Xenicus Gilviventris)

Conservation Genetics of Alpine Rock Wren (Xenicus gilviventris)

Kerry Anne Weston

A thesis submitted for the Degree of Doctor of Philosophy at the University of Otago Dunedin, New Zealand

February 2014

Abstract

Many species occur in naturally subdivided populations due to spatial heterogeneity of the landscape. Such a pattern is especially evident in alpine species, where naturally fragmented habitat forms an ‘alpine archipelago’. High altitude habitat patches and the species they harbour can serve as effective models for monitoring global change processes in sensitive ecosystems. The rock wren (Xenicus gilviventris) is a threatened alpine passerine belonging to the endemic New Zealand wren family (Acanthisittidae). This ancient family was once represented by at least seven species, however due to the impacts of introduced mammalian predators, only two species remain. Conservation management of rock wren has only recently commenced via the translocation of individuals to offshore islands, but genetic considerations are not currently a part of management practices. In this thesis, I investigated the role of genetic factors in the conservation management of rock wren and applied my findings to improve understanding of the species’ ecology and better inform future management efforts. I sampled rock wren (n=221) from throughout their range and using 14 microsatellite markers combined with nuclear and mitochondrial DNA sequence data, describe significant differences in genetic variation and differentiation between rock wren populations across the South Island. A deep North–South genetic divergence was evident (3.7 ± 0.5% at cytochrome b), consistent with the ‘biotic gap’ hypothesis whereby Northern and Southern populations became restricted in ice-free refugia during the Pleistocene era of extensive glaciation c. 2 mya. There is some evidence for a larger refugium within the south of the rock wren’s range, as estimates of genetic variation and long-term effective population size are consistently larger for the Southern lineage. Although this finding may also be indicative of more optimal habitat in the south of the species’ range; supporting a higher density of rock wren long-term. Designation of Northern and Southern rock wren lineages as separate evolutionarily significant units (ESUs) is proposed. Estimates of the long-term effective population size of rock wren are dramatically larger relative to contemporary estimates, indicating that in the past, rock wren sustained a much higher abundance than today. Whilst a genetic signature linking population decline within the Northern lineage to a timeframe of anthropogenic disturbance (i.e. the past c. 100 years) was not detected, there is some evidence for a

recent population bottleneck within this timeframe in the South. This suggests that although natural historical climate fluctuations have clearly played an important role driving patterns of rock wren abundance in the past, these impacts are now being compounded by much more recent anthropogenic impacts, most likely, predation by introduced mammalian predators. Significant fine-scale spatial genetic structure in rock wren was also detected, and a strong pattern of isolation by distance whereby genetic relatedness among neighbouring individuals is significantly greater than that among more distant or randomly located individuals. This pattern of gene flow is indicative of a stepping stone model of dispersal. A potential sex-bias in dispersal, suggestive of male natal philopatry, was also detected which may have further contributed to the strong pattern of fine-scale structuring. The spatial scale of positive genetic structure or ‘genetic patch size’ (i.e. the distance over which individuals were not genetically independent) was unexpectedly large (c.70 km) given the rock wren’s limited flight ability. Asymmetrical gene flow is also evident among populations within the Southern lineage, indicative that source-sink dynamics are operating. The Murchison Mountains appear to be a particularly important source of migrants for other populations. Therefore, management efforts, such as predator control, to ensure this population is conserved should be prioritised. Conversely, the Upper Hollyford and Lake Roe populations appear to be functioning as sink populations, with migration occurring into, but not out of these areas. By improving habitat quality in these areas (e.g. by controlling invasive species) there is potential that they may be converted from sinks into new source populations.

iii

Acknowledgements

From the outset, this PhD has been a team effort and without the support of many people, this project would not have been possible. Firstly, I would like to thank my supervisor Bruce Robertson for embracing the challenge of studying rock wren. Whilst others had been concerned that the logistics of an alpine study of this scale might be insurmountable, Bruce never seemed phased, always trusting that it was achievable and most importantly, realising rock wren were a species in need of scientific attention. So thank you Bruce, for the enthusiasm, confidence and guidance in getting this project off the ground and teaching me everything I needed to know around using genetics as tool for research. To my co-supervisor, Ian Jamieson, thank you for welcoming me into your threatened bird research group and providing me with support, structure and interactions with some excellent role models from which to learn the ropes of academia. Thanks also for being a calming force lurking in the background, always ready to step in should any minor crisis arise, providing moral support and wise words of perspective. Together, Bruce and Ian have been a great supervisory team, an excellent combination of skills and attributes. Thanks also to the other members of my supervisory committee, Sue Michelsen-Heath and Graham Wallis for providing valuable feedback along the way and always being enthusiastic about my work. This research could not have happened without the huge amount of support I received from those volunteering their time with me in the field. Some were there for a day and some for many months, but each and every person was vital for this project to succeed. Ian Clark, you stuck it out for so many months and through so many incredibly tough trips. Although we were often utterly exhausted, freezing, home sick, grumpy or overwhelmed, we had so many great moments out there and your enthusiasm was definitely what carried me through to the end of my final field trip. I hope you had a great time too. Thanks also to the Department of Conservation for all of their logistical support in accessing remote locations, hut use and assistance with sample collection. Thank you to the Bellringer family, especially Christine, for often looking after Lily whilst I went into the field and Ray, for giving me my training wheels in the mountains around Aoraki, fuelling my interest in rock wren and accompanying me on many rock wren catching trips as part of this project.

A special mention to those with whom I spent many hours in the laboratory, Fiona Robertson, Julia Barth and Sam Botting. Initially the genetics laboratory was a completely foreign environment to me, but with friendly faces, good conversation and helpful advice, it became something of a second home and a place where I could really relax and focus. Thanks to all of my special friends in the Zoology Department, especially Nic Dussex and Emily Weiser. Your companionship over the last couple of years was so important to me. Just having others who understood exactly how I felt and who were experiencing the same challenges made life so much better. Nic, you became a real pillar of support for me. Always there for coffee breaks and emergency discussions, you were great to bounce ideas off and a real inspiration. Emily, the endless games of scrabble, exciting weekend bike rides and lengthy online discussions on absolutely everything imaginable often presented the head space I needed from my thesis. Also, just knowing you were always there with your friendly smile and quiet supportive way was of great reassurance. Lily, you are the best thesis-writing companion I could ask for. Always keen to drag me out for a walk when the screen becomes a blur, but equally happy to snuggle in at my feet and snore through weeks of analysis and writing. Whilst you really weren’t much help in catching rock wren, you were great to come home to after every trip and I missed you lots in my time away. I hope now that I have more spare time we can go on many more missions together and kill possums. I’m so glad I found you the month before I started my PhD. Alex, thanks for understanding the importance of this project to me. Thanks for supporting my leave of absence to Dunedin for extended periods to undertake lab work and immerse myself in university life. Thanks for caring for Lily for all those weeks when I would run off to catch rock wren in the field and thank you for all the trips you accompanied me on, some of which constituted our Christmas/New Year holidays. Lastly, thanks for letting me fester on the couch for weeks on end writing up and for providing me with food and shelter when my scholarship payments ceased. Thanks also to my father, Barry Weston, who always seemed to know when to send an emergency support cheque accompanied by some kind words of encouragement. Hopefully now I’ll finally be able to come home and visit again. Also, a big thank you to my brother Shaun Weston, the GIS extraordinaire, for helping me

with all of my mapping and setting up systems so that I could easily visually monitor the progress of our field sampling throughout the season. Lastly, thank you to all of those who provided the much appreciated financial support for this project. The University of Otago Post Graduate Scholarship, JS Watson trust administered through Forest and Bird NZ, Brian Mason Scientific and Technical Trust, Macpac NZ and Mohua Charitable Trust.

Table of Contents

Abstract ...... ii

Acknowledgements ...... iv

Table of Contents ...... vii

Table of Figures ...... x

List of Tables ...... xi

Chapter 1 General Introduction...... 1

Alpine species management in the face of global change ...... 1 Genetic applications for the conservation of alpine species ...... 3 Molecular markers and their utility ...... 6 A brief history of factors shaping the New Zealand biota...... 8 New Zealand rock wren...... 9 Thesis aims and structure ...... 12

Chapter 2 Isolation and characterisation of 14 microsatellites for the New Zealand rock wren Xenicus gilviventris ...... 15

Introduction ...... 15 Materials and methods ...... 15 Results and discussion ...... 17

Chapter 3 Population structure within an alpine archipelago: a strong signature of past climate change within the New Zealand rock wren Xenicus gilviventris .... 21

Introduction ...... 21 Materials and methods ...... 24

Sampling and DNA extraction ...... 24 Mitochondrial DNA sequencing...... 27 Nuclear DNA sequencing ...... 28 Microsatellite genotyping ...... 29 Genetic variation and differentiation ...... 29 Phylogenetic analysis ...... 30 Molecular clock ...... 31

vii

Microsatellite cluster analyses ...... 32

Results ...... 32

Phylogenetic analysis ...... 33 Microsatellite cluster analyses ...... 41 Genetic variation and differentiation ...... 44

Discussion ...... 47

Phylogeographic structure ...... 47 Glacial refugia with subsequent expansion...... 49 Implications for management of a threatened species ...... 50

Chapter 4 Past demography and effective population size of the New Zealand alpine rock wren (Xenicus gilviventris) ...... 55

Introduction ...... 55 Materials and methods ...... 57

Microsatellite genotyping ...... 57 Mitochondrial DNA sequencing ...... 57 Nuclear DNA sequencing ...... 58 Data analysis ...... 58 Past demography ...... 59 Effective population size...... 61

Results ...... 63

Past demography ...... 63 Effective population size...... 66

Discussion ...... 69

Past demography ...... 69 Effective population size...... 70 Conservation implications ...... 71

Chapter 5 Identifying populations for management: fine-scale population structure in the New Zealand alpine rock wren (Xenicus gilviventris) ...... 75

Introduction ...... 75 Materials and methods ...... 77

viii

Sampling and DNA extraction ...... 77 DNA sexing ...... 79 Microsatellite genotyping ...... 79 Genetic variation within populations...... 79 Migration between populations ...... 80 Spatial structure ...... 80 Tests for sex-biased dispersal ...... 82

Results ...... 83

Genetic variation within populations...... 83 Migration between populations ...... 84 Spatial structure ...... 85 Sex-biased dispersal ...... 89

Discussion...... 90

Population-based inferences and spatial structure ...... 90 Management implications ...... 92

Chapter 6 General Discussion ...... 97

Summary of major findings ...... 97 Implications for conservation management of rock wren ...... 100

Northern and Southern rock wren lineages ...... 100 Rock wren management units ...... 101 Genetics and rock wren translocations ...... 102

Future research ...... 104

References ...... 109 Appendix One ...... 129 Appendix Two ...... 130 Appendix Three ...... 131 Appendix Four ...... 134 Appendix Five ...... 137

Table of Figures

Figure 3.1 Map of sampling locations throughout South Island ...... 26 Figure 3.2 Bayesian phylogeny using control region 1st domain...... 34 Figure 3.3 Bayesian phylogeny using cytochrome b with molecular clock ...... 35 Figure 3.4 Bayesian phylogeny using nuclear beta-fibrinogen intron-7 ...... 36 Figure 3.5 Geographic distribution of mtDNA 1st domain control region haplotypes .. 37 Figure 3.6 Maximum parsimony haplotype networks ...... 40 Figure 3.7 Proportional membership to genetic clusters ...... 42 Figure 3.8 Determination of K, the number of genetic clusters ...... 43

Figure 3.9 Isolation by distance using pairwise FST ...... 46 Figure 4.1 Mismatch distributions ...... 65 Figure 4.2 Long-term effective population size estimates ...... 68 Figure 5.1 Map of sampling locations in lower South Island ...... 78 Figure 5.2 Isolation by distance in lower South Island using pairwise relatedness ...... 85 Figure 5.3 Correlogram showing auto correlation across distance classes ...... 86 Figure 5.4 Correlogram showing auto correlation across distance classes for male and female rock wren...... 86 Figure 5.5 Mean pairwise genetic relatedness estimates for populations in the lower South Island ...... 87 Figure 5.6 Determination of K, the number of genetic clusters in lower South Island ...... 88 Figure 5.7 Map of estimated membership to genetic clusters ...... 89

List of Tables

Table 2.1 Characteristics of microsatellite loci ...... 18 Table 3.1 Model of sequence evolution for each locus ...... 30 Table 3.2 Genetic variation in Northern and Southern populations ...... 45

Table 3.3 Pairwise FST values among sampling areas ...... 45 Table 3.4 Genetic variation in threatened New Zealand birds ...... 51 Table 4.1 Indices of population expansion ...... 64 Table 4.2 Contemporary effective population size estimates ...... 66 Table 4.3 Long-term effective population size estimates using microsatellite data...... 67 Table 5.1 Genetic variation within lower South Island populations ...... 84 Table 5.2 Migration rates among lower South Island populations ...... 84 Table 5.3 Tests for sex-biased dispersal ...... 89

Lake Adelaide, Fiordland

Peace for the heart life in the mountains (Santoka)

xii

Chapter 1 General Introduction

The need for effective conservation measures is critical now more than ever in an era of unprecedented species loss driven by human-induced global change (Barnosky et al. 2011). This is particularly true for habitat specialists (Harcourt et al. 2002; Warren et al. 2001), which have long been recognised to be at higher risk of population decline and extinction due to drivers of environmental change such as ongoing habitat loss, fragmentation, the introduction of invasive species and climate change (Brook et al. 2008; Purvis et al. 2000). Alpine habitat specialists present additional challenges for conservation management because the ecosystems within which they live are especially fragile and are already naturally fragmented, with populations typically restricted within small and isolated high altitude patches (Bech et al. 2009; Fedy et al. 2008; Floyd et al. 2005; Segelbacher & Storch 2002; Sekar & Karanth 2013). In this thesis, I investigate genetic factors in the conservation management of the New Zealand alpine rock wren (Xenicus gilviventris) and apply my findings to improve understanding of the ecology of this threatened species and better inform future management strategies.

Alpine species management in the face of global change

Rapidly changing atmospheric conditions and global warming associated with human- induced increases in CO2 levels pose the greatest challenge for alpine species management over the next century, with many predicting large-scale range reductions and changes in alpine biodiversity (Moritz & Agudo 2013; Sala et al. 2000; Yoccoz et al. 2010). There are several ways that species can buffer the effects of climate change; the simplest and most widely recognised response is a geographical shift in range to track a climatic niche (Moritz & Agudo 2013). Consequently, fragmented populations are most susceptible to the effects of anthropogenic climate change due to the constrained opportunity for dispersal and range shifts. Increasing or maintaining connectivity has emerged as the most prevalent conservation strategy to combat the effects of climate change (Heller & Zavaleta 2009; Hodgson et al. 2009). A variety of recommendations have been associated with this strategy such as designing habitat corridors, removing barriers for dispersal, locating reserves within close proximity of each other and habitat restoration such as reforestation (Heller & Zavaleta 2009). Assisted colonisation of

Chapter One: General Introduction individuals from warm-adapted populations into cooler higher latitude and altitude areas has also been proposed to increase the probability of adaptation to changing climate (Hoffmann & Sgró 2011; Loss et al. 2011; McLachlan et al. 2007). Many high- altitude alpine and montane species face range contraction with upward shifts in elevational ranges, and for some, the complete loss of their climatic niche altogether (Chamberlain et al. 2013; Engler et al. 2011; Forister et al. 2010; Pounds et al. 1999; Rubidge et al. 2012; Sekercioglu et al. 2008). Alpine species may serve as effective models for monitoring global change processes in sensitive ecosystems (Ditto & Frey 2007; McDonald & Brown 1992; Pauli et al. 2003; Pounds et al. 1999). However, studies encompassing the entire range of a species, or at least the lower and upper extremes are few and have been focused predominantly on amphibians (Pounds et al. 2006; Pounds et al. 1999), a mammal (Beever et al. 2003; Beever et al. 2011) and butterflies (Parmesan 1996, 2006). The scarcity of whole-range studies is likely attributed to difficulties collecting data at the scale of a species’ range, particularly when working with often patchily-distributed, widespread species at high elevations (Parmesan 2006). Despite these difficulties, some good examples of general upward movement in elevation do exist; such as that of the American Pika (Ochotona princeps), a small talus -dwelling alpine mammal, which has been documented to be moving upslope at an average rate of at least 145 m per decade

(Beever et al. 2011). In addition to geographical shift in range to track the climatic niche associated with higher elevations, resilience to climate change in situ can be also maximised by managing the adaptive potential of species. In a rapidly changing environment, populations that have a diverse gene pool upon which natural selection can act have more potential to respond adaptively to changing conditions (Bakker et al. 2010; Frankham 2005; Lande & Shannon 1996). This may involve maintaining levels of genetic variation and large effective populations sizes (>1000 breeding individuals) or facilitating gene flow between populations by increased management within the surrounding matrix (Heller & Zavaleta 2009; Hoffmann & Sgró 2011). Synergistic interactions among climate and other drivers of global change represent one of the largest areas of uncertainty in projections of change to alpine ecosystems (Sala et al. 2000; Walther et al. 2002). For example, human land use changes and the expansion of disease, novel competitors and predators into higher elevations present additional threats to alpine species isolated atop alpine refugia and 2

Chapter One: General Introduction may lead to even greater isolation for already naturally fragmented alpine populations (Brook et al. 2008; Cox et al. 2013; Hughes 2003; Pounds et al. 2006). Populations of capercaillie (Tetrao urogallus) in mountain ranges across Europe have become increasingly isolated due to anthropogenic habitat destruction in valleys, and the encroachment of farmland further upslope (Segelbacher et al. 2003; Segelbacher & Storch 2002). Long-term monitoring of species composition over elevational gradients in Costa Rica has revealed that lowland birds have begun expanding upslope into montane habitat to breed over the past 20 years, potentially displacing montane birds to even higher elevations (Parmesan 2006; Pounds et al. 1999). There is also strong evidence that increasing temperatures are associated with expanding mosquito-borne diseases in the highlands of Asia, East Africa and Latin America (Epstein et al. 1998). In Hawaii, avian malaria has had devastating effects on native avifauna and it is now predicted that warming temperatures will enable mosquito vectors to invade into the few remaining mountain refugia of several endangered species (Benning et al. 2002; Freed et al. 2005; Van Riper Iii et al. 1986). Vectors of avian malaria in New Zealand are known to be extending in range, signalling a real risk of disease emergence within alpine ecosystems with further climatic warming (Tompkins & Gleeson 2006). Lastly, whilst firm data are currently lacking for shifts in mammalian predators to higher elevations, it seems likely that this is occurring, given the observed elevational range shifts across a wide range of taxa, including small mammalian species (Beever et al. 2011; Parmesan 2006).

Genetic applications for the conservation of alpine species

Obtaining data at the level of sampling resolution required to understand the connectivity, dispersal characteristics and population structure of a widespread alpine species is logistically impractical using direct ecological methods such as observing movements of marked individuals over long time periods (Crochet 1996). Genetics presents an indirect approach that can make a valuable contribution towards understanding the ecology of a species, whilst characterising responses to environmental change (Frankham et al. 2002). Population subdivision via habitat fragmentation for example, typically leads to increased genetic structuring as genetic variation is essentially shifted from within to between subpopulations (Keyghobadi 2007; Segelbacher et al. 2003). Gene flow,

Chapter One: General Introduction together with the accumulation of random mutations, maintains genetic variation within populations by counteracting the evolutionary force of ‘genetic drift’, the process whereby alleles become fixed or lost from a population through the random fluctuation in allele frequencies between generations (Masel 2011). The strength of genetic drift is strongly influenced by population size, with small populations being much more susceptible (Franklin 1980). Genetic data can be effectively utilised to establish a baseline from which to index future changes in connectivity (Broquet & Petit 2009; Lowe & Allendorf 2010). For example, Bech et al. (2013; 2009) found that Rock Ptarmigan (Lagopus muta) populations in the French Pyrenees Mountains have become increasingly fragmented in “sky islands” in association with climate change and anthropogenic disturbance to their alpine habitat during the Holocene. Translocations of Rock Ptarmigan from more genetically diverse populations to genetically isolated and depauperate ones was proposed as a measure to boost genetic variation within isolated fragments (Bech et al. 2009). This technique is commonly known as ‘genetic rescue’ (Ingvarsson 2001). Such augmentations may only need to be at rates as low as one successfully reproducing migrant per generation (Mills & Allendorf 1996; Wang 2004), though this number may need to be >10 where fluctuations in population size are pronounced (Vucetich & Waite 2000). Even if large and genetically variable populations do not exist, which may be the case with many threatened species (Jamieson et al. 2008); individuals from inbred but genetically differentiated populations may be used to increase genetic variation (Heber et al. 2013). Further, it has also been recommended that the number of individuals introduced into recipient populations should not exceed approximately 20% gene flow within the first generation of translocation from the source population(s) to avoid the loss of locally adapted alleles from within the recipient population (Hedrick 1995). It is also important to have a good knowledge of the genetic variation within a species and how it is distributed when establishing new populations via translocations (Weeks et al. 2011).Whilst it is widely recognised that introducing a large number of founders is important (Tracy et al. 2011), maximising genetic variation within the founder population by sourcing individuals from genetically divergent populations may in some situations, be beneficial and assist species in persisting in the face of environmental change (Biebach & Keller 2012; Binks et al. 2007; Kolbe et al. 2007). However, the risks and benefits of these approaches need to be carefully considered, especially where populations have been isolated for a long time (e.g. exchanged no 4

Chapter One: General Introduction genes in the last 500 years, see Frankham et al. 2011), have fixed chromosomal differences or occupy different environments due to the risks of ‘outbreeding depression’ if specialised localised adaptations have evolved (Frankham et al. 2011; Tallmon et al. 2004; Weeks et al. 2011).

Robust estimates of population census size (Nc), i.e. the number of adults in a study area or population, are difficult to achieve for most organisms, as traditional methods involve marking and recapturing animals in studies commonly referred to as ‘mark-recapture’ (Luikart et al. 2010). Even more difficult to estimate through the collection of demographic data is the genetically ‘effective’ population size (Ne) (Waples 2005). All negative genetic effects of small population size are determined by

Ne rather than the population census size (Frankham 1995; Frankham et al. 2002). The parameter Ne can be defined as the size of an ‘idealized’ population that would give rise to the same calculated loss of heterozygosity, inbreeding or variance in allele frequencies as the population under consideration (Wright 1931). Long-term effective population size is typically much lower than census size, with an estimated ratio of

Ne/Nc of just 0.1 (Frankham 1995; but see Palstra & Ruzzante 2008). Factors contributing to this low Ne/Nc ratio include large variations in individual reproductive success, life history traits such as age at maturity and life span, unequal sex-ratios and fluctuations in population size over generations whereby the smallest population size reached continues to dominate genetic processes (Frankham 1995; Hare et al. 2011; Luikart et al. 2010; Waples et al. 2013). Therefore, the effective size of a population represents not only current census size, but also the history of the population. Contemporary genetic data can also be used to estimate the historical or ‘long term’ effective population size, thereby providing insight into the demographic history of populations (Luikart et al. 2010; Palsbøll et al. 2013). Gaining this knowledge is valuable in conservation management as it enables managers to infer whether the current abundance and distribution of populations reflects historical status, in the absence of long-term monitoring data or historic samples. Some of the most well- known examples of these approaches have been applied to obtaining pre-whaling abundance estimates for whale populations in the North Atlantic and Pacific oceans (Alter et al. 2007; Roman & Palumbi 2003). Genetic approaches be utilized to reveal relationships between the timing of population divergence or genetic ‘bottleneck’ events (Amos & Harwood 1998) and known historic climatic events such as glaciations (Hewitt 2004). Understanding the 5

Chapter One: General Introduction effects of historic climatic oscillations on genetic variation contributes to more informed predictions regarding the response of species to future climatic changes (Hoelzel 2010). Resolving taxonomic uncertainties using genetics assists in establishing priorities for conservation (Haig et al. 2011). For example, recent taxonomic resolution of the New Zealand storm-petrel (Fregetta maoriana) enabled conservation status of this species to be changed from data deficient to critically endangered and thus see its recovery prioritised (Robertson et al. 2011). Several already endangered species have also been revealed as cryptic species complexes (two or more distinct species mis- classified as a single species), making each species even more rare and range restricted than previously considered and requiring different conservation strategies (Bickford et al. 2007). Genetic data also play an important role in delineating conservation units within species. These are essentially the population units identified within species used to help guide conservation management strategy (Funk et al. 2012). The two most commonly referred to conservation units are ‘evolutionarily significant units’ (ESUs) (Crandall et al. 2000; Fraser & Bernatchez 2001; Moritz 1994; Ryder 1986; Waples 1991) and ‘management units’ (MUs) (Moritz 1994; Palsbøll et al. 2007). Whilst there are many definitions of both ESUs and MUs (reviewed in Funk et al. 2012), an ESU can be generally defined as “a population or group of populations that warrant separate management or priority for conservation because of high genetic and ecological distinctiveness” whilst MUs are typically much smaller and constitute a population that is demographically independent; there may be many MUs within a single ESU (Funk et al. 2012). There is no clear criterion for assigning either ESU or MU status to populations and although some advocate that MUs at least, should be based upon a predefined level of population genetic divergence (Palsbøll et al. 2007); perhaps the best approach at this stage is to conduct a range of different population genetic assessments using different classes of molecular markers (Funk et al. 2012).

Molecular markers and their utility

There are several types of molecular markers commonly utilized in conservation genetics that serve as a proxy of genome-wide variation (Arif & Khan 2009; Avise 1994). Different markers have different characteristics, making them either more or less

Chapter One: General Introduction suitable for addressing the diverse range of issues relevant to conservation biology (Wan et al. 2004). Further, no single molecular marker can be regarded as optimal for all applications and the combined use of different marker types is necessary to resolve historical, demographic and selective processes occurring at different spatial and temporal scales (Sunnucks 2000). Neutral markers have limited utility for quantifying adaptive divergence. Thus, where genomic information or tools are available, these can be used to complement neutral loci (Allendorf et al. 2010). This approach will be particularly useful if significant adaptive differentiation is expected among populations. For example, where significant environmental variation occurs across a species’ range, migration rates are low (so that gene flow does not constrain adaptive divergence) or where effective population sizes are large and genetic drift is overpowered by selection (Funk et al. 2012). Protein coding DNA sequences from the mitochondrial genome (mtDNA), such as from the cytochrome b gene are also commonly used as markers in conservation genetics for resolving taxonomic uncertainties, by clarifying evolutionary relationships within and among taxa (Avise 1987). The non-coding mtDNA control region (otherwise referred to as the d-loop) is typically considered to be most appropriate for determining the current status of population structure within species studies due it’s relatively faster rate of evolution (Wan et al. 2004). MtDNA is also particularly useful due to the high number of copies throughout the cell, maternal inheritance and haploidy

(Bernt et al. 2013). Microsatellite markers detect typically highly variable, short sequences of nuclear DNA comprised of 1–6 base pair motifs, tandemly-repeated at a single locus on a chromosome (Li et al. 2002). They are assumed to be distributed randomly throughout an organism’s genome and selectively neutral (Schlötterer & Wiehe 1999), though these assumptions may not always be met (Bachtrog et al. 1999; Gemayel et al. 2010; King et al. 1997; Li et al. 2002). Due to high levels of variability, microsatellites markers are particularly useful for assessing levels of relatedness between individuals, genetic variation within and among populations and genetic structure across finer spatial scales (Arif & Khan 2009).The recent development of high-throughput DNA sequencing such as the 454 sequencing platform (454 Life Sciences/Roche FLX) has greatly reduced the time and costs involved with microsatellite marker development and increased our ability to examine a sufficient number of variable markers (Kircher & Kelso 2010). 7

Chapter One: General Introduction

Nuclear sequence data also provide a valuable comparison to mtDNA due to biparental inheritance (Prychitko & Moore 2000; Zhang & Hewitt 2003), though as nuclear DNA (nDNA) is much slower evolving (c. 5 – 10 times the rate of nucleotide substitution) it is generally most useful in resolving relationships among taxa and providing a more historical perspective of evolutionary relationships (Hare 2001; Sunnucks 2000). Phylogenetic reconstruction using nDNA may be impeded by recombination or poor resolution resulting from low mutation rates (Hare 2001). The highest levels of genetic diversity in nDNA are typically found in non-coding regions of the locus known as ‘introns’ (Prychitko & Moore 1997).

A brief history of factors shaping the New Zealand biota

The island archipelago of New Zealand is part of the largely submerged continent of Zealandia, which became isolated from the Gondwanan supercontinent around 65 – 80 million years ago (ma) (Graham 2008). The distinctiveness of the New Zealand biota, characterised by a high proportion of flightless avifauna, an almost complete lack of mammals and widespread endemism, have long been associated with this prolonged isolation (Bellamy et al. 1990; Cooper & Millener 1993; Fleming 1962). However, during the Oligocene, Zealandia underwent almost complete marine submersion and there is uncertainty around whether habitable land persisted during this time (Cooper & Cooper 1995; Landis et al. 2008; Waters & Craw 2006). Given this debate and the sparseness of the fossil record pre-Pliocene for terrestrial fauna, the actual extent of modern lineages that are true Zealandian elements isolated by vicariance, as opposed to much more recent post-Oligocene dispersers, remains largely unresolved (Giribet & Boyer 2010; Trewick & Gibb 2010; Wallis & Trewick 2009). Another driving force in the evolution of New Zealand’s biota was the intensive period of tectonic uplift experienced during the Pliocene 5-6 ma (Batt et al. 2000; Chamberlain & Poage 2000). Pliocene uplift initiated the formation of extensive alpine habitat in New Zealand, including the Southern Alps, the major mountain chain running along the South Island of New Zealand with 19 peaks exceeding 3000 m above sea level (Whitehouse & Pearce 1992). The formation of this new alpine habitat saw the formation of new alpine forms of fauna, particularly insects (Fleming 1975) and even birds such as kea (Nestor notabilis) and presumably, the rock wren (X. gilviventris) (Fleming 1975; Wallis & Trewick 2009). Radiations within a wide range of animal

Chapter One: General Introduction groups such as moa (Baker et al. 2005), skinks (Greaves et al. 2007; O'Neill et al. 2008), fish (Waters & Craw 2008; Waters & Wallis 2001) and insects (Buckley & Simon 2007; Chinn & Gemmell 2004; Trewick et al. 2000) have also been attributed to Pliocene uplift (reviewed in: Wallis & Trewick 2009). Since the late Pliocene to early Pleistocene (2.5 ma), several periods of glaciation have occurred forming vast networks of ice-fields and glaciers along the Southern Alps, most extensively in the central South Island (Newnham et al. 1999; Suggate 1990). Large areas of the South Island remained largely ice-free however – most extensively around Malborough and eastern – central Otago/ Southland. These non-glaciated regions represented valuable refugia for formerly widespread populations forced to retract with the advancing snow and ice (Wallis & Trewick 2009). Most recently, human settlement inflicted dramatic impacts upon New Zealand’s flora and fauna, commencing with colonisation by Polynesians c.700 yrs bp and compounded by the subsequent arrival of Europeans c.200 yrs bp (McGlone 1989; Worthy & Holdaway 2002). Huge reductions in species diversity and density occurred both due to direct exploitation and as a result of large scale habitat changes, including the introduction of several predatory mammal species (Duncan & Blackburn 2004). The majority of New Zealand’s land bird species persist as small, range constricted populations either on offshore islands or in remnants of highly fragmented habitat (Craig 1991; Jamieson et al. 2006). This quandary, combined with the fact that New Zealand now has more introduced mammalian predators than any other archipelago, is responsible for the enduring status of decline of most endemics (Towns et al. 2001).

