similar or lower in the connected site relative to the isolated sites indicating that the wildlife corridor does not buffer populations against the effects of habitat fragmentation. Without the inclusion of populations from isolated sites, the decreased levels of diversity and richness in the connected sites relative to continuous habitats would be known, but not the effectiveness of the corridor in providing any positive benefits to the populations within the connected habitats relative to those in isolated sites.

Admittedly not all corridor systems will have appropriate controls. This is especially the case where corridors are long and/or no continuous habitat is available locally. This is likely to be true in areas where habitat clearing has been so vast that no continuous habitat remains. Suitable controls may also be difficult to locate in areas with a lack of isolated remnants, or in landscapes where land use is varied and in which a uniform matrix cannot be found. In the absence of the required control sites, the use of pairwise comparisons is inappropriate but in some instances, genetic markers may still be applied, either through the investigation of paternity, or via the use of assignment tests (Paetkau et al. 1995, Waser and Strobeck 1998, Berry et al. 2004). Both techniques assess connectivity solely from within the corridor system and do not rely on comparisons with control sites. It must be Chapter 5. Gene flow - 109 - recognized therefore, that while these approaches have the potential to test for connectivity, they do not provide any indication of the effect of the corridor on the populations relative to unconnected sites. While limited by the lack of comparative data, these approaches offer the advantage of being suitable for implementation much sooner than comparative population genetic approaches following corridor construction.

5.4.6 Summary Neither M. cervinipes nor U. caudimaculatus populations derived any apparent genetic benefit from the retention of a wildlife corridor between two remnant patches of rainforest. Populations of both species experienced an erosion of genetic richness and diversity for at least one form of DNA regardless of the presence of a wildlife corridor. Further, populations in connected and isolated sites showed the same extent of population differentiation.

Populations of M. cervinipes appeared to function similarly within the corridor system as they do within continuous forest where they apparently form highly structured social groups that differ in genetic composition from nearby groups. Given this, the presence of a corridor has little chance of offering an avenue for gene flow solely as a consequence of the behavioural attributes of the species. While the wildlife corridor is therefore somewhat irrelevant to the process of gene flow in the species, this should be viewed very differently to the conclusion that the corridor does not work for a species in which gene flow would be expected as was the case for U. caudimaculatus. Gene flow between the connected patches was expected due to the presence of panmixia along a creek within continuous forest but this expectation was not met with significant differentiation found among populations within the connected habitats. - 110 -

6. GENERAL DISCUSSION

6.1 The benefits of an integrated approach

The assessment of wildlife corridor effectiveness is often confounded by the differences in the meaning of the terms “use” and “function”, and the implications that these terms have for data interpretation. Whereas the term function implies an operation or fulfillment of a task, use simply refers to utilization. Studies investigating wildlife corridor effectiveness often attempt to assess and comment on corridor function yet only investigate corridor use. Where the specific aim of a corridor is to provide additional habitat, the terms are interchangeable and this approach is perfectly valid. However, should the desired function of the corridor involve interaction among the populations within the connected patches in either a demographic or genetic manner, i.e. connectivity, the terms have distinctly different implications.

While only based on one study system, the data obtained in this study highlights how inaccurate conclusions can be drawn on the capacity of wildlife corridors to facilitate connectivity in the absence of specific investigation into this function. That the corridor under investigation i) was shown to be structurally similar to the vegetation within the remnant that it connects, ii) contained resident individuals of all demographic classes and iii) supported breeding populations, suggests that the linkage could be viewed as successful in its attempt to link two remnant rainforest patches. This was exemplified by the high numbers of both species at the corridor site which least resembled the natural vegetation and could have been initially perceived as a potential barrier to movement and/or gene flow. Connectivity via a corridor requires animals of breeding age to either live within the corridor, or to travel along the corridor, depending on relative dispersal capabilities. With these requirements met, it could easily be assumed that connectivity between the connected patches would automatically occur. However, the genetic results obtained did not accord with this expectation.

Further, the positive correlation between the proximity of a corridor site to one of the connected patches and the number of U. caudimaculatus individuals could readily Chapter 6. Discussion - 111 - be interpreted as evidence that the connected patch is acting as a source population with animals dispersing into the corridor. Whether this would be viewed as a positive sign that the corridor facilitates movement or a negative indication of the corridor acting as a sink and draining the source population of individuals (Pulliam 1988, Henein and Merriam 1990, Bolen and Robinson 1995) becomes somewhat irrelevant in light of the genetic data. Both genetic markers indicate a high level of population structuring between the connected patch and the corridor sites. That several alleles were also found at the closest corridor site and not in the connected patch, despite the larger sample size at the latter, strongly suggests that the connected patch is not acting as a feeder population for the corridor.

These examples demonstrate the risk of inferring connectivity based solely on habitat use and demographic parameters and the advantage of incorporating a genetic component. It must also be realized that while genetic techniques provide an indication of gene flow, they also indicate the level of connectivity per se. If gene flow is shown to be absent within a given corridor system, then it can be assumed that the breeding subsets of the populations within the connected patches have little to no interaction and that demographic benefits are also unlikely to result from the presence of a corridor.

While genetic data can indicate the level of connectivity, it is the combined use of genetic and ecological methods that allow a more comprehensive understanding of a wildlife corridor system. The use of genetic markers alone is sufficient to determine whether connectivity occurs between remnant patches, and should it be shown to occur, this information may suffice. However, if connectivity is shown to be absent, and the reason is not evident from genetic data, ecological studies may indicate the mechanisms responsible. Such mechanisms can include the discontinuity of suitable habitat, use of the corridor by inappropriate subsets of the population and social structuring of the species. While the ideal method of assessing connectivity would involve experimentally removing all individuals from a habitat and assessing recolonization, this has major ethical considerations and is unlikely to possible in the majority of instances. Thus, an integrated approach to assessing whether wildlife corridors facilitate connectivity is clearly the preferred alternative. This supports calls for integrated studies that were made in the 1990’s, yet were generally not heeded (Lindenmayer 1994, Bierregaard et al. 1997).

Chapter 6. Discussion - 112 -

6.2 The role of scale in wildlife corridor assessment

Selection of the appropriate scale of investigation has been identified as a primary concern for ecologists (Meetenmeyer and Box 1987, Turner et al. 1989, MacNally and Quinn 1998) with proper analysis requiring that the scale of measurement and the response of the organism occur within the same domain (Wiens 1989). Spatial scale-dependency may be either continuous, with every change in scale eliciting an associated ecological change, or it may be confined to domains whereby regions of the continuum contribute equally to an ecological change (Wiens 1989).

Ludwig et al. (2000), in defining a conceptual framework for scaling in landscapes, identified three interacting processes involved in scaling: environmental scales, the scale of organism response and observational scale. All three of these factors are relevant to wildlife corridor assessment and the data collected here highlights the importance of their consideration in the planning of sampling strategies.

6.2.1 Environmental scale The discipline of landscape ecology focuses on the spatially explicit patterns of landscape mosaics and interactions among their elements – primarily at the scale of kilometers (Wiens et al. 1993). Consequently, a commonly adopted approach is to consider the landscape in a binary manner (e.g. Söndgerath and Schröder 2002, Fahrig 2003). This practice involves identifying tracts of “suitable” habitat within the landscape and distinguishing them from the “unsuitable” matrix. Representation of the landscape in this manner enables rapid visualization of the extent of habitat clearing, easy identification of isolated habitats that may benefit from management efforts, and the construction of models that examine the movement of individuals between habitats via the matrix. Although heterogeneity of the habitat is well recognized by landscape ecologists at a large (habitat patches vs matrix ) scale, this method of landscape interpretation usually considers the patch as the smallest unit of investigation (i.e. the finest “grain”) and does not allow for any consideration of within- (but see Fahrig and Merriam 1985, 1994). Highlighting the large scale approach of the discipline, grain size of a landscape mosaic has been described as “the average diameter or area of all patches present” (Forman and Godron 1986). This approach persists despite it being recognized that habitats are heterogeneous at virtually any level of scale (Wiens 1997), that organisms differ in Chapter 6. Discussion - 113 - the scale of their ecological neighbourhood (Wiens 1989), that species perceive the degree of heterogeneity in the landscape differently (Wiens and Milne 1989) and that proper analysis requires that the scale of measurements and the organism’s response fall within the same domain (Wiens 1989).

The field of wildlife corridor research, like that of its parent discipline of landscape ecology, largely views the landscape, and in particular patch/corridor/patch systems in a binary manner. This approach has led to the simplistic view that connected habitats are internally homogenous entities occupied by a single panmictic population. This has resulted in corridor systems being viewed as an area of “suitable” habitat within a matrix and by default it thus represents the smallest scale or “grain” of investigation. It is also apparent from the design of field studies which usually consider the patch as the smallest sampling unit regardless of the size or life-history of the organism investigated (but see Boudjemadi et al. 1999 and Haddad et al. 2003 [who matched lifetime movement ranges of the study species with patch size within an experimental system]). This is despite most corridor studies occurring within fragmented landscapes in which habitat heterogeneity occurs on two spatial scales: the general landscape consisting of preferred habitats within a matrix of unsuitable habitat (habitat patches) and the non-continuous distribution of resources (resource patches) that occur within the habitat patches (Hanski 1994).

In the present study, the identification of processes operating at a within patch scale, which ultimately affect the effectiveness of the corridor, clearly demonstrates that consideration needs to be paid to the scale at which sampling occurs within corridor systems. The prevalent strategy of sampling at the patch level may be inappropriate for many species as it may overlook the processes that are responsible for the pattern that is observed at the next higher level of scale. Conversely, where habitat patches are relatively homogenous, as may occur in monoculture plantations or forests dominated by a single species, patch size may approximate the appropriate “grain” and the location of within patch sampling may not be as crucial.

6.2.2 The scale of organism response It is widely recognized that the response of organisms to corridors will largely depend on the size of the organism relative to the corridor length. Entomologists have noted that the majority of wildlife corridor research concerns highly vagile Chapter 6. Discussion - 114 - species of birds and mammals (Mönkkönen and Mutanen 2003), the results of which may bare little relevance to corridor/ interactions. This study intentionally incorporated two species with similar life history characteristics, but which differed in their dispersal and movement capabilities. Rather than affecting the response of the species to the corridor, vagility appears to be associated with the benefits that the corridor can confer. Both species utilized the corridor yet the differential vagility of the species, which is likely related to body size, affected how the species were likely to benefit from the corridor. Firstly, the larger species, which moved greater distances and was present within the system at lower density presumably due to higher resource requirements, was more likely to benefit from the corridor due to the greater effect habitat fragmentation has had on genetic diversity relative to the smaller and less vagile species. Also, the larger species was more likely to benefit from the presence of the corridor for the purpose of recolonization in the event of local extinction due to the large distances known to be moved by both sexes. While these examples are specific to the investigated corridor system, the concept of differential organism response is likely to be universal. The scale of organism response is therefore crucial to wildlife corridor assessment and planning.

6.2.3 Observational scale The importance of considering observational scale was clearly demonstrated by the inferences that can be obtained from the ecological and genetic data. Rather than being a constant event, gene flow can result from episodic events which may be missed by ecological studies. The use of genetic markers therefore opens the observational “window” (Allen et al. 1993) and allows the cumulative effect of all dispersal events since the formation of the corridor to be viewed. The wider the “window”, the more likely it is that the observations will accurately reflect real organism-environment interactions (Ludwig et al. 2000). Put into a methodological sense, it is more likely that the results from the genetic markers reflect the true historic situation rather than the narrow field of view provided by trapping studies (Sarre 1995). There is a caveat to the use of genetic markers for population analysis in the manner of this study however, as they are only appropriate for use in landscapes that have experienced long term fragmentation and give no indication of the timing and prompts for dispersal.

Chapter 6. Discussion - 115 -

The three factors involved in scaling, as identified by Ludwig et al. (2000), were all encompassed in the design of this current study. Spatial scale was investigated at a landscape and within patch level, variation in organism response was assessed by including two species with different movement capabilities, and the use of ecological and genetic methods provided for a comparison in observational scale. Rather than considering each factor to have a “correct” scale for investigation in a corridor system, it was apparent that investigation at multiple scales can lead to a more comprehensive and informed understanding of the system.

6.3 The suitability of wildlife corridors for other small mammal species

The lack of any apparent physical or geographic barrier between either the populations within the connected habitats, or those resident within the corridor, leaves behavioural distance (Chesser 1983b) as the most likely explanation for the differences in the genetic composition of populations between the sites. The segregation of social units through behavioural characteristics allows for the differentiation of populations despite the apparent ability of individuals to move between the social units (Selander 1970, Chesser 1983b). Rather than being an interesting trait of the species studied in this corridor system, a restriction in gene flow has the potential to occur in wildlife corridors for species that naturally show behaviourally mediated social structure. Interestingly, it is small mammals, on which many wildlife corridor studies have been undertaken, and for which many wildlife corridors are constructed and/or proposed, that commonly display this kind of social behaviour (Chesser 1998). Genetic differentiation over a relatively small scale has been documented for Peromyscus leucopus (500m) (Mossman and Waser 2001) and Clethrionomys glareolus (1km) (Aars et al. 1998) which are two species commonly used in wildlife corridor studies. Yet these results suggest that if social structuring at a scale much less than that of the corridor length is a normal and temporally stable trait of a species, then it must be considered that the inclusion of a corridor between two isolated patches may not automatically lead to an increase in gene flow.to the desired levels. Fine-scale structuring has also been demonstrated in another Australian rodent (Rattus fuscipes - ≤1km), however limited dispersal rather than social structuring was suggested as the reason for the genetic differentiation (Peakall et al. 2003).

Chapter 6. Discussion - 116 -

This is not to say that wildlife corridors are of no conservation benefit to small mammals. Mansergh and Scotts (1989) clearly demonstrated the benefit of a wildlife corridor to the mountain pygmy-possum, Burramys parvus. In this instance a road had bisected the over-wintering habitat of males and females leading to a decrease in population levels. The construction of a corridor in the form of a road underpass enabled movement between the discrete habitats and re-established normal population function. This is a prime example of a corridor having a specific function and the effectiveness of the corridor being measured against the desired outcome. Further, the addition of habitat to remnant patches may allow for larger population sizes which can in turn buffer the effects of isolation (albeit that a non- linear configuration may better meet this objective).

The results of this study suggest that even a limited knowledge of the social biology of a species can be used in a predictive manner for the potential effectiveness of a corridor in terms of connectivity. Although predominantly discussed in relation to mammals due to their propensity for social structuring and high conservation interest, these generalizations may also be applicable to other taxa that show similar traits. While it is acknowledged that the effectiveness of all wildlife corridors will not be constrained by the social behaviour of a species, the data from this study system illustrates that they can have a dramatic effect on gene flow and connectivity. Thus it is prudent for the potential effect of species behaviour on corridor effectiveness to be considered in all situations. Regardless of the species or habitat type, it must be kept in mind, that social interactions can vary with different habitats (Fuller et al. 1997) so any background data for a species that is to be used in a predictive manner should be taken from a habitat type similar to that within the corridor system.

6.4 Applications for management

Many papers have been written emphasizing properties of wildlife corridors that landscape managers should consider in decision making processes (Harris and Scheck 1991, Beier and Loe 1992, Harrison 1992, Lindenmayer and Nix 1992). These have primarily been structure focussed with strong emphasis placed on physical attributes such as corridor width and vegetation structure. These criteria are predominantly concerned with habitat use, which in itself is an important factor, Chapter 6. Discussion - 117 - however, to maximize the probability of a corridor enabling connectivity, additional factors also warrant consideration in the planning stage: i) Does the wildlife corridor have the potential to function in the desired manner? Much has already been written in previous sections about the potential impact of habitat continuity, population structure and social interactions between individuals of a species on corridor success. These points are therefore incorporated into Figure 6.1 but are only summarized here: • Is the habitat suitable for transit and/or residency along the length of the corridor? • Is the corridor used by a subset of the population that has the potential to provide connectivity? • Do all individuals within the connected patch have access to the corridor? i.e. would a corridor potentially link only a subset of the population in the connected patch and if so, is construction of a corridor justified? • Do social interactions in the corridor have the potential to reduce connectivity even if habitat use is continuous along its length? ii) Do the populations that are to be linked display any genetic differentiation? Habitat fragmentation has been shown to reduce genetic diversity and richness in isolated habitat patches. Chesser et al. (1980) suggested that the isolation of populations in remnant habitats may act to conserve a high proportion of the total genetic diversity as it could be expected that by chance alone, different alleles/haplotypes would be conserved in different habitats. The two species in this study showed that habitat fragmentation does indeed reduce genetic diversity in the remnant habitats and the constitution of mtDNA haplotypes for U. caudimaculatus populations in remnants demonstrated that isolated remnants can show remarkable uniformity in genetic structure despite isolation drastically reducing the total haplotypic richness. If the desired effect of a corridor is to increase the diversity of the populations within the connected patches, it is therefore necessary to ensure that the populations that are to be linked display a degree of genetic differentiation. This ensures that gene flow resulting from the provision of a corridor is beneficial to the population by the introduction of new genetic material (Figure 6.1).

A pilot study on the genetic structure of populations of the target species would indicate the degree of differentiation, however, it is recognized that in many Chapter 6. Discussion - 118 -

N

Recolonization unlikely

.

t

p

e c

Y

n

o

c

N through the corridor? the through

Can the species move move species the Can

s

i

h

t

f

o

n

o

i s

Y

s

u

c

s i

Potential d Is the corridorIs the used as habitat?

N

successful successful r

recolonization

o

f

4

.

4

.

5

n

Y o

social structuring?

i

t c

Does the species display Does the species display

e

s

e

e

s

RECOLONIZATION: * but potentially by Recolonization likely animals within patch* N N Connectivity unlikely Connectivity Y Y Connectivity unlikely Connectivity territorial defence? territorial N N length of the corridor? Does the species exhibit exhibit species Does the Can the species move the the move species Can the individual of dispersing of dispersing Genetic benefits unlikely Y Y Potential exclusion N N N Is the corridor used as habitat? capable of future reproduction? future of capable Is the corridor used by individuals individuals used by corridor the Is Y Y Y

Connectivity unlikely

Figure 6.1 Flowcharts outlining aspects of species behaviour and their potential affect on the ability of wildlife corridors to facilitate recolonization and gene flow. Chapter 6. Discussion - 119 - situations this option is not feasible either through lack of resources or genetic techniques not being developed for the species. The prudent course of action would therefore be to link remnant patches with either unfragmented continuous tracts of habitat or very large remnant patches which are likely to sustain high levels of both population numbers and genetic diversity. iii) Are the populations in the patches to be linked large enough to survive further subdivision? The construction of wildlife corridors can be considered as a management strategy for the conservation of endangered species, especially when a healthy population is resident within one isolated patch but not a neighbouring patch. In this instance wildlife corridors can be implemented with the aim of increasing available habitat in the form of the connected patch. However, it must be considered whether the original population has the capacity to provide colonists for the alternate patch both in the original instance and in the event of extinction. The opportunity exists for the healthy population to be jeopardized if independent social groups become established and the original population into effectively divided into two smaller sub- populations. This scenario could provide very different outcomes to those forecast if it was assumed that the animals from both patches would freely interact via the corridor.

iv) Can wildlife corridors be constructed in a manner that reduces the effect of social structuring? Populations of U. caudimaculatus were largely unstructured along the same creek line within the continuous forest yet they displayed significant differences in genetic structure in both connected patches and the corridor in places where the habitat narrows considerably. This suggests that social structuring in this species could be an artifact of an extremely narrow habitat (relative to the movement ability of the species) and that the potential exists for wildlife corridors to be constructed in a manner that circumvents this problem. For instance, a corridor could be constructed such that its minimum width is greater than the maximum diameter of a social unit of the species. This would assist in enabling dispersing individuals to transit around the resident animals without meeting resistance. This concept however, again relies on background information being available on the ecology of the target species and in the instance of a usually panmictic species, may be difficult to determine. This Chapter 6. Discussion - 120 - option is unlikely to work for species that show tight social structuring and are likely to pack social units into all areas of available habitat. v) Constructed versus retained wildlife corridors Given that the construction of wildlife corridors creates new habitat, they have great potential to initially promote connectivity due to the lack of residents. Thus, all new corridors have great potential to facilitate an initial influx of individuals into the newly connected populations. The duration of any connectivity will depend upon the extent of social organization of the species and the time taken for this to become established within the constructed corridor, if at all. Hence, any monitoring programs established to assess the effectiveness of a corridor should report results from a variety of intervals post-construction. The detection of desirable demographic changes and/or gene flow in the initial instance should not be viewed as a fait accompli if their presence over a prolonged duration (eg. decades) is desired.