New Zealand rock wren

The rock wren (X. gilviventris Pelzeln, 1867) is a small, predominantly insectivorous alpine passerine of limited flight, belonging to the endemic New Zealand wren family (Acanthisittidae) (Higgins et al. 2001). Early taxonomic work identified the wrens as a unique group amongst New Zealand’s avifauna in that they are the oldest and most primitive family in the order Passeriformes, placed in an infraorder of their own, the Acanthisittides (Sibley et al. 1982). Subsequent studies have suggested an even greater taxonomic distinctiveness for the wrens in identifying them as a sister group to all other passerines (Barker et al. 2002; Barker et al. 2004; Ericson et al. 2002; Hackett et al. 2008).

Chapter One: General Introduction

Fleming (1975) proposed that within the genus Xenicus, the rock wren was an early derivative of the forest-dwelling bush wren X. longipes, having split during the Ross Glaciation in the Early Quarternary (c. 2.5 ma) when climatic cooling precipitated the evolution of a new cold-adapted form. This proposition placed the rock wren as a relatively young member of the primitive NZ wren family. However, a recent analysis of NZ wren phylogeny based on morphological characters suggested a much deeper coalescence between bush wren and rock wren, dating back to the pre – Miocene (> 19 ma) (Worthy et al. 2010). The only attempt until present to clarify phylogenetic relationships among the New Zealand wrens suggests a 24 – 19 Ma divergence between the bush wren, rock wren and a third family member – the riflemen (Acanthisitta chloris), though this finding is based on molecular data extracted from one specimen of each species (Cooper 1994; Cooper & Cooper 1995). A distinct Southern Fiordland subspecies of rock wren, X. g. rineyi, was proposed in early 1950s based on behavioural (Riney 1953) and morphological characteristics (Falla 1953), though this has never been verified using molecular data. The rock wren is now one of only two remaining extant members of the New Zealand wren family, the other being the riflemen (Higgins et al. 2001). At least six other species, have become extinct, the five most recent extinctions, including that of the bush wren, attributed to destruction of indigenous forest habitat and the introduction of mammalian predators following human settlement in New Zealand (Michelsen- Heath 1989; Worthy et al. 2010). Rock wren are sexually monogamous, maintaining territories ranging in size from c. 0.6 – 4.2 hectares (around 1.4 hectares on average) which vary little between seasons (Michelsen-Heath 1989). Typical habitat comprises of low alpine and subalpine shrubland, herb fields, scree slopes, boulder fields and rocky bluffs (Michelsen 1982). The breeding cycle commences in October – November when nests are built together by the pair in small excavated holes or crevices. A clutch of 1 – 5 eggs are laid and hatch asynchronously after an incubation period of 19 – 20 days, the chicks then develop over a 24 day nestling period. Unlike their confamilials the riflemen, rock wren are solitary nesters exhibiting no cooperative assistance (Michelsen-Heath 1989). Nestlings fledge and become independent of their parents within 14 – 21 days. Juveniles establish territories and find mates by the end of the summer in which they fledge, ready to commence breeding the following spring (Michelsen-Heath 1989). Michelsen-Heath (1989) reported that 75% of 75 juveniles banded as nestlings within 10

Chapter One: General Introduction the Murchison Mountains, Fiordland remained within 500 m from their place of fledging; the others were not resighted. It is not clear whether there are differences in dispersal between the sexes. Records indicate that rock wrens live for at least 5 years, though between-season mortality appears to be quite high (Michelsen-Heath 1989). Rock wren constitute a widespread, naturally fragmented population, occurring in patches of alpine - subalpine habitat, between c. 920 m and 2900 m asl (mostly 1200-2400 m), throughout the length of the South Island (Higgins et al. 2001). Current population size is largely unknown, though estimates range from 2,500 - 10,000 mature individuals (Birdlife International, 2012). Results from recent population studies indicate that both the abundance and range of rock wren have contracted within the past 100 years (Michelsen-Heath & Gaze 2007). Michelsen-Heath & Gaze (2007) collected and collated >2000 rock wren sighting records from the period 1912 – 2005 and mapped past and present distribution. Whilst reported reductions in abundance were largely anecdotal, it is clear that since 1980 rock wren have not been recorded from many areas they were once observed and overall, present a more fragmented distribution throughout their range. Both the IUCN and the Department of Conservation recognise the threatened status of rock wren, classifying the species as Vulnerable and Nationally Endangered, respectively (Robertson et al. 2013). The most significant threat to rock wren is predation by introduced mammals such as house mice Mus musculus and stoats Mustela erminea. Nesting studies have revealed significant egg and chick mortality resulting from predation (Michelsen-Heath, 1989, Department of

Conservation unpublished data). Conservation efforts to date have involved localised trapping of predators by volunteer groups (Stocker et al. 2006) and translocations to predator – free offshore islands. In January 2005, the Department of Conservation relocated 24 individuals from the Murchison Mountains, Fiordland to predator-free Anchor Island in Dusky Sound, Fiordland (Willans & Weston 2005). However, follow-up monitoring of this translocation revealed that only three male birds survived (Weston 2006). Subsequently, between 2008 and 2011, 41 individuals were again translocated from the Murchison Mountains, this time to the larger predator-free Secretary Island in Doubtful Sound, Fiordland. Follow-up monitoring indicates that rock wren are now successfully breeding on Secretary Island (M. Willans pers. comm).

Chapter One: General Introduction

Thesis aims and structure

In this thesis, comprising of an introduction, four data chapters and a general discussion, I use molecular data obtained from contemporary DNA to investigate rock wren phylogeography, population structure and effective population size, and apply my findings to inform conservation efforts for this unique alpine passerine. The first data chapter, Chapter Two, describes the isolation and characterisation of 14 novel microsatellite DNA markers using next generation sequencing. In Chapter Three, I describe the population structure of rock wren throughout their range and measure genetic variation within and between populations to assess connectivity. I investigate the relationship between rock wren phylogeography and past geological and climatic impacts incorporating the use of a molecular clock and describe potential evolutionarily significant units. I also address the taxonomic uncertainty around a separate sub-species proposed during the 1950s. In Chapter Four, I estimate both contemporary and historical effective population size of rock wren and compare these estimates to ascertain whether rock wren were historically less or more abundant than they are today. I also test for recent genetic bottlenecks resulting from a potential rapid decline in population size due to the impacts of introduced mammalian predators. Effective population size between the rock wren lineages identified in Chapter Three is also compared and implications for management are discussed. Finally, in Chapter Five I use microsatellite data to describe migration rates and the spatial scale of non-random genetic structure, i.e. the ‘genetic patch size’ of rock wren. I also investigate source- sink dynamics to identify the most appropriate populations to target for management in the Fiordland region of the species’ range. Each data chapter is presented as a stand-alone scientific paper intended for publication (Chapter Two has already been accepted with Conservation Genetics Resources); as a consequence I have used the first person plural for these chapters as they will be co-authored. Some degree of repetition exists, particularly within the introduction sections, however references have been combined into a single list at the end of the thesis. Although each chapter in this thesis benefited from the input of my supervisors Bruce Robertson and Ian Jamieson, the work is primarily my own as I performed the analyses and drafted each manuscript. Both Ian and Bruce assisted with formalisation of my research questions and provided guidance with the overall structure of the thesis and interpretation of results. My primary supervisor, Bruce Robertson also

Chapter One: General Introduction assisted with the design of the molecular markers, advised on analysis methods and commented on drafts of each manuscript. Graham Wallis, as a member of my PhD committee provided feedback on the interpretation of my results for Chapter Three. I planned and carried out the majority of field work to obtain rock wren blood samples, though a huge number of volunteers helped me during my two field seasons, one of which, Ian Clark, assisted me for a period of six months throughout the South Island. Bruce Robertson and Sabrina Taylor assisted with sample collection in the Darran Mountains. Department of Conservation (DoC) staff, in particular Jo Monks, Colin O’Donnell and Dan Palmer also assisted with sample collection in the Darran Mountains and Haast Range populations.

Searching for rock wren Aoraki/Mount Cook National Park

Chapter 2 Isolation and characterisation of 14 microsatellites for the New Zealand rock wren Xenicus gilviventris1

Introduction

The rock wren (Xenicus gilviventris) is a threatened alpine passerine (Michelsen-Heath & Gaze 2007) belonging to the endemic New Zealand wren family (Acanthisittidae). This largely flightless family was once represented by at least seven species, however due to the impacts of introduced mammalian predators, only two species remain. Conservation management of rock wren has only recently commenced via translocation of individuals to offshore islands, but genetic considerations are not currently a part of management practices. Here we describe 14 microsatellite loci developed to examine population structure, dispersal and gene flow in X. gilviventris throughout their range. This information will be used to identify the most vulnerable and robust populations, delineate management units and inform conservation efforts for the species.

Materials and methods

Genomic DNA was extracted from whole blood from a single rock wren X. gilviventris using a standard 5% Chelex protocol. Five hundred nanograms of genomic DNA was nebulised and used to make a barcoded Rapid Library (Roche, Penzberg, Germany). The rock wren library had an average length of 750 – 800 bp and a concentration of 1.23 x 109 molecules/µl. Emulsion PCR (emPCR) was carried out at a ratio of four copies per bead, producing an average of 218,800 enriched DNA capture beads for sequencing. Of these, 197,500 enriched DNA capture beads were sequenced with two other libraries from different taxa on 1/4 of a 70 x75 PicoTiterPlate using the Titanium XLR70 sequencing system for the Roche 454 GS-FLX instrument. Following sequencing and quality filtering, a total of 52,586 sequenced fragments were obtained. Fragments were further filtered for duplicates and those less than 110 bp were removed,

1 A version of this chapter has been published as: Weston, K.A & Robertson, B.C. 2014. Isolation and characterisation of 14 microsatellites for the New Zealand rock wren Xenicus gilviventris. Conservation Genetics Resources 6, 115-117.

Chapter Two: Microsatellite loci leaving 44,822 sequences between 110 and 686 bp with an average read length of 393 bp. Sequences were screened for di- or trinucleotide microsatellites with a minimum of seven repeats (including combined arrays separated by <5 bp) using MSAT COMMANDER version 0.8.1 (Faircloth 2008). A total of 288 dinucleotide and 42 trinucleotide microsatellites were isolated, within 264 and 41 reads respectively. Thirty-three primer pairs were designed, twenty-seven dinucleotides and six trinucleotides. Forward primers were tagged with M13 sequence (5’- TGTAAAACGACGGCCAGT-3’) at the 5’ end to facilitate the use of universal fluorescent-labeled M13 primers as described in Schuelke (2000). Primers were initially characterised using DNA extracted from seven rock wren blood samples spanning the extent of their range. Loci were assigned to multiplex groups for PCR using MULTIPLEX MANAGER v2 (Holleley & Geerts 2009). Each 2µl PCR reaction contained: 1 µl (10-20 ng) of dried template DNA, 1 µl of Type-it master mix (QIAGEN), 0.08 pM (per loci) of each M13-labelled, locus-specific forward primer, 0.32 pM (per loci) of each locus-specific reverse primer and 0.96 pM of universal M13 primer labelled with fluorescent dye (6FAM, NED, PET or VIC) (Applied Biosystems). Thermocycling conditions were an initial 5 min denaturation at 95ºC, followed by 30 cycles at 95ºC for 30 s, 60ºC for 1 min 30 s and 72ºC for 30 s, 8 cycles at 94ºC for 15 s, 53ºC for 20 s and 72ºC for 35 s and a final elongation step of

72ºC for 15 min (Schuelke, 2000). Electrophoresis of the amplified products was performed using an ABI 3730xl Genetic Analyser (Applied Biosystems) with GeneScanTM 500 LIZ size standard and scored using GENEMAPPER version 4.0 (Applied Biosystems). Tests for Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium were conducted in GENEPOP version 4.1 (Rousset 2008). Genetic diversity based on number of alleles, and expected and observed heterozygosities was calculated in GenAlEx v6.41 (Peakall & Smouse 2006).

Chapter Two: Microsatellite loci

Results and discussion

Twenty-nine of the 33 loci amplified, with 26 of these appearing polymorphic. Of the 26 polymorphic loci, 14 amplified cleanly and consistently. A larger sample of 65 individuals from across the species’ range was genotyped at these 14 loci. A total of 70 alleles were identified across the 14 loci, each locus possessing 2-10 alleles. Overall observed and expected heterozygosities ranged from 0.031 – 0.708 and 0.060 – 0.777, respectively (Table 2.1). Five loci showed significant departure from HWE after significance levels were adjusted for multiple pairwise comparisons using the sequential Bonferroni correction (Table 2.1). This was not due to the presence of null alleles, but rather due to significant population structure (Chapter 3). Linkage disequilibrium was detected between several pairs of loci (13 out of 91 possible pairings) after significance levels were adjusted for multiple pairwise comparisons using the sequential Bonferroni correction. High-throughput sequencing on the 454 GS-FLX instrument was very effective in obtaining the initial 305 reads containing di- and tri-nucleotide microsatellites. Although time and resources only permitted the development of 33 microsatellite primer pairs, there were 94 microsatellites in total for which primer design was suitable, providing ample opportunity for a larger number of loci to be characterised in the future if required. The level of polymorphism within the microsatellites that successfully amplified was high (90%) relative to other threatened bird species, for example 40% in Kea (Nestor notabilis) (Dussex & Robertson 2013) and 54% in blue duck (Hymenolaimus malacorhynchos) (Abdelkrim et al. 2009), though similar to that recently reported for the rock wren’s confamilial, the rifleman Acanthisitta chloris (81%) (Preston et al. 2013). All 14 novel microsatellite loci described here are currently being used as markers in the first genetic study of X. gilviventris aimed to assist the species’ conservation and recovery (Weston et al. unpublished data).

Chapter Two: Microsatellite loci

Table 2.1 Characteristics of 14 microsatellite loci developed for rock wren (Xenicus gilviventris)

* Significant deviation from Hardy–Weinberg equilibrium following Bonferroni correction. All loci were characterized with 65 individuals sampled across the species’ range. Number of alleles (Na), expected and observed Heterozygosity (HE / HO)

Female rock wren Xenicus gilviventris

Chapter 3 Population structure within an alpine archipelago: a strong signature of past climate change within the New Zealand rock wren Xenicus gilviventris

Introduction

Many species occur in naturally subdivided populations due to spatial heterogeneity of the landscape (Watson 2002). This is especially evident in alpine species (Bech et al. 2009; Fedy et al. 2008; Floyd et al. 2005; Galbreath et al. 2009; Goossens et al. 2001; Griffin et al. 2008; Keyghobadi et al. 2005; Koumoundouros et al. 2009; Kruckenhauser & Pinsker 2004; Osborne et al. 2000; Rudá et al. 2010; Segelbacher & Storch 2002; Trizio et al. 2005) where naturally fragmented habitat may be viewed as an ‘alpine archipelago’ (Burkey 1995). These high altitude habitat patches and the species they harbour can provide significant insight into past biogeographical change and serve as useful models for predicting responses to anthropogenic habitat fragmentation (Bech et al. 2013; Bech et al. 2009; Floyd et al. 2005). Habitat specialists have been shown to suffer relatively larger reductions in abundance and range resulting from the synergistic effects of fragmentation and climate change (Warren et al. 2001). In this regard, the study of alpine habitat specialists is particularly pertinent because these species are often more sensitive to temperature increases and suffer greater habitat loss, consequently serving as early indicators of how species may respond to climate change across fragmented landscapes (Koumoundouros et al. 2009). Further, shifts to higher elevations amongst vertebrate predators may present additional threats to alpine species isolated atop alpine refugia (Hughes 2003). New Zealand has a dynamic geological and climatic history making it an interesting setting for the study of phylogeographical patterns (Wallis & Trewick 2009). Isolated from other major landmasses for the past 65 million years, New Zealand experienced almost complete marine submersion during the Oligocene (Cooper &

Cooper 1995; Landis et al. 2008; Waters & Craw 2006) followed by an intensive period of tectonic uplift during the Pliocene (5-6 Mya) (Batt et al. 2000; Chamberlain & Poage 2000). Pliocene uplift initiated the formation of alpine habitat in New Zealand, including the Southern Alps, the major mountain chain running along the South Island of New Zealand (Whitehouse & Pearce 1992). Since the late Pliocene to early

Chapter Three: Population structure and phylogeography

Pleistocene (2.5 Mya), several periods of glaciation have occurred forming vast networks of ice-fields and glaciers along the Southern Alps, most extensively in the central South Island (Newnham et al. 1999; Suggate 1990). Among New Zealand’s alpine taxa, phylogenetic studies support two major radiations: the first correlating with Pliocene uplift and the second with Pleistocene glaciations (Wallis & Trewick 2009). Most examples of Pliocene radiations to date are within broad taxonomic groups, although strong intraspecific structuring has been revealed within alpine invertebrates in the South Island, consistent with alpine radiation during the Pliocene and sustained isolation on geographically isolated mountain ranges (Trewick 2001; Trewick & Wallis 2001). Similarly, the Pleistocene glaciations have also left a strong genetic signature within many invertebrate species (Buckley et al. 2009; Buckley et al. 2001; Leschen et al. 2008; Marske et al. 2011; McCulloch et al. 2010; O'Neill et al. 2009) and vertebrates (Lloyd 2003) as formerly widespread populations were forced to retreat into multiple isolated refugia, or in some cases, became completely extirpated from entire regions (Marshall et al. 2009). Perhaps the best known biogeographic signature of Pleistocene glaciations within New Zealand is the commonly referred to biotic “gap”, whereby the Northern and Southern regions of the South Island are areas of high endemism, separated by a central region of low endemicity (Burrows 1965; Craw 1989; McGlone 1985). This region marks the disjunction in distribution within many taxonomic groups including species within several alpine invertebrate genera (McCulloch et al. 2010; Trewick & Wallis 2001), however this area was first coined the “beech gap,” reflecting the distinct absence of southern beech forest (Nothofagus spp.) (Cockayne 1926). Cockayne (1926) proposed that Northern and Southern beech populations likely persisted in ice-free refugia, with subsequent expansion back towards the centre of the island as ice-fields retreated. It has been proposed that some alpine organisms however, persisted within the biotic gap in refugia along the unglaciated eastern slopes of mountain ranges (Hill et al. 2009; Lockhart et al. 2001; Trewick 2001). The rock wren (Xenicus gilviventris Pelzeln, 1867) is a threatened alpine passerine (Robertson et al. 2013) belonging to the endemic New Zealand wren family (Acanthisittidae). The rock wren constitute a widespread, naturally fragmented population, occurring in patches of suitable habitat over c.900 m in altitude throughout the length of the South Island, New Zealand (Higgins et al. 2001). The dispersal characteristics of rock wren are largely unknown, though given that they have limited 22

Chapter Three: Population structure and phylogeography flight ability (Higgins et al. 2001), long-distance dispersal seems unlikely. Species with lower vagility tend to exhibit higher rates of genetic differentiation and a stronger genetic signature of population fragmentation (Baker et al. 1995; Bech et al. 2013; Burbidge et al. 2003; Burney & Brumfield 2009; Callens et al. 2011). Taxonomic work has identified the Acanthisittid wrens as a unique group amongst New Zealand’s avifauna in that they are the oldest and most primitive family in the order Passeriformes, placed in an infraorder of their own, the Acanthisittides (Sibley et al. 1982). Subsequent studies have suggested an even greater taxonomic distinctiveness for the wrens in identifying them as a sister group to all other passerines (Barker et al. 2002; Barker et al. 2004; Ericson et al. 2002; Hackett et al. 2008). The fact that the rock wren is part of an ancient family, combined with its island-wide distribution and limited dispersal ability makes it an excellent model taxon for uncovering biogeographic patterns. The rock wren is one of only two remaining extant members of the Acanthisittides, the other being the riflemen (Acanthisitta chloris) (Higgins et al. 2001). At least six other species have gone extinct, almost certainly due to destruction of indigenous forest habitat and the introduction of mammalian predators (Michelsen- Heath 1989; Worthy et al. 2010). Results from recent population studies indicate both a decline in rock wren abundance and a contraction in range (Michelsen-Heath & Gaze 2007). Michelsen-Heath & Gaze (2007) collected and collated >2000 rock wren sighting records from the period 1912 – 2005 and mapped past and present distribution. Whilst reported reductions in abundance were largely anecdotal, it is clear that since 1980 rock wren have not been recorded from many areas they were once sighted and overall, present a more fragmented distribution throughout their range. Conservation management of rock wren is currently underway, including the translocation of rock wren to predator-free offshore islands (Weston 2006; Willans & Weston 2005). However, before now, knowledge of genetic structure has not been available to inform management practices. Furthermore, as the climate rapidly changes, management to maintain and enhance species’ adaptive potential to deal with a range of new selection pressures becomes imperative (Hoffmann & Sgró 2011). Assisted colonisation (Loss et al. 2011) to high-altitude refugia may also be necessary for isolated, lower altitude populations that are unable to disperse beyond their current habitat (Hoegh-Guldberg et al. 2008; McLachlan et al. 2007). An understanding of past response to environmental change combined with a thorough knowledge of genetic 23

Chapter Three: Population structure and phylogeography variation and structure would serve to guide conservation action. The present study will develop this knowledge and also identify populations that exhibit high genetic distinctiveness, potentially representing evolutionary significant units (ESUs) (Funk et al. 2012; Moritz 1994). The conservation of ESUs ensures that evolutionary potential within the overall population is maximised (Crandall et al. 2000; Fraser & Bernatchez 2001). Given what is already known with regard to New Zealand’s dynamic geological and climatic history and the response of other alpine taxa to these environmental events, we are able to make predictions regarding the phylogeography of rock wren. Rock wren genetic structure might best be explained by i) a model of alpine radiation whereby rock wren populations became isolated on geographically isolated mountain ranges during the Pliocene or ii) a glacial refugia model where rock wren populations were isolated in one or multiple refugia during the Pleistocene glaciations. Here we distinguish between these predictions by using microsatellite variation, mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) to examine patterns of genetic structuring and divergence time estimates. If the genetic structure of rock wren is best explained by a model of alpine radiation, we expect to find distinct and deeply divergent lineages of rock wren on multiple different mountain ranges, with coalescence among at least some ranges dating back to c. 5 – 6 Ma. Alternatively, under a glacial refugia model, populations are expected to have been effectively connected by gene flow until the onset of the Pleistocene glaciations c. 2.5 Ma. If, in spite of their alpine habit, rock wren were completely eliminated from the heavily glaciated central South Island during the Pleistocene, we would expect to identify a distinctive “gap” phylogeography with disjunct northern and southern South Island lineages.

Materials and methods

Sampling and DNA extraction

Rock wren blood samples (N=134) were collected from throughout their range comprising the mountainous areas of the South Island of New Zealand (c.800 – 2500 m above sea level (asl) (Figure 3.1). Catching effort focused on adult pairs in established territories to avoid sampling relatives. Samples were collected from 2009-2012 during the austral summer

Chapter Three: Population structure and phylogeography breeding season. As the rock wren has limited flight capability (Higgins et al. 2001), birds were attracted using audio playback and captured by ‘herding’ them into mist nets. Resampling of birds was prevented by tagging each bird with a single metal leg band issued by the New Zealand Department of Conservation as part of the national bird banding scheme. Blood samples (<20 μl) were obtained by brachial vein puncture with a sterile needle (25G x 5/8”) and collection of the blood with a glass capillary tube for immediate transfer into 1 ml of Queens lysis buffer (10 mM Tris, 10 mM NaCl, 10 mM Na-EDTA, 1% n-lauroylsarcosine; pH 7.5) (Seutin et al. 1991). DNA was extracted and purified from the whole blood using a standard 5% Chelex protocol (Walsh et al. 1991).

Chapter Three: Population structure and phylogeography

a Lake Clara (Nth Kahurangi) b Henderson Basin, Mt Perry (Nth Kahurangi) c Wangapeka (Sth Kahurangi) d Lake Thompson e Lewis Pass a f Mt Alexander g Bealey, Otira (Arthurs Pass) b h Barker Hut (Arthurs Pass) i Whymper c j Aoraki/Mount Cook k Landsborough l Twin Stream m Ahuriri d n Haast Range o Crucible e p Matukituki q Homer, Gertrude, Monkey Creek, f Lake Adelaide, Lake Marion g r Eyre Mtns h s Murchison Mtns t Lake Roe i u Lake MacArthur j k

l n m p o q

s r

t u

Figure 3.1 Sampling locations of rock wren (Xenicus gilviventris) throughout the mountainous regions of the South Island of New Zealand

Chapter Three: Population structure and phylogeography

Mitochondrial DNA sequencing

Twenty individuals from across the range of rock wren were sequenced at the mitochondrial cytochrome b gene. A 1075 bp fragment of cytochrome b was amplified using primers L14764 and H16064 (Sorenson et al. 1999). PCRs were performed in a

25 µl reaction containing 10 - 30 ng of template DNA, 1 × PCR buffer, 1.5mM MgCl2, 200 µm of each dNTP, 0.5 units of Taq polymerase (Bioline USA, Inc, Randolph MA 02368-4800) and 1 pmol of each primer. The thermal cycling parameters were an initial two minute denaturation at 94°C, followed by a touchdown of 10 cycles at 94°C/20 sec, 60→50°C/25 sec (-1°C per cycle) and 72°C/90 sec, then 25 cycles at 94°C/20 sec, 50°C/25 sec and 72°C/90 sec. PCR products were purified using Acroprep 96 Filter Plates (PALL Corporation). Sequencing was carried out with a BigDye v.3.1 sequencing kit (Applied Biosystems, Foster City, CA) using the two amplification primers as above. Sequence products were then purified using Sephadex-GS50 gel filtration (Amersham Bioscience, New Zealand) and run on an ABI 3730xl DNA Analyser. To aid primer design used to amplify the mitochondrial control region, we used a stand-alone version of blast (blast+ tools) (Camacho et al. 2009) to search the rock wren 454 read data set (Chapter 2) for mitochondrial sequences. Reads were searched for sequence similarities to the Eudyptula minor mitochondrial genome (NCBI accession AF362763). We then converted reads into a nucleotide database using the makeblastdb command and ran Blastn with a moderate E-value cut-off of 1e-3 to find hits against mitochondrial genes and the non-coding control region. One mitochondrial fragment was identified at the tRNAglu end of the control region. We then designed amplification primers for this 454 fragment: L-RW-ND6 (5’-CTTGAAAA GCCACCGTTGTT-3’) and H-RW-12s (5’-TGATACCCGCTCCT AATTGC-3’) using the rifleman (Acanthisitta chloris) genome and rock wren 12S. The thermal cycling parameters were an initial 5 min denaturation at 95ºC, followed by 35 cycles at 95ºC /35 sec, 60ºC /25 sec and 72ºC /120 sec. Three internal sequencing primers were used L-ROWR-CR-internal (5’TGCGCCTCTGGTTCTTAGTT-3’), L-ROWR-CR-1 (5’C CAATGATCTACCAAAATACACCT-3’) and R-ROWR-CR-1 (5’-GGGGGATAGGG TTGTGTGTT-3’) to produce a 1126 bp sequence fragment for 25 individuals (using the same birds as for cyt-b where possible). We had difficulty consistently sequencing forward and reverse through a T-string (c.13 bp) in the middle of the control region

Chapter Three: Population structure and phylogeography fragment and therefore trimmed this region to 4 bp for all sequences. The rock wren control region consisted of two hypervariable domains flanking a conserved central domain (c.500 bp) with few polymorphisms, as is characteristic in vertebrates (Ruokonen & Kvist 2002; Saccone et al. 1987). Therefore, we sequenced in one direction the hypervariable first domain (273 bp) for a further 30 individuals (N=55) using primer L-ROWR-CR-internal.

Nuclear DNA sequencing

Twenty-one individuals from across the range of rock wren (the same birds as used for mtDNA sequencing where possible) were sequenced at the seventh intron of the nuclear β-fibrinogen gene (β –fibint7). An c.1,000 bp fragment containing the entire seventh intron was amplified using primers FIB-BI7U and FIB-BI7L (Prychitko & Moore 1997) in a 25µl reaction as for mtDNA above. The thermal cycling parameters were an initial 5 min denaturation at 95ºC, followed by 35 cycles at 95ºC /35 sec, 60ºC /25 sec and 72ºC /120 sec, with a final extension step of 72ºC for 5 mins. PCR products were purified using Acroprep 96 Filter Plates (PALL Corporation). Using initial rock wren sequence data, internal sequencing primers were designed, F-bfibRW1stinsert (5’- CAAGATAGAAA CCAATGAGAAA-3) and R-bfibRW2ndinsert (5’- GTAAGAA TATTGGTACTCA-3’) and sequencing was carried out with a BigDye v.3.1 sequencing kit (Applied Biosystems, Foster City, CA). Sequence products were purified using Sephadex-GS50 gel filtration (Amersham Bioscience, New Zealand) and run on an ABI 3730xl DNA Analyser. A 531 bp fragment was used in subsequent analyses. All sequences were aligned using progressive pairwise alignment in Geneious (v.6.0.5) and variable sites were confirmed by visual inspection of the chromatograms. Cytochrome b sequences were translated to detect stop codons and reading frame errors (Geneious v.6.0.5, Biomatters, New Zealand. Available from http://www. geneious. com). For the nuclear β -fibint7 gene sequences that contained double peaks indicating heterozygous sites (13 of the 21 individuals), we resolved haplotypes probabilistically using PHASE v.2.1 (Stephens & Donnelly 2003; Stephens et al. 2001), implemented in DnaSP v.5.10.01. Only individuals with phase probabilities >0.75 were included in subsequent analyses (18 of the 21 individuals sequenced). The three individuals with haplotypes what were unable to be reliably reconstructed were from Mt Alexander,

Chapter Three: Population structure and phylogeography

Lake Crucible and Homer Tunnel. Each individual presented one heterozygous site with a phased probability of <0.75.

Microsatellite genotyping

All 134 samples were genotyped at 14 microsatellite loci previously developed for rock wren (see Table 2.1 in Chapter 2). The polymerase chain reaction (PCR) conditions used are as described in Chapter 2. To check for genotyping error, 10% of all samples were run twice for each locus and the error rate per allele was calculated by dividing the number of mismatched alleles by the total number of repeat-genotyped alleles (Hoffman & Amos 2005). Tests for Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium (LD) were conducted for each population in GENEPOP version 4.1 (Rousset 2008) employing the markov chain method with 10,000 dememorisations, 1,000 batches and 10,000 iterations. Significance levels were adjusted for multiple comparisons following standard Bonferroni correction (Rice 1989).

Genetic variation and differentiation

Nucleotide diversity (π) and haplotype diversity (h) indices were calculated overall and for the two mtDNA lineages identified in tree based analyses using ARLEQUIN v.3.5.1.3 (Excoffier et al. 2005). Levels of genetic variation using microsatellite data were calculated overall, among main mtDNA lineages and among sampling areas. Mean number of alleles per locus and observed/expected heterozygosities (Ho/He) were calculated using GenAlEx v.6.41 (Peakall & Smouse 2006). Allelic diversity, calculated as allelic richness to adjust for sample size differences, was also calculated in FSTAT v.2.9.3 (Goudet 2001). Sampling areas with <4 individuals were not included in sampling area analyses. To quantify levels of genetic differentiation between the two main lineages using mtDNA, a pairwise population FST based on pairwise differences for the first domain of the control region was calculated using ARLEQUIN v.3.5.1.3 (Excoffier et al. 2005). Significance was tested over 1,000 random permutations.

For microsatellite data, Weir and Cockerham’s (1984) pairwise FST values were calculated among sampling areas and for the two main lineages in FSTAT v.2.9.3 (Goudet 2001). Significance was tested over 560 permutations and strict Bonferroni corrections for multiple comparisons were applied (Rice 1989).

Chapter Three: Population structure and phylogeography

To test for a correlation between genetic and geographic distance (isolation by distance or ‘IBD’), mantel tests were performed using pairwise FST values among sampling areas in GenAlEx v.6.41 (Peakall & Smouse 2006). We also tested for IBD among sampling areas within each of the two main lineages separately.

Phylogenetic analysis

The most appropriate model of sequence evolution and associated parameters for each dataset were determined using the AIC selection criterion of JMODELTEST v.2.1.1 (Darriba et al. 2012) (Table 3.1). Maximum Likelihood (ML) analyses were performed in PAUP* v.4.0b10 (Swofford 2003) with a heuristic search method comprising 100 repetitions of stepwise addition and node support was calculated using 1,000 bootstrap replicates. Bayesian trees were constructed in MrBayes v.3.2 (Ronquist & Huelsenbeck 2003) in two independent runs for 2,000,000 generations, sampling every 500 generations and discarding the first 25% as burn-in. All priors were left at the default settings. MCMC convergence was assessed by evaluating the standard deviation of the split frequencies and ESS (effective sample size) values in Mr Bayes and examining likelihood plots in Tracer v.1.5 (Rambaut & Drummond 2007).

Table 3.1 Model of sequence evolution for each rock wren (Xenicus gilviventris) locus as determined using the AIC selection criterion in JModelTest

Substitution Locus model Cytochrome b HKY + I

mtDNA control region HKY + G (1st Domain)

mtDNA control region HKY + I (1126 bp)

ß -fibint7 GTR + I

Chapter Three: Population structure and phylogeography

Trees were rooted using the closely related rifleman (Acanthisitta chloris) as an outgroup. All trees were visualised and edited in FigTree v.1.4 (Rambaut 2012). For the control region, we present the gene tree for the hypervariable first domain only. Although 55 individuals were sequenced for the first domain compared with only 25 individuals for the full 1126 bp fragment, the overall topology of the trees remained similar (see Appendix 1 for tree using 1126 bp fragment). Pairwise genetic distances among clades were calculated in MEGA v.5.2.1 using the best-fit substitution model available for each dataset according to the Bayesian Information Criterion (BIC) (Tamura et al. 2011). Bootstrapping with 1,000 replications was used to estimate variance. For comparison with mtDNA, pairwise genetic distances based on the proportion of shared alleles (Dps) were calculated for all genotyped individuals using Microsatellite Analyzer v.4.05 (Dieringer & Schlötterer 2003). Relationships were examined with a neighbor-joining tree constructed in POPULATIONS v.1.2.32 (Langella 2002) and viewed in FigTree v.1.4 (Rambaut 2012). To estimate the genealogical relationships among mtDNA and nuclear haplotypes, haplotype networks were constructed using a 90% statistical parsimony criterion in TCS v.1.21 (Clement et al. 2000). TCS estimates the haplotypes with the highest outgroup probability based on the frequency of the haplotypes in the sample. Here, we define putative ancestral populations as those possessing the outgroup haplotype.

Molecular clock

To estimate the time since divergence among rock wren populations, the cytochrome b dataset was analysed in BEAST v.1.7.4 (Drummond et al. 2012) under a strict molecular clock assumption based on 2.1% sequence divergence per million years (i.e. a substitution rate of 0.0105 per site per million years). This molecular rate has proven to be conserved across a wide range of avian orders for this gene (Weir & Schluter 2008). The HKY + I model of sequence evolution as estimated by JMODELTEST v.2.1.1 was applied with a coalescent constant size tree prior. The length of the Markov chain was 10 million generations, sampling every 200 generations and discarding the first 10% as burn-in. Chain convergence was assessed using the program Tracer v.1.5 (Rambaut & Drummond 2007) through visual inspection of the posterior probability distribution, ESS and traceplot for each sampled parameter. Trees were summarised

Chapter Three: Population structure and phylogeography and the final target tree with posterior support values and 95% highest posterior density (HPD) intervals for divergence estimates was computed using FigTree v.1.4 (Rambaut 2012).

Microsatellite cluster analyses

To investigate whether the number of genetically distinct groups within rock wren based on microsatellite genotypes supports groupings identified using tree-based methods, we used two different Bayesian clustering methods implemented in STRUCTURE v.2.3.3 (Pritchard et al. 2000) and TESS v.2.3.1 (Chen et al. 2007). Both methods assign individuals probabilistically to genetic clusters (K) by calculating membership coefficients per individual per cluster (Q); though spatial coordinates of sampled individuals are utilized as prior information only in TESS. We tested for K = 1 - 8 (eight being the number of broad geographic regions sampled) with twenty replicate runs performed for each value of K. For STRUCTURE, we used an admixture model and assumed correlated allele frequencies among populations. Results are based on 1,000,000 Markov Chain Monte Carlo iterations following a burn-in of 100,000. The true value of K was inferred by evaluating Q values, the mean log-likelihood of the data (ln P(X|K)) and using the Delta-K (∆K) method (Evanno et al. 2005) implemented in Structure Harvester (Earl & vonHoldt 2012). In TESS, we ran the admixture model with 50,000 sweeps following a burn-in of 10,000. To determine the most optimal value of Kmax, the maximal number of parental populations that best suits the data, we plotted the average deviance information criterion (DIC) values against each value of Kmax and selected the Kmax value at the beginning of a plateau (Durand et al. 2009). For both programs, results were averaged across the twenty runs for each of K=2 and 3 using Clumpp v.1.1.2 (Jakobsson & Rosenberg 2007) and visualized using Distruct v.1.1 (Rosenberg 2004).

Results

Chapter Three: Population structure and phylogeography set of individuals were further sub-divided according to sampling area, only the Homer Tunnel population departed from HWE for one locus, Xgilv19, which exhibited heterozygote deficiency and therefore all 14 loci were used. MtDNA sequencing yielded 16 distinct cytochrome b haplotypes from 20 individuals and 18 control region 1st domain haplotypes from 55 individuals. A total of 15 ß-fibint7 haplotypes with a phased probability of >0.75 were identified from 18 individuals, of which 9 individuals were heterozygous.

Phylogenetic analysis

Bayesian and Likelihood analyses of mtDNA datasets revealed a deep and well- supported split between Northern and Southern rock wren (Figures 3.2 and 3.3). The topology of trees was very similar for both methods and therefore only the Bayesian trees are presented with likelihood values also labeled on main nodes. Genetic distance between the Northern and Southern clades was high for both the rapidly evolving control region 1st domain, 13.3 ± 4.9% S.E (Tamura 1992) and for cytochrome b, 3.68 ± 0.5% S.E (Tamura & Nei 1993). The distance detected between rock wren and their sister species the rifleman (8.35± 1.1% S.E) was approximately twice that of the main Northern- Southern split at cytochrome b. Within the Southern lineage, mtDNA revealed a further consistent genetic split between rock wren sampled from deep Fiordland (Lakes MacArthur and Roe) and the rest of the South (Figures 3.2 and 3.3). However, this divergence was shallow (1.2% at cytochrome b) comparative to the main North-South split and overall, genetic distances within each of the two lineages were low (see Appendix 2 for pairwise distance matrices for all three loci). Using a 2.1% sequence divergence per MY calibration for cytochrome b (Weir & Schluter 2008) we estimate the timing of the split between the rock wren and the riflemen at approximately 6.1 Ma (95% HPD: 4.1–8.4 Ma) and the North - South divergence at approximately 2 Ma (95% HPD: 1.4 – 2.7 Ma) (Figure 3.3). Radiation within each lineage appears to have happened much later, approximately 200,000 years ago in the North and 600,000 years ago in the South, starting with the split between the deep Fiordland birds and the rest of the Southern lineage (Figure 3.3). The nuclear gene tree using β -fibint7 did not reveal the same level of phylogeographic structure as the microsatellite and mtDNA datasets (Figure 3.4).

Chapter Three: Population structure and phylogeography

Overall, haplotype diversity was high (h = 0.90) and phylogeographic resolution low. However, a significant sub-structuring was evident in the North of the range, with the individuals from Lake Clara and Lake Thompson forming a relatively well-supported clade (Figure 3.4). Genetic distance within this clade was 0.2% compared with 3.38% between this clade and all other individuals (see Appendix 2 for pairwise distance matrices).

a a b b c c b d 0.88 d e f e g 0.66 f h g i i j 0.90 g n k h l j o m j p k

q l m t s r n p o u 0.89 p 82 q

q s r 0.98 t 71 u

Figure 3.2 Bayesian phylogeny of rock wren (Xenicus gilviventris) using the 1st domain of the control region. Posterior probabilities are shown above main nodes and maximum likelihood bootstrap support values below. The scale depicts the distance corresponding to 0.3 nucleotide substitutions per site. Blue = Northern lineage and red = Southern lineage.

Chapter Three: Population structure and phylogeography

a a b b d c d c 1 e g 57 e g i i j j k n l j o m p k q q s r r u t s 1 n p 92 1 l m 99 o 1 u

95 t

MYA

Figure 3.3 Bayesian cytochrome b phylogeny of rock wren (Xenicus gilviventris) using BEAST. Posterior probabilities are shown above main nodes and maximum likelihood bootstrap support values below. Scale depicts the distance corresponding to 0.6 nucleotide substitutions per site. Time scale axis is in million years before present and bars represent 95% HPD. Blue = Northern lineage and red = Southern lineage.

Chapter Three: Population structure and phylogeography

b b c a 0.85 0.88 d 55 77 g a c 0.73 k b m c 0.59 n u d k e 0.78 g g 62 i e i i j j n k j l o m l q 0.55 p n 59 m

s s r r t t u u

Figure 3.4 Bayesian phylogeny of rock wren (Xenicus gilviventris) using the nuclear beta-fibrinogen intron-7. Posterior probabilities are shown above main nodes and maximum likelihood bootstrap support values below. The scale depicts the distance corresponding to 0.02 nucleotide substitutions per site. Blue = Northern lineage and red = Southern lineage.

Chapter Three: Population structure and phylogeography

One haplotype (in red) is relatively common among the 18 haplotypes identified using the 1st domain of the control region and is shared amongst 16 individuals from sampling areas throughout the South (Figure 3.5). There is less haplotype sharing between sampling areas in the North of the island, with only one haplotype (in yellow) shared by three individuals from Arthurs Pass and Lewis Pass.

Figure 3.5 The geographic distribution of rock wren (Xenicus gilviventris) mtDNA 1st domain control region haplotypes (273 bp). The different colours represent each of the 18 haplotypes and numbers beside charts show number of individuals sampled from each area.

Chapter Three: Population structure and phylogeography

The neighbour-joining microsatellite tree comprised three main clades, one containing all individuals assigned to the Northern lineage using mtDNA (Cytochrome b and the control region), with the exception an individual from Aoraki/Mount Cook (Appendix 3). However, there was no obvious geographic structure within the two Southern clades and overall, bootstrap support was very weak (all less than 73%) indicating this method was not robust in resolving the relationships amongst the 134 genotyped individuals. The mtDNA haplotype networks identified the same main clades as tree-based methods with no sharing of haplotypes between the Northern and Southern lineages (Figure 3.6a and 3.6b). One complete network was not able to be resolved under a 90% parsimony criterion due to the extent of divergence among mtDNA clades. At cytochrome b, it is estimated that 37 mutational steps have occurred between the two main clades (Figure 3.6a), whilst at the control region 1st domain, 17 steps (Figure 3.6b). Both mtDNA datasets estimate the ancestral population within the Southern lineage to be the Eyre Mountains, in the south east of the rock wrens range. However for the control region, the ancestral haplotype (red haplotype in Figure 3.5) also occurs throughout the Southern lineage (Figure 3.6b). Within the Northern lineage, support for the ancestral population is ambiguous. Two populations are estimated as ancestral within the Cytochrome b dataset, the Wangapeka and Bealey Valleys, whereas for the control region 1st domain, two individuals from Lake Thompson and Lewis Pass in the north east of the species’ range are identified as representing the ancestral haplotype (Figures 3.6a and 3.6b). The network constructed using phased nuclear haplotypes is somewhat more complex to interpret (Figure 3.6c). Several haplotypes are shared between individuals from the Northern and Southern lineages, though where this occurs it involves birds identified as admixed by cluster analysis from within the central secondary contact zone around Aoraki/Mount Cook (see microsatellite cluster analyses below). The estimated ancestral population is comprised of several individuals from Aoraki/Twin Stream northwards. Two loops observed within the network indicate that either recurrent mutations (homoplasy) or recombination may have occurred, thus confounding any phylogeographic signal.

Chapter Three: Population structure and phylogeography

Figure 3.6 Haplotype networks of rock wren (Xenicus gilviventris) sequences estimated using a 90% parsimony criterion for a) cytochrome b b) mtDNA control region 1st domain (273 bp) and c) ß- fibrinogen intron 7. Haplotype frequency is indicated by circle size and the putative ancestral haplotype is denoted by an asterisk. Each branch represents a single nucleotide change and black dots between branches represent inferred missing haplotypes. Blue = Northern lineage and red = Southern lineage (see Appendix 4 for specific sampling locations corresponding to each haplotype).

Chapter Three: Population structure and phylogeography

Microsatellite cluster analyses

Cluster analysis using STRUCTURE showed highest support for two genetic populations (K=2) (Figures 3.7 and 3.8), with individual assignments corresponding identically with the two main clades identified by mtDNA tree based analysis (Figures 3.2 and 3.3). The mean log likelihood of the data was lowest for the model K=1 and then increased sharply for K=2, after which there was an overall plateau through to K=8 (Figure 3.8). When lnP(X|K) plateaus or increases only slightly, as is seen here at K=2 (Figure 3.8), then the model with the lowest value of K that captures the major structure of the data is usually the best choice (Pritchard et al. 2010). Delta-K (∆K) cannot be calculated for K=1; however given that ln P(X|K) was lowest at K=1 this indicates that this was not the best model for our data. The highest estimate for ∆K was at K=2 (Figure 3.8), confirming this model was a better choice than any of the models with larger values of K (Evanno et al. 2005). Individual Q values for the K=2 model were high, with an average assignment value to either cluster of 0.97 (Figure 3.7).

Assignments using TESS based on Kmax= 2 were concordant with those of STRUCTURE, and although individual Q values varied slightly between the two methods, the two programs assigned individuals to the two clusters almost identically

(Figure 3.7). However, the model with the lowest DIC value in TESS was where Kmax=

3 (Figure 3.8). Average individual assignment values to either cluster for Kmax= 2 were

0.98. For the Kmax=3 model, average Q values dropped to 0.88 and many individuals within the third cluster (located within the Southern lineage) showed mixed membership (Figure 3.7). The authors of TESS have stated that ‘the actual number of clusters, K, may be less than Kmax’ (Durand et al. 2009). As with tree-based methods, genetic assignments corresponded well with geographic sampling location, with individuals sampled north of Aoraki/Mount Cook National Park assigned to one cluster corresponding to the Northern lineage and individuals sampled south of Aoraki comprising the Southern lineage. Several admixed individuals were present within what appears to be a narrow zone of secondary contact between the two lineages, where Q values ranged from 0.61-0.88 for both TESS and STRUCTURE. This contact zone spans a geographic distance of c.60 km.

Chapter Three: Population structure and phylogeography

1 a)

0.5

0 North South

1 b)

Q 0.5

0 North South

0.5

North South

Figure 3.7 Proportional membership (Q) of rock wren (Xenicus gilviventris) to genetic clusters (K) as estimated using STRUCTURE a) K = 2 and TESS b) Kmax = 2 and c) Kmax = 3. Each vertical bar represents a single individual and individuals are ordered by sampling location from North to South according to latitude. Colours correspond to genetic clusters and map inserts show geographical location of clusters.

Chapter Three: Population structure and phylogeography

Figure 3.8 Determination of K, the number of genetic populations of rock wren (Xenicus gilviventris) from a) STRUCTURE analysis showing mean log-likelihood of the data and ∆K for each value of K and b) TESS analysis showing average DIC for each value of Kmax.

Chapter Three: Population structure and phylogeography

Genetic variation and differentiation

Indices of mtDNA diversity within the overall rock wren population were both high for cytochrome b (π = 0.0202, h = 0.98) and the control region (π = 0.0413, h = 0.89). For ß-fibint7, nucleotide diversity (π = 0.0095), was lower overall than that observed for mtDNA, as would be expected for a slower evolving nuclear gene, though haplotype diversity was relatively high (h = 0.90). For microsatellites, average observed and expected heterozygosities across all loci were 0.426 and 0.564 respectively. However, within each of the Northern and Southern lineages, genetic diversity was substantially lower than when considering the population overall. Furthermore, genetic diversity was lower within the Northern lineage than within the South for both cytochrome b and microsatellite data (Table 3.2). We also detected a total of 39 private alleles within the Southern population compared to just seven in the North and four of the five sampling areas analysed within the Southern lineage had higher levels of genetic variation than each of the three Northern sampling areas (Table 3.2). In contrast, at the faster evolving control region 1st domain, diversity levels were higher in the North (π = 0.0220, h = 0.89) than in the South (π = 0.0098, h = 0.73). st High FST values of 0.783 (p < 0.01) for the mtDNA control region 1 domain and 0.184 (p < 0.01) for microsatellites indicate strong genetic differentiation between the Northern and Southern lineages. A comparison of FST values among the three populations identified by TESS did not support the differentiation of a third cluster, with a relatively low pairwise FST value of 0.085 (p<0.01) between the additional proposed cluster and the rest of the Southern lineage. Average FST values among sampling areas within the Northern population were higher (0.194) relative to those within the Southern population (0.076) (Table 3.3). The majority of pairwise FST values among sampling areas were significantly different from zero (p<0.05) (Table 3.3).

Chapter Three: Population structure and phylogeography

Table 3.2 Genetic variation at 14 microsatellite loci and 1075 bp of cytochrome b in Northern and Southern rock wren (Xenicus gilviventris) populations and 8 geographically distinct sampling areas. N = sample size, NA = mean number of alleles per locus, AR =

allelic richness, HE = expected heterozygosity, HO = observed heterozygosity, H = number of haplotypes, h= haplotype diversity, π = nucleotide diversity.

Microsatellite data MtDNA

Population N NA AR HE HO N H h π Nth Kahurangi 10 2.14 1.99 0.337 0.312 Wangapeka 9 1.71 1.64 0.219 0.202 Arthurs Pass 7 2.93 2.55 0.448 0.400 Northern 35 3.71 3.58 0.395 0.286 8 6 0.93 0.0025 Twin Stream 4 3.00 3.00 0.547 0.625 Haast Range 8 2.71 2.50 0.503 0.393 Homer Tunnel 19 4.14 2.99 0.530 0.485 Murchison Mtns 14 3.93 2.79 0.486 0.444 Lake Roe 15 3.86 2.80 0.484 0.486 Southern 99 6.00 4.96 0.559 0.473 12 10 0.97 0.0055

Table 3.3 Pairwise FST values among rock wren (Xenicus gilviventris) sampling areas for 14 microsatellite loci. Highlighted values indicate values that are not significantly different from zero.

NORTH SOUTH

Location Nth Wangapeka Arthurs Twin Haast Homer Murch. Lake Kahurangi Pass Stream Range Tunnel Mts Roe Nth Kahurangi 0.000 Wangapeka 0.240 0.000 Arthurs P. 0.118 0.223 0.000 Twin Strm 0.242 0.349 0.076 0.000 Haast R. 0.298 0.349 0.114 0.008 0.000 Homer T. 0.268 0.287 0.161 0.096 0.075 0.000 Murchison Mts 0.314 0.354 0.206 0.149 0.118 0.036 0.000 Lake Roe 0.304 0.353 0.181 0.130 0.102 0.036 0.011 0.000

Chapter Three: Population structure and phylogeography

Mantel tests revealed a signiﬁcant pattern of isolation by distance (IBD) across the range of rock wren with geographical distance explaining 63% of variation in pairwise population differentiation among sampling areas (R2 = 0.634; P= 0.001) (Figure 3.9). However, when testing for IBD within the Northern and Southern lineages separately, we found that this relationship was only significant within the Southern lineage (R2 = 0.762; P=0.016) (Appendix 5).

Figure 3.9 Isolation by distance results based on multilocus microsatellite data from rock wren (Xenicus gilviventris) showing the

relationship between pairwise population differentiation (FST/(1-FST) and geographical distance across the South Island, New Zealand. The trend, represented as a solid line, was significant (p=0.001) as determined through Mantel tests with 9,999 permutations.

Chapter Three: Population structure and phylogeography

Discussion

Phylogeographic structure

Both mtDNA and microsatellite data reveal a deep North-South phylogeographic split among rock wren, with the disjunction between the two lineages occurring at Aoraki/Mount Cook National Park in the centre of the South Island. Given that we found a strong signal of genetic differentiation for both the female-inherited mtDNA and bi-parentally inherited microsatellites, there is no strong evidence for sex-biased dispersal at this spatial scale. The level of genetic divergence observed between the Northern and Southern rock wren lineages within the South Island at cytochrome b (3.7 ± 0.5%) and the control region 1st domain (13.3 ± 4.9%) is similar to levels separating North and South Island species of New Zealand robins (Petroica spp.) (5.9 ± 1.0% and 10.3 ± 1.4% for cyt b and the control region respectively: (Miller & Lambert 2006) and greater than that separating North Island Kokako (Callaeas cinerea wilsoni) and South Island Kokako (C. c. cinerea) (4.9 ±1.0% for the control region) (Murphy et al. 2006). However, given that the divergence between rock wren and their sister species the rifleman was 8.4% at cytochrome b, this places the divergence between Northern and Southern rock wren closer to the ‘sub-specific’ level. Within lineages, sub-structuring appears to be shallow relative to the main North-South split (Figures 3.1 and 3.2) and although haplotype diversity is high, low nucleotide diversity values indicate only small differences between haplotypes (Table 3.2). This is also reflected within the haplotype networks which show only 1 - 2 mutational steps amongst individuals from most locations for (Figure 3.6a and 3.6b). The pattern of isolation by distance was significant overall, indicating that gene flow is limited by geographic distance, as would be expected given the rock wrens poor flight ability and the disjunct nature of their specialised alpine habitat (Figure 3.9). However, when addressing each the Northern and Southern lineages separately, a significant pattern of IBD amongst the Northern populations was not detected. Although the number of sampled populations able to be included in the analysis from the North was low (n=3) compared with the South (n=5), we would still expect to find a positive trend if the two lineages shared a similar genetic structure and this does not appear to be the case (Appendix 5). Greater gene flow within the Southern lineage was

Chapter Three: Population structure and phylogeography also indicated by sharing of a control region haplotype (red haplotype in Figure 3.5) between birds from separate mountain ranges throughout much of the southern South Island. In contrast, only one haplotype is shared between geographically isolated areas in the North and that is between three individuals from Lewis and Arthurs passes (approximately 80 km apart). The deepest mtDNA divergence among birds within lineages however, is the divergence between individuals from Southern Fiordland (Lakes Roe and MacArthur) and the rest of the South. Interestingly, a distinct Southern Fiordland subspecies of rock wren, X. g. rineyi, was proposed in the early 1950s based on behavioural (Riney 1953) and morphological characteristics (Falla 1953). Falla collected three specimens from Lake MacArthur, South Fiordland and compared these with nine museum specimens, all from the Northern South Island, describing smaller dimensions and a brighter plumage for the Lake MacArthur birds. Here, we did not detect any notable differences of the southern Fiordland birds either behaviourally or in the hand, though a comprehensive analysis of morphological data would be required to test for more subtle morphological differences. Furthermore, although there was a consistent genetic split between the birds from Lake MacArthur and Lake Roe and the rest of the Southern lineage (Figures 3.2, 3.3 and 3.6a-b), this divergence was shallow (1.2% at cytochrome b) compared to the much deeper North-South split. Pairwise FST values between the south Fiordland birds and other Fiordland populations are also low indicating very little genetic differentiation (Table 3.3). Given that Falla (1953) was comparing the morphology of southern Fiordland specimens to birds from a Northern lineage we have identified in this study, it may be that X. g. rineyi comprises the entire Southern lineage. However, this does not account for the behavioural difference that Riney (1953) noted between the southern Fiordland birds and elsewhere in Fiordland. It would be therefore useful to undertake molecular analysis of the historical specimens of X. g. rineyi collected by Falla (1953) alongside the contemporary samples collected from the type locality of X. g. rineyi during this study, to further clarify the proposed sub-species status of these birds. The phylogenetic signal in the ß -fibint7 dataset appears to be much weaker than for microsatellite and mtDNA, though this is not surprising given that this region evolves at a much slower rate; almost three times slower at cytochrome b in woodpeckers (Prychitko & Moore 2000). Some notable sub-structuring is evident within the North however, with individuals from Lake Clara, Lake Thompson and the 48

Chapter Three: Population structure and phylogeography

Bealey Valley in Arthurs Pass forming a well-supported genetic grouping (Figures 3.4 and 3.6c). Genetic distance within this clade was 0.2% compared with 3.38% between this clade and all other individuals. Interestingly, this differentiation is exactly equivalent to the average distance between the sister-species rifleman and rock wren at β -fibint7, and substantially higher than the average pairwise distance amongst all clades of 1.2% (see Appendix 2 for matrix of pairwise differences). The mtDNA tree based analyses do not identify this same grouping, though error in phylogeographic construction seems unlikely given that the nodes in question are relatively well supported. Incongruence between phylogenetic trees using mtDNA and nDNA is not uncommon (Jeffroy et al. 2006) and may be caused by several evolutionary processes, such as 1) gene duplication or recombination; 2) deep coalescence (otherwise known as incomplete lineage sorting) and 3) introgressive hybridization (Maddison 1997). Some evidence of recombination within this clade is evident in the haplotype network in that the three haplotypes concerned are connected by a loop (Figure 3.6c). Deciphering between incomplete lineage sorting and introgressive hybridisation is difficult given that both these processes generate similar phylogenetic patterns (Holder et al. 2001).

Glacial refugia with subsequent expansion

Using a standard 2.1% sequence divergence per MY calibration at cytochrome b (Weir & Schluter 2008), we estimate the split between the Northern and Southern rock wren lineages to have occurred c. 2 Ma within the Pleistocene era of extensive glaciation, much later than the peak of Pliocene uplift c. 5 – 6 Ma. Consequently, rock wren genetic structure is not consistent with an alpine radiation model, but rather, the North – South divergence among rock wren is best explained by a model of glacial refugia, whereby rock wren populations became isolated in Northern and Southern refugia during the Pleistocene glaciations and subsequently expanded. The genetic structure observed within rock wren also provides further support for the role Pleistocene glaciations in the formation of the biotic gap within the central South Island (Cockayne 1926; McCulloch et al. 2010; Trewick & Wallis 2001). Although on average a 2% mitochondrial clock rate at cytochrome b has been demonstrated to be conserved for passeriformes (Weir & Schluter 2008), caution still applies to this calibration as some variation among different taxonomic groups occurs.

Chapter Three: Population structure and phylogeography

Interestingly, applying this clock rate to the timing of the rock wren – rifleman split, our estimation of c. 6 Ma corresponds with the formation of the Southern Alps during the Pliocene. It seems plausible that adaptive radiation saw the evolution of a new alpine wren at this time. A recent analysis of NZ wren phylogeny based on morphological characters suggests a much deeper coalescence between the rifleman and rock wren, dating back to the pre – Miocene (Worthy et al. 2010). However, the only attempt until present to clarify phylogenetic relationships among the New Zealand wrens based on molecular data suggested a 24 – 19 Ma divergence between the riflemen and rock wren using ND6 and 12S mtDNA extracted from two museum specimens (Cooper 1994; Cooper & Cooper 1995). Low nucleotide and high haplotype diversity as observed within both lineages can be a signature of a rapid demographic expansion (Avise 2000). Evidence for demographic expansion further supports the glacial refugial model as the most likely explanation underlying the rock wrens strong north - south divergence. Tests for population expansion and estimations of effective population size (Chapter 4) will assist in elucidating the historical demography of rock wren. Furthermore, the onset of radiation within each lineage appears relatively recent, commencing with the divergence of the Southern Fiordland birds c. 600 ka. Radiation within the Northern lineage commenced much later at c. 250 ka. There have been numerous glacial cycles within the past 2.5 My in New Zealand, though prior to the last glaciation (Otira) from c.100 – 15 ka, records are fragmentary (Newnham et al. 1999). The timing of the majority of rock wren divergences within both lineages at cytochrome b (c. 20 – 90 ka) correspond with the Otiran period of glaciation when at least five glacial cycles are known to have occurred up until the last glacial maximum 25 – 15 ka (Newnham et al. 1999). It is likely that rock wren populations expanded and contracted during glacial-interglacial periods to match the extent of their low alpine - subalpine shrubland habitat.