With suitable study corridors and appropriate control sites proving difficult to locate, it is not surprising that comparative studies between maintained and constructed corridors have not been conducted at a landscape scale. Coffman et al. (2001) did consider this issue but at a very different temporal and spatial scale. The constructed corridors simply involved transplanting 0.3m x 0.9m sods of ground and laying them across a previously plowed/mown area. Hence, comparisons between an experimental scale study and the late successional rainforest of this present study are not at all meaningful.

The results from this one study system have raised a number of theoretical considerations concerning the successful integration of populations through the provision a wildlife corridor. This is highlighted by the large number of options in Figure 6.1 which have the potential to reduce the amount of connectivity. Merriam (1991, 1995), has long stressed that it is connectivity, not corridors, that is necessary to assist spatially segregated populations. The results of this study add further support to this view and highlight the numerous considerations that researchers and landscape managers face when considering wildlife corridors as a conservation strategy.

Chapter 6. Discussion - 121 -

6.5 Conclusion

Noss (1987) and Soulé (pers. comm. in Noss 1987) expressed concern that wildlife corridors may be prescribed as the answer to every problem. The current study demonstrates that the presence of a wildlife corridor does not necessarily provide connectivity and that they should not be viewed as an automatic remedy for all conservation issues that arise in fragmented habitats.

It is acknowledged that the study was undertaken on only one wildlife corridor system, and that the results can not be extrapolated to all scenarios. However, the concepts identified in this study have the potential to affect all corridors and future corridor assessment and planning may benefit greatly from considering these issues. This is especially so for other riparian corridors of the Atherton Tablelands, either existing or planned, which would be expected to show the same degree of heterogeneity within the rainforest habitat and in which social structuring could be expected to occur.

While wildlife corridors are designed to provide benefits for populations at a landscape scale, their potential effectiveness is largely dependent upon processes that operate at a much finer level of scale. Interactions at the population or sub- population level have the capacity to greatly influence corridor function and can involve behavioural, demographic and genetic parameters. Thus, the successful assessment of wildlife corridor effectiveness, with respect to providing connectivity, ideally requires a multi-disciplinary approach implemented over a range of spatial scales. Such an approach provides an understanding of the processes operating within the corridor system, which may ultimately affect the level of connectivity provided by the corridor.

Appendices - 122 -

Appendix 1. Species identified as present within the corridor system by the collection of fruits and nuts.

Connected R Connected M Corridor SCIENTIFICNAME COMMONNAME 12345 12345 LBRKHKSS Acronychia acidula^* Lemon Aspen + + + + + + Acronychia vestita^ Fuzzy Lemon Aspen + + + Aleurites moluccana^ Candlenut + + + + + + + + + + Beilschmedia bancroftii^* Yellow Walnut + + + + Beilschmedia oligandra Ivory Walnut + Beilschmedia recurva^* Ivory Walnut + + + Calamus moti^ Lawyer vine + Canthium comprosmoides Jilaban Tree + Cardwellia sublimis* Northern Silky oak + Castanospora alphandii^ Brown Tamarind + + + + + Cryptocarya clarksoniana Clarkson's Laurel + + + + + + Cryptocarya claudiana Claudie Laurel + Cryptocarya densiflora Cinnamon Laurel + Cryptocarya hypospodia Northern Laurel + Cryptocarya mackinnoniana^* Rusty Laurel + Cryptocarya melanocarpa -+++ Cryptocarya oblata^* Tarzali Silkwood + + Cryptocarya rhodosperma^ -+ Daphnandra repandula Scentless Sassafras + + + Desmos goezeanus -++ Dysoxylum parasiticum Yellow Mahogany + + Dysoxylum rufum Rusty Mahogany + + + + + Elaeagnus triflora Millaa Millaa Vine + + Elaeocarpus angustifolius^* Blue Quandong + + + + + + + Elaeocarpus largiflorens^ Tropical Quandong + + + Elaeocarpus ruminatus^* Brown Quandong + + + + + + + + + Endiandra hypotephra^ Blue Walnut + + + + Endiandra monothyra^* Rose Walnut + + + Endiandra sankeyana^ Sankey's Walnut + + + + Euodia bonwickii Yellow Evodia + Euphorbiaceae spp -+ Ficus crassipes^* Round leaf Banana Fig + Ficus obliqua var obliqua^* Small-leafed Fig + + + Ficus pleurocarpa^* Banana Fig + Ficus watkinsiana^ Watkins Fig + + Flindersia brayleana Queensland Maple + + + Geissois biagiana Northern Brush Mahogany + + Litsia leefeana^ Bollywood + + + + Maclura cochinchinensis Cockspur Thorn + Mallotus phillipinensis Red Kamala + Melia azedarach^ White Cedar + + + + + + + + + Myristica insipida Australian Nutmeg + + Pararchidendron pruinosum Stinkwood + + + Pilidiostigma tropicum Apricot Myrtle + Pouteria castanospermum^* Yellow Plum + Synima cordierdorum Synima + unknown + unknown + unknown + unknown ++ unknown + unknown + unknown + vine of Menispermaceae +

^ identified as rodent food source (Elmouttie unpublished data) * identified as rodent food source (Dennis and Westcott unpublished data)

Appendices - 123 -

Appendix 2. Species abundances per habitat component of the corridor system. Note that absolute numbers should not be compared across habitat types due to differences in sampling effort.

Connected R Connected M Corridor Species Total No. Species Total No. Species Total No. Aleurites moluccana 250 Endiandra hypotephra 41 Litsia leefeana 37 Elaeocarpus angustifolius 32 Elaeocarpus ruminatus 36 Pilidiostigma tropicum 31 Endiandra monothyra 17 Dysoxylum rufum 31 Endiandra monothyra 16 Castanospora alphandii 16 Daphnandra repandula 28 Acronychia acidula 12 Melia azedarach 14 Endiandra sankeyana 18 Castanospora alphandii 12 Flindersia brayleana 13 Aleurites moluccana 11 Pararchidendron pruinosum 12 Cryptocarya clarksoniana 7 Melia azedarach 10 Elaeagnus triflora 7 Beilschmedia recurva 6 Dysoxylum parasiticum 9 Elaeocarpus angustifolius 6 Elaeocarpus ruminatus 6 Desmos goezeanus 8 Unknown D 6 Dysoxylum rufum 5 Cryptocarya desiflora 7 Acronychia vestita 5 Myristica insipida 5 Cryptocarya oblata 7 Melia azedarach 4 Daphnandra repandula 4 Ficus obliqua var obliqua 6 Unknown F 4 Litsia leefeana 4 Cryptocarya clarksoniana 5 Cardwellia sublimis 3 Acronychia acidula 3 Elaeocarpus largiflorens 5 Aleurites moluccana 2 Beilschmedia bancroftii 3 Beilschmedia recurva 2 Cryptocarya melanocarpa 2 Euodia bonwickii 2 Acronychia acidula 1 Unknown A 2 Ficus obliqua var obliqua 2 Acronychia vestita 1 Unknown B 2 vine of Menispermaceae 2 Beilschmedia bancroftii 1 Canthium comprosmoides 1 Acronychia vestita 1 Beilschmedia oligandra 1 Cryptocarya clarksoniana 1 Endiandra hypotephra 1 Calamus moti 1 Cryptocarya hypospodia 1 Euphorbiaceae spp 1 Castanospora alphandii 1 Cryptocarya oblata 1 Mallotus phillipinensis 1 Cryptocarya claudiana 1 Elaeocarpus largiflorens 1 Beilschmedia oligandra 0 Cryptocarya mackinnoniana 1 Elaeocarpus ruminatus 1 Calamus moti 0 Cryptocarya melanocarpa 1 Ficus pleurocarpa 1 Canthium comprosmoides 0 Cryptocarya rhodosperma 1 Ficus watkinsiana 1 Cardwellia sublimis 0 Cryptocarya spp. 1 Geissois biagiana 1 Cryptocarya claudiana 0 Elaeocarpus angustifolius 1 Maclura cochinchinensis 1 Cryptocarya desiflora 0 Ficus crassipes 1 Pouteria castanospermum 1 Cryptocarya hypospodia 0 Ficus watkinsiana 1 Unknown C 1 Cryptocarya mackinnoniana 0 Geissois biagiana 1 Unknown E 1 Cryptocarya melanocarpa 0 Litsia leefeana 1 Beilschmedia bancroftii 0 Cryptocarya oblata 0 Synima cordierdorum 1 Beilschmedia oligandra 0 Cryptocarya rhodosperma 0 Unknown G 1 Beilschmedia recurva 0 Cryptocarya spp. 0 Canthium comprosmoides 0 Calamus moti 0 Desmos goezeanus 0 Cardwellia sublimis 0 Cryptocarya claudiana 0 Dysoxylum parasiticum 0 Cryptocarya hypospodia 0 Cryptocarya desiflora 0 Elaeagnus triflora 0 Elaeagnus triflora 0 Cryptocarya mackinnoniana 0 Elaeocarpus largiflorens 0 Endiandra monothyra 0 Cryptocarya rhodosperma 0 Endiandra sankeyana 0 Euodia bonwickii 0 Cryptocarya spp. 0 Ficus crassipes 0 Euphorbiaceae spp 0 Daphnandra repandula 0 Ficus pleurocarpa 0 Ficus pleurocarpa 0 Desmos goezeanus 0 Ficus watkinsiana 0 Flindersia brayleana 0 Dysoxylum parasiticum 0 Geissois biagiana 0 Maclura cochinchinensis 0 Dysoxylum rufum 0 Maclura cochinchinensis 0 Mallotus phillipinensis 0 Endiandra hypotephra 0 Pararchidendron pruinosum 0 Myristica insipida 0 Endiandra sankeyana 0 Pilidiostigma tropicum 0 Pararchidendron pruinosum 0 Euodia bonwickii 0 Pouteria castanospermum 0 Pilidiostigma tropicum 0 Euphorbiaceae spp 0 Synima cordierdorum 0 Pouteria castanospermum 0 Ficus crassipes 0 Unknown A 0 vine of Menispermaceae 0 Ficus obliqua var obliqua 0 Unknown B 0 Unknown A 0 Flindersia brayleana 0 Unknown C 0 Unknown B 0 Mallotus phillipinensis 0 Unknown D 0 Unknown C 0 Myristica insipida 0 Unknown E 0 Unknown D 0 Synima cordierdorum 0 Unknown F 0 Unknown E 0 vine of Menispermaceae 0 Unknown G 0 Unknown F 0 Unknown G 0

Total 395 242 176

Appendices - 124 -

Appendix 3. mtDNA procedures and protocols

DNA Extraction Total genomic DNA was extracted from biopsy tissue using a phenol-chloroform protocol. Tissue samples were initially rehydrated in 1ml GTE (100mM glycine, 10mM Tris, 1mM EDTA) for 30 mins at room temperature and then transferred to 500 µl extraction buffer (100mM NaCl, 50mM Tris, 10mM EDTA, 0.5% SDS) and 20µl 20mg/ml Proteinase-K. Tubes were placed in a rotating wheel and incubated overnight at 55oC.

The protocol consisted of two phenol:chloroform (1:1) extractions followed by a final chloroform extraction. Tubes were inverted gently 3 times and centrifuged at 8000rpm for 5 mins after each step. To the final supernatant, 100% cold ethanol (2 x volume) and 3M sodium acetate (0.1 x volume) was added. Tubes were inverted several times before being placed in a –4oC freezer for 1 hour. Samples were centrifuged for 5 mins at 10,500 rpm and the supernatant discarded from the genomic DNA pellet. 500µl of cold 70% ethanol was added to each tube, which was gently vortexed before again being centrifuged at 10,500 for 5 mins. After removal of the ethanol, pellets were left to air dry for approximately 30 mins before being suspended in 30µl or 50 µl of TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0) according to pellet size. Samples were left at room temperature overnight to facilitate dissipation and then stored at -4 oC. The concentration and purity of extracted DNA was estimated using an Eppendorf BioPhotometer and samples were diluted to 200ng/µl in TE buffer where necessary.

DNA Amplification Amplification was performed using a reaction volume of 25 µl consisting of 2.5 µl Boehringer Mannheim 10 x Taq buffer, 2 µl dNTP’s (1mM), 1µl of each primer

(10pmoles), 2 µl MgCl2 (25nM), 0.2µl Boehringer-Mannheim Taq polymerase (5 units/µl), 15.3 µl dH2O and 200ng genomic DNA. A negative control without template DNA was included in every PCR run to indicate the presence of any contamination by foreign DNA. The primers used in the amplification were: MT 15996L 5’CTCCACCATCAGCACCCAAAGC3’ and MT 16498H 5’CCTGAAGTAGGAACCAGATG3’

Appendices - 125 -

Amplification was conducted in a PTC-100 Programmable Thermal Controller (MJ Research Inc.) with the hot bonnet activated, using the following cycle program: 1) 95oC for 30s, 2) 52oC for 45s, 3) 75oC for 1 min 4) repeat steps 1-3 another 29 times, 5) 75oC for 4 mins (modified from Campbell et al. 1995).

PCR products (2 µl with 3 µl loading dye) were visualized using a 1% agarose gel stained with ethidium bromide. A DNA size marker (ø174 HAEIII) was run on every gel which was immersed in 1x TBE buffer and electrophoresed for 20 mins at 100V. DNA was visualized on a UV light transluminator. Products were overlaid with mineral oil and stored at 4 oC.

DNA Screening via TGGE DNA heteroduplexes were formed in reaction volumes of 5.6µl consisting of 1 µl sample DNA (200 ng), 1µl reference DNA (200 ng), 0.6µl 10x ME + dye buffer and 3µl 4M urea. Samples were denatured to 95oC for 5 mins and then re-annealed for 15 min at 50oC. All samples were referenced to the same individual of the appropriate outgroup species (Melomys burtoni or Uromys hadrourus). Haplotypes “B” and “AC” were not distinguishable via TGGE when used in conjunction with M. burtoni as the outgroup. Samples scored as either of these haplotypes were re-run using an individual from a geographically distant M. cervinipes population as an outgroup. The sample used was collected from the Conondales, South-East Queensland, approximately 2000 km south of the Atherton Tablelands.

Using a Diagen GmbH TGGE system, heteroduplexed DNA was electrophoresed through a 5% acrylamide gel comprising of 21.6 g urea, 16.7ml dH2O, 0.9 ml 50X ME buffer (104.65 g MOPS, 9.3 g EDTA, 18 g NaOH in a total volume of 500 ml), 2.25 ml 40% glycerol, 75µl TEMED, 5.6ml 30% acrylamide and 136 µl 10% ammonium persulphate. Gels were poured onto Pagbond films (FMC) and left to polymerise for 60 mins after which they were placed onto the apparatus heating plate. A thin layer of 0.1% Triton was evenly distributed between the plate and the gel to ensure uniform conductivity and heating. Buffer tanks were filled with 980 ml dH2O and 20 ml 50X ME buffer, and synthetic electrode wicks were placed to create buffer bridges to the gel. 5µl of each heteroduplex sample was loaded on the gel which was electrophoresed at 300V for 2 hr 40 mins over a pre-determined optimum temperature gradient of 13-38oC (Campbell et al. 1995).

Appendices - 126 -

DNA was visualized using a silver staining protocol (Qiagen TGGE handbook 1993).

The gel was covered with 135 ml dH2O and 15 ml buffer A (10% ethanol, 0.5% acetic acid) for 3 mins. This step was then repeated, the buffer discarded and the gel immersed in 135 ml dH2O and 20 ml buffer B (1% AgNO3). The gel was washed

twice in dH2O and incubated in a buffer containing 4.5 g NaOH, 300ml dH2O, 1.2 ml formaldehyde and 0.1 g NaBH4 (Buffer C) until the DNA bands were visible. This

buffer was then discarded and the gel fixed in 135 ml dH2O and 15 ml buffer D

(0.75% Na2CO3) for 10 mins. Haplotypes were assigned to individuals according to the banding pattern on the gel with similar haplotypes being re-run in adjacent lanes on subsequent gels to increase scoring accuracy.

Haplotype sequencing Samples from all haplotypes identified via TGGE were prepared for sequencing according to the procedure outlined in the QIAquick PCR Purification Kit (Qiagen) with the following modifications: i) 5µl of 3M (formula for sodium acetate) was added to each PCR product prior to the addition of PB buffer and ii) samples were left to stand at room temperature for 20 mins after the addition of EB buffer prior to centrifuging.

Sequence chromatographs were visualized in Chromas v2.131 and sequences aligned using the gap analysis program of WebANGIS.

1 CHROMAS 2.13 – Technelysium Pty Ltd, Helensvale, Queensland 4212, Australia Appendices - 127 -

Appendix 4. Nuclear DNA procedures and protocols

DNA Extraction This was performed as per “DNA Extraction” in Appendix 3 with genomic DNA stocks diluted to 100ng/µl.

DNA Amplification M. cervinipes: Amplification of microsatellites was performed using a reaction volume of 20µl consisting of: 3µl 10 x Reaction buffer (Fisher Biotech), 0.3µl dNTP’s (1mM), 1µl of

each primer (10pmoles), 1.6µl/2.4µl MgCl2 (for 2mM/3mM reaction), 0.08µl Tth Plus

Taq (Fisher Biotech) (5.5 units/µl), 100ng genomic DNA and dH2O to total volume.

Amplification was conducted in a Mastercycler (Eppendorf) with the hot bonnet activated, using the following cycle program: 1) 94oC for 1 min 15 sec, 2) 94oC for 15 sec, 3) annealing temp for 20 sec, 4) 72oC for 20 sec, 5) repeat steps 2-4 34 times, 5) 72oC for 2 mins (Paetkau pers. comm.). The following hex-labeled primers were used: Mc1K F GCA CAG CAG CCT AGG CAT Mc1K R GGC CTG TGC AGA TAT CTA GT

Mc2B F GCA GGC ATA GGT ATG ATG AC Mc2B R GAG ACA GCA TGA TCA GCA C

Mc2E F ATC AAC ATT CCC TCC GA Mc2E R ATC TTT TTC ACA GCA GGA CT

Mc2O F GTT ATC TAA GAG TTT ACA GTC GGA GGG TGG ACT Mc2O R AGT CAA GGT CAT CAG GCT CA

Mc2P F CTT TCA TAA GTT GCC TTG ATC T Mc2P R ATC TGC TGT TAC CAC TGG AG

The TOUCHDOWN PCR program was trialled as an alternative approach for those M. cervinipes samples that did not amplify using the above protocol. The TOUCHDOWN procedure uses a lower annealing temperature during its initial cycles and is therefore less stringent with products showing a greater number of stutterbands. Despite this, scorable products were produced using this technique for the Mc1K and Mc2B loci. The TOUCHDOWN PCR program involved the following steps: 1) 94oC for 1 min, 2) 92oC for 30sec, 3) 70oC for 40 sec, 4) repeat steps 2-3 19 times decreasing the temperature in step 3 by 0.5 oC each cycle, 5) 92oC for 30 sec, 6) 60oC for 40 sec, 7) repeat steps 5-6 19 times increasing the extension time by 1 sec Appendices - 128 - each cycle. Control samples were amplified using both PCR programs to ensure consistency in results across procedures.