Implications for management of a threatened species

Whilst genetic variation appears relatively high among rock wren overall, this is likely an artefact of the strong North-South structuring detected. Within lineages, mtDNA nucleotide diversity is an order of magnitude lower (π = 0.0025, 0.0055 for the North and South respectively) than overall (π = 0.0202). Similarly, overall allelic richness

Chapter Three: Population structure and phylogeography

(6.45) was almost twice that of the Northern lineage alone (Table 3.4) and expected heterozygosity across all loci was 0.564, compared with 0.559 in the South and a low expected heterozygosity (0.395) for the Northern lineage. This places genetic variation within the Northern and Southern rock wren lineages within the range of other threatened New Zealand endemic birds, all of which have undergone significant population bottlenecks (Table 3.4).

Table 3.4 Levels of microsatellite genetic variation in threatened New Zealand birds

Species N Np NA He

Tieke/S.I. saddleback1 35 6 1.5 0.13 (Big Island) S.I robin1 203 8 3.63 0.34 Mohua2 155 11 3.18 0.35 Northern rock wren 35 14 3.71 0.40 Takahe3 68 5 2 0.42 Kakapo4 90 30 3.3 0.47 Black stilt/kaki5 20 8 3.3 0.54 Kokako6 8-24 4 3.8 0.56 Southern rock wren 99 13 6.00 0.56

N = sample size, Np = number of polymorphic loci, NA = mean number of alleles 1 per locus, He = expected heterozygosity. Taylor et al. 2007 ; Tracy & Jamieson 20112; Grueber 20053; Robertson et al. 20094; Steeves et al. 20085; Hudson et al. 20006

Lower genetic variation within and greater differentiation among Northern rock wren populations is as predicted when gene flow is limited and genetic variation slowly shifts from within to between subpopulations (Keyghobadi 2007). This pattern may be symptomatic of the naturally more fragmented nature of alpine habitat in the Northern region of the rock wren’s range, or, it may indicate a response to anthropogenic impacts. For example, an increase in the abundance of introduced predators within the matrix surrounding the rock wren’s network of alpine habitat patches may be impeding effective dispersal among populations (Ewers & Didham 2006). Furthermore, the fact that dispersal within the Northern lineage appears more restricted than in the south signals increased vulnerability to the effects of rapid climate change (Hoffmann & Sgró 2011).

Chapter Three: Population structure and phylogeography

It appears that at least a limited level of gene flow is occurring between the two subpopulations, given that several admixed individuals were identified by cluster analysis and a positive pattern of isolation by distance was detected overall (Figure 3.9). However, it is clear that in the past, Northern and Southern populations have remained isolated for many thousands of years leaving a strong genetic signature of divergence, in spite of a present-day lack of geographic separation between the two lineages. Interestingly, a male and female captured together in Aoraki/Mount Cook National Park within the secondary contact zone, assigned to either mtDNA lineage (Figures 3.2 and 3.3), though microsatellite analysis did reveal both individuals as admixed (Figure 3.7). It is not known if this pair subsequently bred and produced young. This lack of current geographic separation between the two lineages and a narrow zone of admixture (< 60 km), combined with the preservation of a strong genetic divergence suggests there may be a behavioural or biological barrier to gene flow between the lineages and further studies focused on this central contact zone are warranted. Lastly, given the findings of this study and the deep divergence between the two lineages, the designation of the Northern and Southern rock wren lineages as separate evolutionarily significant units (ESUs) should be considered (Moritz 1994). Ensuring the conservation of both lineages under future management would maximise the genetic variation present within the species and thus evolutionary potential in the face of environmental change. The development of adaptive markers using genomic data would also elucidate whether there are important adaptive differences (e.g. formed through divergent selection) among the historically isolated rock wren lineages (Allendorf et al. 2010).

Searching boulderfields Murchison Mountains, Fiordland

Chapter 4 Past demography and effective population size of the New Zealand alpine rock wren (Xenicus gilviventris)

Introduction

Achieving robust estimates for the population size and dynamics of a small, elusive yet widely distributed alpine species can be logistically unattainable using traditional ecological methods, such as the marking and recapturing of animals (Lukacs & Burnham 2005). However, this information is important to ensure effective conservation management as it helps to gauge change and predict extinction risk (Luikart et al. 2010). Even more difficult to estimate through the collection of demographic data is the genetically ‘effective’ population size (Ne), which represents not only current census size (Nc; the number of adults) (Luikart et al. 2010), but also the demographic history of the population (Waples 2005). The parameter Ne can be defined as the size of an ‘idealized’ population that would give rise to the same calculated loss of heterozygosity, inbreeding or variance in allele frequencies as the population under consideration (Wright 1931). Effective population size can be influenced by a variety of factors including large variations in individual reproductive success, life-history traits, unequal sex-ratios and fluctuations in population size over generations whereby the smallest population size reached continues to dominate genetic processes (Frankham 1995; Hare et al. 2011; Luikart et al. 2010; Waples et al. 2013).

Consequently, Ne is typically much lower than census size, with an estimated Ne/Nc ratio of just 0.1 (Frankham 1995; but see Palstra & Ruzzante 2008). There are a range of methods now available to estimate contemporary (the past one-to-few generations) effective population size utilising genetic samples collected from a single temporal sampling period (reviewed in: Luikart et al. 2010). Recent genetic data can also be used to estimate the historical or ‘long-term’ effective population size (the past tens-to-thousands of generations) (Palsbøll et al. 2013). Further, in addition to comparing estimates of contemporary and long-term effective population size, demographic changes such as past population bottlenecks and expansion can also be tested for over a range of time scales using genetic data from a single temporal sample (Rogers & Harpending 1992; Zachariah Peery et al. 2012). Gaining this knowledge is valuable for conservation management, as it affords insight

Chapter Four: Past demography and effective population size into past responses to historical processes and indicates whether current abundance and distribution of populations reflects historical status, in the absence of long-term monitoring data or historical samples. New Zealand’s dynamic geological and climatic history has strongly shaped the present-day distribution of its’ endemic biota (Wallis & Trewick 2009). Subsequent human settlement also inflicted dramatic impacts, commencing with colonisation by Polynesians c. 700 yrs bp and compounded by the arrival of Europeans c. 200 yrs bp (McGlone 1989; Worthy & Holdaway 2002). Reductions in species diversity, distribution and abundance occurred both due to direct exploitation and as a result of large scale habitat changes, including the introduction of several predatory mammal species (Duncan & Blackburn 2004). The endemic New Zealand wrens (Acanthisittidae) featured prominently among the losses with at least six species going extinct, leaving only two remaining members alive today, the riflemen (Acanthisitta chloris), which is considered common and widespread, and the relatively rarer and threatened alpine rock wren (Xenicus gilviventris) (Higgins et al. 2001; Robertson et al. 2013). Results from recent population studies indicate that both the abundance and range of rock wren have contracted during the past 100 years, most likely due to the impact of introduced mammals (Michelsen-Heath & Gaze 2007). The rock wren population comprises two highly divergent genetic lineages, thought to have become isolated within Northern and Southern refugia during the Pleistocene glaciations c. 2 million years ago (see Chapter 3). Subsequently, the majority of within-lineage rock wren divergences are estimated to have occurred c. 20 – 90 Ka, (Chapter 3) corresponding with the Otiran period of glaciation from c. 100 – 15 Ka when at least 5 glacial cycles are known to have occurred up until the last glacial maximum 25 – 15 Ka (Newnham et al. 1999). Here, we use recently collected genetic samples to evaluate past population patterns of abundance across different timeframes and obtain estimates of both contemporary and long-term effective population size for each lineage. Specifically, we test the following hypotheses: i) rock wren have expanded from glacial refugia since the last glacial maximum 25 – 15 thousand years ago; ii) rock wren constitute a naturally rare species (i.e. they were not abundant historically); iii) a relatively recent (i.e. within the past c. 100 years) population reduction has occurred in rock wren, likely in response to the impacts of introduced mammalian predators. We also compare the

Chapter Four: Past demography and effective population size effective population sizes of each lineage, highlighting potential implications for conservation management of rock wren.

Materials and methods

Rock wren blood samples (N=134) were collected from throughout their range as described in Chapter 3 (see Figure 3.1). Blood samples (<20 μl) were obtained by brachial vein puncture with a sterile needle (25G x 5/8”) and collection of the blood with a glass capillary tube for immediate transfer into 1 ml of Queens lysis buffer (10 mM Tris, 10 mM NaCl, 10 mM Na-EDTA, 1% n-lauroylsarcosine; pH 7.5) (Seutin et al. 1991). DNA was extracted and purified from the whole blood using a standard 5% Chelex protocol (Walsh et al. 1991).

Microsatellite genotyping

All 134 samples were genotyped at 14 microsatellite loci previously developed for rock wren (see Table 2.1 in Chapter 2); comprising a total of 35 mature individuals from the Northern Lineage and 99 from the Southern Lineage (Chapter 3). The polymerase chain reaction (PCR) conditions used are described in Chapter 2. To check for genotyping error, 10% of all samples were run twice for each locus (Hoffman & Amos 2005). Tests for Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium (LD) were conducted for each population in GENEPOP version 4.1 (Rousset 2008) employing the markov chain method with 10,000 dememorisations, 1,000 batches and 10,000 iterations. Significance levels were adjusted for multiple comparisons following standard Bonferroni correction (Rice 1989).

Mitochondrial DNA sequencing

Twenty individuals from across the range of rock wren were sequenced for a 1075 bp fragment of the mitochondrial cytochrome b gene using primers L14764 and H16064 (Sorenson et al. 1999), as described in Chapter 3. Twenty-five individuals (using the same birds as for cyt-b where possible) were sequenced for a 1126 bp fragment of the mitochondrial control region. Primers L-RW- ND6 and H-RW-12s were used for amplification as described in Chapter 3 and L- ROWR-CR-internal, L-ROWR-CR-1 and R-ROWR-CR-1 for sequencing.

Chapter Four: Past demography and effective population size

All sequences were aligned using progressive pairwise alignment in Geneious (v.6.0.5) and variable sites were confirmed by visual inspection of the chromatograms. Cytochrome b sequences were translated to detect stop codons and reading frame errors (Geneious v.6.0.5, created by Biomatters. Available from http://www.geneious.com).

Nuclear DNA sequencing

Data analysis

Population structure can create spurious bottleneck effects and artificially increase effective population size when several strongly differentiated populations are mistakenly analysed together (Chikhi et al. 2010; Roman & Palumbi 2003). Consequently, we conducted analysis on the Northern and Southern rock wren lineages separately due to the strong genetic differentiation detected between these populations. For the ß-fibint7 dataset, statistics were calculated overall given the lack of population structure identified for this gene (Chapter 3). The parameter generation time can be defined as the average age of reproducing females (Lambert et al. 2009). Generation time in rock wren is unknown (Higgins et al. 2001). Surveys conducted during a study of breeding ecology indicated that rock wrens ‘probably live for around 5 years, possibly more’ (Michelsen-Heath 1989), whilst other reports indicate a longevity of at least 8 years (Heather & Robertson 2000). Maximum longevity for the rock wrens closest living relative, the rifleman (A. chloris) is recorded as 6 years, with an average life expectancy of 2.2 years for males and 1.7 years for

Chapter Four: Past demography and effective population size females (Sherley 1985). To account for the uncertainty around this parameter in rock wren, we calculated Ne over a range of estimates of generation time (2 – 6 years inclusive) extending beyond those used for other small passerines such as the New Zealand robin (Petroica longipes); 5 years (Weiser et al. 2013) or the common rosefinch (Carpodacus erythrinus) 2 years (Hung et al. 2013). For all other demographic methods, we assumed the intermediate value within this range of three years.

Past demography

We tested for a recent population reduction using the M ratio method, whereby M is the ratio of the number of alleles (k) in a population to the range in allele size (r). This method is based on the assumption that during a bottleneck, k is reduced faster than r and therefore M is expected to be smaller in recently reduced populations than those at equilibrium (Garza & Williamson 2001). Where a population returns to its initial size rapidly following a bottleneck, M will reach pre-bottleneck levels again after approximately 100 generations, however when a population remains small, M may take hundreds (>500) of generations to reach its pre-reduction value (Garza & Williamson 2001). We used the software M_P_Val and M_Crit developed by Garza and Williamson (2001), to calculate the average M value across loci and compare this with the critical value (Mc) estimated using 10,000 simulations with the same parameters as the data (samples size, no. of loci, θ = 4Neµ (where Ne = effective population size and µ

= mutation rate), ps = mean proportion of one-step mutations (the program actually uses ps-1) and Δg = mean size of >one-step mutations), but assuming the population to be at drift–migration equilibrium. Each of the simulated M values is ranked and Mc is defined as that which only 5% of the simulation values fall below. A bottleneck is inferred when the M ratio value is below the Mc threshold. For the two unknown parameters, we used those suggested by Garza and Williamson (2001): ps-1 = 0.1 and Δg = 3.5. For the third parameter θ, we used the long-term value for each lineage as derived from the Bayesian analysis in the software IMa2 (Hey 2010a, b) (see below). As suggested by Zachariah Peery et al. (2012), we assessed the effects of different parameter levels on Mc, by performing additional runs using θ values of 0.5, 1, 5 or 10, ps-1 = 0.15, 0.22, or 0.25 and Δg = 2.5, 3, or 4.

Chapter Four: Past demography and effective population size

We also tested for a recent bottleneck within either lineage by examining heterozygosity relative to that expected at mutation-drift equilibrium in the program BOTTLENECK v.1.2.02 (Cornuet & Luikart 1996; Piry et al. 1999). This method is likely to be effective in detecting bottlenecks that have occurred within the past 0.2 –

4.0 Ne generations (where Ne is the bottleneck effective population size) (Luikart & Cornuet 1998). Based on the assumption that the number of alleles decreases faster than heterozygosity following a bottleneck, a bottleneck is indicated when expected heterozygosity (He) is greater than heterozygosity estimated from the number of alleles

(Heq). Accordingly, when He < Heq, then demographic growth or expansion may be inferred (Cornuet & Luikart 1996). We ran the two-phase mutuation model (TPM; Di Rienzo et al. 1994) with 95% single-step mutations and a variance of 12 as recommended by Piry et al. (1999). The TPM is considered the most appropriate mutation model for the majority of microsatellite data (Piry et al. 1999). Estimates were based on 1,000 replicates and the Wilcoxon test was used for statistical analysis as is recommended by Piry et al. (1999) when using less than 20 polymorphic loci. We also tested for an even more recent bottleneck in rock wren (assuming a 3 year generation time, this approach is likely to detect bottlenecks which have occurred within the past c. 100 – 200 yrs) using the mode-shift test in BOTTLENECK (Piry et al. 1999). The method evaluates the frequency distribution of alleles, testing for an allele frequency distortion or ‘mode shift’ whereby bottlenecks shift the distribution to fewer alleles in the low frequency classes and more alleles at an intermediate frequency (Luikart et al. 1998). To test for a signature of rapid demographic expansion, the mismatch distribution (i.e. the distribution of the number of observed nucleotide differences between pairs of haplotypes) was generated using ARLEQUIN (Excoffier et al. 2005) for cytochrome b, the control region and ß-fibint7 datasets separately. The shape of the distribution is expected to be unimodal in populations having undergone rapid population expansion. Where the distribution is bimodal/ multimodal, a population with a more stable demographic history may be inferred (Rogers & Harpending 1992; Slatkin & Hudson 1991). Goodness of fit to the expected model of sudden expansion was tested using the sum of squared deviations (SSD) and raggedness indices computed over 1000 bootstrap replicates in ARLEQUIN v.3.5.1.3 (Schneider & Excoffier 1999).

The time to the purported expansion ( ) can be estimated by the equation

Chapter Four: Past demography and effective population size

(Rogers 1995; Rogers & Harpending 1992; Schneider & Excoffier 1999). We estimated for cytochrome b: ; where µ denotes mutations per site per generation based on a standard mutation rate of a 2.1% sequence divergence per million years, a rate proven to be conserved across a wide range of avian orders for cytochrome b (Weir & Schluter 2008). An intermediate generation time estimate of 3 years was assumed and Tau (τ) was calculated from the mismatch distribution in ARLEQUIN.

Effective population size

Contemporary Ne

We estimated contemporary effective population size (Ne) of both the Northern and

Southern rock wren lineages using several one-sample Ne estimation methods. The most well assessed and widely applied of these is the linkage disequilibrium (LD) method (Hill 1981). This method can be used to estimate Ne based on the principle that as Ne decreases, genetic drift acting on few parental individuals generates non-random associations among alleles at two or more loci, that is, linkage disequilbrium (Luikart et al. 2010). The heterozygote excess method (Pudovkin et al. 1996; Pudovkin et al. 2010;

Zhdanova & Pudovkin 2008) is another commonly used approach for estimating Ne utilizing the principle that when the number of breeders giving rise to the next generation is small, allelic frequencies will be different in male and female parents by chance (because of binomial sampling error), resulting in an excess of heterozygotes in the progeny with respect to expectations under Hardy-Weinberg equilibrium (Pudovkin et al. 1996). One more recently developed Ne estimator is that referred to as the

‘molecular coancestry’ method (Nomura 2008), where an estimation of Ne is obtained from a parameter of allele sharing among the sampled individuals. NeEstimator v.2.0 (Do et al. 2013) implements all three of these one-sample methods to estimate Ne, and includes a correction for bias attributed to a monogamous mating system, known to strongly affect the LD method (Hill 1981; Waples 2006; Waples & Do 2010). To account for this bias, we specified the rock wren mating system as monogamous in the model (Michelsen-Heath 1989). No minimum allele frequency was specified. We also used the program ONeSAMP v1.2 to estimate contemporary Ne (Tallmon et al. 2008). ONeSAMP utilises approximate Bayesian

Chapter Four: Past demography and effective population size computation and the linkage disequilibrium (LD) method plus seven other summary statistics, and is considered less biased and more precise (in theory) to conventional LD methods (Luikart et al. 2010). Analyses were submitted online to the ONeSAMP server

(http://genomics.jun.alaska. edu/asp/Default.aspx), specifying a minimum Ne value of 2 and a maximum of 500; these values were selected based on the estimates of Ne obtained using NeEstimator.

Long-term Ne

We estimated historical (long-term) effective population size (Ne) using a Bayesian algorithm implemented in the software IMa2 (Hey 2010a, b). Our dataset consisted of 35 individuals genotyped from the Northern population and 99 individuals genotyped from the Southern population. Given the strong differentiation between the two populations, we estimated Ne for each population separately, as unaccounted for substructure can lead to increased values of theta (, the product of Ne and mutation rate µ, see below) and biased results (Roman & Palumbi 2003; Rosenbaum et al. 2002). Estimates of  were obtained using a stepwise mutational model consisting of 2,000,000 steps and a 10,000 step burn in period (a total of 20,000 genealogies were saved, as recommended in the manual). Metropolis coupling was implemented using 40 hot chains with a geometric incremental heating model and the first and second heating parameters set to 0.975 and 0.750 respectively. Priors for  were empirically derived over a series of short, exploratory runs until clear unimodal posterior distributions were achieved. Mixing of the chains was assessed using a combination of the update rates, trend plots, ESS values and by comparisons within and among five independent runs. The upper and lower bounds of the estimated 95% highest posterior density (HPD) interval are reported together with the mean value of . Given that results from repeat runs were largely congruent, we only present the results from the first independent run. Assuming that the populations are in mutation-drift equilibrium (Curtis et al. 2011), effective population size can be estimated using  by rearranging the equation

 = 4Neµ:

Chapter Four: Past demography and effective population size

We estimated µ based on a range of mutation rates used for passerines in previous studies (Qu et al. 2011) and also over a range of estimates of generation time (2 – 6 years inclusive, see start of methods section). Therefore, µ = (generation time × (1 × 10- 4 or 10-5)).

Results

The mean genotyping error rate across all loci was low, with 0.0131 of alleles resulting in an allelic mismatch. There was no evidence for linkage disequilibrium among the 14 polymorphic microsatellite markers. Two microsatellite loci, Xgilv 19 and 27 showed significant departure from Hardy-Weinberg equilibrium (HWE) with heterozygote deficiency in both the Northern and Southern lineages. These deviations from HWE were not likely to be the result of null alleles, but rather due to finer-scale genetic structuring within each lineage, given that both loci are in equilibrium when tested within localised populations (see Chapter 3).

Past demography

We found no evidence for a recent bottleneck in either lineage using the likelihood- generated  values to calculate M ratios. The average M ratio across loci for the North was 0.832 and for the South, 0.797. In both cases, these values were higher than the calculated critical value (0.803 and 0.765 for the North and South respectively) and the critical value of 0.68 provided by Garza and Williamson (2001) meaning there is no evidence of a significant recent reduction in population size. Across all additional parameter values tested, the M ratio remained higher than the critical value, in the North. In the Southern models however, under lower parameter values of  (0.5 and 1, thus in a scenario of a lower pre-reduction Ne) and a higher Δg value of 4, a significant population reduction was indicated. Significant heterozygosity excess, a signature of a recent bottleneck, was not detected within either the Northern or Southern population (P > 0.05), though significant heterozygosity deficiency was detected in the Northern population (P = 0.02). We found no signature of recent bottleneck when testing for a mode shift in allele frequency, with both lineages exhibiting a normal L-shaped allele frequency distribution (Luikart et al. 1998). Thus overall, we did not detect a signal of recent

Chapter Four: Past demography and effective population size population reduction within the Northern lineage, though there is some evidence for a recent population bottleneck within the Southern lineage. The mismatch distribution showed a reasonable fit between the observed frequencies of pairwise differences and the expected frequencies under a rapid expansion model, with the Northern population exhibiting a unimodal distribution using mtDNA (Figure 4.1). However, the pattern for the Southern population appeared bimodal, which is not consistent with population expansion (Figures 4.1b and 4.1d). Despite the poorer fit for the Southern population, the SSD and Raggedness indices were almost all non-significant (Table 4.1), and therefore the null hypothesis of rapid population expansion could only be rejected for the Southern population control region sample based on a significant SSD p-value (0.023).

Table 4.1 Indices of population expansion in Rock Wren populations with sample number (n), squared sum of deviances (SSD), raggedness index (RI) and associated P-values. P-values in bold support population expansion.

n SSD p RI p Cytochrome b North 8 0.051 0.203 0.202 0.169 South 12 0.059 0.055 0.092 0.194 Control region North 11 0.024 0.411 0.052 0.355 South 14 0.172 0.023 0.017 0.994 β-fibrinogen Overall 28 0.017 0.451 0.036 0.498

Chapter Four: Past demography and effective population size

a) Cytb: North b) Cytb: South

c) Cntrl reg: North d) Cntrl reg: South

Frequency

e) ß-fibrinogen

)

Number of pairwise differences

Figure 4.1 Mismatch distributions in rock wren (Xenicus gilviventris) using cytochrome-b a) North and b) South, the mitochondrial control region c) North and d) South and e) ß-fibrinogen. Bars represent the observed frequencies of pairwise differences and the dashed line the expected frequencies under a model of rapid population expansion.

Chapter Four: Past demography and effective population size

Effective population size

Contemporary Ne

Estimates of contemporary Ne using the LD method in ONeSamp and NeEstimator were largely congruent, with a higher Ne evident for the Southern lineage compared with the Northern one (Table 4.2). However, the calculated 95% CIs indicate the difference between lineages is only significant using the LD method implemented in NeEstimator (Table 4.2). The heterozygote excess method in NeEstimator was unable to be calculated for either lineage and returned infinite estimates for Ne. The molecular co-ancestry method of Nomura (2008) provided an estimated effective number of breeders of 11 in the North and 26 in the South, however precision was poor with the associated CIs overlapping (Table 4.2).

Table 4.2 Estimates of contemporary effective population size (Ne) of Northern and Southern rock wren (Xenicus gilviventris) lineages using one-sample methods. In brackets, 95% credible limits for the posterior distribution (ONeSamp) and 95% confidence intervals (NeEstimator).

ONeSamp LD – NeEstimator Mol. Coancestry- NeEstimator

North 57 (37-156) 59 (35 - 122) 11 (1 - 33) South 166 (107 - 403) 236 (167 - 371) 26 (2 - 82)

Long-term Ne

Long-term effective population size was higher for the Southern lineage than the Northern lineage across all generation times and mutation rates (Table 4.3, Figure 4.2).

The difference in Ne between lineages was statistically significant (Figure 4.2). Given that the estimated effective size differs by a factor of ten depending on the mutation rate used, the actual long-term population estimates for rock wren remain indeterminate.

Chapter Four: Past demography and effective population size

Table 4.3 Estimates of long-term effective population size (Ne) of Northern and Southern rock wren (Xenicus gilviventris) lineages across a range of estimated generation times and two mutation rates, averaged over 14 microsatellite loci using the software IMa2. Theta values shown in brackets.

Generation time µ North South (years) (0.2437) (2.170)

2 2 × 10-4 305 2,713 2 × 10-5 3,046 27,125 3 3 × 10-4 203 1,808 3 × 10-5 2,031 18,083 4 4 × 10-4 152 1,356 4 × 10-5 1,523 13,563 5 5 × 10-4 122 1,085 5 × 10-5 1,219 10,850 6 6 × 10-4 102 904 6 × 10-5 1,015 9,042

Chapter Four: Past demography and effective population size

e N

Generation time (years)

Figure 4.2 Long-term (historical) effective population size across different estimates of generation time for rock wren (Xenicus gilviventris) a) assuming a mutation rate of 10-4 and b) assuming a mutation rate of 10-5. Error bars represent 95% confidence interval.

Chapter Four: Past demography and effective population size

Discussion

Past demography

We found limited evidence of a recent population bottleneck within the Southern lineage and no signal of a recent population bottleneck within the Northern lineage. Statistical power to detect bottlenecks has been shown to be very sensitive to model parameters (Zachariah Peery et al. 2012). For example, the mean proportion of one-step mutations (Ps), is typically much higher than that suggested by Garza and Williamson

(2001) for the M ratio method and underestimating Ps can result in a higher probability of incorrectly detecting bottlenecks in stable populations (Zachariah Peery et al. 2012). Given that we did not detect a bottleneck across a range of parameters, including the higher parameter values for Ps recommended by Zachariah Peery et al. (2012), the possibility of a recent bottleneck in rock wren remains uncertain. Furthermore, power is also low for lower values of , even where large reductions in Ne have taken place, and when sampling is undertaken <5 generations after the bottleneck (Zachariah Peery et al. 2012). Using the heterozygosity excess method, signature of a genetic bottleneck was not detected. The detection of significant heterozygosity deficiency within the Northern population suggests that this population is not at mutation-drift equilibrium and may have either recently expanded in size, or, experienced a recent influx of rare alleles through immigration from the genetically distinct Southern population, thereby mimicking population expansion (or obscuring any population reduction) (Luikart & Cornuet 1998). However, the mismatch distribution supports a hypothesis of rapid population expansion, with a stronger signal for the Northern population. It is important to consider that the genetic signature of the M ratio and heterozygosity excess tests is relatively short-lived. When considering a generation time for rock wren of 3 years for example, the signal would be retained for c. 300 - 1500 years for the M ratio method, depending on the length and intensity of the bottleneck. For the heterozygosity method, where the period of time the bottleneck can be detected for is equal to c. 0.2 – 4.0 Ne generations (where Ne is the bottleneck effective population size), the signal is expected to be retained for a shorter length of time. For example, if we consider a post-bottleneck Ne of c. 50 for the North and c. 200 for the South, this would provide a period of c. 10 – 200 years in the North and 40 – 800 years in the South, up until which the signal may be detected. This period

Chapter Four: Past demography and effective population size corresponds to early European settlement in New Zealand and a bottleneck during this time could be attributed to anthropogenic disturbance. Even when considering a longer generation time of 6 years, the signal would still be retained for a maximum of c. 3000 years. In contrast, the mismatch distribution retains a much more ancient signature of population expansion. Using the values of tau estimated from the mismatch distribution (τ = 2.8 for the North and τ = 3.7, South), we estimate the timing of demographic expansion at 13,781 yrs bp for the North and 18,211 yrs bp for the South. This timing is consistent with the retreat of the South Island glaciers around 15 - 14 ka (Newnham et al. 1999).

Effective population size

Contemporary Ne

All three methods estimating contemporary effective population size were consistent in estimating a larger Ne for the Southern lineage. However, given this result was only significant using the LD method implemented in NeEstimator, it must be interpreted with caution. The fact that an estimate of Ne using the heterozygote excess method was unable to be calculated for the rock wren is somewhat expected given that this approach is generally only useful for polygamous and polygynous populations where the number of breeders is very small (<20) unless both sample size and number of alleles is high (> 60 individuals and >20 multiallelic loci) (Luikart & Cornuet 1999). Similarly, precision was poor for the molecular co-ancestry method, indicating that this method also may not be usefully applied to the rock wren population. As for the heterozygote excess method, the molecular co-ancestry method is generally best applied to populations where the effective number of breeders is very small (Luikart et al. 2010).

Long-term Ne

Estimates of long-term Ne were again consistently larger for the Southern lineage, indicating that differences between rock wren abundance in the North and the South are rooted within a historical timeframe. This may simply imply that habitat in the Southern South Island supports a higher density of rock wren than in the North and this has been the case even prior to any potential anthropogenic impacts. Alternatively, it 70

Chapter Four: Past demography and effective population size may be that the Northern rock wren were confined within much smaller refugia during the Pleistocene glaciations. Estimates were highly sensitive to the parameter values used, for example the effect of a specified generation time of 2 versus 6 years is a three-fold decrease in Ne. Similarly, the estimated effective size differs by a factor of ten depending on which of the commonly used microsatellite mutation rates is used. Given these considerations, the point estimates of long-term Ne should be regarded with caution and rather, the relative differences in long-term effective size between the two lineages and the dramatic discrepancy between long-term and contemporary population estimates should be noted.

Conservation implications

Whilst the difference in effective population size between the Northern and Southern rock wren lineages appear to be historical, the magnitude of difference between the contemporary and long-term Ne suggests that rock wren in both lineages were in the past, much more abundant than they are today. We were unable to detect a genetic signature linking population decline within the Northern lineage to a timeframe of anthropogenic disturbance, though there is some evidence for a recent population bottleneck within this timeframe in the South. This indicates that although natural historical climate fluctuations have clearly played an important role driving patterns of rock wren abundance in the past, these impacts may now be being compounded by much more recent anthropogenic impacts, most likely, predation by introduced mammalian predators.