Repeat Annealing Dilution Primer sequence temp. [MgCl2] factor* Source

O McIK (CA)14T(AC)3 58 C 2mM 1 : 2.5 D. P. (pers comm)^ O Mc2B (CA)23 66 C 2.5mM 1 : 2 D. P. (pers comm)^ O Mc2E (AC)20 56 C 2mM 1 : 2 D. P. (pers comm)^ O Mc2O (CA)25 56 C 2.5mM 1 : 2 D. P. (pers comm)^ O Mc2P (CA)4T(AC)16 56 C 2mM 1 : 1 D. P. (pers comm)^ * pcr product : loading dye ^ Dave Paetkau, Wildlife Genetics International, Nelson, British Columbia

U. caudimaculatus: Amplification of microsatellites was performed using a reaction volume of 20µl consisting of: 3µl 10 x Reaction buffer (Fisher Biotech), 0.5µl dNTP’s (1mM), 0.8µl of each primer (10pmoles), 1.2µl/1.6µl MgCl2 (for 1.5mM/2mM reaction), 0.1µl Tth Plus

Taq (Fisher Biotech) (5.5 units/µl), 100ng genomic DNA and dH2O to total volume.

Amplification was conducted in a Mastercycler (Eppendorf) with the hot bonnet activated, using the following cycle program: 1) 94oC for 5 mins, 2) 94oC for 30 sec, 3) annealing temp for 30 sec, 4) 72oC for 1 min, 5) repeat steps 2-4 34 times, 5) 72oC for 8 mins (Chand pers. comm.). The following hex-labeled primers were used:

UVC19 F CCC TAA ACC CAA CGA CAA GTG UVC19 R AAT TCT GCC TGC CTC ATG TGG

UVC232 F CTC TGA AAC TCT GTA AAT TAG C UVC232 R GGT TTT CAC TGT TGT TGT TGC

UVC238 F TTG GAC TGA TGC AGA AAG ATA CA UVC238 R TCA GAC CAG CTG ACA ACA CTT C

UVC245 F CCT AAA CCC AAC GAC AAG TGT UVC245 R GTA ACT CAA AAT TCT GCC TGC

UVC432 F ATG TTC TCC AAC CCT TCC UVC432 R GTT TTT CCT CCC ATC TCC

UVC452 F TTA ACT ATA CAT ATC GGC AGG G UVC452 R CAT GTG AGG AAG CGG AAA GAC G

Appendices - 129 -

Repeat Annealing Dilution Primer sequence temp. [MgCl2] factor* Source

O UVC19 (TG)11 + (TCTG)8 55 C 2Mm 1 : 2 V. C. (pers comm)^ O UVC232 (CA)13 50 C 1.5mM 1 : 2 V. C. (pers comm) O UVC238 (CA)19 56 C 1.5mM 1 : 1.5 V. C. (pers comm) O UVC245 (GT)12 + (CTGT)8 50 C 2mM 1 : 2 V. C. (pers comm) O UVC432 (GT)27 50 C 1.5mM 1 : 2 V. C. (pers comm) O UVC452 (CA)21 52 C 1.5mM 1 : 2 V. C. (pers comm) * pcr product : loading dye ^ Vincent Chand, Queensland University of Technology, Brisbane, Australia.

DNA Screening PCR products were diluted (see above table) with loading dye (5µl stock bromophenol blue dye in 45µl formamide) (stock dye: 1mg bromophenol blue in 10ml de-ionized formamide) and denatured at 95oC for 3 minutes before being placed on ice. Electrophoresis occurred through a 5% acrylamide gel (60µl of Ammonium Persulphate [10%] and 6µl of Temed to 10µl of filtered gel mix [84g urea, 12ml 10x TBE, 25ml 40% acrylamide in a total volume of 200ml]) at 1200V. Gels were run on a Gel-Scan 2000 (Corbett Research) in 0.6x TBE buffer. A size marker (T350, Applied BioSystems) and a reference individual were run on every gel to ensure consistent scoring across gels. Digital images from Gel-Scan 2000 were scored using One-DScan (Scanalytics Inc.). All output from One-DScan was manually cross-checked for accuracy. Appendices - 130 -

Appendix 5. mtDNA haplotype frequencies for M. cervinipes.

5

H

0

------

.

A

0

4

G

0

------

.

A

0

4 3 0

F

0 0 1

------

. . .

A

0 0 0

6 9 0

1

E

0 1 0 1

------

------

. . . .

A

0 0 0 0

3 6 2 7

D

0 0 0 0

------

------

. . . .

A

0 0 0 0

7 4 3 0 3 8 9

C

1 0 4 2 0 0 1

------

......

A

0 0 0 0 0 0 0

3 6 7 9 1 9 0 2 3 0

B

4 0 1 0 2 0 3 0 0 1

------

......

A

0 0 0 0 0 0 0 0 0 0

3 9 4

A

0 1 0

------

. . .

A

0 0 0

9

0

------

Z

.

0

5

0

------

Y

.

0

7 4 9 9 5

0 0 0 0 0

------

X

. . . . .

0 0 0 0 0

3

0

------

.

W

0

3

0

------

V

.

0

3

0

------

U

.

0

3

0

------

T

.

0

3 9

0 0

------

S

. .

0 0

e

8 5

p

1 0

------

R

. .

y

t

0 0

o

l

4 3 6 9 0

p

0 0 0 0 1

------

a

Q

. . . . .

H

0 0 0 0 0

7 3 6 5

0 0 0 0

------

P

. . . .

0 0 0 0

0

1

------

O

.

0

3 3 4 4

0 0 0 0

------

N

. . . .

0 0 0 0

3 8 7 6

0 2 0 0

------

. . . .

M

0 0 0 0

8 8 2 0 4 7 8 4

4 2 1 1 0 0 0 0

------

L

......

0 0 0 0 0 0 0 0

7 4 4

1

0 0 1 0

------

K

. . . .

0 0 0 0

5

0

------

J

.

0

6

I 0

------

.

0

8 3 9 7 8 0 4 6

0 0 1 0 0 1 0 0

------

H

......

0 0 0 0 0 0 0 0

7 0 4 1 3 9 3 8 8

0 1 0 2 0 1 2 0 0

------

G

......

0 0 0 0 0 0 0 0 0

7 4 7 4 7 6 7 6 4 7 6 3 3 7 0 6 7 8

0 0 0 0 0 1 1 0 0 2 2 1 2 1 1 0 0 0

- - - -

F

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1

0 0 6 6 9 5 6 9 7 1 3 7 9 7 8 9 6 0 3

2

2 3 2 0 0 0 0 3 3 4 3 3 0 0 3 5 0 1 0

- -

E .

......

0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 3 2 5

0 0 1 0

------

D

. . . .

0 0 0 0

1 4 3 7 6 0 4 3 9 5 3 4 8 9 9 3 2 0

2 0 1 0 0 2 2 1 0 3 0 0 0 2 3 0 2 1

- - - -

C

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

4 9 0 3 9 5

0 2 1 3 0 0

------

B

......

0 0 0 0 0 0

7 2 7 0 3 7 6 3 7 4 7 8 9 0 0 7 3 9 5

0 0 4 0 2 0 1 2 0 1 1 1 0 1 1 1 1 0 0

- - -

A

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

9 0 0 3 8 8 5 1 7 2 4 3 0 6 2 9 1 0 4 0

1 1

n

2 3 3 3 1 1 1 2 3 2 2 2 1 2 3 2 1 4 3 3 2 2

1 2 3 4 5

A A

1 2 3

1 1 2 2 3 4

R R R M M M M M

H S

R / / s s s s s s

d d d d d d d d

n

u u u u u u

L B K K S

e e e e e e e e

o

t t t t t t t t o o o o o o

i

r r r r r

B L W

t

c c c c c c c c

u u u u u u

o o o o o

a

e e e e e e e e e e e

l n n n n n n

d d d d d

t t t i i i i i i

i i i i i

n n n n n n n n t t t t t t

u

r r r r r

a a a

l l l

n n n n n n n n n n r r r r r n n n n

p

o o o

o o o o o o o o o o o o o o o o o o o

o

s s s

P C C C C C C C C C C C C C C C I I I C C C C Appendices - 131 -

Appendix 6. mtDNA haplotype frequencies for U. caudimaculatus.

Haplotype Population n A B C D E F G H I J K L M N Connected R1 32 0.84 0.16 ------Connected R3 11 0.91 - 0.09 ------Connected M1 30 0.84 0.03 0.10 ------0.03 - - - - Corridor L191.00------Corridor BR90.330.450.22------Corridor K/H61.00------Corridor K/S11.00------Corridor S31.00------Isolate B380.580.030.360.03------Isolate W 28 0.82 - - - - - 0.18 ------Continuous 1 28 0.21 0.07 - 0.25 - 0.10 0.10 0.07 0.04 - 0.04 0.04 0.08 - Continuous 2 21 0.23 0.10 - 0.14 - 0.10 0.24 - 0.05 - - - 0.14 - Continuous 4 18 0.17 0.11 - 0.11 0.06 0.32 0.17 ------0.06 Appendices - 132 -

Appendix 7. mtDNA sequences: M. cervinipes

Melomys burtoni (outgroup)

CATAAAATTACATAGTACAATAAAACATTAATGTATATCGTACATTAAATTATTTACCCCTA GCATATAAGCATGTATAATAAATCAATGAATTAAGACATTAATCAAATTTTAACTTAAAATT TATTAACACATGGATATCCAATAAAACATTTTAATTAATGCTTTATAGACATATCTGTGTT ATTAGACATACACCATTAAGTCATAAACCCTTCTCTTCCACACGACTATCCCCTGCTACA TTAATCTATAGTTCTACCATCCTCCGTGAAATCATCAACCCGCCCACTTAATGCCCCTCT TCTCGCTCCGGGCCCATAAAACTTGGGGGTAGCTAAAGTGAAACTTTATCAGA

Melomys cervinipes – haplotype A (bold text indicates variable sites)

CATAAAATTATATAGTACAATAGAACATTAATGTATATCGTACATTAAATTATTTACCCCTA GCATATAAGCATGTATAAAAAGTTAATGAATTAAGACATTAATCAAACTTCAACTTGAAA TTTACAAACACATGAATATTCATAAAAACATTTTAATTAATGTTTCTTAGACATAACTGTG TTATTAGACATACACCATTAAGTCATAAACCCTTCTCTTCCACATGACTATCCCCTGCTA CATTAATCTATAGTTCTACCATCCTCCGTGAAATCAACAACCCGCCCACTTAATGCCCCT CTTCTCGCTCCGGGCCCATGAAACTTGGGGGTAGCTAAACTGAAACTTTATCAGA

Polymorphic sites in D-Loop sequences for all unique haplotypes. “.” indicates conformity to haplotype “A”. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 5 7 8 8 8 9 9 9 0 0 1 1 1 1 2 2 2 2 2 2 2 3 3 4 4 4 4 5 5 6 7 7 8 8 9 9 0 0 1 3 5 5 6 8 8 9 9 9 0 0 0 2 2 4 5 8 9 0 1 9 4 1 4 5 2 4 7 3 7 0 1 2 4 0 1 3 4 6 7 9 3 9 4 5 7 8 1 7 7 1 5 1 6 1 2 5 6 9 0 1 4 0 1 4 1 2 7 0 4 5 7 8 1 0 A TCG TCTA G - A T A TCA CTCG A A - TTCCA TCTA -TG C G A G T A T- CCTTG T A G CTA CCG A G - B ...... A ...... T ...... C ...... A .....T . A ..A ...... C ...... T ...... C...... D ...... A ...... T ..A ...... T ...... T ...... T...A . E ...... A ...... T ..A ...... T ...... T..... F ...... A ....C ..T ..A ...... G ...TA...... T..... G ...... A .....T . T ..A ...... G ...T ...... T..... H ...... A .....T . A ..A ...... T ...... C...... I ...... A ...... T ..A ...... T .....T ...... T..... J ...... A ..C ..A . T ..A ...... T ...... T ...C ...... T..... K ...... A ...... T ..A ...... T ...... T..... L ...... AA...... T ...... T.... M ...... A ...... T ...... T ...... C...... N ...... T ..A ...... G ...TA...... T..... O ....T ..A ...... T ..A ...... C . T ...... T..... P ...... A ...... G ...... T ...... Q ...... A .....T ...... T ...... C ...... T.... R ...... A ...... T ...... T ...... S .....C . A . G ...T ...... T ...... T ...C ...... C ...... T..... T ...... A ..C ....T ..A ...C ..T ...... T ...... T..... U ...... A ...... T ...... V . TCG ...A ...... T ...... T ...... W ...... A .....T . T ...... T ...C .....C ...... T.... X .....C . A .....T ...... T ...... C ...... C A ...... T..... Y ...... A ...... T ...... T ...... G A C...... Z C....C ...... TTACTA...CCT . G . A G . T ..T . TA. G CCTT.....A T. - . . A C. T AA ....T ..A ...... T ..A .....T ...... T ...... T..... AB ...... T ...... AC ...... A ...... G .....T ...... T ...... T..... AD ...... A ...... T ..A ...... T ...... T..... AE ....T ..A ...... A ...... T ...... T...A . AF C....CT....G CTTACTA....CT . G . A G . T ..T . TA. G CCTT.....A TC- . . A ..T AG ...... A ...... T ..A ..A ...... G ...T ...... T..... AH ...... A .....T . A ..A . G ...... C ...... T ...... C...... Appendices - 133 -

Appendix 8. mtDNA sequences: U. caudimaculatus

Uromys hadrourus (outgroup)

AACTACTTCTTGAACAGTACATAAAATTATATAGTACAATAAGACATTAATGTATATTGTA CATTAAATTATTTACCCCTAGCATATAAGCACGAATAATAAATTAATGAACTAAGACATTA ATCAAATTTTGACTTAAAATTTATAAACACATGAATATCAAACCAAACATTTTAATTAATGT TTTACAGACATATCTGTGTTATTAGACATACACCATTAAGTCATAAACCTTTCTCTTCCAT ATGACTATCCCCTCCCACATCAATCTATATTTCTACCATCCTCCGTGAAATCAACAACCC GCCCACCTATGCCCCTCTTCTCGCTCCGGGCCCATAAAACTTGGGGGTAGCTAAACTG AAACTTTAT

Uromys caudimaculatus – haplotype A (bold text indicates variable sites)

AACTACTTCTTGACTAGTACATAAAATTACATAGTACAATAGAACATTAATGTATATCGTA CATTAAATTATTTACCCCTAGCATATAAGCAAGAATAATAAATTAATGAATTAAGACATTA ATTAAAATTTAACTTAAAATTTATAAACTCATGAATATTAACCTAAACATTTTAATTAATGT TTTAAGGACATATCTGTGTTATTAGACATACACCATTAAGTCATAAACTCTTCTCTTCCAT ATGACTATCCCCTTCCACATCAATCTATATTTCTACCATCCTCCGTGAAATCAACAACCC GCCCACCTATGCCCCTCTTCTCGCTCCGGGCCCATAAAACTTGGGGGTAGCTAAACTG AAACTTTAT

Polymorphic sites in D-Loop sequences for all unique haplotypes. “.” indicates conformity to haplotype “A”.

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 1 1 1 3 7 8 9 0 1 2 2 3 3 4 4 5 5 6 6 6 6 8 8 9 9 0 3 3 4 4 6 9 4 5 6 2 5 2 3 5 2 4 5 0 2 4 6 1 6 2 4 6 7 4 9 0 7 6 2 3 1 5 6 6 ACTATA GATTTTTTTTTAACTATAGTTCTTTCT B ...... T...... C ....C...... G...... A...... C D AC. . . .CC. .C. .CCC.GACG.CA.C.CCCT. E .....C...... C..T..C.A...... C F ...... C...... G ...... C.C....C.....C.A...... C H AC . . C . CC . . C . CCCC . GAC . . CA . C . CCCT . I AC..T.CC.C...CCC.GAC..TA..TCCCT. J ..-...... C..T....A...... C K AC..C.CC..CC.CCC.GACG.CA.C.CCCT. L GC . CT . C . . . C . . CCC . GAC . . TA . CTCCCT . M AC..T.CC..C..CCC.GAC..TA..TCCCT. N ...... C...... C...... Appendices - 134 -

Appendix 9. The relative frequency of microsatellite alleles per population per locus for M. cervinipes. Numbers beneath site names represent sample size.

Connected Corridor Isolate Continuous R1 R2 R3 M1 M2 M3 M4 M5 L BR KH KS S B L W 1 1A 2 2A 3 4 Mc1K nAB 28182626253022261123322612493130282431181122 ------0.010------C 0.0710.0830.038-0.0600.1000.114----0.0380.125--0.0670.0540.0420.0650.028-0.027 -----0.017------0.0830.071----0.032-0.0450.027 DEFG 0.411 -0.0830.0580.038-0.0830.045-0.136-0.031-0.083----0.0420.016-0.091- H 0.333 0.089I 0.269 0.028J 0.731 0.500K - - 0.467L 0.409 - 0.018M 0.019 0.500 0.071----0.017-0.019------0.0160.0670.0710.0830.0320.1390.0450.068 0.083 0.232 0.200N 0.500 - 0.056 0.115 0.111 0.089 0.117 0.391O 0.038 0.019 0.308 0.139 0.182 0.018 0.031 0.115 0.115 0.096 0.083 0.404 ----0.020------0.1290.3330.0890.0420.0650.028-0.027 -Mc2B 0.020 - 0.019 - 0.058 0.458 0.100 0.080n - 0.194 - 0.159 0.038 0.333A 0.113 - 0.135 - 0.040 - 0.023 0.367B 26172424253022261123322611463130282432171122 0.091 0.333 0.077 0.393 ------0.077--0.031------0.020C 0.283 - 0.045 0.354 0.017 0.017 ------0.018----- 0.203D 0.130 - 0.274 0.045 - 0.328 0.115E 0.109 0.417 0.077 0.173 - ---0.021-0.167-0.019---0.019------0.016--- 0.096 0.227 0.136F 0.019 - 0.038 0.042 - 0.269 0.318 0.065G 0.091 0.214 0.019 0.147 0.016 - ----0.020------0.0180.0420.0160.088-0.023 - 0.130 0.019 0.113H 0.188 0.133 0.019 0.077 0.141 0.060 ------0.019------0.018----- 0.300 0.019-0.0210.0630.0800.1670.091----0.019--0.2260.150-0.0210.016--0.045 I 0.167 - 0.113 0.059 0.092 0.077 0.125 0.080 0.117J - 0.021 0.032 0.083 0.167 - - 0.167 0.054K 0.125 0.067 --0.063---0.045----0.019----0.0360.0210.1250.029-0.023 0.235 0.065 0.136 0.042 0.120 - 0.054 0.161L 0.250 0.010 0.058 0.023 0.097 0.063 0.200 0.063 0.091M 0.227 0.019 0.028 0.019 0.159 0.129 0.045 0.078 - 0.019 - 0.239N 0.182 0.042 0.135 0.056 0.058 0.188 0.159 -O 0.091 0.091 0.059 - - 0.077 0.054 0.077 - 0.043P 0.083 0.154 0.088 - 0.218 - Q 0.125 0.206 - 0.063 - 0.021 0.017 0.327 -R 0.146 0.250 0.104 - 0.054 0.176 0.227 0.154 - - 0.021 0.220 0.020S 0.185 0.146 0.042 0.147 0.060 ----0.020--0.019-----0.022------0.0450.023 0.267 0.167 0.032T 0.021 0.065 0.010 - 0.146 - - 0.100 - 0.023 0.068 0.100 0.017 ---0.021--0.0230.019--0.031------0.140 0.323 0.042U 0.114 - 0.173 0.058 0.045 - 0.179 0.133 0.133 0.080 - ---0.021------0.028V 0.038 0.273 0.045 0.058 0.019 0.029 0.104 0.091 0.018 0.033 ------0.045- -W 0.239 0.109 0.083 0.136 0.042 0.036 0.125 0.192 0.088 0.104 0.091 ------0.021---- 0.047 0.016 0.031 0.174 -X 0.021 0.118 0.182 0.063 0.163 0.038 ------0.0360.021---0.023 0.019 0.100 0.045 0.130Y - 0.045 - 0.083 - 0.068 - 0.328 - 0.022 ------0.018----0.023 - - 0.068 - - 0.091 0.077 0.109 ------0.018------0.115 0.273 0.096 0.182 0.019 0.036 0.060 0.091 - 0.037 0.182 0.043 0.022 0.021 - 0.167 0.068 0.031 0.177 0.174 - 0.177 0.145 0.031 - 0.023 0.115 0.028 0.283 0.016 0.100 0.050 0.029 0.019 0.091 0.091 0.107 0.133 0.089 0.086 0.182 0.043 0.043 0.087 - 0.027 0.250 0.107 0.083 0.021 0.023 0.161 0.065 - 0.156 0.208 0.094 0.047 - 0.100 0.033 0.021 0.294 0.156 0.029 0.107 0.089 0.091 0.176 0.045 - 0.104 0.021 0.019 - 0.114 0.045 - 0.045 0.125 0.078 0.182 0.136 0.176 - 0.068 0.273 - - - 0.182 0.091 - 0.077 0.045 0.068 0.045 0.022 - - - 0.100 0.036 0.033 0.063 0.016 - - - - - 0.023 0.059 - 0.068 Locus28182626253022261123322612493130282432181122 Appendices - 135 -