If we apply a Ne/Nc ratio of 0.1 to the Ne estimates obtained for rock wren (Frankham 1995), then current census sizes approximate 600 rock wren in the Northern lineage and 2400 rock wren in the South, based on estimates using the LD method in NeEstimator. This represents the lower end of the census size estimated for rock wren in the IUCN threatened species lists (2,500-10,000 mature individuals) (Birdlife International, 2012). It has been proposed, as a rule of thumb, that the effective population size should not fall below 50 in the short term to reduce the risk of extinction due to the harmful effects of inbreeding depression and in the long term, should be at least 500 to retain sufficient genetic variation for adaptive potential (Franklin 1980). However it is

Chapter Four: Past demography and effective population size now generally acknowledged that the 50/500 rule should be regarded as ‘a guiding principle’ as opposed to an absolute threshold for assessing minimum viable population size (Jamieson & Allendorf 2012). Nevertheless, genetic concerns are clearly a relevant issue for rock wren management given that estimates of contemporary effective population size for both lineages are well below 500 and in the North, effective size is estimated at <50. Furthermore, given that less gene flow and hence lower connectivity was detected within the Northern lineage (Chapter 3), it is unlikely that this lineage represents a coherent ‘population’; consequently, the prognosis for the Northern lineage may actually be worse. In contrast, long-term estimates of effective population size of both lineages are well above these levels across most estimates, indicating that the carrying capacity of rock wren may have been a lot higher historically than it is today. Given that adaptive potential is particularly important during periods of rapid climate change, particularly for alpine habitat specialists such as the rock wren which are often more sensitive to temperature increases, the low effective population sizes detected here will likely constrain adaptive response to current anthropogenic climate change (Hoffmann & Sgró 2011; Warren et al. 2001). Further, as avian diseases continue to expand in distribution throughout the range of rock wren (Howe et al. 2012;

Tompkins & Gleeson 2006), low estimates of Ne may render rock wren populations highly susceptible to novel disease outbreaks. These issues warrant further attention.

Mist netting rock wren Xenicus gilviventris Murchison Mountains, Fiordland

Chapter Five: Fine-scale population structure

Chapter 5 Identifying populations for management: fine-scale population structure in the New Zealand alpine rock wren (Xenicus gilviventris)

Introduction

Understanding the dispersal dynamics and connectivity among populations of threatened species is crucial in determining the appropriate units and spatial scale for conservation management (Funk et al. 2012; Waples & Gaggiotti 2006). Genetic structuring is typically low in birds and expected to occur only at large geographic spatial scales due to a high capacity for dispersal (Abbott et al. 2002; Crochet 2000). In species where flight is limited however, we might expect fine-scale structuring to occur over relatively smaller spatial scales of just a few kilometres, resembling more the spatial structure of small mammals (Baker et al. 1995; Peakall et al. 2003). Furthermore, avian species that occupy naturally fragmented habitats, such as the alpine ‘sky islands’ (Bech et al. 2009) have been demonstrated to exhibit strong genetic structuring (Caizergues et al. 2003; Segelbacher & Storch 2002). The spatial heterogeneity of the alpine landscape and patchiness of suitable habitat may also lead to source-sink population dynamics (O'Keefe et al. 2009). In ‘sink’ populations, reproduction is insufficient to balance local mortality, however these populations may be sustained through immigration from more productive ‘source’ populations (Pulliam 1988). Source-sink dynamics have important implications for conservation management and can be recognised through asymmetries in dispersal and the effects on genetic variation within and among populations (Andreasen et al. 2012; King et al. 2000). Genetic structuring may also arise from a sex-bias in dispersal. Several hypotheses have been proposed to account for differences in dispersal behaviour between the sexes: competition for resources (Greenwood 1980), competition for mates (Perrin & Mazalov 2000) and inbreeding avoidance (Pusey 1987). However, the best predictor of sex-biased dispersal can be derived from a species’ mating system. In birds, which are usually monogamous, females tend to disperse further from their place of birth whilst males exhibit stronger philopatry (i.e. remain at their natal site to breed) (Greenwood 1980). Consequently, it can be expected that stronger fine-scale genetic structuring will generally be detected for males in birds.

Chapter Five: Fine-scale population structure

Here, we investigate the spatial structuring of rock wren (Xenicus gilviventris Pelzeln, 1867), a threatened alpine passerine belonging to the endemic New Zealand wren family (Acanthisittidae). Rock wren constitute a widespread, naturally fragmented population, occurring in patches of suitable habitat over c.900 m in altitude throughout the length of the South Island, New Zealand (Higgins et al. 2001). Several areas of high population density have been identified in localised areas of the lower South Island; areas which have consequently become the focus for conservation management efforts. However, the genetic status of these ‘sub-populations’ and the extent to which they are connected by dispersal has until now, been entirely unknown. By gaining an understanding of the spatial structuring of rock wren and patterns of fine-scale gene flow, we aim to identify appropriate populations for conservation management. Phylogeographic analysis in Chapter 3 identified two genetically distinct lineages of rock wren, each spanning approximately 370 km, within which only relatively shallow-sub-structuring was detected. This finding, combined with a strong pattern of isolation by distance among populations (Chapter 3) suggests dispersal between geographically isolated areas is occurring, in spite of the rock wrens poor flight ability. Rock wren are sexually monogamous, maintaining territories ranging in size from c. 0.6 – 4.2 hectares (around 1.4 hectares on average) which vary little between seasons (Michelsen-Heath 1989). Typical habitat comprises of low alpine and subalpine shrubland, herb fields, scree slopes, boulder fields and rocky bluffs (Michelsen 1982). Juveniles establish territories and find mates by the end of the summer in which they fledge, ready to commence breeding the following spring (Michelsen-Heath 1989). Michelsen-Heath (1989) reported that of 75 juveniles banded as nestlings within the Murchison Mountains, Fiordland 75% remained within 500 m of their place of fledging with the remainder not re-sighted (sex bias and time interval were not recorded). Very little is known with regard to the migration pathways of rock wren, though they have almost never been recorded below the treeline (pers. obs). Using 13 polymorphic microsatellite DNA loci, we apply both individual and population based analyses to investigate fine-scale patterns of dispersal in rock wren. We aim to (i) establish the spatial scale of non-random genetic structure, i.e. the ‘genetic patch size’ of rock wren; (ii) investigate source-sink dynamics; iii) identify the most appropriate populations for conservation management in the lower South Island.

Chapter Five: Fine-scale population structure

Materials and methods

Sampling and DNA extraction

Rock wren blood samples (n =101) were collected from throughout the lower South Island with a maximum distance between samples of approximately 265 km (Figure 5.1). All samples were collected from 2010-2012 during the austral summer breeding season. As the rock wren has limited flight capability (Higgins et al. 2001), birds were attracted using audio playback and captured by ‘herding’ them into mist nets. The capture location of each individual was directly recorded using a Garmin handheld GPS. Resampling of birds was prevented by tagging each bird with a single metal leg band issued by the New Zealand Department of Conservation as part of the national bird banding scheme. Nestlings or fledglings as identified in obvious family groups were not sampled. Blood samples (<20 μl) were obtained by brachial vein puncture with a sterile needle (25G x 5/8”) and collection of the blood with a glass capillary tube for immediate transfer into 1 ml of Queens lysis buffer (10 mM Tris, 10 mM NaCl, 10 mM Na-EDTA, 1% n-lauroylsarcosine; pH 7.5) (Seutin et al. 1991). DNA was extracted and purified from the whole blood using a standard 5% Chelex protocol (Walsh et al. 1991).

Chapter Five: Fine-scale population structure

Hr C

M E

R Mc

Figure 5.1 Sampling locations of rock wren (Xenicus gilviventris) throughout the mountainous regions of the lower South Island. Sampling areas are labelled: Hr = Haast Range n = 17; C = Lake Crucible n=4; Ma = Matukituki n=6; H= Upper Hollyford n = 25, Lake Adelaide n = 6, Lake Marion n = 2 and Monkey Creek n= 4; M = Murchison Mtns n=14; E = Eyres n=2; R = Lake Roe n=16; Mc = Lake MacArthur n = 5.

Chapter Five: Fine-scale population structure

DNA sexing

Although rock wren are sexually dimorphic (Higgins et al. 2001), the differences can be subtle and vary throughout their range (pers. obs). Therefore, the sex of each bird was reaffirmed using a DNA test. In birds, females are heterogametic, i.e. have two different sex chromosomes (ZW) and males are homogametic (ZZ). Consequently, detection of the female-specific W chromosome differentiates the sexes (Dubiec & Zagalska-Neubauer 2006; Ellegren & Sheldon 1997). We used the 2550F/2718R primer pair developed to amplify sex-specific versions of introns in the chromo-helicase-DNA binding protein (CHD) (Fridolfsson & Ellegren 1999). PCRs were performed in a 10 µl reaction containing 10 - 30 ng of template DNA, 1 × PCR buffer, 1.5mM of MgCl2, 200 µM of each dNTP, 0.5 unit of Taq polymerase (Bioline USA, Inc, Randolph MA

02368-4800) and 0.5 pmol of each primer (Robertson & Stephenson 2008). The thermal cycling parameters were an initial three minute denaturation at 94°C, followed by a touchdown of 10 cycles at 94°C/15 sec, 55→45°C/25 sec (-1°C per cycle) and 72°C/35 sec, then 25 cycles at 94°C/15 sec, 45°C/25 sec and 72°C/35 sec. PCR products were separated in 2% agarose gels, run in standard TAE buffer, and visualized using SYBR® Safe DNA Gel Stain (Life Technologies).

Microsatellite genotyping

All 101 samples were genotyped at 13 microsatellite loci previously developed for rock wren (see Chapter 2 Table 2.1, Xgilv25 was not used as it is monomorphic among southern rock wren). The polymerase chain reaction (PCR) conditions used were as described in Chapter 2. To check for genotyping error, 10% of all samples were run twice for each locus. The error rate per allele was calculated by dividing the number of mismatched alleles by the total number of repeat-genotyped alleles (Hoffman & Amos 2005).

Genetic variation within populations

Tests for Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium (LD) were conducted in GENEPOP version 4.1 (Rousset 2008) employing the markov chain method with 10,000 dememorisations, 1,000 batches and 10,000 iterations.

Chapter Five: Fine-scale population structure

Significance levels were adjusted for multiple comparisons among loci following standard Bonferroni correction (Rice 1989). Levels of genetic variation were evaluated for each of the four populations where n > 10; Haast Range, Upper Hollyford, Murchison Mountains and Lake Roe. Mean number of alleles per locus and observed/expected heterozygosity (Ho/He) were calculated using GenAlEx v.6.41 (Peakall & Smouse 2006). Allelic richness, a measure of allelic diversity that accounts for differences in sample size, and the inbreeding coefficient FIS were calculated in FSTAT v.2.9.3 (Goudet 2001). This software was also used to calculate the significance of FIS values based on a significant deviation from HWE (i.e. heterozygote deficiency for positive values) over 1300 randomisations. Significance levels were adjusted for multiple comparisons among loci following standard Bonferroni correction (Rice 1989).

Migration between populations

To assess gene flow among populations, recent migration rates (over the last 2 – 3 generations) were estimated using Bayesian inference implemented in BayesAss 3.0 (Wilson & Rannala 2003). Rates are provided as the proportion of individuals in population one that are migrants derived from populations one through to four, per generation. The programme was run with a Bayesian Markov Chain Monte Carlo (MCMC) length of 10,000,000 iterations, a burn-in period of 2,000,000 iterations and sampling every 100 generations. Various mixing parameter values were trialled to achieve optimal acceptance rates of 20 – 60% for migration rates (m), allele frequency (a) and inbreeding coefficients (f); in the final model parameter values were set at: m = 0.3, a = 0.5 and f = 0.6. Convergence was confirmed through the assessment of the plotted log-probability using Tracer 1.5 (Rambaut & Drummond 2007) and by comparing the posterior mean parameter estimates amongst multiple runs for concordance.

Spatial structure

To test for a positive correlation between genetic relatedness and linear (Euclidian) geographic distance (isolation by distance or ‘IBD’), mantel tests were performed using coefficients of genetic relatedness (Queller & Goodnight 1989) between each pair of

Chapter Five: Fine-scale population structure individuals sampled. Significance of correlations was estimated over 9,999 replicates using GenAlEx v.6.41 (Peakall & Smouse 2006). We also tested for spatial autocorrelation, the correlation of genotypes between pairs of individuals separated by similar geographic distances or which fall into specified ‘distance classes’, using a multivariate approach in GenAlEx v.6.41 (Peakall & Smouse 2006). Twelve arbitrary distance classes were constructed to contain approximately equal numbers of samples. A correlogram is produced showing the autocorrelation coefficient r as a function of distance class. Where fine-scale structure exists, r is typically greater than zero for shorter distance classes and less than zero for larger distance classes. The lowest distance at which the correlogram crosses the x-axis may be interpreted as the extent of non-random (positive) genetic structure or ‘genetic patch size’ i.e., the distance at which samples become genetically independent (Smouse & Peakall 1999). However, the authors caution that interpretation of the x-intercept in this regard needs to consider the interplay of the distance class sizes chosen, sample sizes for each distance class and the true (but unknown) extent of genetic structure. For this reason, we experimented with different distance class intervals to ensure approximately equal but adequate sample sizes and that our sampling intervals were not greater than the scale of genetic structure. Two methods are used to test the significance of the autocorrelation in GenAlEx: random permutation and bootstrap estimates of r. Random permutations (9,999) are used to generate a distribution of permuted (rp) values under the assumption of no spatial structure, from which a 95% null confidence interval is calculated. If the observed r value falls outside of this confidence interval, then significant spatial genetic structure can be inferred. Bootstrapping is used to place a confidence interval around each observed estimate of r. The bootstrap autocorrelation coefficient (rbs) is calculated for each of 9,999 bootstraps and when the estimated bootstrap confidence interval does not straddle r = 0, significant spatial structure is inferred. The overall significance of spatial autocorrelation across all distance classes is also tested over 9,999 permutations under the null hypothesis that the entire correlogram is ‘flat’ against the alternative hypothesis that it is not. Nonparametric heterogeneity tests were also performed to test for significant differences in spatial genetic structure patterns between the sexes (Smouse et al. 2008). The tests are performed at each distance class where the null hypothesis predicts overlap in the 95% CI’s between the sexes and also at the whole

Chapter Five: Fine-scale population structure correlogram level under the null hypothesis that the correlograms from both sexes are homogeneous. Mean pairwise relatedness among individuals within each population was also calculated in GenAlEx using the Queller and Goodnight (1989) index of relatedness (r). The overall index ranges between -1 and 1, with values of r approximating 0.5 for parent-offspring or full-siblings, c.0.25 for half-siblings and zero for unrelated individuals (Queller & Goodnight 1989). The 95% confidence intervals about the mean values were estimated using 10,000 bootstraps. To further investigate genetic structure and the geographic location of any genetic discontinuities within the lower South Island, we utilized the Bayesian clustering program Geneland (Guillot et al. 2005). Geneland uses the Individual GPS coordinates as prior information to assign individuals probabilistically to genetic clusters (K). We tested for K = 1 - 7 (seven being the number of broad geographic areas sampled) over 1,000,000 Markov Chain Monte Carlo (MCMC) iterations. The uncorrelated allele frequency model was used as this model is less prone to algorithm instabilities and departure from model assumptions (e.g. presence of isolation-by-distance) (Guillot 2012). Ten independent runs were performed to assess the results for consistency across runs in terms of the estimated value of K. Post processing was then carried out on the run with the highest average posterior probability to compute the posterior probability of population membership for every pixel in the spatial domain (150 x 250 pixels, 2,000 iterations discarded as burn-in).

Tests for sex-biased dispersal

We conducted five independent tests for sex-biased dispersal using FSTAT v.2.9.3

(Goudet 2001). Firstly, we compared values of Weir & Cockerham’s (1984) FST estimator ST among populations for males and females. It is expected that the dispersing sex will have a lower ST given that dispersal homogenises allelic frequencies. Differences in the inbreeding coefficient FIS between the sexes were also tested for. The dispersing sex should display a higher FIS than the more philopatric sex indicative of greater genetic structuring among sub populations (Goudet et al. 2002). This underlying genetic structure leads to a heterozygote deficit known as the ‘Wahlund effect’ and produces an effect analogous to inbreeding (Goudet et al. 2002). This is because individuals sampled from a single location will be a mixture of both

Chapter Five: Fine-scale population structure residents and immigrants in the dispersing sex, meaning that homozygotes will be more frequent and heterozygotes less frequent than expected based on the calculated allele frequency for the combined population. FSTAT v.2.9.3 calculates Hamilton's (1971) index of relatedness r = 2FST/(1 + FIT), using an estimator strictly equivalent to Queller and Goodnight's (1989). Logically, relatedness amongst members of the dispersing sex is predicted to be lower (Goudet et al. 2002). Lastly, we calculated the Assignment Index statistic (AI) and tested for differences in the mean (mAIc) and variance (vAIc) between the sexes. AIc values are centred on zero and calculate the probability for each genotype to be represented in the sampled population. Resident individuals are likely to have genotypes with an above-average likelihood of occurring in the sample, which should lead to positive AIc values. Consequently, mAIc should be higher in the philopatric sex and vAlc should be higher in the dispersing sex because of the inclusion of both immigrant and resident genotypes (Helfer et al. 2012; Lampert et al. 2003).

Results

The mean genotyping error rate across all 13 loci was low, with 0.0116 of alleles resulting in an allelic mismatch. There was no evidence for linkage disequilibrium between any pair of loci, or significant consistent departure from HWE for any locus after significance levels were adjusted for multiple pairwise comparisons using the sequential Bonferroni correction.

Genetic variation within populations

Levels of genetic diversity were similar among the four populations (He: 0.524 – 0.557), however the Haast Range population exhibiting notably lower levels of allelic diversity (Table 5.1). Values of FIS were positive for all populations, indicative of inbreeding, with the highest values for the Haast Range and Murchison Mountain populations, though this value was only significantly different from zero for the Haast Range (Table 5.1).

Chapter Five: Fine-scale population structure

Table 5.1 Genetic variation in rock wren (Xenicus gilviventris) within four lower South Island populations using 13 microsatellite markers

N Na NR Ho He FIS Haast Range 17 3.39 3.31 0.454 0.557 0.214* Upper Hollyford 25 4.54 4.19 0.520 0.553 0.081 Murch Mtns 14 4.15 4.15 0.478 0.524 0.124 Lake Roe 16 4.46 4.34 0.534 0.531 0.027

N = sample size, Na = mean number of alleles, NR = allelic richness, Ho = observed heterozygosity, He = expected heterozygosity, Fis = Inbreeding coefficient, *Significant deviation from Hardy–Weinberg equilibrium (p ≤ 0.05) after Bonferroni correction

Migration between populations

The mean posterior probabilities of migration rates showed that the majority of individuals originated from the population within which they were sampled, with the Haast Range having the highest proportion of non-migrants (m = 0.94) (Table 5.2). The Murchison Mountain population was the only significant source of migrants for all populations (m = 0.15 – 0.27), though migration was asymmetrical with almost no gene flow into the Murchison Mountains from any other population (Table 5.2).

Table 5.2 Inferred migration rates among four rock wren (Xenicus gilviventris) populations sampled from the lower South Island generated using BayesAss 3.0.3

Putative source population Sample location Haast Range Upper Holly Murch Mtns Lake Roe Haast Range 0.94 (0.03) 0.02 (0.02) 0.03 (0.02) 0.02 (0.02) Upper Holly 0.04 (0.03) 0.80 (0.05) 0.15 (0.05) 0.01 (0.01) Murch Mtns 0.03 (0.02) 0.03 (0.03) 0.91 (0.04) 0.04 (0.03) Lake Roe 0.02 (0.02) 0.03 (0.02) 0.27 (0.03) 0.68 (0.02)

Values are the inferred (posterior mean) migration rates per generation, standard deviation of the marginal posterior distribution for each estimate are in parentheses. The receiving populations are listed in rows and columns represent populations from which migrants are derived.

Chapter Five: Fine-scale population structure

Spatial structure

Across the lower South Island, a weak but significant IBD relationship of decreasing pairwise relatedness over increasing geographic distance was detected (R² = 0.080, P < 0.001) (Figure 5.2).

1.0 y = -0.0009x + 0.0727 0.8 R² = 0.080 0.6 0.4

0.2 r 0.0 -0.2 -0.4 -0.6 -0.8 0 50 100 150 200 250 300 Geographic distance (km)

Figure 5.2 Isolation by distance relationship between pairwise relatedness and geographical distance of rock wren (Xenicus gilviventris) across the lower South Island. The trend, represented as a solid line, was significant (p < 0.001) as determined through Mantel tests with 9,999 permutations.

Spatial autocorrelation was highly significant (P < 0.001) overall indicating that rock wren are non-randomly distributed across the landscape (Figure 5.3). Autocorrelation is positive and significant (P < 0.001) for the first six distance classes thus providing evidence that individuals < 50 km apart are on average more genetically similar than spatially random pairs. The correlogram then crosses the X-axis at c.70 km, and individuals separated by > 70 km are less genetically similar than expected at random (Figure 5.3). The heterogeneity tests revealed significant overall differences among the correlograms of males and females (P = 0.01). Males exhibited higher positive autocorrelation than females up to distances of 25 km (Figure 5.4). However, these differences were only significant at the 1 – 2 km and 5 – 10 km distance classes (P < 0.05). Beyond 25 km, the pattern generally reversed with males exhibiting lower levels of genetic similarity than females, though differences were only significant at the 70 – 100 km distance class (P < 0.05). 85

Chapter Five: Fine-scale population structure

0.20

0.15

0.10

0.05

r r 0.00 U -0.05 L -0.10

-0.15 1 2 5 10 25 50 70 100 125 150 200 265 Distance Class (km)

Figure 5.3 Correlogram showing auto correlation across distance classes for rock wren (Xenicus gilviventris) sampled in the lower South Island. The endpoint of each distance class is labelled on the x-axis. Error bars represent bootstrapped 95% confidence intervals about mean r values. Dashed lines represent upper (U) and lower (L) 95% confidence limits under the null hypotheses that genotypes are randomly distributed across the landscape.

0.25 0.20 0.15

0.10

r 0.05 0.00 M -0.05 F -0.10 -0.15 1 2 5 10 25 50 70 100 125 150 200 265 Distance Class (km)

Figure 5.4 Correlogram comparing auto correlation across distance classes for male and female rock wren (Xenicus gilviventris) sampled in the lower South Island. The endpoint of each distance class is labelled on the x-axis. Error bars represent bootstrapped 95% confidence intervals about mean r values.

Levels of inter-individual relatedness differed significantly among populations (Figure 5.5). The Haast Range, Murchison Mountain and Lake Roe populations all exhibited

Chapter Five: Fine-scale population structure relatively high levels of relatedness (mean = 0.090 – 0.111), whilst individuals within the Upper Hollyford were unrelated (mean = 0.025, 95 % CI -0.002 - 0.029).

0.20

0.15

0.10

r 0.05

0.00

-0.05

-0.10 Haast Range Upper Hollyford Murchison Mtns Lake Roe

Figure 5.5 Mean pairwise genetic relatedness estimates for four rock wren (Xenicus gilviventris) populations in the lower South Island. Error bars are the 95 % CI as determined by bootstrapping. Upper Hollyford n = 25; Haast Range n = 17; Murchison Mtns n = 14; Lake Roe n = 16

Chapter Five: Fine-scale population structure

The Bayesian cluster analysis using Geneland showed strongest support for two genetic populations (K=2). Figure 5.6a shows good mixing around K = 2 and Figure 5.6b shows a clear mode at this value. Cluster assignments corresponded with spatial locations (Figure 5.7). Cluster 1 (green) encompassed all individuals from the Haast Range (and Mt Aspiring National Park); Cluster 2 (white) encompassed all other populations in the lower South Island. Individual assignment probabilities to each cluster were high (0.99 on average).

a) b)

0 1000000 1 2 3 4 5 6 7

Number of MCMC K

Figure 5.6 Number of rock wren (Xenicus gilviventris) genetic populations simulated from the posterior distribution generated in GENELAND visualised as a) trace of number of populations (K) along one million MCMC iterations and b) histogram showing simulated number of populations (K)

Chapter Five: Fine-scale population structure

Hr 5100000

Northing 5000000 5000000

4900000 4900000 1100000 1200000 1300000 Easting

Figure 5.7 Map of estimated rock wren (Xenicus giliventris) membership to two genetic clusters as generated by GENELAND. Black dots denote sampling locations. Sampling areas used in population analysis (i.e. where N > 10) are labelled: Hr = Haast Range, H = Upper Hollyford, M = Murchison Mtns and R = Lake Roe.

Sex-biased dispersal

Results from two of the tests for sex-biased dispersal indicated a female-bias dispersal system (females: lower ST, lower relatedness (r)). However the assignment indices (mAIc, vAIc) supported a male-bias (males: lower mean assignment index and higher variance in assignment index). Results from all five tests were not significant (P > 0.05) (Table 5.3).

Table 5.3 Tests for sex-biased dispersal in rock wren (Xenicus gilviventris)

n ST FIS r mAIc vAIc Females 32 0.05 0.11 0.09 0.41 9.11 Males 40 0.08 0.11 0.13 -0.33 11.56 P value 0.22 0.53 0.22 0.82 0.72

Chapter Five: Fine-scale population structure

Discussion

We evaluated the genetic structure of rock wren over spatial scales ranging up to c.250 km and detected a large genetic patch size for rock wren (c. 70 km), significant differences in genetic relatedness within populations and asymmetrical gene flow among populations indicative of source-sink dynamics.

Population-based inferences and spatial structure

Levels of genetic diversity were similar among the three southern-most populations, though low levels of allelic diversity together with high levels of inbreeding and relatedness were detected within the Haast Range population. It is unlikely that family groups were unintentionally included during sampling of the Haast Range thereby inflating relatedness, because this area was sampled late in the season – i.e. end of February - early March by which time most juveniles are likely to have dispersed (Michelsen-Heath 1989). Further, all individuals sampled were identified as adults (seven females, ten males) in the same manner as the other populations. Juvenile rock wren are distinguishable (albeit subtly) as they do not moult to adult plumage until c.3 months post-fledging, by which time they have typically paired and established territories of their own (Michelsen-Heath 1989). Another explanation for the high inbreeding in the Haast Range could be its peripheral location, right at the Western- most extent of their distribution (Figure 5.1). Being an edge population means naturally fewer neighbours to exchange genes with compared to more core populations. Consequently, peripheral populations often have less gene flow and hence greater genetic isolation, lower genetic variation and smaller effective population sizes (Bouzat & Johnson 2004; Keyghobadi 2007; Segelbacher & Storch 2002; Trizio et al. 2005). Cluster analysis did however group individuals from the Haast Range and Mount Aspiring National Park (i.e. Lake Crucible and Matukituki) together indicating high gene flow among these areas, which are located a maximum of c.40 km from each other. A genetic discontinuity was identified between the Haast Range/Mount Aspiring area and the Upper Hollyford. It is possible that this discontinuity is associated with a lack of landscape connectivity between the Matukituki Valley and the Upper Hollyford, with Lake Wakatipu and several major river systems (e.g. the

Chapter Five: Fine-scale population structure

Arawhata, Hollyford and Pyke Rivers) in the area forming a barrier to gene flow. This discontinuity is also reflected in the migration rates among populations, with gene flow only detected among the three southern-most populations; the Upper Hollyford, Murchison Mountains and Lake Roe. Relatively higher differentiation between the Haast Range population and the other Southern populations was also demonstrated by the pairwise-FST values calculated among these populations (see Chapter 3, Table 3.3). Rivers have been identified as a barrier to gene flow in several taxonomic groups from small mammals to large ungulates and birds (Bush et al. 2011; Coulon et al. 2004; Lugon-Moulin & Hausser 2002; Storfer et al. 2010; Voelker et al. 2013). It is important to consider that in this study, we applied Euclidian (straight-line) distances between individuals/populations. Another approach would be to calculate the ‘effective distance’ using landscape genetic methods such as least-cost distance, which measures the path between two locations which avoids barriers and minimises the ‘resistance’ of the matrix habitat (Adriaensen et al. 2003). This method would provide an interesting comparison for rock wren, given that they are an alpine species only very rarely recorded below the bushline and therefore, the effective distance travelled above the bushline and along ridges may differ substantially from the Euclidian distance conventionally used in tests of isolation by distance and spatial autocorrelation. A genetic patch size of c. 70 km for rock wren initially seems unexpectedly large, especially given their poor flight ability (Higgins et al. 2001), small territory size (c. 0.6 – 4.2 hectares; Michelsen-Heath 1989) and previous observation that the majority of juveniles disperse to within 500 m of their place of fledging (Michelsen- Heath 1989). However, given that rock wren populations demonstrate a pattern of isolation by distance suggesting a ‘stepping stone’ model of gene flow among populations (Kimura & Weiss 1964), genes can be dispersed over distances far greater than their dispersal ability would predict, as has been shown for other sedentary alpine bird species (Fedy et al. 2008; Segelbacher & Storch 2002). Further, although a majority (75%) of juvenile birds were observed dispersing short distances by Michelsen-Heath (1989), the remaining 25% were not resighted. Therefore, it is possible that at least some of these birds dispersed over much greater distances. Spatial autocorrelation has not often been used to assess avian fine-scale population genetic structure and where it has been applied, genetic patch size has been found to be highly variable in passerines (< 200 m – c.15 km) and strongly influenced

Chapter Five: Fine-scale population structure by social system, demography and landscape features (Beck et al. 2008; Double et al. 2005; Liebgold et al. 2013; Nelson-Flower et al. 2012; Vangestel et al. 2013; Wilson et al. 2011). We might expect long-distance dispersal to be selected for in a patchy and dynamic environment such as the alpine zone in order for birds to locate and colonise new habitat and for populations to persist (Johst et al. 2002). This could also partly explain the relatively large genetic patch size in rock wren, compared to forest/woodland species for example, where resources and habitat may be more homogenous (Beck et al. 2008; Peakall et al. 2003; Vangestel et al. 2013). There is some evidence indicating that non-random mating is contributing to the fine-scale population structure observed in rock wren with higher positive spatial genetic autocorrelation among males over shorter distances (significantly from 1 – 10 km) ; combined with a higher FST and higher relatedness for males within populations. Whilst results for the latter tests were indicative, they were not significant, though these methods are known to have limited power unless the bias in dispersal is extreme (e.g. an 80:20 sex ratio of dispersers ) (Goudet et al. 2002). Nevertheless, taken together our results are suggestive of natal philopatry in male rock wren and support the predominant finding of female-biased dispersal in birds (Greenwood 1980). The fact that the difference in positive autocorrelation between the sexes declined over larger distances in rock wren is interesting, though not entirely unexpected given growing evidence suggesting that selective pressures (e.g. access to food and mates) affect males and females differentially, resulting in sex-specific and often antagonistic dispersal strategies over different spatial scales (Vangestel et al. 2013; Yannic et al. 2012). Long-distance dispersal, for example, is more likely to be motivated by the opportunity to colonise empty habitat patches and is therefore, unlikely to be sex-specific (Goudet et al. 2002).

Management implications

We focused this study on populations within the lower South Island due to the higher sampling coverage in this area and because conservation management is already underway in the southern region. With further sampling, this approach could also be applied to other areas, for example within the North of the rock wren’s range where differences in habitat and other environmental factors may differentially affect gene

Chapter Five: Fine-scale population structure flow. Understanding the spatial scale of gene flow in rock wren has several important implications for management. Firstly, patterns of gene flow suggest that there are important source-sink dynamics operating within the Southern populations. The Murchison Mountains appear to be a particularly important source of migrants for other populations. Therefore, management efforts to ensure this population is conserved should be prioritised. The Murchison Mountains is already under intensive conservation management as it constitutes a Special Area reserved for the endangered Takahe (Porphyrio hochstetteri). Since the discovery of the Takahe there in 1948 there has been various degrees of predator control and in 2002, a stoat trapping programme was established over a 15 000 ha block in the southeastern sector (Wickes et al. 2009). Subsequently, trapping has been extended to encompass the entire 50,000 hectares of the Murchison Mountains Special Area. These management practices may be benefiting the rock wren by boosting reproductive success, thereby contributing to the important role of the Murchison Mountains as a source population. Conversely, for the Upper Hollyford and Lake Roe populations, migration is occurring into, but not out of these areas. Further, the lack of pairwise relatedness among individuals within the Upper Hollyford is also indicative of this population being a sink (King et al. 2000). This is because where a large proportion of recruitment results from immigration; fewer close relatives should be present, and in contrast, higher relatedness should be detected in closed and self-sustaining populations (Peery et al. 2008). Sinks typically occur in habitats of poorer quality, where productivity is insufficient to balance mortality (Dias 1996; Pulliam 1988). Consequently, without migration from outside sources, sink populations would be expected to eventually become extinct (Pulliam 1988). High rates of predation within these populations may be contributing to their demographic instability, especially given the findings of a recent breeding success study within the Upper Hollyford where 100% of monitored rock wren nests (n=20) failed as a result of stoat (Mustela erminea) predation (DoC , pers. comm). Anthropogenic-induced mortality, such as predation by introduced species, has been previously implicated in source-sink dynamics (Balogh et al. 2011; Woodford & McIntosh 2010). If invasive predators were to be controlled within these habitats, there is potential that they may be converted from sinks into new source populations (Woodford & McIntosh 2010).

Chapter Five: Fine-scale population structure

There is also evidence for significant genetic differentiation between populations north and south of Lake Wakatipu. This may reflect impaired connectivity in this area due to natural landscape barriers such the lake and several major rivers. Alternatively, or synergistically; there may be anthropogenic factors in force such as high levels of predation on dispersing individuals. In the case of the latter, predator-controlled areas functioning as ‘habitat corridors’ could be an option to increase connectivity (Mech &

Hallett 2001).

Male rock wren Xenicus gilviventris

Chapter 6 General Discussion

In this thesis, I have described the genetic structuring of rock wren (Xenicus gilviventris), within both a phylogeographic and landscape context, and also estimated temporal changes in rock wren demography and effective population size. This study constitutes the first genetic study of the species throughout its range. The questions posed relate directly to the conservation of this unique and threatened alpine passerine and also illustrate the wider application of genetic techniques to threatened species conservation.

Summary of major findings

To investigate the genetic structure and diversity in rock wren first required the development of rock-wren specific DNA markers. We successfully used 454 next generation sequencing to isolate and characterise 14 novel microsatellite markers for rock wren (Chapter 2). Using the read data set obtained from next generation sequencing we also isolated the rock wren mitochondrial control region (Chapter 3). We were able to isolate and amplify the mitochrondrial cytochrome b gene and the seventh intron of the nuclear β-fibrinogen gene using non-specific primer pairs (Chapter 3). Using both mitochrondrial DNA and microsatellites, a deep north-south phylogeographic split among rock wren was identified, with the disjunction between the two lineages occurring at Aoraki/Mount Cook National Park in the centre of the South Island (Chapter 3). This divergence appears to have formed vicariantly approximately two million years ago, within the Pleistocene era of extensive glaciation. Consequently, the genetic structure detected within rock wren provides strong support for the role of Pleistocene glaciations in the formation of isolated northern and southern refugia in the South Island of New Zealand. This central South Island region marks the disjunction in distribution within many taxonomic groups, including species within several alpine invertebrate genera (McCulloch et al. 2010; Trewick & Wallis 2001), however the area was first coined the “beech gap,” reflecting the distinct absence of southern beech forest (Nothofagus spp.) (Cockayne 1926). Cockayne (1926) proposed that northern and southern beech populations likely persisted in ice-free refugia, with subsequent expansion back towards the centre of the island as ice-fields retreated.

Chapter Six: General Discussion

Radiation within each the Northern and Southern rock wren lineages looks to have commenced much later than the initial North – South split; corresponding with the Otiran period of glaciation from c.100 – 15 Ka when at least 5 glacial cycles are known to have occurred up until the last glacial maximum 25 – 15 Ka. A genetic signature of demographic expansion c.14 – 18 Ka was also detected, consistent with the retreat of the South Island glaciers around 15 – 14 Ka (Newnham et al. 1999) (Chapter 4). There is some evidence for a larger refugium within the south of the rock wren’s range, given that the estimates of genetic variation (Chapter 3) and long-term effective population size were consistently larger for the Southern lineage (Chapter 4). However this may also be indicative of more optimal habitat in the south of the species range, supporting a higher density of rock wren than in the north. It is interesting that the patterns detected in rock wren genetic structure and past demography indicate range contraction and decreased gene flow during glacial periods, given they are a cold-adapted alpine species. Many cold-adapted species such as those from alpine or arctic ecosystems in the Northern Hemisphere experienced range expansion and increased gene flow during glacial periods, exhibiting a genetic signature of recent population decline during the Holocene that is interpreted as range retraction since the LGM (Galbreath et al. 2009; Hewitt 2000). However, unlike rock wren, many of these species appear to be associated with savannah or open grassland, a habitat known to have expanded during the cooler, drier climate of the Quaternary glacial periods (Hewitt 2000). In New Zealand, the extent of glaciations was not as large as in the Northern Hemisphere and as much as two-thirds of the total land area may have been dominated by scrub- and grassland-dominated communities (McGlone et al. 2010). Presumably, it was within scrub communities that rock wren were able to persist. It appears that the long-term effective population size of rock wren is dramatically larger relative to contemporary estimates (Chapter 4). This finding indicates that it is unlikely rock wren constitute a ‘naturally rare’ species (Kunin & Gaston 1993), given that in the past, they clearly sustained much higher abundance. Therefore, with the appropriate management applied (e.g. intensive predator control) the carrying capacity of rock wren has the potential to be increased substantially. It has also been of considerable value to date the divergence of rock wren and rifleman (A. chloris) at approximately 6 Ma (Chapter 3). This finding corroborates the pre-existing assumption that rock wren diverged from rifleman with the formation of the Southern Alps during the Pliocene period of tectonic uplift, an assumption that had 98

Chapter Six: General Discussion not been previously ratified using molecular data. The only attempt to clarify this relationship suggested a 24 – 19 Ma divergence between the riflemen and rock wren using limited ND6 and 12S mtDNA sequence extracted from two museum specimens (Cooper 1994; Cooper & Cooper 1995). Species with lower dispersal ability are not only expected to retain stronger signals of vicariance (McCulloch et al. 2010), but also exhibit fine-scale spatial genetic structure (Peakall et al. 2003), i.e. non-random spatial distribution of genotypes across the landscape. Almost nothing is known of the dispersal patterns of rock wren, other than a previous observation from the Murchison Mountains, Fiordland, that the majority of juveniles dispersed to within 500 m of their place of fledging (Michelsen- Heath 1989). In the present study, a strong pattern of isolation by distance was identified, where genetic relatedness among neighbouring individuals is significantly greater than that among more distant or random individuals (Chapter 5). A potential sex-bias in dispersal, suggestive of male natal philopatry, was also detected which may have further contributed to the strong pattern of fine-scale structuring. However, the scale of positive genetic structure, i.e. the distance over which individuals were not genetically independent, was unexpectedly large (c.70 km) considering the rock wren’s poor flight, small territory size (c. 0.6 – 4.2 hectares) and previous observations of juvenile dispersal (Michelsen-Heath 1989) (Chapter 5). This is not entirely surprising however, given that gene flow among rock wren populations indicates a ‘stepping stone’ model of dispersal and therefore, genes can be dispersed over distances far greater than their dispersal ability would predict, as has been shown for other sedentary alpine bird species (Fedy et al. 2008; Segelbacher & Storch 2002). Asymmetrical gene flow among populations within the Southern lineage was also identified, indicative that source-sink dynamics are operating (Chapter 5). This finding has important implications for rock wren conservation management. Finally, it is interesting to note that the zone of admixture between the Northern and Southern rock wren lineages identified in Chapter 3 was of a similar distance (c.60 km) to the genetic patch size subsequently detected using spatial auto-correlation (c.70 km) (Chapter 5).

Chapter Six: General Discussion

Implications for conservation management of rock wren

There is no formalised management strategy in place for rock wren, though a draft management statement has been prepared (Gaze 2004). Within this document, two key management objectives are outlined: “to secure the species….by translocation to a predator-free island” and “to control predators in four key locations to a level that allows these populations to support more individuals than at present.” Key management locations or units were not specified however, and genetics is yet to be incorporated within management practices. Based on our findings from throughout the range of rock wren, some general recommendations are proposed:

Northern and Southern rock wren lineages

This study reveals a degree of divergence between Northern and Southern rock wren lineages similar to levels separating North and South Island species of New Zealand passerines such as robins (Petroica spp.) (Miller & Lambert 2006) and Kokako (Callaeas spp.) (Murphy et al. 2006). However, the level of divergence between rock wren and their sister species the rifleman (c. 8% at cytochrome b) is approximately twice that of the Northern and Southern rock wren (c. 4% at cytochrome b), thus placing these rock wren lineages at around the ‘sub-specific’ level. Genetic evidence did not support the designation of birds from deep Fiordland as a distinct sub-species X. g. rineyi as was proposed in the early 1950s based on behavioural (Riney 1953) and morphological characteristics (Falla 1953). However, it could be considered that X. g. rineyi comprises the entire Southern lineage (Chapter 3). Given that the Northern and Southern lineages appear to have formed approximately two million years ago and a strong signal of both genetic and geographical divergence has been retained, the Northern and Southern rock wren should at the least be regarded as distinct Evolutionarily Significant Units (ESUs). Consequently, management of rock wren should focus on conserving viable representative populations from both of these lineages.

100

Chapter Six: General Discussion

Rock wren management units

With regard to the management objective: “to control predators in four key locations to a level that allows these populations to support more individuals than at present”, the following approach is proposed based on the molecular data. Given that rock wren populations demonstrate a pattern of isolation by distance suggesting a ‘stepping stone’ model of gene flow; managing just two populations within each lineage will not be enough to maintain genetic connectivity throughout their range. We identified the genetic patch size (i.e. the spatial scale of non-random genetic structure) of rock wren to be c. 70 km. Therefore ideally, several populations within each lineage or ‘ESU’, spatially separated by no more than this spatial scale need to be managed to retain this genetic structure. Within the Southern lineage, the Murchison Mountains was identified as an important source of migrants for other Southern populations (Chapter 5). This area is already managed for conservation as it comprises a Special Area for the threatened Takahe (Porphyrio hochstetteri) (Wickes et al. 2009). Ongoing management of this area, increased to a scale appropriate for rock wren conservation should be considered. Two other areas - Lake Roe and the Upper Hollyford - exhibited relatively high allelic and genetic diversity and low levels of inbreeding. These areas, particularly the Upper Hollyford, were also identified as potential ‘sink’ populations that may benefit from improvements to habitat in the form of intensive predator control. Other populations within the Southern lineage that should be prioritised for rock wren conservation management include those in Mt Aspiring National Park and northwards through to Aoraki/Mount Cook National Park. Within the Northern lineage, Arthurs Pass exhibited the highest genetic variation of the three populations we assessed, followed by Northern Kahurangi (Chapter 3). With more intensive sampling of populations within the Northern lineage, migration rates and fine-scale gene flow among populations could be assessed in a similar manner to the approach demonstrated for Southern populations in Chapter 5. This would further assist the identification of management units within the Northern lineage. The Lewis Pass and Nelson Lakes areas represent an important spatial link between Arthurs Pass and Kahurangi, but unfortunately rock wren numbers were so low that we were unable to sample comprehensively from these areas. Although we were only able to sample two birds from Aoraki/Mount Cook due to the low density of

101

Chapter Six: General Discussion birds present, this area represents a key population for conservation. This is because it is located within the zone of admixture between the two lineages and may form an important dispersal link for gene flow between the North and the South, especially given the stepping stone pattern of gene flow we have detected in rock wren. Maintaining natural connectivity between these two lineages is also pertinent during this time of rapid climate change to facilitate both the potential transfer of adaptations among lineages and the migration of individuals southwards into relatively cooler climes (Walther et al. 2002). Furthermore, densities in Aoraki/Mount Cook were much higher in the past (Michelsen 1982) and their decline appears to have coincided with the arrival of mammalian predators, most notably mustelids, into the park within the past c.30 years according to long-term staff based within the village (R. Bellringer, DoC. pers. comm.). Therefore, it seems that the carrying capacity of Aoraki/Mount Cook for rock wren could potentially be increased to former densities with effective predator control.

Genetics and rock wren translocations

To date, a total of 41 rock wren have been established on predator-free Secretary Island, Fiordland, all of which were sourced from the Murchison Mountains. There may be opportunity for more individuals to be released (M. Willans, pers. comm.). Based on the high level of divergence between the Northern and Southern lineages, we would advise against translocating individuals from the Northern lineage onto Secretary Island at this time. Within the Southern lineage, we detected two distinct genetic groupings in the lower South Island. A prudent approach would be to only translocate individuals from populations within the same genetic grouping as the Murchison Mountains (e.g from Lake Roe, Lake MacArthur, Eyre Mountains or the Upper Hollyford). Low sample sizes from Lake MacArthur and the Eyre Mountains meant we were unable to assess levels of genetic variation within these areas. However, Lake Roe and the Upper Hollyford exhibited relatively high levels of genetic variation and lower levels of inbreeding relative to the Murchison Mountain population (Chapter 5). Therefore, these populations would serve as a good option for any future ‘top-up’ translocations to Secretary Island aimed at boosting numbers of birds on the island (Tracy et al. 2011). Individuals from outside the genetic patch of Secretary Island could also be considered to boost genetic variation within the founder population (Weiser et al. 2012), such as

102

Chapter Six: General Discussion individuals from locations north of Lake Wakatipu e.g. the Haast Range or Mount Aspiring National Park. Whilst the Haast Range population exhibited the lowest levels of allelic diversity and highest inbreeding, this population does contain genetic variation not present within populations south of Lake Wakatipu. Previous translocations of New Zealand South Island robin (Petroica australis) populations have demonstrated that inbred donor populations can also effectively increase levels of genetic variation in small isolated island populations (Heber et al. 2013). Alternatively, individuals from Mount Aspiring National Park could be sourced, but due to low sample sizes, we were not able to assess levels of genetic variation in any populations from this region. Further, low densities of rock wren in this area would present a logistical challenge in obtaining enough individuals for translocation. Lastly, climate change is predicted to have a disproportionate impact on alpine specialists and management against this threat for rock wren presents additional challenges that much be acknowledged (Dirnböck et al. 2011; Galbreath et al. 2009; Rubidge et al. 2012). One emerging and widely debated approach is the assisted colonisation of individuals from small, isolated populations marooned atop of shrinking alpine islands in warmer locations to colder locations harbouring more favourable habitat, thereby improving their chances of survival as the climate changes (Hoegh- Guldberg et al. 2008; McLachlan et al. 2007; Ricciardi & Simberloff 2009; Richardson et al. 2009; Schlaepfer et al. 2009; Seddon 2010). This approach could be considered for rock wren populations here identified as exhibiting genetic signals of small population size and limited connectivity, such as those at the very northern end of the rock wrens range in Northern Kahurangi and the Wangapeka (see Chapter 3). Further, translocations of rock wren from areas harbouring high levels of genetic diversity into novel regions from which they have not been previously recorded, i.e. beyond the current southern extent of their range, could also buffer populations against climate warming. For example, predator-free Southern Islands such as those within the Sub- Antarctic group could be considered. Several authors outline a framework for decision- making around this issue (Hoegh-Guldberg et al. 2008; McLachlan et al. 2007; Richardson et al. 2009), which could be worked through in further detail as an option for future rock wren management.

103

Chapter Six: General Discussion

Future research

Whilst we were not able to detect a genetic signature linking population decline within the Northern lineage to a timeframe of anthropogenic disturbance, there is some evidence for a recent population bottleneck within the Southern lineage. Given that a variety of influences are known to potentially obscure genetic bottleneck signals and limit power of detection, a study incorporating the use of historic DNA samples would be a valuable extension to this work (Zachariah Peery et al. 2012). Early museum specimens collected during the 1800s would provide comparison of genetic variation prior to the proliferation of mammalian predators throughout the South Island; over 40 rock wren specimens have been located in several museums worldwide. Recent assessments of temporal changes in genetic variation and structure in other New Zealand avian endemics have revealed a clear association with human settlement (Boessenkool et al. 2009; Tracy & Jamieson 2011). For example, Tracey & Jamieson (2011) demonstrated that genetic structure among contemporary populations of the threatened passerine mohua (Mohoua ochrocephala) was not indicative of historical, pre-fragmentation patterns of mohua gene flow c.100 years ago when minimal genetic structure existed. Further, as mentioned in Chapter 3, it would be useful to undertake molecular analysis of the historical X. g. rineyi specimens collected by Falla (1953), alongside the contemporary samples collected from the type locality during this study. This would further clarify the sub-specific status of the deep Fiordland birds. Given the high degree of genetic and geographic divergence detected between the Northern and Southern lineages, it is highly possible that local adaptation exists (Hereford 2009; Kawecki & Ebert 2004; Margraf et al. 2007). In this study, we developed and applied putatively neutral markers as proxies for genome-wide variation. Selectively neutral markers are most effective for inferring population connectivity and relatedness; however their suitability for detecting adaptive variation and as proxies for genetic variation in fitness-related loci remains controversial (Alcaide et al. 2008; Grueber et al. 2013). Recent improvements in genomic tools are now enabling genome- wide studies to examine the genetics of local adaptation in a wider range of non-model species (Savolainen et al. 2013). Therefore, molecular tools such as quantitative trait locus (QTL) mapping could be combined with an analysis of morphometric data collected during this study to explore adaptive variation in rock wren in the future (Yeaman & Whitlock 2011). Recent findings also indicate greater losses of diversity in

104

Chapter Six: General Discussion adaptive major histocompatibility complex (MHC) genes compared to neutral genetic diversity following a genetic bottleneck (Sutton et al. 2011). The analysis of adaptive genes, including MHC and others associated with individual fitness such as Toll-like receptors (TLRs) (Brownlie & Allan 2011; Grueber & Jamieson 2013) in future studies would therefore be valuable to ensure accurate estimation of the true extent of functional genetic diversity and adaptive potential remaining within rock wren. Due to low densities of birds within the central region of the South Island, we were not able to sample comprehensively in many areas. However, the results clearly showed a narrow zone of genetic admixture between the Northern and Southern lineages around Aoraki/Mount Cook. As previously mentioned (Chapter 3), it is interesting that such a high degree of divergence has been maintained between the lineages, despite their geographic proximity and an apparent lack of landscape barriers to gene flow. Detailed behavioural and fine-scale genetic studies of birds occupying this secondary contact zone are now required to elucidate the mechanisms involved in the maintenance of genetic divergence between the Northern and Southern rock wren lineages (Barton & Hewitt 1985; Hewitt 2001; Phillips et al. 2004). In this study, the mating system and social structure of rock wren were not investigated. By utilising the neutral microsatellite markers developed here, these aspects of rock wren ecology could be elucidated. For example, it is not known whether extra pair paternity exists in rock wren, though given that extra-pair offspring have been found in approximately 90% of avian species assessed, it is highly likely to occur and is directly linked to the amount of genetic variation maintained within populations and the effective population size (Griffith et al. 2002; Petrie & Kempenaers 1998). This is because variation in individual reproductive success may be reduced if unpaired males gain opportunity to mate, or increased, if a disproportionate number of offspring are sired by just a few dominant individuals (Castro et al. 2004; O'Connor et al. 2006). Finally, it important to acknowledge that whilst the genetic information presented here offers a valuable contribution to rock wren conservation management, it is just a piece of the overall puzzle. Demographic information on rock wren is urgently required. For example, robust estimates of generation length, survival estimates across different age classes and reproductive rates are critical for informing further management. The impact of predators on alpine systems also warrants much further research, particularly with respect to predicted synergistic interactions with climate change (Hughes 2003; Sala et al. 2000). 105

Chapter Six: General Discussion

106

Homeward bound Lewis Pass

107

108

References

Abbott CL, Põldmaa T, Lougheed S, Clarke M, Boag PT (2002) Hierarchical analysis of genetic population structure in the noisy miner using DNA microsatellite markers. Condor 104, 652-656. Abdelkrim J, Robertson BC, Stanton JAL, Gemmell NJ (2009) Fast, cost-effective development of species-specific microsatellite markers by genomic sequencing. BioTechniques 46, 185-192. Adriaensen F, Chardon JP, De Blust G, et al. (2003) The application of 'least-cost' modelling as a functional landscape model. Landscape and Urban Planning 64, 233-247. Alcaide M, Edwards SV, Negro JJ, Serrano D, Tella JL (2008) Extensive polymorphism and geographical variation at a positively selected MHC class II B gene of the lesser kestrel (Falco naumanni). Molecular Ecology 17, 2652- 2665. Allendorf FW, Hohenlohe PA, Luikart G (2010) Genomics and the future of conservation genetics. Nature Reviews Genetics 11, 697-709. Alter SE, Rynes E, Palumbi SR (2007) DNA evidence for historic population size and past ecosystem impacts of gray whales. Proceedings of the National Academy of Sciences of the United States of America 104, 15162-15167. Amos W, Harwood J (1998) Factors affecting levels of genetic diversity in natural populations. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 353, 177-186. Andreasen AM, Stewart KM, Longland WS, Beckmann JP, Forister ML (2012) Identification of source-sink dynamics in mountain lions of the Great Basin. Molecular Ecology 21, 5689-5701. Arif IA, Khan HA (2009) Molecular markers for biodiversity analysis of wildlife animals: A brief review. Animal biodiversity and conservation 32, 9-17. Avise JC (1987) Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual review of ecology and systematics. Vol. 18, 489-522. Avise JC (1994) Molecular Markers, Natural History and Evolution Chapman & Hall, New York. Avise JC (2000) Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, Massachusetts. Bachtrog D, Weiss S, Zangerl B, Brem G, Schlötterer C (1999) Distribution of dinucleotide microsatellites in the Drosophila melanogaster genome. Molecular Biology and Evolution 16, 602-610. Baker AJ, Daugherty CH, Colbourne R, McLennan JL (1995) Flightless brown kiwis of New Zealand possess extremely subdivided population structure and cryptic species like small mammals. Proceedings of the National Academy of Sciences of the United States of America 92, 8254-8258. Baker AJ, Huynen LJ, Haddrath O, Millar CD, Lambert DM (2005) Reconstructing the tempo and mode of evolution in an extinct clade of birds with ancient DNA: The giant moas of New Zealand. Proceedings of the National Academy of Sciences of the United States of America 102, 8257-8262. Bakker J, van Rijswijk MEC, Weissing FJ, Bijlsma R (2010) Consequences of fragmentation for the ability to adapt to novel environments in experimental Drosophila metapopulations. Conservation Genetics 11, 435-448.

109

References

Balogh AL, Ryder TB, Marra PP (2011) Population demography of Gray Catbirds in the suburban matrix: Sources, sinks and domestic cats. Journal of Ornithology 152, 717-726. Barker FK, Barrowclough GF, Groth JG (2002) A phylogenetic hypothesis for passerine birds: Taxonomic and biogeographic implications of an analysis of nuclear DNA sequence data. Proceedings of the Royal Society B: Biological Sciences 269, 295-308. Barker FK, Cibois A, Schikler P, Feinstein J, Cracraft J (2004) Phylogeny and diversification of the largest avian radiation. Proceedings of the National Academy of Sciences of the United States of America 101, 11040-11045. Barnosky AD, Matzke N, Tomiya S, et al. (2011) Has the Earth's sixth mass extinction already arrived? Nature 471, 51-57. Barton NH, Hewitt GM (1985) Analysis of hybrid zones. Annual Review of Ecology and Systematics 16, 113-148. Batt GE, Braun J, Kohn B, McDougall I (2000) Thermochronological analysis of the dynamics of the Southern Alps, New Zealand. Bulletin of the Geological Society of America 112, 250-266. Bech N, Barbu CM, Quéméré E, et al. (2013) Pyrenean ptarmigans decline under climatic and human influences through the Holocene. Heredity 111, 402-409. Bech N, Boissier J, Drovetski S, Novoa C (2009) Population genetic structure of rock ptarmigan in the 'sky islands' of French Pyrenees: Implications for conservation. Animal Conservation 12, 138-146. Beck NR, Peakall R, Heinsohn R (2008) Social constraint and an absence of sex-biased dispersal drive fine-scale genetic structure in white-winged choughs. Molecular Ecology 17, 4346-4358. Beever EA, Brussard PF, Berger J (2003) Patterns of apparent extirpation among isolated populations of Pikas (Ochotona princeps) in the Great Basin. Journal of Mammalogy 84, 37-54. Beever EA, Ray C, Wilkening JL, Brussard PF, Mote PW (2011) Contemporary climate change alters the pace and drivers of extinction. Global Change Biology 17, 2054-2070. Bellamy D, Springett B, Hayden P (1990) Moa's Ark: The Voyage of New Zealand Viking, New York. Benning TL, LaPointe D, Atkinson CT, Vitousek PM (2002) Interactions of climate change with biological invasions and land use in the Hawaiian islands: Modeling the fate of endemic birds using a geographic information system. Proceedings of the National Academy of Sciences of the United States of America 99, 14246-14249. Bernt M, Braband A, Schierwater B, Stadler PF (2013) Genetic aspects of mitochondrial genome evolution. Molecular Phylogenetics and Evolution 69, 328-338. Bickford D, Lohman DJ, Sodhi NS, et al. (2007) Cryptic species as a window on diversity and conservation. Trends in Ecology and Evolution 22, 148-155. Biebach I, Keller LF (2012) Genetic variation depends more on admixture than number of founders in reintroduced Alpine ibex populations. Biological Conservation 147, 197-203. Binks RM, Kennington WJ, Johnson MS (2007) Rapid evolutionary responses in a translocated population of intertidal snail (Bembicium vittatum) utilise variation from different source populations. Conservation Genetics 8, 1421-1429.

110

References

Birdlife_International (2012) Xenicus gilviventris. In: IUCN 2013. IUCN Red List of Threatened Species. Version 2013.2. www.iucnredlist.org. Boessenkool S, Austin JJ, Worthy TH, et al. (2009) Relict or colonizer? Extinction and range expansion of penguins in southern New Zealand. Proceedings. Biological sciences / The Royal Society 276, 815-821. Bouzat JL, Johnson K (2004) Genetic structure among closely spaced leks in a peripheral population of lesser prairie-chickens. Molecular Ecology 13, 499- 505. Brook BW, Sodhi NS, Bradshaw CJA (2008) Synergies among extinction drivers under global change. Trends in Ecology and Evolution 23, 453-460. Broquet T, Petit EJ (2009) Molecular estimation of dispersal for ecology and population genetics. Annual Review of Ecology, Evolution, and Systematics 40, 193-216. Brownlie R, Allan B (2011) Avian toll-like receptors. Cell and Tissue Research 343, 121-130. Buckley TR, Marske KA, Attanayake D (2009) Identifying glacial refugia in a geographic parthenogen using palaeoclimate modelling and phylogeography: The New Zealand stick insect Argosarchus horridus (White). Molecular Ecology 18, 4650-4663. Buckley TR, Simon C (2007) Evolutionary radiation of the cicada genus Maoricicada Dugdale (Hemiptera: Cicadoidea) and the origins of the New Zealand alpine biota. Biological Journal of the Linnean Society 91, 419-435. Buckley TR, Simon C, Chambers GK (2001) Phylogeography of the New Zealand cicada Maoricicada campbelli based on mitochondrial DNA sequences: Ancient clades associated with cenozoic environmental change. Evolution 55, 1395- 1407. Burbidge ML, Colbourne RM, Robertson HA, Baker AJ (2003) Molecular and other biological evidence supports the recognition of at least three species of brown kiwi. Conservation Genetics 4, 167-177. Burkey TV (1995) Extinction rates in Archipelagoes: Implications for populations in fragmented habitats. Conservation Biology 9, 527-541. Burney CW, Brumfield RT (2009) Ecology predicts levels of genetic differentiation in neotropical birds. American Naturalist 174, 358-368. Burrows CJ (1965) Some discontinuous distributions of plants within New Zealand and their ecological significance. Part II: Disjunctions between Otago-Southland and Nelson-Marlborough and related distribution patterns. Tuatara 13, 9-29. Bush KL, Dyte CK, Moynahan BJ, et al. (2011) Population structure and genetic diversity of greater sage-grouse (Centrocercus urophasianus) in fragmented landscapes at the northern edge of their range. Conservation Genetics 12, 527- 542. Caizergues A, Bernard-Laurent A, Brenot JF, Ellison L, Rasplus JY (2003) Population genetic structure of rock ptarmigan Lagopus mutus in Northern and Western Europe. Molecular Ecology 12, 2267-2274. Callens TOM, Galbusera P, Matthysen E, et al. (2011) Genetic signature of population fragmentation varies with mobility in seven bird species of a fragmented Kenyan cloud forest. Molecular Ecology 20, 1829-1844. Camacho C, Coulouris G, Avagyan V, et al. (2009) BLAST+: Architecture and applications. BMC Bioinformatics 10.