Connected Corridor Isolate Continuous R1 R2 R3 M1 M2 M3 M4 M5 L BR KH KS S B L W 1 1A 2 2A 3 4 67556023 29504902468 19 Mc2E nA 22292510342930222426187 B 25152525182422211121202511392929272027131121 ------0.024 CDE 0.020F 0.060G - 0.020H - 0.100 0.080I 0.020 0.020 0.033 0.160J 0.020 0.060 0.220 0.067 0.180 0.040 -K 0.120 0.100 0.120 0.111 - 0.222L 0.120 - 0.100 0.063 0.188 0.280 0.080M 0.100 0.250 0.114 - 0.133 0.060 0.056 -----0.063---0.0240.1500.040------0.360 0.100 0.190 0.095 - 0.320 0.367 0.083 0.020 ------0.017------0.306 0.056Mc2O 0.364 0.140 0.180 0.114 0.100 0.021 0.250 0.104 0.238n - 0.056 0.060 0.024 0.227 0.250 0.091 0.150A 0.194 0.167 0.045 - 0.381 0.048 - 0.240 0.024 0.042 0.068 0.071B 0.318 - 26172525263022237 0.045 0.050 0.119 0.091 0.075 ------0.020----0.023----- 0.238C 0.040 - 0.192 0.048 0.182 0.260 0.450D 0.345 - 0.262 0.091 0.273 - 0.140 - 0.086 0.048 -E 0.071 0.075 0.141 0.318 0.115 0.204 0.050 ------0.071------0.017------0.200 0.190F 0.231 0.150 0.020 0.019 0.021 0.024 0.121 - -G 0.276 - 0.222 0.182 0.059 - 0.056 - 0.448H 0.128 - 0.125 - 0.204I - - 0.220 0.093 0.300 0.205 - - - 0.058 - -J 0.077 0.095 0.222 0.020 0.034 0.182 0.227 0.080 -K 0.231 0.155 - - 0.173 0.071 0.333 0.069 0.167 0.182 0.204 -L - - 0.118 - - 0.093 0.017 - 0.190 0.200 0.028M 0.100 0.023 0.148 0.019 -N - 0.111 - 0.103 0.022 - 0.154 0.017 - 0.088 0.269 - -O 0.231 0.069 0.413 - - 0.136 0.160 0.176 0.088 0.077 0.091 0.034 0.182P 0.071 0.020 0.100 0.120 - 0.023 0.058 0.080 0.088 0.119 0.086 0.130 0.052 - - 0.136Q 0.250 0.060 0.022 0.118 0.060 0.100 0.050 - 0.020 0.019 - 0.183 -R 0.071 0.060 0.111 0.200 0.056 - 0.096 0.205 - 0.023 - 0.120S - 0.115 - 0.020 0.206 - 0.130 0.077 0.121 - - 0.139 ------0.034------0.091 -T 0.045 0.060 0.037 0.143 0.029 0.060 0.019 0.117 - - 0.050 0.119 -0.029------0.023------0.019 0.240U 0.029 0.091 0.060 0.060 0.182 - 0.017 0.100 0.167 --0.040------0.083-0.021--0.071- - 0.069 0.040V - 0.080 0.109 - - - 0.033 - 0.023 0.317 - - 0.040 ------0.017------0.028 - 0.111W 0.143 0.068 - 0.017 - 0.114 0.067 0.050 ------0.017------0.074 0.046X 0.397 0.022 0.020 0.087 0.217 - 0.136 0.045 - 0.114 ---0.020------0.077 0.121 - 0.333 0.259Y 0.071 - - 0.239 0.087 0.045 0.040 ------0.065---0.0200.050------0.023 - - 0.071 - - 0.071 0.214 - 0.045 0.100 0.088 - 0.034 - 0.034 ------0.017------0.046 0.050 0.159 0.136 0.059 0.103 0.074 0.060 0.037 0.121 0.138 0.114 0.083 - 0.025 - 0.021 0.100 0.023 0.038 0.140 0.180 - 0.155 - - 0.182 - - - 0.019 0.154 0.015 0.050 0.200 0.260 0.063 0.077 0.034 - 0.103 0.100 - 0.038 0.046 0.067 - - - 0.015 - 0.083 - - - 0.052 0.136 - - - - 0.286 0.024 0.020 0.104 0.023 0.052 0.053 - - 0.103 0.173 0.100 0.023 0.017 0.068 - 0.040 0.083 0.100 0.056 0.015 - - 0.040 - 0.063 0.200 - 0.045 0.067 0.286 0.056 0.190 - 0.058 0.088 0.132 0.159 0.105 - 0.167 0.191 - 0.111 0.138 0.104 - 0.023 - 0.052 0.050 - 0.096 - 0.208 0.083 0.117 0.114 - - 0.139 0.096 0.115 0.091 0.026 0.188 - 0.143 0.104 - 0.25 0.019 0.105 0.053 0.091 - 0.135 - - 0.083 0.071 0.034 0.017 0.111 0.056 0.079 0.060 0.050 0.017 0.143 - 0.100 0.017 - 0.079 - - 0.045 - 0.105 - 0.021 0.079 0.105 0.021 - - 0.019 - 0.052 0.028 - 0.050 0.068 - - - - 0.026 0.023 0.026 0.038 - - 0.019 0.026 - - 0.026 Locus28182626253022261123322612493130282432181122 Appendices - 136 - 014 0.014 0.017 0.083 -042 0.047 0.042 0.100 0.050 0.091 0.042 0.050 0.023 0.141 0.067 0.091 - 083 0.056 0.056 0.050 0.042 0.045 0.078 0.033 - 0.025 .109 0.042 0.028 0.028 0.050 0.042 0.023 0.063 0.167 - 0.050 0.174 0.2500.239 0.028 0.125 0.028 0.250 0.250 0.250 0.104 0.133 0.205 0.229 0.250 0.091 0.100 0.031 0.091 0.133 0.300 0.045 0.275 4871 - 0.043 0.167 - 0.028 0.028 0.069 0.083 0.069 0.104 0.083 0.068 0.021 0.063 0.045 0.067 0.016 0.182 0.033 0.020 - 0.075 024 - 0.083 0.111 0.111 - 0.021 0.159 - 0.067 0.045 - .012 0.283 - -.071 0.065 0.042 0.333 - 0.333 0.133 - 0.146 0.205 0.063 0.234 0.045 0.033 0.182 0.100 - 0.067 0.091 - Connected Corridor Isolate Continuous R1 R2 R3 M1 M2 M3 M4 M5 L BR KH KS S B L W 1 1A 2 2A 3 4 Mc2P nAB ------0.063------C ------0.024------29D ------0.024------E ------0.0210.045--0.0450.025 15F ------0.025 G ------0.042----- 26H ---0.154------0.042------0.025 I 26JK 19L - 0.019M 30N - 0.074 -O 20 0.019 -P 0.067 - 0.019Q - - - 25 0.037 0.019R 0.259 0.115 0.038 - 0.038 - 0.200S 0.054 0.026 0.333 9 - - 0.135 0.100 0.033T 0.233 0.074 0.054 - 0.115 0.019 0.025 0.327 0.267 0.167U 0.058 0.158 0.096 19 0.083 - 0.173 0.115 0.125 0.058V 0.100 0.158 0.100 - - 0.111 0.211 0.058 0.063 -0.1330.077--0.0330.025--0.026-----0.017-0.023-0.033-- 0.026 0.050 0.033 0.025 - 0.133 0.125 0.158 21 0.074 0.067 0.167 0.100 0.063 --0.038------0.067 0.026 0.275 - 0.056 0.067 0.133 - 0.222 0.146 0.038 0.104 0.050 0.135 0.210 - 0.167 23 - 0.096 - 0.222 0.063 - 0.019 - 0.019 0.079 0.381 0.105 0.079 0.053 0.056 0.038 0.024 0 0.042 0.033 0.095 0.100 12 0.184 0.054 0.024 - 0.095 0 0.056 0.025 - 0.021 0.0 0.067 0.065 0.042 0.200 - 0.067 36 0.056 - 0.021 - - 0.023 0.056 - - 0. 0.0 24 0.184 - 0 0.022 - - 0.167 30 - 0. - - 26 - 0.111 0.132 0. 22 - - - 0. 32 - - 15 0.042 0.042 13 - 0.133 0.042 0.016 20 0.023 0.100 0.063 0.091 - - 0.045 - Locus 28 18 26 26 25 30 22 26 11 23 32 26 12 49 31 30 28 24 32 18 11 22 Appendices - 137 -

Appendix 10. The relative frequency of microsatellite alleles per population per locus for U. caudimaculatus. Numbers beneath site names represent total sample size. N.B. UVC19 was significantly linked to UVC245 and was not included in any analyses.

Connected Corridor Isolate Continuous R1 R3 M1 L BR KH KS S B W 1 2 4 Locus 32 11 29 19 9 6 1 3 29 30 30 22 18 UVC19 n3211 28 19 9 6 1 3 28 30 29 22 18 A0.0460.045 0.089 0.184 - 0.083 - 0.167 0.138 0.150 0.017 0.068 0.056 B-- - - - 0.083 - - - 0.017 - - 0.028 C0.4060.455 0.393 0.474 0.389 0.167 0.500 0.333 0.241 0.150 0.138 0.250 0.389 D-- 0.018 ------0.083 0.017 - - E-- 0.018 - - - - - 0.034 0.083 0.052 0.023 - F0.2810.045 0.357 0.316 0.044 0.333 0.500 - 0.276 0.333 0.052 0.023 0.056 G-0.091 - - - - - 0.167 - 0.017 0.017 0.091 0.028 H0.2660.364 0.125 0.026 0.167 0.250 - 0.333 0.276 0.100 0.172 0.159 0.167 I-- - - - 0.083 - - - 0.017 0.052 0.023 - J------0.069 0.045 0.167 K------0.190 0.136 - L------0.034 - 0.121 0.091 0.083 M------0.052 0.068 0.028 N------0.017 0.023 - O------0.050 0.034 - -

UVC232 n 31 11 27 19 9 6 1 2 28 28 29 19 17 A 0.065 0.045 0.037 0.026 0.056 - - - 0.036 - - 0.026 - B 0.065 ------0.036 - - 0.053 0.029 C 0.065 - - - - - 0.500 - 0.036 - 0.034 0.079 0.176 D 0.032 0.045 ------0.036 - 0.017 0.026 0.088 E 0.113 0.136 0.167 0.105 0.333 - - - 0.125 0.267 0.190 0.158 0.147 F 0.194 0.136 0.315 0.132 0.111 - 0.500 0.250 0.036 0.321 0.086 0.079 0.088 G 0.339 0.318 0.352 0.684 0.222 0.750 - 0.250 0.607 0.286 0.034 0.079 0.206 H 0.113 0.182 0.074 0.052 0.278 0.250 - - 0.036 0.125 0.052 0.079 0.059 I 0.016------J - - 0.556 - - - - 0.500 - - 0.052 0.053 0.029 K ------0.018-0.017-0.059 L ------0.034-- M ------0.0520.026- N ------0.1900.105- O ------0.069 0.053 0.029 P ------0.155 0.105 0.088 Q ------0.0170.053- R ------0.026- S ------0.036----

UVC238 n 32 11 29 19 9 6 1 3 28 30 27 21 17 A 0.125 0.136 0.017 0.026 - - - - 0.328 - - 0.024 0.028 B 0.172 0.091 0.017 0.026 - - - - 0.034 0.183 0.019 0.024 0.028 C ------0.019-0.056 D 0.016 ------0.034 0.033 0.074 0.190 0.222 E 0.219 0.273 0.138 0.158 0.111 0.083 0.500 0.667 0.224 0.333 0.111 0.119 0.361 F 0.359 0.050 0.672 0.368 0.778 0.833 0.500 0.167 0.207 0.333 0.056 0.071 0.083 G 0.031 - 0.017 0.026 - - - 0.167 - - 0.037 0.071 0.056 H 0.078 - 0.138 0.395 0.111 0.083 - - 0.138 0.067 0.037 0.095 0.028 I ------0.0500.0370.024- J ------0.130-0.083 K ------0.0740.214- L ------0.0190.024- M ------0.1480.048- N ------0.017 - 0.056 0.024 0.028 O ------0.0930.048- P ------0.017 - 0.074 0.024 0.028 Q ------0.019-- Appendices - 138 -

Connected Corridor Isolate Continuous R1 R3 M1LBRKHKSSBW124 Locus 32 11 29 19 9 6 1 3 29 30 30 22 18 UVC245 n 311129189 6 1 3 2827302218 A ---0.139-0.083------B 0.048 0.450 0.103 0.111 - 0.083 - 0.170 0.161 0.185 0.017 0.045 0.056 C ---0.083------0.028 D 0.419 0.455 0.379 0.250 0.667 0.167 0.500 0.330 0.196 0.167 0.133 0.205 0.389 E --0.034------0.0190.0500.023- F 0.274 0.045 0.379 0.333 0.333 0.333 0.500 - 0.303 0.456 0.100 0.023 0.056 G ------0.056-0.023- H 0.242 0.455 0.103 0.083 - 0.333 - 0.500 0.286 0.148 0.150 0.318 0.167 I 0.016------0.018-0.083-0.028 J ------0.0330.0230.167 K ------0.2000.1820.083 L ------0.036-0.1170.0450.028 M ------0.0500.091- N ------0.0330.023- O ------0.033--

UVC432 n 321129199 6 1 3 2828302118 A ------0.083 B 0.094 0.045 0.086 - - - - - 0.054 0.036 0.033 0.071 0.028 C 0.109 0.182 0.310 0.026 0.167 - - 0.167 0.250 0.267 0.050 - - D 0.016-0.0520.026------E ------0.017-- F --0.1720.026-----0.0180.033-- G 0.078 - 0.155 0.158 0.222 - - - 0.161 0.125 0.050 0.095 0.028 H 0.047 0.045 0.086 0.105 0.111 0.417 0.500 - 0.054 0.089 - 0.048 0.083 I 0.0310.091-0.184-----0.0710.0670.1190.028 J 0.109 0.318 0.172 0.105 0.222 - 0.500 0.167 0.018 0.036 0.050 0.167 0.194 K 0.375 - 0.086 0.132 0.111 0.250 - 0.500 0.268 0.250 0.150 0.119 0.111 L 0.094 0.045 - 0.026 - 0.083 - - 0.036 0.071 0.167 0.119 0.194 M 0.081 0.273 - 0.211 0.167 0.250 - 0.167 0.143 0.018 0.233 0.095 0.139 N 0.016------0.018-0.0830.0480.028 O ------0.0180.0170.0240.083 P --0.172------0.0170.071- Q ------0.0330.024-

UVC452 n 321128189 6 1 3 2828292217 A --0.017------B --0.017-0.111-----0.017-0.029 C 0.016------0.023- D 0.109 - 0.034 0.053 0.222 - - - 0.017 0.086 0.069 0.136 0.147 E 0.016----0.417--0.0170.0340.0340.023- F 0.156 0.136 0.103 0.053 - - 0.500 - 0.293 - 0.034 0.045 0.088 G 0.266 0.136 0.551 0.500 0.333 0.250 0.500 0.333 0.121 0.241 0.138 0.114 0.412 H 0.281 0.227 0.121 0.026 - - - - 0.086 0.017 0.155 0.205 0.176 I 0.109 0.455 0.069 0.211 0.222 - - 0.167 0.431 0.586 0.034 0.091 0.029 J 0.031 0.045 0.086 0.105 0.111 0.333 - 0.500 - 0.034 - - - K 0.016------0.052-- L ------0.017-- M - - - 0.026 - - - - 0.017 - 0.293 0.114 - N ------0.0340.0910.088 O ------0.017-0.0860.0910.029 P ---0.026------0.0340.068- Appendices - 139 -

Appendix 11. Hardy-Weinberg estimates for M. cervinipes. A = number of microsatellite alleles, Ho = observed heterozygosity, He = expected heterozygosity. * represents significant departures from Hardy-Weinberg equilibrium after sequential Bonferroni correction.

Locus Population n Mc1K Mc2B Mc2E Mc2O Mc2P Mean ( ± s.e.) Connected R1 28 A 8 9 10 12 9 9.60 ± 0.68 p 0.656 0.591 0.869 0.052 0.732 - Ho 0.714 0.769 0.920 0.730 0.741 0.77 ± 0.04 He 0.764 0.856 0.851 0.881 0.806 0.83 ± 0.02 Connected R2 18 A 9 9 8 10 7 8.60 ± 0.51 p 0.692 0.819 0.335 0.062 0.952 - Ho 0.944 0.882 1.000 0.647 0.800 0.85 ± 0.06 He 0.849 0.886 0.830 0.911 0.835 0.86 ± 0.02 Connected R3 26 A 8 12 8 12 12 10.40 ± 0.98 p 0.808 0.079 0.012 0.226 0.425 - Ho 0.885 0.750 0.960 0.840 0.846 0.86 ± 0.03 He 0.812 0.907 0.817 0.900 0.846 0.86 ± 0.02 Connected M1 26 A 8 13 9 10 12 10.40 ± 0.93 p 0.027 0.034 0.251 <0.001* 0.924 - Ho 0.384 0.750 0.800 0.480 0.960 0.67 ± 0.11 He 0.485 0.892 0.816 0.873 0.907 0.79 ± 0.08 Connected M2 25 A 9 12 7 8 11 9.40 ± 0.93 p 0.856 0.564 0.107 0.316 0.584 - Ho 0.720 0.920 0.833 0.778 1.000 0.85 ± 0.05 He 0.708 0.901 0.821 0.861 0.887 0.84 ± 0.03 Connected M3 30 A 11 12 10 10 13 11.2 ± 0.58 p 0.027 0.481 0.446 0.002* 0.903 - Ho 0.670 0.833 0.791 0.600 0.933 0.77 ± 0.06 He 0.751 0.867 0.880 0.853 0.881 0.85 ± 0.02 Connected M4 22 A 8 13 8 10 11 10.0 ± 0.95 p 0.002* 0.349 0.008* 0.062 0.610 - Ho 0.636 0.955 1.000 0.818 1.000 0.88 ± 0.07 He 0.792 0.921 0.847 0.898 0.863 0.87 ± 0.02 Connected M5 26 A 10 15 8 10 13 11.20 ± 1.24 p 0.019 0.007* 0.898 0.670 0.178 - Ho 0.653 0.731 0.810 0.870 0.875 0.79 ± 0.04 He 0.721 0.910 0.800 0.874 0.917 0.84 ± 0.04 Corridor L 11 A 6 7 5 9 9 7.20 ± 0.80 p 0.276 0.501 1.000 1.000 0.514 - Ho 0.727 0.818 0.818 1.000 0.889 0.85 ± 0.05 He 0.727 0.848 0.784 0.934 0.895 0.84 ± 0.04 Corridor BR 23 A 5 8 9 13 9 8.80 ± 1.28 p 0.006* 0.783 0.563 0.907 0.830 - Ho 0.869 0.913 0.905 0.863 0.947 0.90 ± 0.02 He 0.745 0.841 0.823 0.906 0.872 0.84 ± 0.03 Corridor K/H 32 A 9 9 7 14 12 10.20 ± 1.24 p 0.522 0.046 0.556 0.001* 0.002* - Ho 0.719 0.781 0.750 0.710 0.714 0.74 ± 0.01 He 0.820 0.806 0.755 0.918 0.827 0.83 ± 0.03 Appendices - 140 -