111

References

Castro I, Mason KM, Armstrong DP, Lambert DM (2004) Effect of extra-pair paternity on effective population size in a reintroduced population of the endangered hihi, and potential for behavioural management. Conservation Genetics 5, 381-393. Chamberlain CP, Poage MA (2000) Reconstructing the paleotopography of mountain belts from the isotopic composition of authigenic minerals. Geology 28, 115- 118. Chamberlain DE, Negro M, Caprio E, Rolando A (2013) Assessing the sensitivity of alpine birds to potential future changes in habitat and climate to inform management strategies. Biological Conservation 167, 127-135. Chen C, Durand E, Forbes F, François O (2007) Bayesian clustering algorithms ascertaining spatial population structure: A new computer program and a comparison study. Molecular Ecology Notes 7, 747-756. Chikhi L, Sousa VC, Luisi P, Goossens B, Beaumont MA (2010) The confounding effects of population structure, genetic diversity and the sampling scheme on the detection and quantification of population size changes. Genetics 186, 983-995. Chinn WG, Gemmell NJ (2004) Adaptive radiation within New Zealand endemic species of the cockroach genus Celatoblatta Johns (Blattidae): A response to Plio-Pleistocene mountain building and climate change. Molecular Ecology 13, 1507-1518. Clement M, Posada D, Crandall KA (2000) TCS: A computer program to estimate gene genealogies. Molecular Ecology 9, 1657-1659. Cockayne L (1926) Vegetation of New Zealand. 2nd ed. Engelmann Press, Leipzig, Germany. Cooper A (1994) Ancient DNA sequences reveal unsuspected phylogenetic relationships within New Zealand wrens (Acanthisittidae). Experientia 50, 558- 563. Cooper A, Cooper RA (1995) The oligocene bottleneck and New Zealand biota: Genetic record of a past environmental crisis. Proceedings of the Royal Society B: Biological Sciences 261, 293-302. Cooper RA, Millener PR (1993) The New Zealand biota: historical background and new research. Trends in Ecology & Evolution 8, 429-433. Cornuet JM, Luikart G (1996) Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144, 2001- 2014. Coulon A, Cosson JF, Angibault JM, et al. (2004) Landscape connectivity influences gene flow in a roe deer population inhabiting a fragmented landscape: An individual-based approach. Molecular Ecology 13, 2841-2850. Cox WA, Thompson FR, Reidy JL, Faaborg J (2013) Temperature can interact with landscape factors to affect songbird productivity. Global Change Biology 19, 1064-1074. Craig JL (1991) Are small populations viable? Acta XX Congressus Internationalis Ornithologici 4, 2546-2551. Crandall KA, Bininda-Emonds ORR, Mace GM, Wayne RK (2000) Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290-295. Craw R (1989) New Zealand biogeography: a panbiogeographic approach. New Zealand Journal of Zoology 16, 527-547. Crochet PA (1996) Can measures of gene flow help to evaluate bird dispersal? Acta Oecologica 17, 459-474.

112

References

Crochet PA (2000) Genetic structure of avian populations - Allozymes revisited. Molecular Ecology 9, 1463-1469. Curtis C, Stewart BS, Karl SA (2011) Genetically effective population sizes of Antarctic seals estimated from nuclear genes. Conservation Genetics 12, 1435- 1446. Darriba D, Taboada GL, Doallo R, Posada D (2012) JModelTest 2: More models, new heuristics and parallel computing. Nature Methods 9, 772. Dias PC (1996) Sources and sinks in population biology. Trends in Ecology and Evolution 11, 326-330. Dieringer D, Schlötterer C (2003) MICROSATELLITE ANALYSER (MSA): A platform independent analysis tool for large microsatellite data sets. Molecular Ecology Notes 3, 167-169. Dirnböck T, Essl F, Rabitsch W (2011) Disproportional risk for habitat loss of high- altitude endemic species under climate change. Global Change Biology 17, 990- 996. Ditto AM, Frey JK (2007) Effects of ecogeographic variables on genetic variation in montane mammals: implications for conservation in a global warming scenario. Journal of Biogeography 34, 1136-1149. Do C, Waples RS, Peel D, et al. (2013) NeEstimator v2: re-implementation of software for the estimation of contemporary effective population size (Ne) from genetic data. Molecular Ecology Resources. Double MC, Peakall R, Beck NR, Cockburn A (2005) Dispersal, philopatry, and infidelity: Dissecting local genetic structure in superb fairy-wrens (Malurus cyaneus). Evolution 59, 625-635. Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29, 1969-1973. Dubiec A, Zagalska-Neubauer M (2006) Molecular techniques for sex identification in birds. Biological Letters 43, 3-12. Duncan RP, Blackburn TM (2004) Extinction and endemism in the New Zealand avifauna. Global Ecology and Biogeography 13, 509-517. Durand E, Chen C, Francois O (2009) Tess version 2.1 - Reference Manual. Available from: http://membres-timc.imag.fr/Olivier.Francois/tess.html Dussex N, Robertson BC (2013) Isolation and characterization of 8 microsatellites and for the Kea Nestor notabilis Conservation Genetic Resources 5, 1063-1065. Earl D, vonHoldt B (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4, 359-361. Ellegren H, Sheldon BC (1997) New tools for sex identification and the study of sex allocation in birds. Trends in Ecology and Evolution 12, 255-259. Engler R, Randin CF, Thuiller W, et al. (2011) 21st century climate change threatens mountain flora unequally across Europe. Global Change Biology 17, 2330- 2341. Epstein PR, Diaz HF, Elias S, et al. (1998) Biological and physical signs of climate change: Focus on mosquito-borne diseases. Bulletin of the American Meteorological Society 79, 409-417. Ericson PGP, Christidis L, Cooper A, et al. (2002) A Gondwanan origin of passerine birds supported by DNA sequences of the endemic New Zealand wrens. Proceedings of the Royal Society B: Biological Sciences 269, 235-241.

113

References

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology 14, 2611-2620. Ewers RM, Didham RK (2006) Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews of the Cambridge Philosophical Society 81, 117-142. Excoffier L, Laval G, Schneider S (2005) Arlequin (Version 3.0). An integrated software package for population genetic data analysis. Evolutionary Bioinformatics Online 1, 47-50. Faircloth BC (2008) MSATCOMMANDER: Detection of microsatellite repeat arrays and automated, locus-specific primer design. Molecular Ecology Resources 8, 92-94. Falla RA (1953) Description of a new form of New Zealand wren. Notornis 5, 142-143. Fedy BC, Martin K, Ritland C, Young J (2008) Genetic and ecological data provide incongruent interpretations of population structure and dispersal in naturally subdivided populations of white-tailed ptarmigan (Lagopus leucura). Molecular Ecology 17, 1905-1917. Fleming CA (1962) New Zealand biogeography. A palaeontologist's approach. Tuatara 10, 53-108. Fleming CA (1975) The geological history of New Zealand and its biota. In: Biogeography and Ecology in New Zealand (ed. Kuschel G), pp. 1-86. Dr W Junk b.v. Publishers, The Hague. Floyd CH, Van Vuren DH, May B (2005) Marmots on Great Basin mountaintops: Using genetics to test a biogeographic paradigm. Ecology 86, 2145-2153. Forister ML, McCall AC, Sanders NJ, et al. (2010) Compounded effects of climate change and habitat alteration shift patterns of butterfly diversity. Proceedings of the National Academy of Sciences of the United States of America 107, 2088- 2092. Frankham R (1995) Effective population size/adult population size ratios in wildlife: a review. Genetical Research 66, 95-107. Frankham R (2005) Stress and adaptation in conservation genetics. Journal of Evolutionary Biology 18, 750-755. Frankham R, Ballou JD, Briscoe DA (2002) Introduction to conservation genetics University of Cambridge, Cambridge. Frankham R, Ballou JD, Eldridge MDB, et al. (2011) Predicting the Probability of Outbreeding Depression. Conservation Biology 25, 465-475. Franklin IR (1980) Evolutionary change in small populations. In: Conservation Biology: An Evolutionary–Ecological Perspective (eds. Soule´ ME, Wilcox BA), pp. 135–150. Sinauer Associates. Fraser DJ, Bernatchez L (2001) Adaptive evolutionary conservation: Towards a unified concept for defining conservation units. Molecular Ecology 10, 2741-2752. Freed LA, Cann RL, Goff ML, Kuntz WA, Bodner GR (2005) Increase in avian malaria at upper elevation in Hawai'i. Condor 107, 753-764. Fridolfsson AK, Ellegren H (1999) A simple and universal method for molecular sexing of non-ratite birds. Journal of Avian Biology 30, 116-121. Funk WC, McKay JK, Hohenlohe PA, Allendorf FW (2012) Harnessing genomics for delineating conservation units. Trends in Ecology & Evolution 27, 489-496. Galbreath KE, Hafner DJ, Zamudio KR (2009) When cold is better: Climate-driven elevation shifts yield complex patterns of diversification and demography in an alpine specialist (american pika, ochotona princeps). Evolution 63, 2848-2863.

114

References

Garza JC, Williamson EG (2001) Detection of reduction in population size using data from microsatellite loci. Molecular Ecology 10, 305-318. Gaze P (2004) Recovery statement for rock wren (Draft). Internal Document, Department of Conservation, Nelson. Gemayel R, Vinces MD, Legendre M, Verstrepen KJ (2010) Variable tandem repeats accelerate evolution of coding and regulatory sequences, pp. 445-477. Giribet G, Boyer SL (2010) 'Moa's Ark' or 'Goodbye Gondwana': Is the origin of New Zealand's terrestrial invertebrate fauna ancient, recent or both? Invertebrate Systematics 24, 1-8. Goossens B, Chikhi L, Taberlet P, Waits LP, Allainé D (2001) Microsatellite analysis of genetic variation among and within Alpine marmot populations in the French Alps. Molecular Ecology 10, 41-52. Goudet J (2001) FSTAT, a program to estimate and test gene diversities and fixation indices (Version 2.9.3). Available at http://www2.unil.ch/popgen/softwares/ fstat.html Goudet J, Perrin N, Waser P (2002) Tests for sex-biased dispersal using bi-parentally inherited genetic markers. Molecular Ecology 11, 1103-1114. Graham IJ (2008) A continent on the move: New Zealand geoscience into the 21st century. The Geological Society of New Zealand in association with GNS Science, Wellington. Greaves SNJ, Chapple DG, Gleeson DM, Daugherty CH, Ritchie PA (2007) Phylogeography of the spotted skink (Oligosoma lineoocellatum) and green skink (O. chloronoton) species complex (Lacertilia: Scincidae) in New Zealand reveals pre-Pleistocene divergence. Molecular Phylogenetics and Evolution 45, 729-739. Greenwood PJ (1980) Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour 28, 1140-1162. Griffin SC, Taper ML, Hoffman R, Mills LS (2008) The case of the missing marmots: Are metapopulation dynamics or range-wide declines responsible? Biological Conservation 141, 1293-1309. Griffith SC, Owens IPF, Thuman KA (2002) Extra pair paternity in birds: A review of interspecific variation and adaptive function. Molecular Ecology 11, 2195-2212. Grueber CE, Jamieson IG (2013) Primers for amplification of innate immunity toll-like receptor loci in threatened birds of the Apterygiformes, Gruiformes, Psittaciformes and Passeriformes. Conservation Genetic Resources 5, 1043- 1047. Grueber CE, Wallis GP, Jamieson IG (2013) Genetic drift outweighs natural selection at toll-like receptor (TLR) immunity loci in a re-introduced population of a threatened species. Molecular Ecology 22, 4470-4482. Guillot G (2012) Population genetic and morphometric data analysis using R and the GENELAND program, User Manual. Guillot G, Mortier F, Estoup A (2005) GENELAND: A computer package for landscape genetics. Molecular Ecology Notes 5, 712-715. Hackett SJ, Kimball RT, Reddy S, et al. (2008) A phylogenomic study of birds reveals their evolutionary history. Science 320, 1763-1768. Haig SM, Bronaugh WM, Crowhurst RS, et al. (2011) Genetic Applications in Avian Conservation. The Auk 128, 205-229. Hamilton WD (1971) Selection of selfish and altruistic behaviour in some extreme models. In: Man and Beast: Comparative Social Behavior (eds. JF E, WS D), pp. 57-91. Smithsonian Institute Press, Washington, D.C.

115

References

Harcourt AH, Coppeto SA, Parks SA (2002) Rarity, specialization and extinction in primates. Journal of Biogeography 29, 445-456. Hare MP (2001) Prospects for nuclear gene phylogeography. Trends in Ecology and Evolution 16, 700-706. Hare MP, Nunney L, Schwartz MK, et al. (2011) Understanding and Estimating Effective Population Size for Practical Application in Marine Species Management. Conservation Biology 25, 438-449. Heather BD, Robertson HA (2000) The Field Guide to the Birds of New Zealand. Viking, Auckland. Heber S, Varsani A, Kuhn S, et al. (2013) The genetic rescue of two bottlenecked south island robin populations using translocations of inbred donors. Proceedings of the Royal Society B: Biological Sciences 280. Hedrick PW (1995) Gene flow and genetic restoration: The Florida panther as a case study. Conservation Biology 9, 996-1007. Helfer V, Broquet T, Fumagalli L (2012) Sex-specific estimates of dispersal show female philopatry and male dispersal in a promiscuous amphibian, the alpine salamander (Salamandra atra). Molecular Ecology 21, 4706-4720. Heller NE, Zavaleta ES (2009) Biodiversity management in the face of climate change: A review of 22 years of recommendations. Biological Conservation 142, 14-32. Hereford J (2009) A quantitative survey of local adaptation and fitness trade-offs. The American Naturalist 173, 579-588. Hewitt G (2000) The genetic legacy of the quaternary ice ages. Nature 405, 907-913. Hewitt GM (2001) Speciation, hybrid zones and phylogeography - Or seeing genes in space and time. Molecular Ecology 10, 537-549. Hewitt GW (2004) Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society of London B: Biological Sciences 359, 183-195. Hey J (2010a) The divergence of chimpanzee species and subspecies as revealed in multipopulation isolation-with-migration analyses. Molecular Biology and Evolution 27, 921-933. Hey J (2010b) Isolation with migration models for more than two populations. Molecular Biology and Evolution 27, 905-920. Higgins PJ, Peter JM, Steele WK (2001) Handbook of Australian, New Zealand and Antarctic Birds. Oxford University Press, Melbourne. Hill KBR, Simon C, Marshall DC, Chambers GK (2009) Surviving glacial ages within the Biotic Gap: Phylogeography of the New Zealand cicada Maoricicada campbelli. Journal of Biogeography 36, 675-692. Hill WG (1981) Estimation of effective population size from data on linkage disequilibrium. Genetical Research 38, 209-216. Hodgson JA, Thomas CD, Wintle BA, Moilanen A (2009) Climate change, connectivity and conservation decision making: Back to basics. Journal of Applied Ecology 46, 964-969. Hoegh-Guldberg O, Hughes L, McIntyre S, et al. (2008) Assisted colonization and rapid climate change. Science 321, 345-346. Hoelzel AR (2010) Looking backwards to look forwards: Conservation genetics in a changing world. Conservation Genetics 11, 655-660. Hoffman JI, Amos W (2005) Microsatellite genotyping errors: Detection approaches, common sources and consequences for paternal exclusion. Molecular Ecology 14, 599-612.

116

References

Hoffmann AA, Sgró CM (2011) Climate change and evolutionary adaptation. Nature 470, 479-485. Holder MT, Anderson JA, Holloway AK (2001) Difficulties in detecting hybridization. Systematic Biology 50, 978-982. Holleley CE, Geerts PG (2009) Multiplex Manager 1.0: A cross-platform computer program that plans and optimizes multiplex PCR. BioTechniques 46, 511-517. Howe L, Castro IC, Schoener ER, et al. (2012) Malaria parasites (Plasmodium spp.) infecting introduced, native and endemic New Zealand birds. Parasitology Research 110, 913-923. Hughes L (2003) Climate change and Australia: Trends, projections and impacts. Austral Ecology 28, 423-443. Hung CM, Drovetski SV, Zink RM (2013) Recent allopatric divergence and niche evolution in a widespread Palearctic bird, the common rosefinch (Carpodacus erythrinus). Molecular Phylogenetics and Evolution 66, 103-111. Ingvarsson PK (2001) Restoration of genetic variation lost - The genetic rescue hypothesis. Trends in Ecology and Evolution 16, 62-63. Jakobsson M, Rosenberg NA (2007) CLUMPP: A cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23, 1801-1806. Jamieson IG, Allendorf FW (2012) How does the 50/500 rule apply to MVPs? Trends in Ecology and Evolution 27, 580-584. Jamieson IG, Grueber CE, Waters JM, Gleeson DM (2008) Managing genetic diversity in threatened populations: a New Zealand perspective. New Zealand Journal of Ecology 32, 130-137. Jamieson IG, Wallis GP, Briskie JV (2006) Inbreeding and endangered species management: Is New Zealand out of step with the rest of the world? Conservation Biology 20, 38-47. Jeffroy O, Brinkmann H, Delsuc F, Philippe H (2006) Phylogenomics: the beginning of incongruence? Trends in Genetics 22, 225-231. Johst K, Brandl R, Eber S (2002) Metapopulation persistence in dynamic landscapes: The role of dispersal distance. Oikos 98, 263-270. Kawecki TJ, Ebert D (2004) Conceptual issues in local adaptation. Ecology Letters 7, 1225-1241. Keyghobadi N (2007) The genetic implications of habitat fragmentation for animals. Canadian Journal of Zoology 85, 1049-1064. Keyghobadi N, Roland J, Strobeck C (2005) Genetic differentiation and gene flow among populations of the alpine butterfly, Parnassius smintheus, vary with landscape connectivity. Molecular Ecology 14, 1897-1909. Kimura M, Weiss GH (1964) The stepping stone model of population structure and the decrease of genetic correlation with distance. Genetics 49, 561-576. King DG, Soller M, Kashi Y (1997) Evolutionary tuning knobs. Endeavour 21, 36-40. King TM, Williams M, Lambert DM (2000) Dams, ducks and DNA: Identifying the effects of a hydro-electric scheme on New Zealand's endangered blue duck. Conservation Genetics 1, 103-113. Kircher M, Kelso J (2010) High-throughput DNA sequencing - Concepts and limitations. BioEssays 32, 524-536. Kolbe JJ, Glor RE, Schettino LR, et al. (2007) Multiple sources, admixture, and genetic variation in introduced Anolis lizard populations. Conservation Biology 21, 1612-1625.

117

References

Koumoundouros T, Sumner J, Clemann N, Stuart-Fox D (2009) Current genetic isolation and fragmentation contrasts with historical connectivity in an alpine lizard (Cyclodomorphus praealtus) threatened by climate change. Biological Conservation 142, 992-1002. Kruckenhauser L, Pinsker W (2004) Microsatellite variation in autochthonous and introduced populations of the Alpine marmot (Marmota marmota) along a European west-east transect. Journal of Zoological Systematics and Evolutionary Research 42, 19-26. Kunin WE, Gaston KJ (1993) The biology of rarity: Patterns, causes and consequences. Trends in Ecology & Evolution 8, 298-301. Lambert DM, Shepherd LD, Huynen L, et al. (2009) The molecular ecology of the extinct New Zealand Huia. PLoS ONE 4. Lampert KP, Rand AS, Mueller UG, Ryan MJ (2003) Fine-scale genetic pattern and evidence for sex-biased dispersal in the túngara frog, Physalaemus pustulosus. Molecular Ecology 12, 3325-3334. Lande R, Shannon S (1996) The role of genetic variation in adaptation and population persistence in a changing environment. Evolution 50, 434-437. Landis CA, Campbell HJ, Begg JG, et al. (2008) The waipounamu erosion surface: Questioning the antiquity of the New Zealand land surface and terrestrial fauna and flora. Geological Magazine 145, 173-197. Langella O (2002) POPULATIONS v.1.2.32 http://bioinformatics.org/~tryphon/populations/#ancre_telechargement Leschen RAB, Buckley TR, Harman HM, Shulmeister J (2008) Determining the origin and age of the Westland beech (Nothofagus) gap, New Zealand, using fungus beetle genetics. Molecular Ecology 17, 1256-1276. Li YC, Korol AB, Fahima T, Beiles A, Nevo E (2002) Microsatellites: Genomic distribution, putative functions and mutational mechanisms: A review. Molecular Ecology 11, 2453-2465. Liebgold EB, Gerlach NM, Ketterson ED (2013) Similarity in temporal variation in sex-biased dispersal over short and long distances in the dark-eyed junco, Junco hyemalis. Molecular Ecology 22, 5548-5560. Lloyd BD (2003) Intraspecific phylogeny of the New Zealand short-tailed bat Mystacina tuberculata inferred from multiple mitochondrial gene sequences. Systematic Biology 52, 460-476. Lockhart PJ, McLenachan PA, Havell D, et al. (2001) Phylogeny, radiation, and transoceanic dispersal of New Zealand alpine buttercups: Molecular evidence under split decomposition. Annals of the Missouri Botanical Garden 88, 458- 477. Loss SR, Terwilliger LA, Peterson AC (2011) Assisted colonization: Integrating conservation strategies in the face of climate change. Biological Conservation 144, 92-100. Lowe WH, Allendorf FW (2010) What can genetics tell us about population connectivity? Molecular Ecology 19, 3038-3051. Lugon-Moulin N, Hausser J (2002) Phylogeographical structure, postglacial recolonization and barriers to gene flow in the distinctive Valais chromosome race of the common shrew (Sorex araneus). Molecular Ecology 11, 785-794. Luikart G, Allendorf FW, Cornuet JM, Sherwin WB (1998) Distortion of allele frequency distributions provides a test for recent population bottlenecks. Journal of Heredity 89, 238-247.

118

References

Luikart G, Cornuet JM (1998) Empirical evaluation of a test for identifying recently bottlenecked populations from allele frequency data. Conservation Biology 12, 228-237. Luikart G, Cornuet JM (1999) Estimating the effective number of breeders from heterozygote excess in progeny. Genetics 151, 1211-1216. Luikart G, Ryman N, Tallmon DA, Schwartz MK, Allendorf FW (2010) Estimation of census and effective population sizes: The increasing usefulness of DNA-based approaches. Conservation Genetics 11, 355-373. Lukacs PM, Burnham KP (2005) Review of capture-recapture methods applicable to noninvasive genetic sampling. Molecular Ecology 14, 3909-3919. Maddison WP (1997) Gene Trees in Species Trees. Systematic Biology 46, 523-536. Margraf N, Verdon A, Rahier M, Naisbit RE (2007) Glacial survival and local adaptation in an alpine leaf beetle. Molecular Ecology 16, 2333-2343. Marshall DC, Hill KBR, Fontaine KM, Buckley TR, Simon C (2009) Glacial refugia in a maritime temperate climate: Cicada (Kikihia subalpina) mtDNA phylogeography in New Zealand. Molecular Ecology 18, 1995-2009. Marske KA, Leschen RAB, Buckley TR (2011) Reconciling phylogeography and ecological niche models for New Zealand beetles: Looking beyond glacial refugia. Molecular Phylogenetics and Evolution 59, 89-102. Masel J (2011) Genetic drift. Current Biology 21, R837-R838. McCulloch GA, Wallis GP, Waters JM (2010) Onset of glaciation drove simultaneous vicariant isolation of alpine insects in New Zealand. Evolution 64, 2033-2044. McDonald KA, Brown JH (1992) Using montane mammals to model extinctions due to global change. Conservation Biology 6, 409-415. McGlone MS (1985) Plant biogeography and the late Cenozoic history of New Zealand. New Zealand Journal of Botany 23, 723-749. McGlone MS (1989) The Polynesian settlement of New Zealand in relation to environmental and biotic changes. New Zealand Journal of Ecology 12, 115- 129. McGlone MS, Richardson SJ, Jordan GJ (2010) Comparative biogeography of New Zealand trees: Species richness, height, leaf traits and range sizes. New Zealand Journal of Ecology 34, 137-151. McLachlan JS, Hellmann JJ, Schwartz MW (2007) A framework for debate of assisted migration in an era of climate change. Conservation Biology 21, 297-302. Mech SG, Hallett JG (2001) Evaluating the effectiveness of corridors: A genetic approach. Conservation Biology 15, 467-474. Michelsen-Heath S (1989) The breeding biology of the rock wren, Xenicus gilviventris, in the Murchison Mountains, Fiordland National Park, New Zealand MSc thesis, University of Otago. Michelsen-Heath S, Gaze P (2007) Changes in abundance and distribution of the rock wren (Xenicus gilviventris) in the South Island, New Zealand. Notornis 54, 71- 78. Michelsen S (1982) Habitat requirements and feeding behaviour of the rock wren, Xenicus gilviventris in Mount Cook National Park, New Zealand Post-Grad. Dip. Wildlife Management thesis, University of Otago. Miller HC, Lambert DM (2006) A molecular phylogeny of New Zealand's Petroica (Aves: Petroicidae) species based on mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 40, 844-855. Mills LS, Allendorf FW (1996) The one-migrant-per-generation rule in conservation and management. Conservation Biology 10, 1509-1518.

119

References

Moritz C (1994) Defining 'evolutionarily significant units' for conservation. Trends in Ecology and Evolution 9, 373-375. Moritz C, Agudo R (2013) The future of species under climate change: Resilience or decline? Science 341, 504-508. Murphy SA, Flux IA, Double MC (2006) Recent evolutionary history of New Zealand's North and South Island Kokako (Callaeas cinerea) inferred from mitochondrial DNA sequences. Emu 106, 41-48. Nelson-Flower MJ, Hockey PAR, O'Ryan C, Ridley AR (2012) Inbreeding avoidance mechanisms: Dispersal dynamics in cooperatively breeding southern pied babblers. Journal of Animal Ecology 81, 876-883. Newnham RM, Lowe DJ, Williams PW (1999) Quaternary environmental change in New Zealand: A review. Progress in Physical Geography 23, 567-610. Nomura T (2008) Estimation of effective number of breeders from molecular coancestry of single cohort sample. Evolutionary Applications 1, 462-474. O'Connor KD, Marr AB, Arcese P, et al. (2006) Extra-pair fertilization and effective population size in the song sparrow Melospiza melodia. Journal of Avian Biology 37, 572-578. O'Keefe K, Ramakrishnan U, Van Tuinen M, Hadly EA (2009) Source-sink dynamics structure a common montane mammal. Molecular Ecology 18, 4775-4789. O'Neill SB, Buckley TR, Jewell TR, Ritchie PA (2009) Phylogeographic history of the New Zealand stick insect Niveaphasma annulata (Phasmatodea) estimated from mitochondrial and nuclear loci. Molecular Phylogenetics and Evolution 53, 523- 536. O'Neill SB, Chapple DG, Daugherty CH, Ritchie PA (2008) Phylogeography of two New Zealand lizards: McCann's skink (Oligosoma maccanni) and the brown skink (O. zelandicum). Molecular Phylogenetics and Evolution 48, 1168-1177. Osborne MJ, Norman JA, Christidis L, Murray ND (2000) Genetic distinctness of isolated populations of an endangered marsupial, the mountain pygmy-possum, Burramys parvus. Molecular Ecology 9, 609-613. Palsbøll PJ, Bérubé M, Allendorf FW (2007) Identification of management units using population genetic data. Trends in Ecology & Evolution 22, 11-16. Palsbøll PJ, Zachariah Peery M, Olsen MT, Beissinger SR, Bérubé M (2013) Inferring recent historic abundance from current genetic diversity. Molecular Ecology 22, 22-40. Palstra FP, Ruzzante DE (2008) Genetic estimates of contemporary effective population size: What can they tell us about the importance of genetic stochasticity for wild population persistence? Molecular Ecology 17, 3428- 3447. Parmesan C (1996) Climate and species' range. Nature 382, 765-766. Parmesan C (2006) Ecological and evolutionary responses to recent climate change. Annual Review of Ecology, Evolution, and Systematics 37, 637-669. Pauli H, Gottfried M, Grabherr G (2003) Effects of climate change on the alpine and nival vegetation of the Alps. Journal of mountain ecology 7, 9-12. Peakall R, Ruibal M, Lindenmayer DB (2003) Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes. Evolution 57, 1182-1195. Peakall R, Smouse PE (2006) GENALEX 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes 6, 288- 295.

120

References

Peery MZ, Beissinger SR, House RF, et al. (2008) Characterizing source sink dynamics with genetic parentage assignments. Ecology 89, 2746-2759. Perrin N, Mazalov V (2000) Local competition, inbreeding, and the evolution of sex- biased dispersal. American Naturalist 155, 116-127. Petrie M, Kempenaers B (1998) Extra-pair paternity in birds: explaining variation between species and populations. Trends in Ecology & Evolution 13, 52-58. Phillips BL, Baird SJE, Moritz C (2004) When vicars meet: A narrow contact zone between morphologically cryptic phylogeographic lineages of the rainforest skink, Carlia rubrigularis. Evolution 58, 1536-1548. Piry S, Luikart G, Cornuet JM (1999) BOTTLENECK: A computer program for detecting recent reductions in the effective population size using allele frequency data. Journal of Heredity 90, 502-503. Pounds JA, Bustamante MR, Coloma LA, et al. (2006) Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439, 161- 167. Pounds JA, Fogden MPL, Campbell JH (1999) Biological response to climate change on a tropical mountain. Nature 398, 611-615. Preston SAJ, Dawson DA, Horsburgh GJ, Hatchwell BJ (2013) Characterisation of microsatellite loci in the rifleman (Acanthisitta chloris, Acanthisittidae, AVES) and their predicted genome locations. Conservation Genetics Resources 5, 555- 560. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155, 945-959. Pritchard JK, Wen X, Falush D (2010) Documentation for structure software: Version 2.3. Available from: http://pritch.bsd.uchicago.edu/structure.html Prychitko TM, Moore WS (1997) The Utility of DNA Sequences of an Intron from the β-Fibrinogen Gene in Phylogenetic Analysis of Woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8, 193-204. Prychitko TM, Moore WS (2000) Comparative evolution of the mitochondrial cytochrome b gene and nuclear β-fibrinogen intron 7 in woodpeckers. Molecular Biology and Evolution 17, 1101-1111. Pudovkin AI, Zaykin DV, Hedgecock D (1996) On the potential for estimating the effective number of breeders from heterozygote-excess in progeny. Genetics 144, 383-387. Pudovkin AI, Zhdanova OL, Hedgecock D (2010) Sampling properties of the heterozygote-excess estimator of the effective number of breeders. Conservation Genetics 11, 759-771. Pulliam HR (1988) Sources, sinks and population regulation. American Naturalist 132, 652-661. Purvis A, Gittleman JL, Cowlishaw G, Mace GM (2000) Predicting extinction risk in declining species. Proceedings of the Royal Society B: Biological Sciences 267, 1947-1952. Pusey AE (1987) Sex-biased dispersal and inbreeding avoidance in birds and mammals. Trends in Ecology and Evolution 2, 295-299. Qu Y, Luo X, Zhang R, et al. (2011) Lineage diversification and historical demography of a montane bird Garrulax elliotii - Implications for the Pleistocene evolutionary history of the eastern Himalayas. BMC Evolutionary Biology 11. Queller DC, Goodnight KF (1989) Estimating relatedness using genetic markers. Evolution 43, 258-275.