Locus Population n Mc1K Mc2B Mc2E Mc2O Mc2P Mean ( ± s.e.) Corridor K/S 26 A 9 12 8 14 8 10.20 ± 1.20 p 0.026 0.404 0.671 0.362 0.260 - Ho 0.692 0.923 0.880 1.000 0.739 0.85 ± 0.06 He 0.805 0.854 0.884 0.879 0.834 0.85 ± 0.02 Corridor S 12 A 8 6 6 9 9 7.60 ± 0.68 p 0.796 0.711 0.360 0.853 0.873 - Ho 0.750 0.909 0.636 0.900 1.000 0.84 ± 0.07 He 0.783 0.836 0.840 0.926 0.884 0.85 ± 0.02 Isolate B 49 A 10 9 6 10 11 9.2 ± 0.86 p 0.053 0.570 0.872 <0.001* 0.510 - Ho 0.776 0.935 0.923 0.471 0.722 0.77 ± 0.08 He 0.843 0.792 0.832 0.794 0.811 0.81 ± 0.01 Isolate L 31 A 8 8 6 11 11 8.80 ± 0.97 p 0.100* 0.026 0.024 <0.001* 0.418 - Ho 0.968 0.903 0.966 0.482 0.875 0.84 ± 0.09 He 0.827 0.847 0.769 0.877 0.843 0.83 ± 0.02 Isolate W 30 A 8 10 10 15 11 10.8 ± 1.60 p 0.183 0.472 0.272 0.040 0.686 - Ho 0.867 0.933 0.690 0.833 0.933 0.85 ± 0.05 He 0.755 0.859 0.760 0.927 0.877 0.84 ± 0.03 Continuous 1 28 A 11 16 8 13 14 12.40 ± 1.36 p 0.156 0.987 0.544 <0.001* 0.070 - Ho 0.714 0.964 0.778 0.545 0.791 0.76 ± 0.07 He 0.892 0.920 0.863 0.919 0.910 0.90 ± 0.01 Continuous 1A 24 A 11 14 8 12 13 11.60 ± 1.03 p 0.800 0.152 0.641 0.150 0.262 - Ho 0.833 1.000 0.800 0.917 0.954 0.90 ± 0.04 He 0.854 0.875 0.837 0.891 0.889 0.87 ± 0.01 Continuous 2 32 A 11 13 10 12 11 11.40 ± 0.51 p 0.226 0.620 0.125 0.165 0.291 - Ho 0.839 1.000 0.889 0.731 0.813 0.85 ± 0.04 He 0.867 0.899 0.863 0.912 0.856 0.88 ± 0.01 Continuous 2A 18 A 9 9 8 11 13 10.00 ± 0.89 p 0.025 0.722 0.973 0.022 0.889 - Ho 1.000 0.823 1.000 0.889 0.933 0.93 ± 0.03 He 0.759 0.864 0.871 0.894 0.948 0.87 ± 0.03 Continuous 3 11 A 10 10 8 6 11 9.00 ± 0.89 p 1.000 0.312 0.836 0.963 0.829 - Ho 1.000 0.727 0.909 0.857 0.909 0.88 ± 0.05 He 0.909 0.931 0.892 0.846 0.939 0.90 ± 0.02 Continuous 4 22 A 10 16 9 14 11 12.00 ± 1.30 p 0.183 0.905 0.305 0.544 0.536 - Ho 0.863 0.954 1.000 0.895 0.850 0.91 ± 0.03 He 0.872 0.915 0.830 0.842 0.856 0.86 ± 0.02 mean (± s.e.) # alleles 8.96 ± 0.33 11.05 ± 0.61 8.09 ± 0.34 11.00 ± 0.48 11.05 ± 0.39

Appendices - 141 -

Appendix 12. Hardy-Weinberg estimates for U. caudimaculatus. A = number of microsatellite alleles, Ho = observed heterozygosity, He = expected heterozygosity. * represents significant departures from Hardy-Weinberg equilibrium after sequential Bonferroni correction.

Locus Population n UVC232 UVC238 UVC245 UVC432 UVC452 Mean ( ± s.e.) Connected R1 32 A 4 7 5 11 9 7.20 ± 1.28 p 0.006* 0.531 0.958 0.392 0.106 - Ho 0.68 0.88 0.71 0.75 0.94 0.79 ± 0.05 He 0.78 0.78 0.70 0.82 0.81 0.78 ± 0.02 Connected R3 11 A 5 6 4 7 5 5.40 ± 0.51 p 0.041 0.13 0.38 0.416 0.845 - Ho 0.72 0.73 0.46 0.73 0.91 0.71 ± 0.07 He 0.87 0.69 0.65 0.83 0.74 0.76 ± 0.04 Connected M1 29 A 6 6 5 9 8 6.80 ± 0.73 p 0.092 0.191 0.589 <.001* 0.956 - Ho 0.78 0.48 0.66 0.83 0.69 0.69 ± 0.06 He 0.75 0.52 0.7 0.84 0.67 0.70 ± 0.05 Corridor L 19 A 5 6 6 10 8 7.0 ± 0.89 p 0.489 0.814 0.052 0.39 0.078 - Ho 0.47 0.63 0.83 0.9 0.74 0.71 ± 0.08 He 0.55 0.7 0.8 0.88 0.71 0.73 ± 0.06 Corridor BR 9 A 5 3 2 6 5 4.20 ± 0.73 p 0.089 1.000 0.181 0.958 0.221 - Ho 1.00 0.44 0.56 1.00 0.89 0.78 ± 0.12 He 0.79 0.39 0.71 0.87 0.84 0.72 ± 0.09 Corridor K/H 7 A 2 3 5 4 3 3.40 ± 0.51 p 1.000 1.000 1.000 1.000 1.000 - Ho 0.50 0.33 0.83 0.83 0.83 0.66 ± 0.11 He 0.41 0.32 0.80 0.76 0.71 0.60 ± 0.10 Corridor K/S 1 A 2 2 2 2 2 2.0 ± 0.00 p------Ho------He------Corridor S 3 A 3 3 3 4 3 3.20 ± 0.20 p 1.000 1.000 1.000 0.611 1.000 - Ho 1.00 0.67 0.67 0.67 1.00 0.80 ± 0.08 He 0.83 0.60 0.73 0.93 0.73 0.76 ± 0.06 Isolate B 29 A 10 8 6 9 8 8.20 ± 0.66 p 0.159 0.206 0.021 0.050 0.363 - Ho 0.57 0.83 0.93 0.86 0.76 0.79 ± 0.06 He 0.62 0.79 0.78 0.83 0.72 0.75 ± 0.04 Isolate W 30 A 4 6 6 11 6 6.60 ± 1.17 p 0.435 0.168 0.083 0.792 0.569 - Ho 0.93 0.90 0.82 0.89 0.55 0.82 ± 0.07 He 0.74 0.76 0.75 0.84 0.6 0.74 ± 0.04 Continuous 1 30 A 14 16 12 14 13 13.80 ± 0.66 p <.001* 0.056 <.001* <.001* <.001* - Ho 0.93 0.89 0.87 0.97 0.86 0.90 ± 0.02 He 0.90 0.93 0.90 0.89 0.87 0.90 ± 0.01 Continuous 2 22 A 15 14 11 12 11 12.60 ± 0.80 p 0.063 0.041 <.001* 0.427 <.001* - Ho 0.79 0.76 0.50 0.91 0.82 0.76 ± 0.07 He 0.95 0.91 0.84 0.92 0.90 0.90 ± 0.02 Continuous 4 18 A 11 11 9 11 8 10.00 ± 0.63 p 0.149 0.286 0.211 0.174 0.242 - Ho 0.77 0.83 0.78 0.78 0.82 0.80 ± 0.01 He 0.90 0.84 0.80 0.90 0.78 0.84 ± 0.03 mean (± s.e.) # alleles 6.61 ± 1.22 7.00 ± 1.20 5.85 ± 0.87 8.46 ± 1.00 6.85 ± 0.90 Appendices - 142 -

Appendix 13. Pairwise comparisons for all populations of M. cervinipes – mtDNA.

FST values above diagonal, exact test p-values below diagonal. Values in bold represent those pairwise comparisons referred to in Chapter 5. * = p < 0.001.

2 1 7 9 6 4 7 7 9 5 7 3 5 5 7 8 9 4 8 4 3 5 0

3 0 5 5 7 9 3 6 3 6 5 8 5 0 3 2 0 7 0 4 0 3 0

4

1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 2 0 9 9 7 0 8 2 8 3 1 0 6 1 1 5 6 0 3 4 0 5

6 5 5 7 8 8 6 8 6 7 6 2 9 0 4 9 6 2 2 7 1 0 6

3

1 1 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 0 0 0 7

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 7 6 1 3 9 6 8 7 9 2 0 6 9 7 0 1 9 9 7 4

5 9

s

8 7 4 9 2 6 4 3 2 3 2 2 0 9 0 0 3 8 5 7 8

1 1

u

A

1 1 0 0 0 0 0 0 0 0 0 1 1 0 1 1 0 1 0 0 0 2 0

o

2

......

u

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

n

i

t

8 6 3 9 9 5 6 9 4 9 5 7 6 7 4 7 0 1 1 5 6 1

8

n

4 2 4 6 9 8 6 9 8 8 4 8 7 7 2 4 8 0 5 0 3 6

1

o

2

1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 9 0 0 0 1 2

......

C

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

7 3 9 8 5 0 9 6 2 6 1 8 4 0 5 2 0 4 7 6 6

0

6 9 9 0 2 6 0 2 2 2 9 2 6 1 2 3 0 3 0 0 1

1

A

*

1 1 0 1 1 1 1 1 1 1 1 0 1 1 2 1 1 0 0 0 0 0

1

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 4 1 1 8 3 3 8 4 7 0 2 4 4 1 0 9 4 0 5 7 1 1

4 2 3 6 7 8 6 8 6 7 5 9 6 0 5 5 7 8 6 7 3 5 5

1

1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 4 0 0 1

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

8 3 7 9 5 8 4 6 9 5 2 7 4 0 1

7 5 8

7 3 2 2 0 7 0 9 6 6 7 9 3 6 3

1 1 1

* * * * *

1 2 1 1 1 1 1 0 1 1 1 0 1 1 1 1 1 4

W

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

e

8 4 7 3 6 9 7 6 9 5 4 3 8 4 2 4 3 3

t

2 2 8 3 3 0 8 6 5 6 3 5 4 3 4 5 4 0

a

l

* * * * *

L

2 2 0 1 0 1 0 0 0 0 0 1 1 1 1 1 2 0

......

o

s

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

I

1 7 1 9 5 5 5 4 2 4 8 1 8 4 3

0 3 6 4 7 1 4 6 7 6 6 3 8 9 3

* * * * * * * *

B

2 2 1 1 1 2 1 1 1 1 1 1 1 1 2

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 5 9 6 6 1 0 5 9 9 0 0 1 3 4 3 7 7 7 3

4 3 9 3 2 3 2 0 4 1 0 5 6 0 0 0 3 2 0 2

* * *

S

2 2 0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

4 5 9 5 6 6 5 8 3 4 0 9 2

8

2 9 8 2 5 3 3 4 7 4 0 0 3

1

S

* * * * * * * * *

2 1 0 1 0 0 0 0 0 0 0 1 1 6

K

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0

r

o

4 5 2 4 9 4 2 4 3 7 7 5 5 1

d

5 3 0 6 0 9 3 0 6 7 6 8 0 0

i

H

* * * * * * * * *

r

1 0 1 0 1 0 0 1 0 0 0 0 0 0

r

......

K

o

0 0 0 0 0 0 0 0 0 0 0 0 0 0

C

1 9 6 8 9 0 7 1 1 9 3 9 2

7

5 9 2 0 8 3 6 8 9 5 0 0 0

1

R

* * * * * * * * *

1 1 1 1 0 1 0 0 1 0 0 0 0 0

......

B

0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 6 8 8 0 8 0 0 0 0 9 6 7 4 0 8 5 1 9 1 2

1

9 2 4 7 0 3 0 0 0 0 5 6 2 7 8 0 2 1 4 5 7 1

*

L

1 1 0 0 0 0 0 0 0 0 0 0 1 5 0 0 2 0 2 1 0 2

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

)

l

9 1 9 5 0 1 0 0 0 5 1 2 6

l

6 4 5 8 0 2 0 0 0 3 0 0 9

a

* * * * * * * * *

(

1 1 0 0 0 0 0 0 0 6 0 0 4

......

0 0 0 0 0 0 0 0 0 0 0 0 0

M

2 8 6 9 0 8 0 0 3 4 4 1 4 7 3 8

7 9 5 6 0 4 0 0 8 3 0 9 0 2 0 2

5

* * * * * * *

1 0 0 0 0 0 0 0 5 4 0 0 0 0 0 0

......

M

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

4 8 9 1 0 0 0 8 1 1 7 2 4 0 5 1 1 5 0

9 7 6 0 0 4 0 1 9 7 1 0 0 7 0 0 0 3 1

4

* * * *

1 1 0 1 0 0 0 8 9 6 0 0 0 3 0 0 0 0 0

......

M

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

3 7 8 2 7 3 4 0 8 7 0 0 9 7 2 1 5

2 7 4 4 0 3 4 3 4 3 9 9 0 9 0 0 0

3

* * * * * *

1 0 0 0 0 0 7 6 5 6 0 0 0 6 0 0 0

M ......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 6 5 3 4 0 8 1 8 4

7 3

d

2 1 0 6 4 0 1 1 3 0 0

1 1

e

* * * * * * * * * * *

2 2 0 1 0 0 0 0 0 0 1 0

t

......

M

c

0 0 0 0 0 0 0 0 0 0 0 0

e

n

7 4 6 4 3 8 6 4 4 1 3 6 0 2 4

n

9 8 6 0 0 0 5 4 4 9 0 0 2 0 2

1

o

* * * * * * * *

1 1 0 1 0 2 9 3 9 6 0 2 0 0 0

......

M

C

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

)

l

9 0 1 4 2 1 1 5

l

1 5 3 1 0 6 0 1

a

* * * * * * * * * * * * * * *

(

0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0

R

1 5 6 1 3 3 4 4 3 9 8 8 4

8

2 4 8 0 0 0 0 1 7 4 1 1 0

3 1

* * * * * * * * *

1 1 0 0 0 0 0 0 1 0 0 0 0 0

R

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 7 6 2 0 2

4 8 3 0 1 3 0

2

* * * * * * * * * * * * * * * *

1 1 0 0 0 1 0

R

......

0 0 0 0 0 0 0

4

1

1

* * * * * * * * * * * * * * * * * * * * * *

2

R

.

0

)

)

l

l

l

l

a

a

(

(

1 2 3 4 5

A A

1 2 3

1 1 2 2 3 4

R R R R M M M M M M

s R H S s s s s s

d d d d d d d d d d

u u u u u u

L B K K S

e e e e e e e e e e

o t t t t t t t t t t o o o o o

r r r r r

B L W

c c c c c c c c c c

u u u u u u

o o o o o

e e e e e e e e e e e e e

n n n n n n

d d d d d

t t t i i i i i i

i i i i i

t n n n n n n n n n n t t t t t

r r r r r

a a a

l l l

n n n n n n n n n n n r r r r r n n n n n

o o o

o o o o o o o o o o o o o o o o o o o o o

s s s

C C C C C C C C C C C C C C C C I I I C C C C C Appendices - 143 -

Appendix 14. Pairwise comparisons for all populations of U. caudimaculatus – mtDNA. FST values above diagonal, exact test p-values below diagonal. Values in bold represent those pairwise comparisons referred to in Chapter 5. * = p < 0.001.

7 7 7 2 8 4 0 2 4 7 5

1

5

7 4 1 6 9 5 0 7 4 3 1 0

1

4

3 3 4 3 4 1 3 0 2 2 3 0 0

......

s

0 0 0 0 0 0 0 0 0 0 0 0 0

u

o

3 7 1 0 7 7 4 0 2 1 9 0 9

u

0 7 4 9 1 8 8 0 0 9 4 0 0

n

2

i

3 2 3 2 4 0 2 0 2 1 2 0 4

t

......

n

0 0 0 0 0 0 0 0 0 0 0 0 0

o

5 9 0 1 0 5 8 0 4 7 4 8 1

C

9 6 3 8 9 9 7 0 0 8 6 5 7

1

2 2 3 2 3 0 2 0 2 1 2 8 3

......

0 0 0 0 0 0 0 0 0 0 0 0 0

1 4 5 7 4 3 0 0 5

5

6 2 5 3 6 2 0 0 8

1

* * *

0 0 0 0 1 3 0 0 0 1

W

e

......

t

0 0 0 0 0 0 0 0 0 0

a

l

o

2 5 5 1 4 9 0

4 2

s

9 2 9 8 3 8 0

1 1

I

* * * *

B

1 1 1 1 2 1 1 0 1

......

0 0 0 0 0 0 0 0 0

0 0 0 0 0 3 0 0 5 0 0 7 7

0 0 0 0 0 7 0 0 9 0 7 7 6

S

0 0 0 0 0 2 0 0 5 0 6 5 2

......

0 0 0 0 0 0 0 0 0 1 0 0 0

8

0 0 0 0 0 0 0 0 0 0 0 0

8

0 0 0 0 0 0 0 0 0 0 0 0

S

6

0 0 0 0 0 0 0 0 0 0 0 0

.

K

......

0

0 0 0 0 0 0 0 1 1 1 1 1

r

5 0 0 0 0 9 0 0 7 4 6

o

1 0

0 0 0 0 0 8 0 0 1 7 3 3

d

1

H

i

0 0 0 0 0 3 0 0 3 5 2 1 0

r

......

K

r

0 0 0 0 0 0 1 1 0 0 0 0 0

o

C

9 3 8 4 6 5 0 6 1 6 7

1

1 5 6 2 9 3 0 1 0 3 4 2

R

*

3 3 3 3 5 0 0 2 0 0 0 0

......

B

0 0 0 0 0 0 1 0 0 0 0 0

1 3 4 0 0 0 0

9 5 5 6 0 0 0

* * * * * *

L

0 0 0 0 0 0 0

......

0 0 0 0 1 1 1

3 0 0 4 0 0 0 2

1 1

1 0 0 1 0 0 0 0 2 1

1

* * *

0 0 0 5 0 0 0 0 0 0

......

M

0 0 0 0 0 1 1 1 0 0

)

l

0 0 5 9 3 0 0 0 3

l

d

0 0 6 1 0 0 0 0 0

a

e

* * * *

(

0 0 1 4 0 0 0 0 0

t

......

c

0 0 0 0 0 1 1 1 0

R

e

n

6 0 0 7 0 0 0 2 8 9 5 1

1

n

2 6 0 6 1 0 0 0 8 0 0 0 0

3

o

1 2 0 3 0 0 0 0 1 1 0 0 0

R

......

C

0 0 1 0 0 1 1 1 0 0 0 0 0

4 5 9 8 3 8 0 0 3

9 5 6 3 0 6 0 0 0

1

* * * *

0 8 0 1 0 5 0 0 0

R

......

0 0 0 0 0 0 1 1 0

)

l

l

a

(

1

1 3

1 2 4

R R R M

R H S s s s

d d d d

u u u

L B K K S

e e e e

t t t t o o o

r r r r r

B W

c c c c

u u u

o o o o o

e e e e e e

n n n

d d d d d

t t i i i

i i i i i

n n n n t t t

r r r r r

a a

l l

n n n n r r r r r n n n

o o

o o o o o o o o o o o o

s s

C C C C C C C C C I I C C C Appendices - 144 -

Appendix 15. Pairwise comparisons for all populations of M. cervinipes – nuclear

DNA. FST values above diagonal, exact test p-values below diagonal. Values in bold represent those pairwise comparisons referred to in Chapter 5. * = p < 0.001.

2 9 2 2 4 1 3 5 0 6 5 9 6 6 8 0 7 8 5 2 5 7 5

2 1 1 1 4 2 3 2 3 2 0 2 4 2 2 6 3 2 0 1 0 2 1

4

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 6 4 6 0 0 0 3 4 5 1 8 4 3 2 4 5 5 0 1 7 2 2

4 2 2 2 4 2 1 1 2 1 2 2 4 2 3 6 4 3 0 2 0 3 4

3

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

7 7 7 8 2 1 0 4 9 4 5 5 6 5 6 8 2 0 9 2 6

s

2 2 4 2 5 3 3 2 1 2 4 5 5 4 1 4 7 2 0 0 2

u

A

* *

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

o

2

......

u

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

n

i

t

0 3 0 9 8 1 4 6 9 0 7 7 2 5 3 6 4 2 9 6 4 0

n

4 3 3 2 4 3 3 3 2 3 1 2 4 3 3 6 4 2 0 0 2 4

*

o

2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

......

C

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 0 4 6 6 1 8 5 7 1 8 5 8 3 4 3 2 5 5 9 6 9 9

3 3 3 2 4 2 2 2 1 2 1 3 3 3 1 5 4 0 0 0 2 0 3

A

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 1

1

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 3 8 3 3 6 0 9 8 9 3 1 2 1 5 6 2 5 3 0 8 3 2

2 1 1 1 2 0 1 0 1 0 0 2 5 2 2 4 5 2 5 2 5 5 6

1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 1

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 2 5 6 2 4 6 0 1 6 4 0 5 8 2 6 5 9

5 5 5 4 5 2 3 3 2 2 3 5 3 4 3 6 4 1

* * * * *

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

W

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

e

7 5 6 6 1 7 8 7 0 9 6 3 6 9 3 3

t

7 7 6 6 0 5 7 6 7 6 4 6 5 6 7 6

a

l

* * * * * * *

L

0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0

......

o

s

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

I

1 7 6 0 3 1 8 9 4 3 8 0 7 5 6

5 6 5 5 0 7 7 5 8 7 5 6 5 7 7

* * * * * * * *

B

0 0 0 0 1 0 0 0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

8 7 0 7 4 6 3 4 1 7 6 3 0 2 2 2 5

3 2 6 3 3 2 1 3 2 1 3 5 5 3 0 0 0

* * * * * *

S

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2 3 2 2 8 6 5 8 4 3 4 8 1

3 4 4 3 3 2 2 2 3 2 2 3 5

S

* * * * * * * * * *

0 0 0 0 0 0 0 0 0 0 0 0 0

K

......

0 0 0 0 0 0 0 0 0 0 0 0 0

r

o

1 8 8 0 4 0 3 7 1 8 0 5

d

6 6 6 6 9 0 6 5 5 5 5 5

i

H

* * * * * * * * * * *

r

0 0 0 0 0 0 0 0 0 0 0 0

r

......

K

o

0 0 0 0 0 0 0 0 0 0 0 0

C

3 7 8 2 0 4 7 1 3 5 0

3 3 1 2 3 3 2 4 3 2 0

R

* * * * * * * * * * * *

0 0 0 0 0 0 0 0 0 0 0

......

B

0 0 0 0 0 0 0 0 0 0 0

6 4 6 7 4 3 6 6 2 6 7 3 9 3 7 1

2 3 1 1 3 0 1 2 1 0 7 0 5 0 5 3

* * * * * * *

L

0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 1

......

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

)

l

7 2 9 1 7 0 0 6 0 7 1

l

3 3 3 3 0 0 0 0 0 6 2

a

* * * * * * * * * * * *

(

0 0 0 0 0 0 0 0 0 0 0

......

0 0 0 0 0 0 0 0 0 0 0

M

5 5 1 2 8 0 2 5 4 6 0

4 4 5 4 1 1 1 1 1 2 1

5

* * * * * * * * * * * *

0 0 0 0 0 0 0 0 4 0 0

M ......

0 0 0 0 0 0 0 0 0 0 0

4 8 5 0 8 9 7 7 9

3 3 3 3 3 0 1 3 0

4

* * * * * * * * * * * * * *

0 0 0 0 0 0 0 2 0

......

M

0 0 0 0 0 0 0 0 0

3 9 1 3 6 6 7 6 3 2 4 1 0

4 3 4 3 1 0 0 0 2 2 0 0 5

3

* * * * * * * * * *

0 0 0 0 0 0 0 0 9 0 0 0 1

M ......

0 0 0 0 0 0 0 0 0 0 0 0 0

9 9 0 4 0 7 6 2 5 1

d

4 2 4 3 2 0 1 0 0 0

2

e

* * * * * * * * * * * * *

0 0 0 0 0 0 0 0 4 0

t

......

M

c

0 0 0 0 0 0 0 0 0 0

e

n

8 4 7 6 1 4 9 4

n

4 5 5 4 0 5 4 1

1

o

* * * * * * * * * * * * * * *

0 0 0 0 0 1 2 0

M ......

C

0 0 0 0 0 0 0 0

)

l

0 0 0 6

l

0 0 0 0

a

* * * * * * * * * * * * * * * * * * *

(

0 0 0 0

. . . .

0 0 0 0

R

8 0 9

2

1 4 1

3 1

* * * * * * * * * * * * * * * * * * *

0 1 6 1

R

. . . .

0 0 0 0

8 7 1 2 3

1 0 7 0 0

2

* * * * * * * * * * * * * * * * * *

0 0 6 0 0

R

. . . . .

0 0 0 0 0

5 2

9 4

1

* * * * * * * * * * * * * * * * * * * * *

4 0

R

. .

0 0

)

)

l

l

l

l

a

a

(

(

1 2 3 4 5

A A

1 2 3

1 1 2 2 3 4

R R R R M M M M M M

R H S s s s s s s

d d d d d d d d d d

u u u u u u

L B K K S

e e e e e e e e e e

t t t t t t t t t t o o o o o o

r r r r r

B L W

c c c c c c c c c c

u u u u u u

o o o o o

e e e e e e e e e e e e e

n n n n n n

d d d d d

t t t i i i i i i

i i i i i

n n n n n n n n n n t t t t t t

r r r r r

a a a

l l l

n n n n n n n n n n r r r r r n n n n n n

o o o

o o o o o o o o o o o o o o o o o o o o o

s s s C C C C C C C C C C C C C C C I I I C C C C C C Appendices - 145 -

Appendix 16. Pairwise comparisons for all populations of U. caudimaculatus - nuclear DNA. FST values above diagonal, exact test p-values below diagonal. Values in bold represent those pairwise comparisons referred to in Chapter 5. * = p < 0.001.

2 3 0 8 4 7 8 0 8 1 1 4 1

4 6 4 9 4 9 5 0 3 0 0 4 2

4

0 0 0 0 0 0 1 0 0 1 1 0 0

......

s

0 0 0 0 0 0 0 0 0 0 0 0 0

u

o

6 2 3 4 8 8 3 0 7 7 1 7 2

u

5 6 5 2 0 9 5 0 4 9 8 0 3

n

2

i

0 0 0 1 1 0 1 0 0 0 0 0 0

t

......

n

0 0 0 0 0 0 0 0 0 0 0 0 0

o

0 0 1 7 2 9 8 5 7

6 6 3

C

8 0 8 3 2 6 0 8 0

1 1 1

*

1

0 1 0 1 1 1 1 0 0 1 1 3

......

0 0 0 0 0 0 0 0 0 0 0 0

4 4 0 7 1 7 7 9 4 5

5 7 5 7 8 6 5 2 9 5

* * *

0 0 0 0 0 0 1 0 0 0

W

e

......

t

0 0 0 0 0 0 0 0 0 0

a

l

o

0 1 4 9 4 6 1 9 4

s

5 6 4 0 6 0 3 6 0

I

* * * *

B

0 0 0 1 0 1 1 0 1

......

0 0 0 0 0 0 0 0 0

8 6 1 7 9 3 4 1 4 9 0

9 7

4 6 4 4 8 5 0 0 0 2 9

1 1

S

0 0 0 1 1 1 1 0 0 0 0 1 1

......

0 0 0 0 0 0 0 0 0 0 0 0 0

1

0 0 0 0 0 7 6 2 1 5 6

1

6

0 0 0 0 1 0 5 4 9 5 2 3

S

9

0 0 0 0 0 0 1 8 2 1 4 9

.

K

......

0

0 0 0 0 0 0 0 0 0 0 0 0

r

4 5 8 3 7 6 3

o

1

0 3 9 8 9 6 1

d 1

H

i

* * * * *

1 1 0 1 0 0 2 0

r

......

K

r

0 0 0 0 0 0 0 0

o

C

1 8 2 7 7 3 4

5 5 4 1 6 5 0

R

* * * * * *

0 0 0 0 0 3 0

......

B

0 0 0 0 0 0 0

1 7 1 5 2 8 9

6 9 6 5 1 8 0

* * * * * *

L

0 0 0 0 0 4 0

......

0 0 0 0 0 0 0

6 6 7 7 8

1

4 8 4 2 5 1

1

* * * * * * *

0 0 0 0 5 0

......

M

0 0 0 0 0 0

)

l

0 8 6

1

l

d

0 1 9 3

a

e

* * * * * * * * *

(

0 0 8 0

t

. . . .

c

0 0 0 0

R

e

n

1 7 4 0

n

4 0 3 2

3

o

* * * * * * * * *

0 3 7 0

R

. . . .

C

0 0 0 0

3 0 9 0

0 0 1 5

1

* * * * * * * * *

0 0 9 0

R

. . . .

0 1 0 0

)

l

l

a

(

1

1 3

1 2 4

R R R M

R H S s s s

d d d d

u u u

L B K K S

e e e e

t t t t o o o

r r r r r

B W

c c c c

u u u

o o o o o

e e e e e e

n n n

d d d d d

t t i i i

i i i i i

n n n n t t t

r r r r r

a a

l l

n n n n r r r r r n n n

o o

o o o o o o o o o o o o

s s C C C C C C C C C I I C C C

- 146 -

REFERENCES

Aars, J. and Ims, R. A. (1999) The effect of habitat corridors on rates of transfer and interbreeding between vole demes. Ecology 80(5): 1648-1655.

Aars, J., Ims, R. A., Liu, H-P., Mulvey, M. and Smith, M. H. (1998) Bank voles in linear habitats show restricted gene flow as revealed by mitochondrial DNA (mtDNA). Molecular Ecology 7: 1383-1389.

Aars, J., Johannesen, E. and Ims, R. A. (1999) Demographic consequences of movements in subdivided root vole populations. Oikos 85: 204-216.

Adler, G. H. and Levins, R. (1994) The island syndrome in rodent populations. Quarterly Review of Biology 69: 473-490.

Allendorf, F. W. (1983) Isolation, gene flow and genetic differentiation among populations. In: Genetics and Conservation. A reference for managing wild animal and plant populations (Eds. Schoenwald-Cox, C. M., Chambers, S. M, MacBryde, B., and Thomas, W. L.) 51-65. The Benjamin/Cummings Publishing Company, Inc., California.

Allendorf, F. W. and Leary, R. F. (1986) Heterozygosity and fitness in natural populations of animals. In: Conservation Biology: The Science of Scarcity and Diversity (Eds. Soulé, M.) 57-76. Sinauer, Sunderland, MA.

Amarasekare, P. (1994) Spatial population structure in the banner-tailed kangaroo rat, Dipodomys spectabilis. Oecologia 100: 166-176.

Amos, W. and Harwood, J. (1988) Factors affecting levels of genetic diversity in natural populations. Philosophic Transactions of the Royal Society of London B 353: 177-186.

Anderson, G. S. and Danielson, B. J. (1997) The effects of landscape composition and physiognomy on metapopulation size: the role of corridors. Landscape Ecology 12: 261-271.

Anderson, P. K. (1970) Ecological structure and gene flow in small mammals. Symposium of the Zoological Society of London 26: 299-325. References - 147 -

Andreassen, H. P., Hertxberg, K. and Ims, R. A. (1998) Space-use responses to habitat fragmentation and connectivity in the root vole Microtus oeconomus. Ecology 79(4): 1223-1235.

Ballard, J. W. O. and Whitlock, M. C. (2004) The incomplete natural history of mitochondria. Molecular Ecology 13: 729-744.

Balloux, F. and Lugon-Moulin, N. (2002) The estimation of population differentiation with microsatellite markers. Molecular Ecology 11: 155-165.

Barton, N. H. (1992) The genetic consequences of dispersal. In: Animal dispersal: small mammals as a model (Eds. Stenseth, N. C. and Lidicker Jr., W. Z.) 37- 60. Chapman and Hall, London.

Baudry, J. and Merriam, G. (1988) Connectivity and connectedness: functional vs structural patterns in landscapes. In: Connectivity in Landscape Ecology. Proceedings of the 2nd seminar of the International Association for Landscape Ecology (Eds. Schreiber, K. F.) 23-28. Munster,

Beier, P. (1995) Dispersal of juvenile in fragmented habitat. Journal of Wildlife Management 59(2): 228-237.

Beier, P. and Loe, S. (1992) A checklist for evaluating impacts to wildlife movement corridors. Wildlife Society Bulletin 20: 434-440.

Beier, P. and Noss, R. F. (1998) Do habitat corridors provide connectivity? Conservation Biology 12(6): 1241-1252.

Bennett, A. F. (1990a) Habitat corridors: their role in wildlife management and conservation. Department of Conservation and Environment, Melbourne.

Bennett, A. F. (1990b) Habitat corridors and the conservation of small mammals in a fragmented forest environment . Landscape Ecology 4(2): 109-122.

Bennett, A. F. (1999) Linkages in the Landscape. The role of corridors and connectivity in wildlife conservation. IUCN, Gland, Switzerland.

Bennett, A. F., Henein, K. H. and Merriam, G. (1994) Corridor use and the elements of corridor quality: chipmunks and fencerows in a farmland mosaic. Biological Conservation 68: 155-165. References - 148 -

Bentley, J. M. and Catterall, C. P. (1997) The use of bushland, corridors and linear remnants by birds in southeastern Queensland, Australia. Conservation Biology 11(5): 1173-1189.

Bentley, J. M., Catterall, C. P. and Smith, G. C. (2000) The effects of araucarian vine forest fragmentation on small mammal communities: the role of habitat specificity. Conservation Biology 14(4): 1075-1087.

Berggren, A., Birath, B. and Kindvall, O. (2002) Effect of corridors and habitat edges on dispersal behavior, movement rates, and movement angles in Roesels' bush-cricket (Metrioptera roesli). Conservation Biology 16(6): 1562- 1569.

Berry, O., Tocher, M. D. and Sarre, S. D. (2004) Can assignment tests measure dispersal? Molecular Ecology 13(3): 551-561.

Bierregaard Jr. R.O., Laurance, W. F., Sites, J. W., Lynam, A. J., Didham, R. K., Anderson, M., Gascon, C., Tocher, M. D., Smith, A. P., Viana, V. M., Lovejoy, T. E., Sieving, K. E., Kramer, E. A., Restrepo, C. and Moritz, C. (1997) Key priorities for the study of fragmented tropical ecosystems. In: Tropical Forest Remnants. Ecology, Management and Conservation of Fragmented Communities. (Eds. Laurance, W. F. and Bierregaard, R. O.) 515-525. The University of Chicago Press, Chicago.

Birky, C. W., Fuerst, P. and Maruyama, T. (1989) Organelle gene diversity under migration, mutation, and drift: equilibrium expectations, approach to equilibrium, effects of heteroplasmic cells, and comparison to nuclear genes. Genetics 121: 613-627.

Bolen, E. G. and Robinson, W. L. (1995) Wildlife ecology and management. Prentice Hall, New Jersey.

Bolger, D. T., Scott, T. A. and Rotenberry, J. T. (2001) Use of corridor-like landscape structures by and small mammal species. Biological Conservation 102: 213-224.

Boudjemadi, K., Lecomte, J. and Clobert, J. (1999) Influence of connectivity on demography and dispersal in two contrasting habitats: an experimental approach. Journal of Animal Ecology 68: 1207-1224. References - 149 -

Bowen, B. S. (1982) Temporal dynamics of microgeographic structure of genetic variation in Microtus californicus . Journal of Mammalogy 63: 625-638.

Bowne, D. R., Peles, J. D. and Barrett, G. W. (1999) Effects of landscape spatial structure on movement patterns of the hispid cotton rat (Sigmodon hispidus). Landscape Ecology 14: 53-65.

Brandt, C. A. (1992) Social factors in immigration and emigration. In: Animal dispersal: small mammals as a model (Eds. Stenseth, N. C. and Lidicker Jr, W. Z.) 96-144. Chapman and Hall, London.

Brooker, L., Brooker, M. and Cale, P. (1999) Animal dispersal in fragmented habitat: measuring habitat connectivity, corridor use, and dispersal mortality. Conservation Ecology [online] 3(4).

Brooks, C. P. (2003) A scalar analysis of landscape connectivity. Oikos 102(2): 433-439.

Brooks, R. A. and Bell, S. S. (2001) Mobile corridors in marine landscapes: enhancement of faunal exchange at seagrass/sand ecotones. Journal of Experimental Marine Biology and Ecology 264: 67-84.

Brown, J. H. and Kodric-Brown, A. (1977) Turnover rates in : effect of immigration on extinction. Ecology 58: 445-449.

Burbrink, F. T., Phillips, C. A. and Heske, E. J. (1998) A riparian zone in southern Illinois as a potential dispersal corridor for and . Biological Conservation 86: 107-115.

Burel, F. (1989) Landscape structure effects on carabid beetles spatial patterns in western France. Landscape Ecology 2(4): 215-226.

Burgman, M. A. and Lindenmayer, D. B. (1998) Conservation biology for the Australian environment. Surrey Beatty and Sons, Chipping Norton, NSW.

Burkey, T. V. (1995) Extinction rates in archipelagos: implications for populations in fragmented habitats. Conservation Biology 9(3): 527-541.

References - 150 -

Campbell, N. J. H. (1996) Mitochondrial Control Region variation in two genera of Australian rodents; Melomys and Uromys. Application to phylogenetics, phylogeography and conservation. PhD Thesis, Southern Cross University, Lismore.

Campbell, N. J. H., Harriss, M. S., Elphinstone, M. S. and Baverstock, P. R. (1995) Outgroup heteroduplex analysis using temperature gradient gel electrophoresis: high resolution, large scale, screening of DNA variation in the mitochondrial control region. Molecular Ecology 4: 407-418.

Chambers, G. K. and MacEvoy, E. S. (2000 ) Microsatellites: consensus and controversy . Comparative Biochemistry and Physiology Part B 126: 455- 476.

Chesser, R. K. (1983a) Genetic variability within and among populations of the black-tailed prairie dog. Evolution 37: 320-331.

Chesser, R. K. (1983b) Isolation by distance: relationship to the management of genetic resources. In: Genetics and conservation. A reference for managing wild animal and plant populations (Eds. Schoenwald-Cox, C. M., Chambers, S. M, MacBryde, B., and Thomas, W. L.) 66-77. The Benjamin/Cummings Publishing Company, Inc., California.

Chesser, R. K. (1998) Relativity of behavioural interactions in socially structured populations. Journal of Mammalogy 79(3): 713-724.

Chesser, R. K., Smith, M. H. and Brisbin, I. L. (1980) Management and maintenance of genetic variability in endangered species. International Zoology Yearbook 20.

Cockburn, A. (1992) Habitat heterogeneity and dispersal: environmental and genetic patchiness. In: Animal dispersal: small mammals as a model (Eds. Stenseth, N. C. and Lidicker Jr, W. Z.) 65-85. Chapman and Hall, London.

Coffman, C. J., Nichols, J. D. and Pollock, K. H. (2001) Population dynamics of Microtus pennsylvanicus in corridor-linked patches. Oikos 93(1): 3-21.

References - 151 -

Colgan, D. J., Brown, S., Major, R. E., Christie, F., Gray, M. R. and Cassis, G. (2002) Population genetics of spiders of fragmented habitat in the wheat belt of New South Wales. Molecular Ecology 11(11): 2295-2305.