121

References

Rambaut A (2012) FigTree v 1.4. Available from http://tree.bio.ed.ac.uk/ software/figtree/ Rambaut A, Drummond AJ (2007) Tracer v 1.5. Available from http://beast.bio.ed .ac.uk/Tracer Ricciardi A, Simberloff D (2009) Assisted colonization is not a viable conservation strategy. Trends in Ecology and Evolution 24, 248-253. Rice WR (1989) Analysing tables of statistical tests. Evolution 43, 223-225. Richardson DM, Hellmann JJ, McLachlan JS, et al. (2009) Multidimensional evaluation of managed relocation. Proceedings of the National Academy of Sciences of the United States of America 106, 9721-9724. Riney T (1953) Notes on habitat and behaviour of the rock wren species Xenicus gilviventris rineyi (Falla). Notornis 5, 186-188. Robertson BC, Stephenson BM (2008) DNA sexing of the critically endangered New Zealand storm petrel (Oceanites maorianus, or Pealeornis maoriana). Notornis 55, 209-211. Robertson BC, Stephenson BM, Goldstien SJ (2011) When rediscovery is not enough: Taxonomic uncertainty hinders conservation of a critically endangered bird. Molecular Phylogenetics and Evolution 61, 949-952. Robertson HA, Dowding JE, Elliott GP, et al. (2013) Conservation Status of New Zealand Birds, 2012. In: New Zealand Threat Classification Series 4, p. 22. Department of Conservation, Wellington. Rogers AR (1995) Genetic evidence for a pleistocene population expansion. Evolution 49, 608 - 615. Rogers AR, Harpending H (1992) Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9, 552-569. Roman J, Palumbi SR (2003) Whales before whaling in the North Atlantic. Science 301, 508-510. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574. Rosenbaum HC, Weinrich MT, Stoleson SA, et al. (2002) The effect of differential reproductive success on population genetic structure: Correlations of life history with matrilines in humpback whales of the Gulf of Maine. Journal of Heredity 93, 389-399. Rosenberg NA (2004) DISTRUCT: A program for the graphical display of population structure. Molecular Ecology Notes 4, 137-138. Rousset F (2008) GENEPOP'007: A complete re-implementation of the GENEPOP software for Windows and Linux. Molecular Ecology Resources 8, 103-106. Rubidge EM, Patton JL, Lim M, et al. (2012) Climate-induced range contraction drives genetic erosion in an alpine mammal. Nature Climate Change 2, 285-288. Rudá M, Žiak D, Kocian L, Martínková N (2010) Low genetic variability in a mountain rodent, the Tatra vole. Journal of Zoology 281, 118-124. Ruokonen M, Kvist L (2002) Structure and evolution of the avian mitochondrial control region. Molecular Phylogenetics and Evolution 23, 422-432. Ryder OA (1986) Species conservation and systematics: the dilemma of subspecies. Trends in Ecology & Evolution 1, 9-10. Saccone C, Attimonelli M, Sbisà 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. Sala OE, Chapin Iii FS, Armesto JJ, et al. (2000) Global biodiversity scenarios for the year 2100. Science 287, 1770-1774.

122

References

Savolainen O, Lascoux M, Merilä J (2013) Ecological genomics of local adaptation. Nature Reviews Genetics 14, 807-820. Schlaepfer MA, Helenbrook WD, Searing KB, Shoemaker KT (2009) Assisted colonization: evaluating contrasting management actions (and values) in the face of uncertainty. Trends in Ecology & Evolution 24, 471-472. Schlötterer C, Wiehe T (1999) Microsatellites, a neutral marker to infer selective sweeps. In: Microsatellites: Evolution and Applications (eds. Schlötterer C, Goldstein DB), pp. 238-248. Oxford University Press, Oxford. Schneider S, Excoffier L (1999) Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: Application to human mitochondrial DNA. Genetics 152, 1079-1089. Schuelke M (2000) An economic method for the fluorescent labeling of PCR fragments. Nature Biotechnology 18, 233-234. Seddon PJ (2010) From Reintroduction to Assisted Colonization: Moving along the Conservation Translocation Spectrum. Restoration Ecology 18, 796-802. Segelbacher G, H glund , Storch I (2003) From connectivity to isolation: Genetic consequences of population fragmentation in capercaillie across Europe. Molecular Ecology 12, 1773-1780. Segelbacher G, Storch I (2002) Capercaillie in the Alps: Genetic evidence of metapopulation structure and population decline. Molecular Ecology 11, 1669- 1677. Sekar S, Karanth P (2013) Flying between Sky Islands: The Effect of Naturally Fragmented Habitat on Butterfly Population Structure. PLoS ONE 8. Sekercioglu CH, Schneider SH, Fay JP, Loarie SR (2008) Climate change, elevational range shifts, and bird extinctions. Conservation Biology 22, 140-150. Seutin G, White BN, Boag PT (1991) Preservation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology 69, 82-90. Sherley GH (1985) The breeding system of the South Island rifleman (Acanthisitta chloris) at Kowhai Bush, Kaikoura, New Zealand. Unpublished PhD thesis. University of Canterbury, Christchurch. Sibley CG, Williams GR, Ahlquist JE (1982) The relationships of the New Zealand wrens (Acanthisittidae) as indicated by DNA-DNA hybridization. Notornis 29, 113-130. Slatkin M, Hudson RR (1991) Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics 129, 555-562. Smouse PE, Peakall R (1999) Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure. Heredity 82, 561-573. Smouse PE, Peakall R, Gonzales E (2008) A heterogeneity test for fine-scale genetic structure. Molecular Ecology 17, 3389-3400. Sorenson MD, Ast JC, Dimcheff DE, Yuri T, Mindell DP (1999) Primers for a PCR- Based Approach to Mitochondrial Genome Sequencing in Birds and Other Vertebrates. Molecular Phylogenetics and Evolution 12, 105-114. Stephens M, Donnelly P (2003) A Comparison of Bayesian Methods for Haplotype Reconstruction from Population Genotype Data. American Journal of Human Genetics 73, 1162-1169. Stephens M, Smith NJ, Donnelly P (2001) A new statistical method for haplotype reconstruction from population data. American Journal of Human Genetics 68, 978-989.

123

References

Stocker, Petyt C, Milne A (2006) A five-year study of rock wren in Henderson Basin, Kahurangi National Park. Occasional publication, Department of Conservation, Nelson. Storfer A, Murphy MA, Spear SF, Holderegger R, Waits LP (2010) Landscape genetics: Where are we now? Molecular Ecology 19, 3496-3514. Suggate RP (1990) Late pliocene and quaternary glaciations of New Zealand. Quaternary Science Reviews 9, 175-197. Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology and Evolution 15, 199-203. Sutton JT, Nakagawa S, Robertson BC, Jamieson IG (2011) Disentangling the roles of natural selection and genetic drift in shaping variation at MHC immunity genes. Molecular Ecology 20, 4408-4420. Swofford DL (2003) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Tallmon DA, Koyuk A, Luikart G, Beaumont MA (2008) ONeSAMP: A program to estimate effective population size using approximate Bayesian computation: COMPUTER PROGRAMS. Molecular Ecology Resources 8, 299-301. Tallmon DA, Luikart G, Waples RS (2004) The alluring simplicity and complex reality of genetic rescue. Trends in Ecology & Evolution 19, 489-496. Tamura K (1992) Estimation of the number of nucleotide substitutions when there are strong transition-transversion and G+C-content biases. Molecular Biology and Evolution 9, 678-687. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10, 512-526. Tamura K, Peterson D, Peterson N, et al. (2011) MEGA5: Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28, 2731- 2739. Tompkins DM, Gleeson DM (2006) Relationship between avian malaria distribution and an exotic invasive mosquito in New Zealand. Journal of the Royal Society of New Zealand 36, 51-62. Towns DR, Daugherty CH, Cree A (2001) Raising the prospects for a forgotten fauna: a review of 10 years of conservation effort for New Zealand reptiles. Biological Conservation 99, 3-16. Tracy LN, Jamieson IG (2011) Historic DNA reveals contemporary population structure results from anthropogenic effects, not pre-fragmentation patterns. Conservation Genetics 12, 517-526. Tracy LN, Wallis GP, Efford MG, Jamieson IG (2011) Preserving genetic diversity in threatened species reintroductions: How many individuals should be released? Animal Conservation 14, 439-446. Trewick SA (2001) Scree weta phylogeography: Surviving glaciation and implications for Pleistocene biogeography in New Zealand. New Zealand Journal of Zoology 28, 291-298. Trewick SA, Gibb GC (2010) Vicars, tramps and assembly of the New Zealand avifauna: A review of molecular phylogenetic evidence. Ibis 152, 226-253. Trewick SA, Wallis GP (2001) Bridging the "beech-gap": New Zealand invertebrate phylogeography implicates pleistocene glaciation and pliocene isolation. Evolution 55, 2170-2180.

124

References

Trewick SA, Wallis GP, Morgan-Richards M (2000) Phylogeographical pattern correlates with Pliocene mountain building in the alpine scree weta (Orthoptera, Anostostomatidae). Molecular Ecology 9, 657-666. Trizio I, Crestanello B, Galbusera P, et al. (2005) Geographical distance and physical barriers shape the genetic structure of Eurasian red squirrels (Sciurus vulgaris) in the Italian Alps. Molecular Ecology 14, 469-481. Van Riper Iii C, Van Riper SG, Goff ML, Laird M (1986) The epizootiology and ecological significance of malaria in Hawaiian land birds. Ecological Monographs 56, 327-344. Vangestel C, Callens T, Vandomme V, Lens L (2013) Sex-Biased Dispersal at Different Geographical Scales in a Cooperative Breeder from Fragmented Rainforest. PLoS ONE 8. Voelker G, Marks BD, Kahindo C, et al. (2013) River barriers and cryptic biodiversity in an evolutionary museum. Ecology and Evolution 3, 536-545. Vucetich JA, Waite TA (2000) Is one migrant per generation sufficient for the genetic management of fluctuating populations? Animal Conservation 3, 261-266. Wallis GP, Trewick SA (2009) New Zealand phylogeography: Evolution on a small continent. Molecular Ecology 18, 3548-3580. Walsh PS, Metzger DA, Higuchi R (1991) Chelex® 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. BioTechniques 10, 506-513. Walther GR, Post E, Convey P, et al. (2002) Ecological responses to recent climate change. Nature 416, 389-395. Wan QH, Wu H, Fujihara T, Fang SG (2004) Which genetic marker for which conservation genetics issue? Electrophoresis 25, 2165-2176. Wang J (2004) Application of the one-migrant-per-generation rule to conservation and management. Conservation Biology 18, 332-343. Waples RS (1991) Pacific salmon, Oncorhynchus spp., and the definition of "species" under the Endangered Species Act. Marine Fisheries Review 53, 11-22. Waples RS (2005) Genetic estimates of contemporary effective population size: To what time periods do the estimates apply? Molecular Ecology 14, 3335-3352. Waples RS (2006) A bias correction for estimates of effective population size based on linkage disequilibrium at unlinked gene loci. Conservation Genetics 7, 167-184. Waples RS, Do C (2010) Linkage disequilibrium estimates of contemporary Ne using highly variable genetic markers: A largely untapped resource for applied conservation and evolution. Evolutionary Applications 3, 244-262. Waples RS, Gaggiotti O (2006) What is a population? An empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Molecular Ecology 15, 1419-1439. Waples RS, Luikart G, Faulkner JR, Tallmon DA (2013) Simple life-history traits explain key effective population size ratios across diverse taxa. Proceedings of the Royal Society B: Biological Sciences 280. Warren MS, Hill JK, Thomas JA, et al. (2001) Rapid responses of British butterflies to opposing forces of climate and habitat change. Nature 414, 65-69. Waters JM, Craw D (2006) Goodbye Gondwana? New Zealand biogeography, geology, and the problem of circularity. Systematic Biology 55, 351-356. Waters JM, Craw D (2008) Evolution and biogeography of New Zealand's longjaw galaxiids (Osmeriformes: Galaxiidae): The genetic effects of glaciation and mountain building. Freshwater Biology 53, 521-534.

125

References

Waters JM, Wallis GP (2001) Mitochondrial DNA phylogenetics of the Galaxias vulgaris complex from South Island, New Zealand: Rapid radiation of a species flock. Journal of Fish Biology 58, 1166-1180. Watson DM (2002) A conceptual framework for studying species composition in fragments, islands and other patchy ecosystems. Journal of Biogeography 29, 823-834. Weeks AR, Sgro CM, Young AG, et al. (2011) Assessing the benefits and risks of translocations in changing environments: A genetic perspective. Evolutionary Applications 4, 709-725. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution 38, 1358-1370. Weir JT, Schluter D (2008) Calibrating the avian molecular clock. Molecular Ecology 17, 2321-2328. Weiser EL, Grueber CE, Jamieson IG (2012) Allele Retain: A program to assess management options for conserving allelic diversity in small, isolated populations. Molecular Ecology Resources 12, 1161-1167. Weiser EL, Grueber CE, Jamieson IG (2013) Simulating retention of rare alleles in small populations to assess management options for species with different life histories. Conservation Biology. Weston K (2006) Post-translocation monitoring of rock wren Xenicus gilviventris on Anchor Island. Internal report, Department of Conservation, Te Anau. Whitehouse IE, Pearce AJ (1992) Shaping the Mountains of New Zealand. In: Landforms of New Zealand (eds. Soons JM, Selby MJ), pp. 144–160. Longman Paul, Auckland. Wickes C, Crouchley D, Maxwell J (2009) Takahe (Porphyrio hochstetteri) recovery plan 2007 - 2012 Department of Conservation, New Zealand. Willans M, Weston K (2005) Translocation of rock wren Xenicus gilviventris from Murchison Mountains to Anchor Island, Dusky Sound, Fiordland. Department of Conservation Internal Report. Te Anau. Wilson AG, Arcese P, Chan YL, Patten MA (2011) Micro-spatial genetic structure in song sparrows (Melospiza melodia). Conservation Genetics 12, 213-222. Wilson GA, Rannala B (2003) Bayesian inference of recent migration rates using multilocus genotypes. Genetics 163, 1177-1191. Woodford DJ, McIntosh AR (2010) Evidence of source–sink metapopulations in a vulnerable native galaxiid fish driven by introduced trout. Ecological Applications 20, 967-977. Worthy TH, Hand SJ, Nguyen JMT, et al. (2010) Biogeographical and phylogenetic implications of an early Miocene Wren (Aves: Passeriformes: Acanthisittidae) from New Zealand. Journal of Vertebrate Paleontology 30, 479-498. Worthy TH, Holdaway RN (2002) The Lost World of the Moa: Prehistoric Life of New Zealand Indiana University Press, Bloomington. Wright S (1931) Evolution in Mendelian populations. Genetics 16, 97-159. Yannic G, Basset P, Büchi L, Hausser J, Broquet T (2012) Scale-specific sex-biased dispersal in the valais shrew unveiled by genetic variation on the y chromosome, autosomes, and mitochondrial dna. Evolution 66, 1737-1750. Yeaman S, Whitlock MC (2011) The genetic architecture of adaptation under migration-selection balance Evolution 65, 1897-1911. Yoccoz NG, Delestrade A, Loison A (2010) Impact of climatic change on alpine ecosystem:inference and prediction. Journal of Alpine Research 98.

126

References

Zachariah Peery M, Kirby R, Reid BN, et al. (2012) Reliability of genetic bottleneck tests for detecting recent population declines. Molecular Ecology 21, 3403- 3418. Zhang DX, Hewitt GM (2003) Nuclear DNA analyses in genetic studies of populations: Practice, problems and prospects. Molecular Ecology 12, 563-584. Zhdanova OL, Pudovkin AI (2008) Nb_HetEx: A Program to Estimate the Effective Number of Breeders. Journal of Heredity 99, 694-695.

127

128

References

Appendix One

84 1 100 1

99 96 1 1

Bayesian phylogeny of rock wren using control region (1126 bp). Posterior probabilities and maximum likelihood bootstrap support values are shown on main nodes. Scale indicates substitutions per site. Colours correspond to the Northern (blue) and Southern (red) populations identified in cluster and mtDNA analysis (Chapter 3)

129

Appendix Two

Pairwise sequence divergence at Cytochrome b (1075 bp) among one riflemen and 20 rock wren individuals. The number of base substitutions per site from between sequences are shown below diagonal. Standard error estimates are shown above the diagonal. Analyses were conducted using the Tamura-Nei model in MEGA v.5.

Rif Cla Hen Wan Tho Lew Bea Why Aor1 Aor2 Lan Twi Ahu Cru Haa Mat Hom Eyr Mur Roe Mac Rif 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.011 0.011 0.011 0.011 0.011 0.011 0.011 0.011 0.011 0.011 0.011 0.011 Cla 0.084 0.000 0.001 0.001 0.002 0.001 0.002 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Hen 0.084 0.000 0.001 0.001 0.002 0.001 0.002 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Wan 0.086 0.002 0.002 0.001 0.001 0.000 0.002 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Tho 0.084 0.002 0.002 0.002 0.002 0.001 0.002 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Lew 0.087 0.003 0.003 0.001 0.003 0.001 0.001 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Bea 0.086 0.002 0.002 0.000 0.002 0.001 0.002 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 Why 0.085 0.003 0.003 0.003 0.003 0.002 0 .003 0.001 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.005 0.006 0.005 0.005 Aor1 0.085 0.005 0.005 0.005 0.005 0.004 0.005 0.002 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.006 0.005 0.006 Aor2 0.097 0.038 0.038 0.038 0.038 0.039 0.038 0.037 0.039 0.000 0.002 0.002 0.002 0.002 0.002 0.002 0.001 0.002 0.003 0.003 Lan 0.097 0.038 0.038 0.038 0.038 0.039 0.038 0.037 0.039 0.000 0.002 0.002 0.002 0.002 0.002 0.002 0.001 0.002 0.003 0.003 Twi 0.099 0.038 0.038 0.038 0.038 0.039 0.038 0.037 0.039 0.004 0.004 0.001 0.002 0.002 0.002 0.002 0.001 0.002 0.003 0.003 Ahu 0.097 0.038 0.038 0.038 0.038 0.039 0.038 0.037 0.039 0.004 0.004 0.002 0.001 0.002 0.002 0.002 0.001 0.002 0.003 0.003 Cru 0.095 0.039 0.039 0.039 0.039 0.040 0.039 0.038 0.040 0.005 0.005 0.003 0.001 0.002 0.002 0.001 0.002 0.002 0.003 0.003 Haa 0.097 0.036 0.036 0.036 0.036 0.037 0.036 0.035 0.037 0.004 0.004 0.004 0.004 0.005 0.000 0.002 0.001 0.002 0.003 0.003 Mat 0.097 0.036 0.036 0.036 0.036 0.037 0.036 0.035 0.037 0.004 0.004 0.004 0.004 0.005 0.000 0.002 0.001 0.002 0.003 0.003 Hom 0.096 0.037 0.037 0.037 0.037 0.038 0.037 0.036 0.038 0.003 0.003 0.003 0.003 0.002 0.003 0.003 0.001 0.002 0.003 0.003 Eyr 0.097 0.036 0.036 0.036 0.036 0.037 0.036 0.035 0.037 0.002 0.002 0.002 0.002 0.003 0.002 0.002 0.001 0.001 0.003 0.003 Mur 0.097 0.038 0.038 0.038 0.038 0.039 0.038 0.037 0.039 0.004 0.004 0.004 0.004 0.005 0.004 0.004 0.003 0.002 0.003 0.003 Roe 0.093 0.035 0.035 0.035 0.035 0.036 0.035 0.034 0.036 0.011 0.011 0.011 0.011 0.012 0.011 0.011 0.010 0.009 0.011 0.001 Mac 0.095 0.036 0.036 0.036 0.036 0.037 0.036 0.035 0.037 0.012 0.012 0.012 0.012 0.013 0.012 0.012 0.011 0.010 0.012 0.001

130

Pairwise sequence divergence at the mtDNA control region (1119 bp) among 25 rock wren individuals. The number of base substitutions per site from between sequences are shown below diagonal. Standard error estimates are shown above the diagonal. Analyses were conducted using the Tamura 3- parameter model in MEGA v.5.

Cla Per Hen Wan Tho Lew1 Lew2 Ale Bea Why Aor1 Aor2 Twi Lan Ahu Gre Cru Mat Hom Eyr Mur1 Mur2 Roe1 Roe2 Mac Cla 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 Per 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.03 0.02 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.03 0.03 0.02 0.02 Hen 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.03 0.03 0.03 0.03 Wan 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 Tho 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.03 0.02 0.02 0.02 0.02 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 Lew1 0.01 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 Lew2 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.03 0.02 0.02 0.02 0.02 0.03 0.03 0.02 0.02 0.02 0.02 0.02 0.02 Ale 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.02 0.02 Bea 0.01 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 Why 0.03 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.00 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 Aor1 0.03 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.00 0.02 0.03 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02 0.02 0.02 Aor2 0.08 0.07 0.08 0.08 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Twi 0.09 0.08 0.08 0.09 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Lan 0.08 0.07 0.08 0.08 0.07 0.06 0.07 0.07 0.06 0.07 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Ahu 0.08 0.08 0.08 0.08 0.08 0.07 0.08 0.08 0.07 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Gre 0.08 0.08 0.08 0.08 0.08 0.07 0.08 0.08 0.07 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Cru 0.08 0.08 0.08 0.08 0.08 0.07 0.08 0.08 0.07 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Mat 0.09 0.08 0.08 0.09 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Hom 0.09 0.09 0.09 0.09 0.08 0.07 0.08 0.08 0.07 0.08 0.08 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.01 0.01 0.01 Eyr 0.08 0.07 0.07 0.08 0.07 0.06 0.07 0.07 0.06 0.07 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Mur1 0.08 0.08 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Mur2 0.08 0.08 0.08 0.08 0.08 0.07 0.08 0.08 0.07 0.08 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Roe1 0.09 0.08 0.08 0.09 0.08 0.07 0.08 0.09 0.07 0.07 0.07 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.00 0.00 Roe2 0.09 0.07 0.08 0.09 0.07 0.07 0.07 0.08 0.07 0.06 0.07 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.00 0.00 Mac 0.08 0.07 0.08 0.08 0.07 0.06 0.07 0.07 0.06 0.06 0.06 0.02 0.02 0.02 0.03 0.03 0.03 0.03 0.03 0.02 0.02 0.02 0.00 0.00

131

Pairwise sequence divergence at nuclear beta-fibrinogen intron-7 (526 bp) among one rifleman and 28 rock wren individuals. Number of base substitutions per site between sequences are shown below diagonal, S.E estimates above diagonal. Analyses conducted using Tamura 3-parameter model in MEGA v.5.

Rif Cla Per Hen Wan1 Wan2 Tho1 Tho2 Lew Bea1 Bea2 Why1 Why2 Aor1 Aor1 Aor2 Twi Lan1 Lan2 Ahu1 Ahu2 Haa1 Haa2 Eyr1 Mur1 Mur2 Roe Mac1 Mac2

Rif 0.01 0.02 0.02 0.02 0.02 0.01 0.01 0.02 0.01 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.01 0.01 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.02 0.01

Cla 0.04 0.01 0.01 0.01 0.01 0.00 0.00 0.01 0.00 0.02 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Per 0.04 0.03 0.00 0.00 0.00 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Hen 0.04 0.03 0.00 0.00 0.00 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Wan1 0.04 0.02 0.01 0.01 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.01

Wan2 0.04 0.03 0.00 0.00 0.01 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Tho1 0.04 0.00 0.03 0.03 0.02 0.03 0.00 0.01 0.00 0.02 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Tho2 0.03 0.00 0.03 0.03 0.02 0.03 0.00 0.01 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Lew 0.04 0.03 0.00 0.00 0.01 0.00 0.03 0.03 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Bea1 0.04 0.01 0.03 0.03 0.02 0.03 0.01 0.01 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Bea2 0.04 0.04 0.00 0.00 0.01 0.00 0.04 0.03 0.00 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.01 0.01

Why1 0.04 0.03 0.00 0.00 0.01 0.00 0.03 0.03 0.00 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Why2 0.04 0.04 0.00 0.00 0.01 0.00 0.04 0.03 0.00 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.01 0.01

Aor1 0.03 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Aor1 0.04 0.03 0.00 0.00 0.01 0.00 0.03 0.03 0.00 0.03 0.00 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Aor2 0.04 0.03 0.00 0.00 0.01 0.00 0.03 0.03 0.00 0.03 0.00 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Twi 0.04 0.03 0.00 0.00 0.01 0.00 0.03 0.03 0.00 0.03 0.00 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Lan1 0.04 0.02 0.01 0.01 0.00 0.01 0.02 0.02 0.01 0.02 0.02 0.01 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Lan2 0.04 0.03 0.01 0.01 0.01 0.01 0.03 0.02 0.01 0.02 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Ahu1 0.04 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.01

Ahu2 0.04 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00

Haa1 0.04 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.00 0.01 0.00 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.01

Haa2 0.04 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.00

Eyr1 0.03 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00

Mur1 0.03 0.03 0.00 0.00 0.01 0.00 0.03 0.02 0.00 0.02 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.00

Mur2 0.03 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00

Roe 0.03 0.03 0.01 0.01 0.01 0.01 0.03 0.03 0.01 0.03 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00

Mac1 0.04 0.02 0.02 0.02 0.01 0.02 0.02 0.01 0.02 0.01 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.00 0.01 0.00 0.01 0.00 0.01 0.01 0.01 0.01 0.01 0.00

Mac2 0.04 0.03 0.01 0.01 0.01 0.01 0.03 0.02 0.01 0.02 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01

132

Appendix Three

Neighbor-Joining tree based on proportion of shared alleles across 14 microsatellite loci for 134 rock wren individuals. Colours correspond to the Northern (blue) and Southern (red) populations identified in cluster and mtDNA analysis in Chapter 3.

133

Appendix Four

Cytochrome b haplotypes for 20 individuals sampled from throughout the range of rock wren (Xenicus giliventris)

Haplotype n Sampling location H1 2 Lake Clara + Henderson Basin H2 2 Wangapeka + Bealey H3 1 Lake Thompson H4 1 Lewis Pass H5 1 Whymper H6 1 Aoraki 1 H7 2 Aoraki 2 + Landsborough H8 1 Twin Stream H9 1 Ahuriri H10 1 Lake Crucible H11 2 Haast Range + Matukituki H12 1 Homer Basin H13 1 Eyre Mountains H14 1 Murchison Mountains H15 1 Lake Roe H16 1 Lake MacArthur

134

MtDNA control region 1st domain haplotypes for 55 individuals sampled from throughout the range of rock wren (Xenicus gilviventris)

Haplotype n Sampling location H1 2 Lake Thompson + Lewis Pass2 H2 3 Mt Perry1 & 2, Wangapeka1 H3 6 Lake Clara1 & 2, Henderson Basin1 & 2, Wangapeka 2 & 3 H4 3 Lewis Pass1, Otira, Bealey2 H5 1 Bealey1 H6 2 Mt Alexander1 & 2 H7 2 Whymper1 & 2 H8 3 Otira, Barker1 & 2 H9 1 Aoraki1 H10 16 Aoraki2, Twin Stream2, Landsborough1 & 2, Ahuriri1 & 2,Haast Range1, Lake Crucible1 & 2, Matukituki2, Lake Adelaide, Monkey Creek 1 & 2, Eyre Mtns 1 & 2, Murchison Mtns3 H11 2 Haast Range2, Matukituki1 H12 2 Murchison Mtns1 & 2 H13 1 Twin Stream1 H14 1 Matukituki3 H15 5 Lake Adelaide2, Lake Marion, Homer 1 - 3 H16 3 Lake MacArthur1 & 2, Lake Roe2 H17 1 Lake Roe1 H18 1 Lake Roe3

135

Nuclear beta-fibrinogen intron-7 rock wren (Xenicus gilviventris) haplotypes with a phased probability of >0.75. Identified from 18 individuals, of which 9 individuals were heterozygous. N = number of individuals with haplotype

Haplotype n Sampling location H1 2 Lake Clara, Lake Thompson_a H2 7 Mt Perry, Lake Henderson, Lewis Pass, Wangapeka_b, Whymper_a, Aoraki2, Twin Stream H3 1 Wangapeka_a H4 1 Lake Thompson_b H5 1 Bealey_a H6 2 Bealey_b, Whymper_b H7 4 Aoraki1_a, Eyre Mtns, Murchison Mtns_b, Lake Roe H8 1 Aoraki1_b H9 1 Landsborough_a H10 1 Landsborough_b H11 2 Ahuriri_a, Haast Range_a H12 2 Ahuriri_b, Haast Range_b H13 1 Murchison Mtns_a H14 1 Lake Macarthur_a H15 1 Lake Macarthur_b

136

Appendix Five

a) 0.35 y = -0.0008x + 0.3591 R² = 0.5823

0.30

0.25

FST) -

0.20 FST/(1

0.15

0.10 0 50 100 150 200 250 Distance between sampling areas (km)

b) 0.20 y = 0.0006x - 0.0142 0.18 R² = 0.7618 0.16

0.14

0.12 FST) - 0.10

0.08 FST/(1 0.06 0.04 0.02 0.00 50 100 150 200 250 300 350 Distance between sampling areas (km)

Isolation by distance results based on multilocus microsatellite data from rock wren (Xenicus gilviventris) showing the relationship between pairwise population differentiation (FST/(1-FST) and geographical distance across a) the northern South Island and b) the southern South Island. The trend, represented as a solid line, was not significant (p=0.168) in the North and was significant (0.016) in the South as determined through Mantel tests with 9999 permutations.

137