Cook, W. M., Anderson, R. M. and Schweiger, E. W. (2004) Is the matrix really inhospitable? Vole runway distribution in an experimentally fragmented landscape. Oikos 104(1): 5-14.

Cox, M. P., Dickman, C. R. and Hunter, J. (2003) Effects of rainforest fragmentation on non-flying mammals of the Eastern Dorrigo Plateau, Australia. Biological Conservation 15(2): 175-189.

Crome, F. H. J. (1997) Researching tropical forest fragmentation: shall we keep on doing what we're doing? In: Tropical Forest Remnants. Ecology, Management and Conservation of Fragmented Communities. (Eds. Laurance, W. F. and Bierregaard, R. O.) 485-501. The University of Chicago Press, Chicago.

Crome, F., Isaacs, J. and Moore, L. (1994) The utility to birds and mammals of remnant riparian vegetation and associated windbreaks in the tropical Queensland uplands. Pacific Conservation Biology 1: 328-343.

Csuti, B. (1991) Conservation corridors: countering habitat fragmentation. Introduction. In: Landscape linkages and biodiversity (Eds. Hudson, W. E) 81- 90. Island Press, Washington.

Cunningham, M. and Moritz, C. (1998) Genetic effects of forest fragmentation on a rainforest restricted lizard (Scinidae: Gnypetoscincus queenslandiae). Biological Conservation 83(1): 19-30.

Danielson, B. J. and Hubbard, M. W. (2000) The influence of corridors on the movement behaviour of individual Peromyscus polionotus in experimental landscapes. Landscape Ecology 15: 323-331.

Davis-Born, R. and Wolff, J. O. (2000) Age- and sex-specific responses of the gray-tailed vole, Microtus canicaudus, to connected and unconnected habitat patches. Canadian Journal of Zoology 78: 864-870.

References - 152 - de Lima, M. G. and Gascon, C. (1999) The conservation value of linear forest remnants in central Amazonia. Biological Conservation 91: 241-247.

Debinski, D. M. and Holt, R. D. (2000) A survey and overview of habitat fragmentation experiments. Conservation Biology 14(2): 342-355.

Doak, D. F. and Mills, L. S. (1994) A useful role for theory in conservation. Ecology 75: 615-626.

Dobson, F. S. (1982) Competition for mates and predominant juvenile male dispersal in mammals. Animal Behaviour 30: 1183-1192.

Downes, S. J., Handasyde, K. A. and Elgar, M. A. (1997a) The use of corridors by mammals in fragmented Australian eucalypt forests. Conservation Biology 11(3): 718-726.

Downes, S. J., Handasyde, K. A. and Elgar, M. A. (1997b) Variation in the use of corridors by introduced and native rodents in south-eastern Australia. Biological Conservation 82: 379-383.

Earn, D. J. D., Levin, S. A. and Rohani, P. (2000) Coherence and Conservation. Science 290: 1360-1364.

Ellegren, H (2000) Microsatellite mutations in the germline. Trends in Genetics 16(12): 551-558.

Elphinstone, M. S. and Baverstock, P. R. (1997) Detecting mitochondrial genotypes by Temperature Gradient Gel Electrophoresis and Heteroduplex Analysis. BioTechniques 23: 982-986.

Excoffier, L., Smouse, P. E. and Quattro, J. (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131(2): 479-491.

Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual review of Ecology and Systematics 34: 487-515.

Fahrig, L. and Merriam, G. (1985) Habitat patch connectivity and population survival. Ecology 66(6): 1762-1768. References - 153 -

Fahrig, L. and Merriam, G. (1994) Conservation of fragmented populations. Conservation Biology 8(1): 50-59.

Forman, R. T. T. and Godron, M. (1986) Landscape Ecology. John Wiley, New York.

Frankel, O. H. and Soulé, M. E. (1981) Conservation and evolution. Cambridge University Press, Cambridge.

Frankham, R. and Ralls, K. (1998) Inbreeding leads to extinction. Nature 392: 441-442.

Franklin, I. R. (1980) Evolutionary change in small populations. In: Conservation Biology (Eds. Soulé, M. E. and Wilcox, B. A.) 135-149. Sinauer, Sunderland.

Fraser, D. F., Gilliam, J. F., MacGowan, M. P., Arcaro, C. M. and Guillozet, P. H. (1999) Habitat quality in a hostile river corridor. Ecology 80(2): 597-607.

Frawley, K. J. (1983) A history of forest and land management in Queensland, with particular reference to the North Queensland rainforest. Report to the Rainforest Conservation Society of Queensland. Brisbane.

Friend, J. A. (1987) Local decline, extinction and recovery: Relevance to mammal populations in vegetation remnants. In: Nature conservation: the role of remnants of native vegetation (Eds. Saunders, D. A., Arnold, G. W., Burbidge, A. A., and Hopkins, A. J. M.) 53-64. Surrey Beatty and Sons Pty Ltd, Sydney.

Fuller, S. J., Wilson, J. C. and Mather, P. B. (1997) Patterns of differentiation among wild rabbit populations Oryctolagus cuniculus L. in arid and semiarid ecosystems of north-eastern Australia. Molecular Ecology 6(2): 145-153.

Gaggiotti, O. E., Lange, O., Rassmann, K. and Gliddon, C. (1999) A comparison of two methods for estimating average levels of gene flow using microsatellites data. Molecular Ecology 8: 1513-1520.

Gaines, M. S., Diffendorfer, J. E. , Tamarin, R. H. and Whittam, T. S. (1997) The effects of habitat fragmentation on the genetic structure of small mammal populations. Journal of Heredity 88: 294-304.

References - 154 -

Garza, J. C. and Williamson, E. G. (2001) Detection of reduction in population size using data from microsatellite loci. Molecular Ecology 10: 305-318.

Gilpin, M. E. and Soulé, M. E. (1986) Minimum viable populations: processes of species extinction. In: Conservation Biology: the science of scarcity and diversity (Eds. Soulé, M. E.) . Schaller Associates, Sunderland, MA.

Goosem, M. (2001) Effects of tropical rainforest roads on small mammals: inhibition of crossing movements. Wildlife Research 28(4): 351-364.

Goudet, J., Raymond, M., De Meeüs, T. and Rousset, F. (1996) Testing differentiation in diploid populations. Genetics 144: 1933-1940.

Guo, S. W. and Thompson, E. A. (1992) Performing the exact test of Hardy- Weinberg proportions for multiple alleles. Biometrics (48): 361-372.

Haddad, N. M. (1999a) Corridor and distance effects on interpatch movements: a landscape experiment with butterflies. Ecological Applications 9(2): 612-622.

Haddad, N. M. (1999b) Corridor use predicted from behaviours at habitat boundaries. The American Naturalist 153(2): 215-227.

Haddad, N. M. (2000) Corridor length and patch colonization by a butterfly, Junonia coenia. Conservation Biology 14: 738-745.

Haddad, N. M. and Baum, K. A. (1999) An experimental test of corridor effects on butterfly densities. Ecological Applications 9(2): 623-633.

Haddad, N. M., Bowne, D. R., Cunningham, A., Danielson, B. J., Levey, D. J., Sargent, S. and Spira, T. (2003) Corridor use by diverse taxa. Ecology 84(3): 609-615.

Haddad, N. M., Rosenberg, D. K. and Noon, B. R. (2000) On experimentation and the study of corridors: response to Beier and Noss. Conservation Biology 14(5): 1543-1545.

Hale, M. L., Lurz, P. W., Shirley, M. D. F., Rushton, S., Fuller, R. M. and Wolff, K. (2001) Impact of landscape management on the genetic structure of red squirrel populations. Science 293(2246-2248). References - 155 -

Hanski, I. (1994) Spatial scale, patchiness and population dynamics on land. Phil. Trans. R. Soc. Lond. B 343: 19-25.

Harris, L. D. and Scheck, J. (1991) From implications to applications: the dispersal corridor principle applied to conservation of biological diversity. In: Nature Conservation (Eds. D.A. Saunders and R.J. Hobbs) 189-220. Surrey Beatty and Sons Pty Ltd, Sydney.

Harrison, R. G. (1989) Animal mitochondrial DNA as a genetic marker in population and evolutionary biology. Trends in Ecology and Evolution 4(1): 6-11.

Harrison, R. L. (1992) Toward a theory of inter-refuge corridor design. Conservation Biology 6(2): 293-295.

Hemiview 2.1 (1999) Canopy analysis software. Delta-T Devices Ltd.

Henderson, M. T, Merriam, G. and Wegner, J. (1985) Patchy environments and species survival: chipmunks in an agricultural mosaic. Biological Conservation 31: 95-105.

Henein, K. and Merriam, G. (1990) The elements of connectivity where corridor quality is variable. Landscape Ecology 4(2): 157-170.

Hill, C. J. (1995) Linear strips of rain forest vegetation as potential dispersal corridors for rain forest . Conservation Biology 9(6): 1559-1566.

Hilty, J. A. and Merenlender, A. M. (2004) Use of riparian corridors and vineyards by mammalian predators in northern California. Conservation Biology 18(1): 126-135.

Hobbs, R. J. (1992) The role of corridors in conservation: solutions or bandwagon? Trends in Ecology and Evolution 7(11): 389-392.

Hobbs, R. J. (2002) Habitat networks and biological conservation. In: Applying landscape ecology in biological conservation (Ed. K.J. Gutzwiller) 150-170. Springer, New York.

Hobbs, R. J. and Saunders, D. A. (1990) Nature Conservation: the role of corridors. Ambio 19(2): 94-95. References - 156 -

Hobbs, R. J. and Wilson, A.-M. (1998) Corridors: theory, practice and the achievement of conservation objectives. In: Key Concepts in Landscape Ecology (Eds. Dover, J.W. and Bunce, R.G.H.) 265-279. IALE, Preston.

Horn, H. S. (1966) Measurements of "overlap" in comparative ecological studies. American Naturalist 100: 419-424.

Hudgens, B. R. and Haddad, N. M. (2003) Predicting which species will benefit from corridors in fragmented landscapes from population growth models. The American Naturalist 161(5): 808-820.

Hughes, C. (1998) Integrating molecular techniques with field methods in studies of social behavior: a revolution results. Ecology 79(2): 383-399.

Hyland, B. P. M., Whiffin, T., Christophel, D. C., Gray, B., and Elick, R. W. (1999) Australian tropical rain forest trees and shrubs. CSIRO Publishing, Melbourne.

Ims, R. A. and Andreassen, H. P. (1999) Effects of experimental habitat fragmentation and connectivity on root vole demography. Journal of Animal Ecology 68: 839-852.

Inglis, G. and Underwood, A. J. (1992) Comments on some designs proposed for experiments on the biological importance of corridors. Conservation Biology 6(4): 581-586.

Jarne, P. and Lagoda, P. J. L. (1996) Microsatellites, from molecules to populations and back. Trends in Ecology and Evolution 11(10): 424-429.

Johnsingh, A. J. T., Prasad, S. N. and Goyal, S. P. (1990) Conservation status of the Chila-Motichur Corridor for movement in Rajaji-Corbett National Parks Area, India. Biological Conservation 51: 125-138.

Jongman, R. H. G. (1995) Nature conservation planning in Europe: developing ecological networks. Landscape and Urban Planning 32: 169-183.

Jordan, F. (2000) A reliability-theory approach to corridor design. Ecological Modelling 128: 211-220.

References - 157 -

Kaiser, J. (2001) Bold corridor project confronts political reality. Science 293(5538): 2196-2199.

Keith, K. (1970) The occurrence of a second species of naked-tailed rat genus Melomys in New South Wales. Wildlife Research 15: 19-26.

Kirchner, F., Ferdy, J.-B., Andalo, C., Colas, B. and Moret, J. (2003) Role of corridors in plant dispersal: an example with the endangered Ranunculus nodiflorus. Conservation Biology 17(2): 401-410.

Knox, E. (1978) A note on the identification on Melomys species (Rodentia: Muridae) in Australia. Journal of Zoology London 185: 276-277.

Koenig, W. D., Van Vuren, D. and Hooge, P. N. (1996) Detectability, philopatry, and the distribution of dispersal distances in vertebrates. Trends in Ecology and Evolution 11(12): 514-517.

Kozakiewicz, M. (1993) Habitat isolation and ecological barriers - the effect on small mammal populations and communities. Acta Theriologica 38(1): 1-30.

Krebs, C. J. (1989) Ecological Methodology. Harper Collins Publishers, New York.

Kurt, F. (1991) Das Reh in der Kulturlandschaft. Parey, Berlin.

La Polla, V. N. and Barrett, G. W. (1993) Effects of corridor width and presence on the population dynamics of the meadow vole (Microtus pennsylvanicus). Landscape Ecology 8(1): 25-37.

Lacy, R. C. (1987) Loss of genetic diversity from managed populations: interacting effects of drift, mutation, immigration, selection and population subdivision. Conservation Biology 1(2): 143-158.

Lande, R. (1988) Genetics and demography in biological conservation. Science 241: 1455-1460.

Lande, R and Barrowclough, G. F. Effective population size, genetic variation and their use in population management. In: Viable populations for conservation (Eds. Soule, M. E.) . Cambridge University Press, Cambridge.

References - 158 -

Laurance, S. G. and Laurance, W. F. (1999) Tropical wildlife corridors: use of linear rainforest remnants by arboreal mammals. Biological Conservation 91: 231- 239.

Laurance, W. F. (1990) Comparative responses of five arboreal marsupials to tropical forest fragmentation. Journal of Mammalogy 71(4): 641-653.

Laurance, W. F. (2001) Letter to The Independent (London) 02/02/2001.

Lessa, E. P. and Applebaum, G. (1993) Screening techniques for detecting allelic variation in DNA sequences. Molecular Ecology 2: 119-129.

Leung, L. K., Dickman, C. R. and Moore, L. A. (1993) Genetic variation in fragmented populations of an Australian rainforest rodent, Melomys cervinipes. Pacific Conservation Biology 1: 58-65.

Levins, R. (1969) Some demographic and genetic consequences of environmental heterogeneity for biological control. Bulletin of the Entomological Society America 15: 237-240.

Lidicker Jr., W. Z. (1995) Landscape approaches in mammalian ecology and conservation. University of Minnesota Press, Minneapolis.

Lidicker Jr., W. Z. (1999) Responses of mammals to habitat edges: an overview. Landscape Ecology 14(4): 333-343.

Lidicker Jr., W. Z. and Patton, J. L. (1987) Patterns of dispersal and genetic structure in populations of small rodents. In: Mammalian dispersal patterns. The effects of social structure on population genetics. (Eds. Chepko-Sade, B. D. Halpin Z. T.) 144-161. The University of Chicago Press, Chicago.

Lidicker Jr., W. Z. and Stenseth, N. C. (1992) To disperse or not to disperse: who does it and why? In: Animal dispersal: small mammals as a model (Eds. Stenseth, N. C. and Lidicker Jr., W. Z.) 21-36. Chapman and Hall , London.

Liedloff, A. (1999) Mantel version 2.0. Mantel non-parametric test calculator. www.sci.qut.edu.au/NRS/Mantel.htm .

References - 159 -

Lindenmayer, D. B. (1994) Wildlife corridors and the mitigation of logging impacts on fauna in wood-production forests in south-eastern Australia. Wildlife Research 21: 323-340.

Lindenmayer, D. B., Cunningham, R. B. and Donnelly, C. F. (1993) The conservation of arboreal marsupials in the montane ash forests of the Central Highlands of Victoria, south-east Australia, IV. The presence and abundance of arboreal marsupials in retained linear habitats (wildlife corridors) within logged forest. Biological Conservation 66: 207-221.

Lindenmayer, D. B., Cunningham, R. B., Donnelly, C. F., Triggs, B. J. and Belveders, M. (1994) The conservation of arboreal marsupials in the montane ash forests of the central highlands of Victoria, south-eastern Australia V. Patterns of use and the microhabitat requirements of the mountain brushtail possum Trichosurus caninus Ogilby in retained linear habitats (wildlife corridors). Biological Conservation 68: 43-51.

Lindenmayer, D. B. and Nix, H. A. (1993) Ecological principles for the design of wildlife corridors. Conservation Biology 7(3): 627-630.

Lindenmayer, D. and Peakall, R. (2000) The Tumut experiment - integrating demographic and genetic studies to unravel fragmentation effects: a case study of the native bush rat. In: Genetics, demography and viability of fragmented populations (Eds. Young, A. G. and Clarke, G. M.) 173-201 . Cambridge University Press, Cambridge.

Lomolio, M. V. and Perault, D. R. (2000) Assembly and disassembly of mammal communities in a fragmented temperate rain forest. Ecology 8(16): 1517- 1532.

Lord, J. M. and Norton, D. A. (1990) Scale and the spatial concept of fragmentation. Conservation Biology 4(2): 197-202.

Ludwig, J. A., Wiens, J. A. and Tongway, D. J. (2000) A scaling rule for landscape patches and how it applies to conserving soil resources in topical savannas. Ecosystems 3: 84-97.

References - 160 -

Lynch, J. F. and Saunders, D. A. (1991) Responses of bird species to habitat fragmentation in the wheatbelt of Western Australia: interiors, edges and corridor. In: Nature conservation 2: the role of corridors (Eds. Saunders, D. A. and Hobbs, R. J.) 143-158. Surrey Beattie, Chipping Norton, NSW.

Mabry, K. E. and Barrett, G. W. (2002) Effects of corridors on home range sizes and interpatch movements of three small mammal species. Landscape Ecology 17(4): 629-636.

Mabry, K. E., Dreelin, E. A. and Barrett, G. W. (2003) Influence of landscape elements on population densities and habitat use of three small-mammal species. Journal of Mammalogy 84(1): 20-25.

MacArthur, R. H. and Wilson, E. O. (1967) The theory of island biogeography. Princeton University Press, Princeton.

MacMahon, J. A. and Holl, K. D. (2001) Ecological restoration. A key to conservation biology's future. In: Conservation Biology. Research priorities for the next decade (Eds. Soule, M. E. and Orians, G. H.) 245-264. Island Press, Washington.

MacNally, R. and Quinn, G. (1998) Symposium introduction: the importance of scale in ecology. Australian Journal of Ecology 23: 1-7.

Mansergh, I. M. and Scotts, D. J. (1989) Habitat continuity and social organization of the mountain pygmy-possum restored by . Journal of Wildlife Management 53(3): 701-707.

Mantel, N. (1967) The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209-220.

McCaughley, D. E. (1993) Genetic consequences of extinction and recolonization in fragmented habitats. In: Biotic interactions and global change. (Eds. Kareiva, P. M., Kingsolver, J. G., and Huey, R. B.) 217-233. Sinauer Associates, Mass.,

Mech, S. G. and Hallett, J. G. (2001) Evaluating the effectiveness of corridors: a genetic approach. Conservation Biology 15(2): 467-474.

References - 161 -

Meetenmeyer, V. and Box, E. O. (1987) Scale effects in landscape studies. In: Landscape heterogeneity and (Eds. Turner, M. G.) 15-34 . Springer-Verlag, New York.

Menkhorst, P. and Knight, F. (2001) A field guide to the mammals of Australia. Oxford University Press, South Melbourne.

Merriam, G. (1984) Connectivity: A fundamental ecological characteristic of landscape pattern. In: Proceedings of the first international seminar on methodology in landscape ecological research and planning (Eds. Brandt, J. and Agger, P. A.) 5-15. Rosskilde University Centre, Rosskilde, Denmark.

Merriam, G. (1991) Corridors and connectivity: animal populations in heterogeneous environments. In: Nature Conservation 2: The role of corridors (Eds. D.A. Saunders and R.J. Hobbs) 133-142. Surrey Beatty and Sons Pty Ltd, Sydney.

Merriam, G. (1995) Movement in spatially divided populations: responses to landscape structure. In: Landscape approaches in mammalian ecology and conservation (Eds. Lidicker Jr., W. Z.) 64-77. University of Minnesota Press, Minneapolis.

Merriam, G. and Saunders, D. A. (1993) Corridors in restoration of fragmented landscapes. In: Nature Conservation 3: Reconstruction of fragmented ecosystems (Eds. Saunders, D. A., Hobbs, R. J., and Ehrlich, P. R.) 71-87. Surrey Beatty & Sons, Sydney.

Metzgar, L. H. (1979) Dispersion patterns in a Peromyscus population. Journal of Mammalogy 60: 129-145.

Metzgar, L. H. (1980) Dispersion and numbers in Peromyscus populations. American Midland Naturalist 103: 26-31.

Meyer, A., Kocher, T. D., Basasibwaki, P. and Wilson, A. C. (1990) Monophyletic origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA sequences. Nature 347: 550-553.

Mills, L. S. and Smouse, P. E. (1994) Demographic consequences of inbreeding in small remnant populations. American Naturalist 144: 412-431. References - 162 -

Mönkkönen, M. and Mutanen, M. (2003) Occurrence of moths in boreal forest corridors. Conservation Biology 17(2): 468-475.

Moritz, C. (1994) Applications of mitochondrial DNA analysis in conservation: a critical review. Molecular Ecology 3: 401-411.

Moritz, C., Dowling, T. E. and Brown, W. M. (1987) Evolution of animal mitochondrial DNA: relevance for population biology and systematics. Annual Review of Ecology and Systematics 18: 269-292.

Mossman, C. A. and Waser, P. M. (2001) Effects of habitat fragmentation on population genetic structure in the white-footed mouse (Peromyscus leucopus). Canadian Journal of Zoology 79(2): 285-295.

Murcia, M. (1995) Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution 10(2): 58-62.

Murphy, D. (1989) Conservation and confusion: wrong species, wrong scale, wrong conclusions. Conservation Biology 3(1): 82-84.

Mwalyosi, R. B. B. (1991) Ecological evaluation for wildlife corridors and buffer zones for Lake Manyara National Park, Tanzania, and its immediate environment. Biological Conservation 57: 171-186.

Nei, M. (1987) Molecular Evolutionary Genetics. Columbia University Press, New York.

Neigel, J . (2002) Is FST obsolete? Conservation Genetics 3: 167-173.

Newman, D. and Tallmon D. A. (2001) Experimental evidence for beneficial fitness effects of gene flow in recently isolated populations. Conservation Biology 15(4): 1054-1063.

Newmark, W. D. (1993) The role and design of wildlife corridors with examples from Tanzania. Ambio 22: 500-504.

Nicholls, A. O. and Margules, C. R. (1991) The design of studies to demonstrate the biological importance of corridors. In: Nature Conservation 2: The role of corridors (Eds. D.A. Saunders and R.J. Hobbs) 49-61. Surrey Beatty and Sons Pty Ltd, Sydney. References - 163 -

Nicholls, C. I., Parrella, M. and Altieri, M. A. (2001) The effects of a vegetational corridor on the abundance and dispersal of insect biodiversity within a northern California organic vineyard. Landscape Ecology 16: 133-146.

Noss, R. F. (1987) Corridors in real landscapes: a reply to Simberloff and Cox. Conservation Biology 1: 159-164.

Noss, R. F. (1992) The Wildlands project: land conservation strategy. Wild Earth (Special Issue): 10-25.

Noss, R. F. and Beier P. (2000) Arguing over little things: response to Haddad et al. Conservation Biology 15(4): 1546-1548.

Opdam, P. (2002) Assessing the conservation potential of habitat networks. In: Applying landscape ecology in biological conservation (Ed. K.J. Gutzwiller) 381-427. Springer, New York.

Paetkau, D., Calvert, W., Sterling, I. and Strobeck, C. (1995) Microsatellite analysis of population structure in Canadian polar . Molecular Ecology 4: 347- 354.

Peakall, R., Ruibal, M. and Lindenmeyer, D. B. (2003) Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes. Evolution 57(5): 1182-1195.

Perault, D. R. and Lomolino, M. V. (2000) Corridors and mammal community structure across a fragmented old-growth forest landscape. Ecological Monographs 70(3): 401-422.

Pollard, J. H. (1971) On distance estimators of density in randomly distributed forests. Biometrics 27: 991-1002.

Pope, T. R. (1992) The influence of dispersal patterns and mating system on genetic differentiation within and between populations of the red howler monkey (Alouatta seniculus). Evolution 46: 1112-1128.

Primack, R. B. (1993) Essentials of conservation biology. Sinauer Associates, Inc., Massachusetts.

References - 164 -

Prugnolle, F. and de Meeus, T. (2002) Inferring sex-biased dispersal from population genetic tools: a review. Heredity 88: 161-165.

Pulliam, H. R. (1988) Sources, sinks and population regulation. American Naturalist 132: 652-661.

Pullin, A. S (2002) Conservation Biology. Cambridge University Press, Cambridge.

Puth, L. M. and Wilson, K. A. (2001) Boundaries and corridors as a continuum of ecological flow control: lessons from rivers and . Conservation Biology 15(1): 21-30.

Queller, D. C., Strassman, J. E. and Hughes, C. R. (1993) Microsatellites and Kinship. Trends in Ecology and Evolution 8(8): 285-288.

Rainforest Conservation Society of Queensland (1986) Tropical rainforests of north Queensland: their conservation significance. Australian Government Publishing Service, Canberra.

Randi, E. (2000) Mitochondrial DNA. In: Molecular methods in ecology (Eds. Baker, A. J.) 136-167. Blackwell Science Ltd, Oxford.

Raymond, M. and Rousset, F. (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248-249.

Rice, W. R. (1989) Analyzing tables of statistical tests. Evolution 43: 223-225.

Ricketts, T. H. (2001) The matrix matters: effective isolation in fragmented landscapes. The American Naturalist 158(1): 87-99.

Rosenbaum, V. and Reissner, D. (1987) Temperature gradient gel electrophoresis: thermodynamic analysis of nucleic acids and proteins in purified form and in cellular extracts. Biophysical Chemistry 26: 235-246.

Rosenberg, D. K., Noon, B. R., Megahan, J. W. and Meslow, E. C. (1998) Compensatory behaviour of Ensatina eschscholtzii in biological corridors: a field experiment. Canadian Journal of Zoology 76: 117-133.

References - 165 -

Rousset, F. (1997) Genetic differentiation and estimation of gene flow from F- statistics under isolation by distance. Genetics 145: 1219-1228.

Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W. and Hanski, I. (1998) Inbreeding and extinction in a butterfly metapopulation. Nature 392: 491-494.

Saccone, C., Attimonelli, M. and Sbisa, E. (1987) Structural elements highly preserved during the evolution of the D-loop-containing region in vertebrate mitochondrial DNA. Journal of Molecular Evolution 26: 205-211.

Sarre, S. (1995) Mitochondrial DNA variation among populations of Oedura reticulata (Gekkonidae) in remnant vegetation: Implications for metapopulation structure and population decline. Molecular Ecology 4(4): 395-405.

Saunders, D. A. and Hobbs, R. J. (1991) The role of corridors in conservation: what do we know and where do we go? In: Nature Conservation 2: The role of corridors (Eds. Saunders, D. A. and Hobbs, R. J.) 421-427. Surrey Beatty and Sons, Sydney.

Saunders, D. A., Hobbs, R. J. and Arnold, G. W. (1993) The Kellerberrin project in fragmented landscapes: A review of current information. Biological Conservation 64: 185-192.

Saunders, D. A, Hobbs, R. J. and Margules, C. R. (1991) Biological consequences of ecosystem fragmentation: A review. Conservation Biology 5(1): 18-32.

Saunders, D. A. and Ingram, J. A. (1987) Factors affecting the survival of breeding populations of Carnaby's cockatoo Calyptohynchus funereus latirostris in remnants of native vegetation. In: Nature Conservation: the role of remnants of native vegetation (Eds. Saunders, D. A., Arnold, G. W., Burbridge, A. A., and Hopkins, A. J. M.) 249-258. Surrey Beatty, Chipping Norton, NSW.

Schneider, S., Roessli, D. and Excoffier, L. (2000) Arlequin ver. 2.000: A software for population genetics data analysis. Genetics and Biometry Laboratory, University of Geneva, Switzerland.

Selander, R. K. (1970) Behavior and genetic variation in natural populations. American Zoologist 10: 53-66. References - 166 -

Sherwin, W. B. and Moritz, C. (2000) Managing and monitoring . In: Genetics, Demography and Viability of Fragmented Populations (Eds. Young, A. G. and Clarke, G. M.) 9-34. Cambridge University Press, Cambridge.

Sieving, K. E., Willson, M. F. and De Santo, T. L. (2000) Defining corridor functions for endemic birds in fragmented south-temperate rainforest. Conservation Biology 14(4): 1120-1132.

Simberloff, D. and Cox, J. (1987) Consequences and costs of conservation corridors. Conservation Biology 1: 63-71.

Simberloff, D., Farr, J. A., Cox, J. and Mehlman, D. W. (1992) Movement corridors: conservation bargains or poor investments? Conservation Biology 6(4): 493- 504.

Simonsen, B. T., Siegismund, H. R. and Arctander, P. (1998) Population structure of African buffalo inferred from mtDNA sequences and microsatellite loci: high variation but low differentiation. Molecular Ecology 7(2): 225-238.

Slatkin, M. (1985) Gene flow in natural populations. Annual Review of Ecology and Systematics 16: 393-340.

Slatkin, M. (1987) Gene flow and the geographic structure of natural populations. Science 236: 787-792.

Slatkin, M. (1994) Gene flow and population structure. In: Ecological genetics (Eds. Real, L. A.) 3-17. Princeton University Press, Princeton.

Smith, G. C. (1985) Biology and habitat usage of sympatric populations of the fawn footed melomys (Melomys cervinipes) and the grassland melomys (M. burtoni) (Rondentia: Muridae). The Australian Zoologist 21(7): 551-563.

Soulé, M. E. (1991) Conservation corridors: countering habitat fragmentation. Theory and strategy. In: Landscape linkages and biodiversity (Eds. Hudson, W. E) 91-105. Island Press, Washington.

Soulé, M. E. and Gilpin, M. E. (1991) The theory of wildlife corridor capability. In: Nature Conservation 2: The role of corridors (Eds. D.A. Saunders and R.J. Hobbs) 3-8. Surrey Beatty and Sons Pty Ltd, Sydney. References - 167 -

Soulé, M. E. and Mills, L. S. (1998) No need to isolate genetics. Science 282: 1658-1659.

Soulé, M. E. and Simberloff, D. (1986) What do genetics and ecology tell us about the design of nature reserves? Biological Conservation 35: 19-40.

Southwood, T. R. E. and Henderson, P. A. (2000) Ecological Methods. Blackwell Science Ltd., Oxford.

Stenseth, N. C. and Lidicker Jr., W. Z. (1992) The study of dispersal: a conceptual guide. In: Animal dispersal: small mammals as a model (Eds. Stenseth, N. C. and Lidicker Jr., W. Z.) 5-18. Chapman and Hall, London.

Storz, J. F. (1999) Genetic consequences of mammalian social structure. Journal of Mammalogy 80(2): 553-569.

Strahan, R. (1983) Complete book of Australian mammals. Angus and Robertson Publishers, Sydney.

Suckling, G. C. (1984) Population ecology of the sugar glider, Petaurus breviceps, in a system of fragmented habitats. Australian Journal of Wildlife Research 11: 49-75.

Sugg, D. W, Chesser, R. K., Dobson, F. S. and Hoogland, J. L. (1996) Population genetics meets behavioural ecology. Trends in Ecology and Evolution 11(8): 338-342.

Sumner, J., Jessop, T., Paetkau, D. and Moritz, C. (2004) Limited effect of anthropogenic habitat fragmentation on molecular diversity in a rain forest skink, Gnypetoscincus queenslandiae. Molecular Ecology 13(2): 259-269.

Sunnucks, P. (2000) Efficient gene markers for population biology. Trends in Ecology and Evolution 15(5): 199-203.

Sutcliffe, O. L. and Thomas, C. D. (1996) Open corridors appear to facilitate dispersal by ringlet butterflies (Aphantopus hyperantus) between woodland clearings. Conservation Biology 10(5): 1359-1365.

Tajima, F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585-595. References - 168 -

Tallmon, D. A., Draheim, H. M., Mills, L. S. and Allendorf, F. W. (2002) Insights into recently fragmented vole populations from combined genetic and demographic data. Molecular Ecology 11: 699-709.

Taylor, A. (1999) Genetic elucidation of mammalian population processes in fragmented habitats: the greater glider at Tumut. Australian Biologist 12(2): 154.

Templeton, A. R., Shaw, K., Routman, E. and Davis, S. K. (1990) The genetic consequences of habitat fragmentation. Annals of Missouri Botanical Garden 77: 13-27.

Tewes, M. E. (1994) Habitat connectivity: its importance to ocelot management and conservation. In: Conservation corridors in the Central American region (Ed. Vega, A.) 290-296. Tropical Research and Development Inc., Florida.

Tewksbury, J. J., Levey, D. J., Haddad, N. M., Sargent, S., Orrock, J. L., Weldon, A., Danielson, B. J., Brinkerhoff, J., Damschen, E. I. and Townsend, P. (2002) Corridors affect , animals, and their interaction in fragmented landscapes. Proceedings of the National Academy of Science 99(20): 12923- 12926.

Thompson, J. D., Higgins, D. G. and Gibson, T. J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673-4680.

Tigas, L. A., Van Vuren, D. H. and Sauvajot, R. M. (2002) Behavioural responses of bobcats and coyotes to habitat fragmentation and corridors in an urban environment. Biological Conservation 108: 299-306.

Tracey, J. G. (1982) The vegetation of the humid tropical region of North Queensland. CSIRO, Melbourne.

Tracey, J.G. and Webb, L.J. (1975) Vegetation of the humid tropical region of North Queensland. CSIRO, Indooroopilly.

References - 169 -

Travis, S. E., Slobodchikoff, C. N. and Keim, P. (1995) Ecological and demographic effects on intraspecific variation in the social system of prairie dogs. Ecology 76: 1794-1803.

Tucker, N. I. J. (2000) Linkage restoration: Interpreting fragmentation theory for the design of a rainforest linkage in the humid Wet Tropics of north-eastern Queensland. Ecological Management and Restoration 1(1): 35-41.

Turner, M. G., Dale, V. H. and Gardner, R. H. (1989) Predicting across scales: theory development and testing. Landscape Ecology 3: 245-252.

Usher, M. B. (1987) Effects of fragmentation on communities and populations: A review with applications to wildlife conservation. In: Nature conservation: The role of remnants of native vegetation (Eds. Saunders, D. A., Arnold, G. W., Burbidge, A. A., and Hopkins, A. J. M) 103-121. Surrey Beatty and Sons Pty Ltd, Sydney.

Várkonyi, G., Kuussaari, M. and Lappalainen, H. (2003) Use of forest corridors by boreal Xestia moths. Oecologia 137: 466-474.

Vernes, K. (2003) Fine-scale habitat preferences and habitat partitioning by three mycophagous mammals in tropical wet sclerophyll forest, north-eastern Australia. Austral Ecology 28(5): 471-479.

Vuilleumier, S. and Prélaz-Droux, R. (2002) Map of ecological networks for landscape planning. Landscape and Urban Planning 58:157-170.

Wang, M. and Schreiber, A. (2001) The impact of habitat fragmentation and social structure on the population genetics of roe deer (Capreolus capreolus L.) in Central Europe. Heredity 86: 703-715.

Waser, P. M. and Strobeck, C. (1998) Genetic signatures of interpopulation dispersal. Trends in Ecology and Evolution 13(2): 43-44.

Watts, C. H. S. and Aslin, H. J. (1981) The rodents of Australia. Angus and Robertson, Sydney.

Watts, C. H. S. and Kemper, C. M. (1989) Muridae. In: Fauna of Australia: Volume 1B, Mammalia. (Eds. Walton, D. W. and Richardson, B. J.) 939-956. Australian Government Publishing Service, Canberra. References - 170 -

Webb, L. J. and Tracey, J. G. (1981) Australian rainforests: patterns and change. In: Ecological Biogeography of Australia (Eds. Keast, A.) 605-694. Junk, The Hague.

Webb, N. R. (1997) The development of criteria for ecological restoration. In: and sustainable development (Eds. Urbanska, K. M., Webb, N. R., and Edwards, P. J.) 133-158. Cambridge University Press, Cambridge.

Wegner, J. F. and Merriam, G. (1979) Movements by birds and small mammals between a wood and adjoining farmland. Journal of Applied Ecology 16: 349- 357.

Weir, B. S. and Cockerham, C. C. (1984) Estimating F-statistics for the analysis of population structure. Evolution 38(6): 1358-1370.

Wellesley-Whitehouse, H. L. (1981) A study on the comparative ecologies of two rainforest murid rodents, Uromys caudimaculatus and Rattus fuscipes coriaius in North Queensland. Honours thesis, Department of Zoology, James Cook University, Townsville,

Wet Tropics Management Authority (2003) Draft Wet Tropics Conservation Strategy. Wet Tropics Management Authority, Cairns.

Whitlock, M. C. and McCaughley, D. E. (1999) Indirect measures of gene flow and

migration: FST =/1/(4Nm+1). Heredity 82: 117-125.

Wiens, J. A. (1989) Spatial scaling in ecology. Functional Ecology 3: 385-397.

Wiens, J. A. (1997) Metapopulation dynamics and landscape ecology. In: Metapopulation biology (Eds. Hanski, I. A. and Gilpin, M. E.) 43-61. Academic Press, San Diego.

Wiens, J. A. and Milne, B. T. (1989) Scaling of "landscapes" in landscape ecology, or, landscape ecology from a beetle's perspective. Landscape Ecology 3(2): 87-96.

Wiens, J. A., Stenseth, N. C., Van Horne, B. and Ims, R. A. (1993) Ecological mechanisms and landscape ecology. Oikos 66: 369-380. References - 171 -

Wilcove, D. S., McLellan, C. H. and Dobson, A. P. (1986) Habitat fragmentation in the temperate zone. In: Conservation biology, the science of scarcity and diversity (Eds. Soule, M. E.) 237-256. Sinauer Associates, Sunderland.

Williams, B., Brawn, J. D. and Paige, K. N. (2003) Landscape scale genetic effects of habitat fragmentation on a high gene flow species: Speyeria idalia (Nymphalidae). Molecular Ecology 12(1): 11-20.

Williams, K. A. W. (1999) Native plants of Queensland . K.A.W. Williams, North Ipswich, Queensland.

Williams, S. E., Pearson, R. G. and Walsh, P. J. (1996) Distributions and biodiversity of the terrestrial vertebrates of Australia's Wet Tropics: a review of current knowledge. Pacific Conservation Biology 2: 327-362.

Wilson, A.-M. and Lindenmayer, D.B.L. (1996a) Wildlife corridors - their potential role in the conservation of biodiversity in rural Australia. Australian Journal of Soil and Water Conservation 9(2): 22-28

Wilson, A.-M. and Lindenmayer, D.B.L. (1996b) Wildlife corridors and the conservation of biodiversity: A review. Greening Australia Ltd. Canberra.

Wilson, E. O. and Willis, E. O. (1975) Applied biogeography . In: Ecology and evolution of communities (Eds. Cody, M. L. and Diamond, J. M.) 522-534. Belknap Press, Cambridge.

Winter, J. W. (1991) Northeastern Queensland: some conservation issues highlighted by forest mammals. In: Conservation of Australia's forest fauna (Eds. Lunney, D.) 113-118. Royal Zoological Society of NSW, Mosman.

Winter, J. W. (1997) Responses of non-volant mammals to Late Quaternary climatic changes in the wet tropics region of north-eastern Australia. Wildlife Research 24(5): 493-512.

Winter, J. W., Bell, F. C., Pahl, L. I. and Atherton, R. G. (1987) Rainforest clearfelling in northeastern Australia. Proceedings Royal Society Queensland 98: 41-57.

Wolda, H. (1981) Similarity indices, sample size and diversity. Oecologia 50: 296- 302. References - 172 -

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

Wright, S. (1978) Evolution and the genetics of populations, Vol. 4. Variability within and among natural populations. University of Chicago Press, Chicago.

Young, A. G. and Clark, G. M. (2000) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge.