CONSERVATION GENETICS OF ROOSEVELT ELK IN BRITISH COLUMBIA IAN F. GAZELEY Bachelor of Science, University of Lethbridge, 2017 A thesis submitted in partial fulfilment of the requirements for the degree of MASTER OF SCIENCE in BIOLOGICAL SCIENCES Department of Biological Sciences University of Lethbridge LETHBRIDGE, ALBERTA, CANADA © Ian F. Gazeley, 2021 CONSERVATION GENETICS OF ROOSEVELT ELK IN BRITISH COLUMBIA IAN FREDERICK GAZELEY Date of Defense: April 21, 2021 Dr. T. Burg Professor Ph.D. Thesis Supervisor Dr. C. Goater Professor Ph.D. Thesis Examination Committee Member Dr. A. Iwaniuk Associate Professor Ph.D. Thesis Examination Committee Member Dr. R. Laird Associate Professor Ph.D. Chair, Thesis Examination Committee DEDICATION For my wife Debra, without whose support I would never have attempted a return to school. Your bravery and perseverance, in the face of adversity, inspired me to reach beyond my preconceptions and self-imposed limits. And for my parents, Fred and Nadine. Your constant love and support throughout my early years, despite the challenges, is appreciated more than you know… but I’d bet you would never have thought you would see the day! iii ABSTRACT Species reintroductions have the potential to cause bottleneck events resulting in increased genetic drift, reduced genetic diversity and increased inbreeding, with potentially negative fitness consequences. Wildlife managers must consider how a species’ ecology may affect its genetic diversity. Roosevelt elk, once widespread along the West Coast, were extirpated from the mainland and experienced a substantial population bottleneck on Vancouver Island. The species was reintroduced to the BC mainland in the 1980s, and their descendants used for subsequent reintroductions within the region. To understand genetic diversity in extant and reintroduced populations of Roosevelt elk, we analyzed genetic variation in 355 elk from 13 populations. Molecular analyses showed reduced genetic diversity, genetic isolation of southern Vancouver Island, increased genetic drift resulting in significant differentiation between source and reintroduced herds, and very low effective population size in multiple populations indicating a potential for inbreeding and associated negative fitness consequences. iv ACKNOWLEDGEMENTS First and foremost, I need to thank Dr. Theresa Burg who has been more patient than could be expected. Her wisdom in dealing with the bureaucracy of the University, and her timely sage advice in challenging times, helped get me here today. When I was considering continuing on beyond my undergraduate studies, and was thinking about researching Roosevelt Elk, Darryl Reynolds’ excitement and willingness to help made everything work. He pointed me to the right people in the BC Government that I needed to talk to, provided immeasurable advice, and assisted with collecting samples along the way. Thanks to Dr. Cameron Goater and Dr. Andrew Iwaniuk for helping Theresa keep me from getting off track, and reeling me back in after taking on too much. My lab-mates in the Burg Lab deserve a special thanks; when asked, Brendan Graham has always provided a guiding word, or two, or more… while Dilini, Kanishka and Amanda always laughed at my jokes. Okay, maybe 2/3 of my jokes. Matson’s Laboratory and Carolyn Nistler in Montana for going above and beyond to help me access samples. John Kelly, Helen Schwantje, Cliff Neitvelt, Billy Wilton and many others with the BC Ministry of Forests, Lands and Natural Resource Operations and Rural Development who helped me with planning, sample collection, data access and advice. v TABLE OF CONTENTS DEDICATION iii ABSTRACT iv ACKNOWLEDGEMENTS v LIST OF TABLES viii LIST OF FIGURES ix LIST OF ABBREVIATIONS x CHAPTER 1: General Introduction 1 1.1 Background 1 1.1.1 Conservation Biology 1 1.1.2 Population Genetics 2 1.1.2.1 Evolution 2 1.1.2.2 Influences on Genetic Structure 3 1.1.2.3 Species at Risk 6 1.2 Study Area 6 1.3 Study Species – Ecology, Distribution and History 8 1.4 Molecular Markers 13 1.5 Study Design 14 1.6 Thesis Overview 15 1.7 Literature Cited 16 CHAPTER 2: Multi-marker Genetic Analysis Delineates Metapopulation 23 Structure in Roosevelt Elk in British Columbia 2.1 Introduction 23 2.2 Methods 27 2.2.1 Sample Acquisition 27 2.2.2 DNA Extraction and Amplification 28 2.2.3 Sequence Analyses 31 2.2.4 Microsatellite Analyses 31 2.3 Results 35 2.3.1 Sampling 35 2.3.2 Sequencing 35 2.3.3 Microsatellites 36 2.4 Discussion 41 2.4.1 Population Structure 41 2.4.2 Reintroduction Effects 43 2.5 Conclusions 51 2.6 Literature Cited 54 CHAPTER 3: General Discussion 73 3.1 Management Implications 74 3.2 Future Directions 75 3.3 Conclusions 78 3.4 Literature Cited 79 APPENDIX I: Translocation History of Roosevelt Elk 81 APPENDIX II: Summary of Samples 83 v i APPENDIX III: Primers Used in Study 90 APPENDIX IV: Hardy-Weinberg Equilibrium Table 92 APPENDIX V: Linkage Disequilibrium 93 APPENDIX VI: Microsatellite Allele Frequency Table 94 vii LIST OF TABLES Table 2.1: Mitochondrial Haplotypes in Roosevelt Elk 61 Table 2.2: Summary of Mitochondrial Diversity 61 Table 2.3: Population Groupings for Analyses 62 Table 2.4: Population Genetic Diversity Statistics 63 Table 2.5: Microsatellite Loci Diversity Statistics 64 Table 2.6: Pairwise Genetic Difference Between Populations 65 Table 2.7: Bottleneck Analysis Results 66 Table 2.8: Effective Population Sizes 67 vi ii LIST OF FIGURES Figure 1.1: Bottleneck Events and Drift 20 Figure 1.2: Historic and Current Distribution of Roosevelt Elk 21 Figure 1.3: Photographs of Roosevelt Elk with Genetic Anomalies 22 Figure 2.1: Roosevelt Elk Translocation History – Vancouver Island - Mainland 68 Figure 2.2: Mainland Translocations 2001 – 2013 69 Figure 2.3: Elk Samples and Population Groupings Map 70 Figure 2.4A-B: Haplotype Map, Network and Phylogenetic Tree 71 Figure 2.5A-B: Bar Plots of Ancestry Assignments 72 ix LIST OF ABBREVIATIONS ˚C Degrees Celsius μL Microlitre μM Micromolar 1x One times 5x Five times χ2 Chi squared statistic a Alpha Ar Allelic richness avg. Average BC British Columbia bp Base pair BRIT Brittain-Skwawka group BSA Bovine serum albumin CC Central Coast group (Deserted-Narrows-Clowholm) D Jost’s measure of genetic distance df Degrees of freedom DNA Deoxyribonucleic acid EPU Elk population unit FDR Benjamini-Höchberg false discovery rate FLNRO British Columbia Ministry of Forests, Lands, Natural Resource Operations and Rural Development Fst Fixation index Gis Inbreeding coefficient He Expected heterozygosity Ho Observed heterozygosity Ht Total heterozygosity expected across all populations uHe Unbiased expected heterozygosity HO Homathko-Southgate-Orford group HWE Hardy-Weinberg Equilibrium IAM Infinite alleles model K Genetic cluster ∆K Delta K Prob(K) Probability of K km Kilometres LD Linkage disequilibrium m Metres MgCl2 Magnesium chloride mM Millimolar mtDNA Mitochondrial DNA n / N Sample size / Census population estimate Na Number of observed alleles Ne Effective population size P Probability x PCR Polymerase Chain Reaction PITT Pitt River group PRN Powell River North group PRS Powell River South group RG Rainy-Gray group SMM Step-wise mutation model SP Sechelt Peninsula group SQ Squamish-Mamquam-Indian group STV Stave-Chehalis-Lower Lillooet group TOBA Toba group TPM Two-phase mutation model VIN Vancouver Island North group VIS Vancouver Island South group xi CHAPTER 1: GENERAL INTRODUCTION 1.1 Background A common definition of biodiversity is the abundance, evenness and / or richness of organisms in a defined space: e.g. a single quadrat frame, sampling area, habitat, or ecosystem. However, biodiversity is more than measures of the number and distribution of species; at its fundamental root, biodiversity is genetic (Frankham 2005, Ralls et al. 2018). In viewing any population, be it bacterial, fungal, plant or animal, its greatest level of diversity is contained within the combined genomes of its individuals. Understanding the diversity characteristics of many species and their populations is critical for their long-term conservation, therefore, we must consider how management choices affect the genetic diversity of species we choose to manage (Frankham 2005, Charlesworth and Willis 2009, Balkenhol et al. 2016, Palsbøll et al. 2007, Funk et al. 2012). 1.1.1 Conservation Biology An overarching goal for ecological conservation is maintaining viable, healthy and productive populations of native species (Koenig 1988, Frankham 2005). An important consideration in conserving species and ecosystem function, is the sustainability of individual populations (a group of interbreeding individuals of a species). Small populations are more vulnerable to catastrophic events such as drought, heavy snow, flooding and disease (Koenig 1988, Young 1994). The best mitigation against population loss is to maintain large populations. Connectivity between smaller demes (groups of interbreeding individuals of the same species; Gilmour & Gregor, 1939) through the dispersal of individuals, increases the overall effective population size (Nei & Tajima, 1981, Nei, Maruyama, & Chakraborty, 1975, Reed & Frankham, 2003, Wright, 1931). 1 Appropriately, wildlife managers have increasingly focussed on preserving suitable ‘connected habitat’ to ensure that populations have the greatest number of potential interbreeding individuals. Such connectivity is often assumed, without supporting empirical evidence of metapopulation dynamics (Lowe and Allendorf 2010, Betts et al. 2015). While direct observations of dispersal, such as sightings of marked individuals or global positioning system (GPS) collar tracking data, can verify individual movements, they rarely confirm if those individuals have successfully reproduced. At a fundamental level, conservation biology is focussed on understanding the long term viability of populations (Young 1994). 1.1.2 Population Genetics 1.1.2.1 Evolution The field of population genetics is an extension of evolutionary theory and conservation biology, wherein our understanding of biodiversity includes the genetic similarities and differences within and among individuals, demes, populations and species or what we call genetic ‘structure’. Population genetics involves understanding how the mechanisms of evolution (mutation, gene flow, drift, non-random mating and selection; Wright, 1931) individually and collectively influence the genetic structure of populations and species. Mutation is the fundamental agent of genetic diversity, providing variation in the genomes of organisms upon which all other evolution mechanisms act (Wright 1931). The processes involved in mutation, while underlying much of what is discussed herein, are beyond the scope of this thesis; therefore, mutation will only be discussed in as much detail as is required for the understanding of a general audience. Gene-flow is the 2 dispersal of gametes among demes or populations and is synonymous with connectivity. Gene flow is affected by various characteristics of the organism in question such as fecundity or dispersal ability and interaction with its environment, including physical barriers to dispersal or distance to conspecifics. Drift is the stochastic change in frequency of alleles (specific variants of a genetic sequence) within a deme or population due to the effect of random sampling of gametes in reproduction, eventually leading to the loss or fixation of alleles. Selection is the differential survival and fitness (lifetime reproductive success) of individuals with different alleles interacting with their environment (climate, food, parasites, other individuals, etc.). Non-random mating can be due to isolation and small population, or assortative (preferential) mating. A fundamental concept in evolutionary theory is Hardy-Weinberg Equilibrium (HWE). For populations in HWE allele frequencies can be predicted from one generation to the next. Such populations will have five characteristics: they are large (mitigating drift), individuals breed at random, there is no net migration or mutation, and alleles are not under selection (all genotypes equally fit). Departures from HWE may indicate a violation of one or more of the above characteristics. Figuring out which requires researchers to evaluate overall patterns, and carefully examine potential causes (Waples 2015). 1.1.2.2 Influences on Genetic Structure The five main agents of evolution each play a role in the genetic structure and levels of variation observed in natural populations, however, in small populations drift may act much faster and with greater magnitude (Nei & Tajima, 1981; Nei et al., 1975). A large reduction in a population is referred to as a ‘population bottleneck’. Even when a 3 reduction in population size is of short duration, drift can alter allele frequencies quickly, creating a ‘genetic bottleneck’. Consider a small population of 24 diploid organisms (Figure 1.1) each with two of four possible alleles at a single locus (blue, yellow, purple or green). If the population is randomly reduced by 75%, as might happen during a severe winter storm, the loss of alleles is also random. While it is statistically probable that alleles of high frequency will be present in the surviving individuals, changes in rare allele frequencies are more unpredictable. In Figure 1.1 the purple allele happened to have two copies carried into the new population, increasing its frequency from 6% to 16%, while the yellow allele (initially more than twice as common as purple) was lost. Let us assume that the surviving individuals all successfully breed and rebuild to the pre- bottleneck population within one year. Within two generations, one allele has been lost from the population, a rare allele became relatively common and then rare again, and the frequencies of the two most common alleles have flipped. Like drift, the effects of non-random mating, mutation, and migration is increased in small populations. An important consideration in population genetics is the difference between the census population, or number of individuals (N), and the effective population (Ne), the number of breeding individuals in an idealized population (Frankham, 1995, Kimura & Crow, 1963, Wright, 1931). Ne is often much smaller than N, and this has important conservation implications. Small populations, as low values of Ne suggest, are greatly affected by drift and populations below a threshold may not be sustainable. Widely accepted guidelines for species conservation suggest populations that fall below Ne = 50 may be at high risk of inbreeding depression, and populations should be maintained over the long term at Ne >500 to maintain adaptive diversity, known as the 4 50/500 rule (Franklin 1980, Soulé 1980). Since the 50/500 rule was adopted by the International Union for Conservation of Nature (IUCN), new research has suggested that these numbers are much too low for most wild populations (Keller and Waller 2002, Frankham 2005) and Ne values of 100/1000 likely represents reasonable minimum targets to conserve species at risk (Frankham et al. 2014). The Ne:N ratio provides a metric for evaluating levels of genetic variation in a population (Hedrick 2005). High values (~1) would be indicative of populations with high diversity and therefore less vulnerable to stochastic events. In a review of published data in which both N and Ne estimates were available, Frankham (1995) found that natural populations across taxa (102 species of birds, mammals, insects, molluscs, amphibians, reptiles and plants), some at risk and many common, exhibited an average Ne:N value of 0.10-0.11, or we can consider it as one ‘genetic individual’ for every 9-10 counted individuals in a population. Such a low value has far reaching implications in managing small populations when we recall that long term sustainability goals require an Ne of 100 short term, and >1000 long term. The most significant influences on Ne estimates were fluctuations in population size, variance in reproduction (as expected in polygynous systems), and the census population value used (all individuals, only adults, only breeding individuals). 5 1.1.2.3 Species at Risk Species at risk are, by definition, small populations. To ensure their continued survival, it is imperative to increase Ne as much as possible by maintaining or increasing gene flow between fragmented populations (Palsbøll et al. 2007, Frankham 2015, Ralls et al. 2018). Unfortunately, many populations are fragmented due to anthropogenic disturbance that is not easily remedied. Where populations have poor connectivity due to physical distance, using reintroductions to establish new populations between existing ones may be beneficial. Where connectivity is low due to the unsuitability of the landscape between populations, whether temporary or permanent, translocation of individuals among populations in a genetically informed approach can provide important new diversity to small populations (Frankham 2015, Giglio et al. 2018). However, increasing the overall population of a species is the pre-eminent goal (Allendorf and Luikart 2007). 1.2 Study Area The southwest coast of British Columbia (BC) is divided into two management areas in relation to wildlife and forestry management, the West Coast (WCR) and South Coast (SCR) Natural Resource Regions (Figure 1.2). The WCR encompasses Vancouver Island and the mainland coast and islands from near Phillips Arm to north of Bella Coola and includes Haida Gwaii. The portion of the WCR included in the study area is limited to Vancouver Island, and a few elk population units (EPUs) on the adjacent mainland (Figure 1.2). The SCR is comprised of mainland areas and numerous small islands from the US border up to the north side of Bute Inlet. The southern coastal areas of BC are highly heterogenous landscapes, characterized by topographic and climatic extremes. The 6 region is typified by temperate rainforests, alluvial flood plains, subalpine forests, alpine meadows, glaciers and broad icefields, with mountain peaks exceeding 2500 m and broad icefields less than 10 km from the ocean. The mainland is a mosaic of high elevation rock and ice, broad glacier-formed forested valleys, and deep fjords reaching far inland from the Salish Sea. Large terrestrial mammals are found throughout the regions, including black-tailed deer (Odocoileus hemionus columbianus), cougar (Puma concolor), black bear (Ursus americanus), gray wolf (Canis lupus) and Roosevelt elk (Cervus canadensis roosevelti) in both regions, with mountain goat (Oreamnos americanus), grizzly bear (U. arctos horribilis) and occasional moose (Alces alces) only occurring on the mainland. Vancouver Island has a population of ~870,000 people, with almost half living in and around Victoria, and the remainder are mostly distributed in communities along the east coast (BC Government 2020). The SCR is home to ~2.8 million people, with most living in the highly developed Fraser River floodplain and surrounding uplands (Metro Vancouver Regional District and Fraser Valley Regional District). Approximately 90,000 live in the Squamish-Lillooet Regional District and developed coastline areas (BC Government 2020). Outside of urban and suburban development resource extraction, mainly forestry, has been the main disturbance agent the last two centuries. More recently the construction and operation of small hydroelectric facilities, and increasing wilderness recreation have created significant disturbance, both direct and indirect, on the landscape and wildlife of the SCR (Mountain Goat Management Team 2010). Vegetation in the region is dominated by dense temperate rainforests of western hemlock (Tsuga heterophylla), red cedar (Thuja plicata), Sitka spruce (Picea sitchensis) in fluvial soils near sea level and Douglas fir (Pseudotsuga menziesii) occurring in drier, 7 well drained sites. Mountain hemlock (T. martensiana) and yellow cedar (Chamaecyparis nootkatensis) dominate subalpine forests at higher elevations giving way to alpine tundra and vast glaciers and icefields on the mainland, while deciduous forests of black cottonwood (Populus trichocarpa), red alder (Alnus rubra) and big leaf maple (Acer macrophylum) occur in pure and mixed stands on disturbed sites at low elevation and along numerous river valleys (Meidinger & Pojar 1991). Annual precipitation varies considerably throughout the region from ~700 mm on southern Vancouver Island to over 3500 mm at nearby Port Renfrew, and from 1200 mm at Powell River to over 2800 mm near Squamish on the mainland (Environment and Climate Change Canada 2020). 1.3 Study Species - Ecology, Distribution and History Roosevelt Elk (C. c. roosevelti) are one of four disputed extant North American elk (C. canadensis Erxleben 1777) subspecies collectively known as wapiti. The other subspecies being: Rocky Mountain (C. c. nelsoni), tule (C. c. nannodes) and Manitoban (C. c. manitobensis) elk, along with two extinct subspecies, the Merriam (C. c. merriami) and Eastern (C. canadensis) elk. Until recently, wapiti were considered sub-species of the Eurasian red deer (Cervus elaphus Linnaeus 1758). Studies on pre and postzygotic isolation in C. canadensis and C. elaphus hybrids (Dratch 1986) and modern genetic analyses have shown that all C. canadensis taxa (which includes the Asian C. c. sibericus, C. c. wallichi and C. c. songaricus) are monophyletic and share a most recent common ancestor, and are sister clade to the Sika deer (C. nippon) of eastern Asia, not a subspecies or sister clade to C. elaphus, as long accepted (Polziehn and Strobeck 1998, Lorenzini and Garofalo 2015). In the following discussion the term wapiti is used when 8 discussing all subspecies collectively, and by the common term elk in reference to a singular population, e.g. Roosevelt Elk, Vancouver Island Elk, Sechelt Peninsula Elk, etc. Recent research by Speller et al. (2014), using both contemporary and archaeological genetic samples, suggests that prior to their extirpation from most of North America wapiti occurred as a relatively continuous population from the Rocky Mountains eastward through the Great Plains, calling into question the validity of separating the Manitoban and Rocky Mountain subspecies. Using both mitochondrial and microsatellite loci Pohlzein et al. (1998; 2000) supported the subspecific designation of Roosevelt elk from Vancouver Island, as well as tule elk in California, under the phylogenetic species concept. The authors noted that Roosevelt elk in the northwest United States showed genetic introgression with elk translocated from Yellowstone National Park (NP) in the early 1900s, leaving the BC population as the only remaining ‘pure’ Roosevelt elk. Wapiti shared a common fate with much of North America’s wildlife; as Old World diseases killed the majority of the indigenous human population (Jones 2014), large ungulates that had been heavily utilized initially experienced a release from human predation and exploded in numbers to where descriptions from early European explorers defied belief (McHugh 1979, Gray 1995). Regardless of the numerical accuracy of those reports, it cannot be argued that central and western North America was awash with large mammals by the early 1800s. As the west was further explored, exploited and settled, wildlife was initially viewed as a resource to feed local settlers and communities, and a rapidly expanding urban population in eastern regions (Gray 1995). As cattle were moved into the west and land was used for crops, wildlife was persecuted as competitors of 9 livestock, pests upon agriculture, and to deprive remaining free indigenous peoples a way to continue to exist outside of government control (McHugh 1979). Due to widespread anthropogenic change, by the turn of the twentieth century the legendary abundance of much of North America’s wildlife was nothing but a recent memory. Many species that were once numerous beyond comprehension were now at risk of extinction; the buffalo (Bison bison) once numbered between 30-60 million 200 years after Europeans first set foot on the continent. By 1883, an estimated 325 plains bison (B. b. bison) remained, with only 25 on public lands, and approximately 500 wood bison (B. b. athabascae) near Slave Lake, NWT (McHugh 1979). Unfortunately other legendary species had already been lost, such as the passenger pigeon (Ectopistes migratorius) and Eskimo curlew (Numenius borealis) (Gray 1995, Barsness 2000). Many species of large ungulates were extirpated from former habitats and survived only as remnant populations in protected areas, like Yellowstone NP in the United States and Riding Mountain NP in Canada (Gray 1995, O’Gara and Dundas 2002). Other species, though greatly reduced, found refuge by virtue of their remoteness or difficulty of terrain. The Roosevelt elk fell into that category. Roosevelt elk once ranged widely along the west coast of North America from northern California to southern BC (Figure 1.1; Spalding 1992, O’Gara and Dundas 2002). By the early 1900s only a handful of populations remained. In the province they were extirpated from the mainland, and the total population was reduced to as few as 375 individuals in four herds on Vancouver Island; Shawnigan Lake, Strathcona Park, Kyuquot Sound and Quatsino Sound (Spalding, 1992; Figure 1.2). Though no estimates exist of pre-colonial populations for the sub-species, historic records from the late 1700s 10 to late 1800s report elk occurring at low densities along coastal areas of the province from the Fraser River delta as far as 52° N (near Bella Coola), wherever suitable habitats (major river valleys, riparian zones, coastal estuaries and plains) were available (Brunt 1990, Spalding 1992). The territorial government in BC began enacting various piecemeal legislation regarding the taking of fish and wildlife as early as 1859 (Begg 2007), however, the exploitation of wildlife continued, resulting in large reductions of many species until the early 1900s. To help stem the ongoing loss of wildlife, BC followed many other North American jurisdictions in introducing a consolidated wildlife act. The Game Act (1914) provided legislation to address and control the widespread exploitation of endemic species (Begg 2007). When introduced, the act immediately closed the hunting of elk on Vancouver Island, allowing the population to slowly increase from a substantial population bottleneck (Spalding 1992). The population of elk on Vancouver Island increased moderately until the 1970s. Illegal harvest of elk had always been a concern on the island, however poaching increased substantially in the late 1970s and early 1980s, and combined with the extensive loss of low elevation old growth forests critical to winter survival, wildlife managers became worried about the species’ persistence (Spalding 1992, Quayle and Brunt 2003). To address conservation concerns provincial biologists translocated 22 Roosevelt elk from the Qualicum and Campbell River areas on Vancouver Island to the Sechelt Peninsula on the mainland in the late 1980s, and moved an additional five elk in the mid 1990s (Figure 1.2; Spalding 1992). The new mainland populations grew rapidly. So much so that conflict with local residents became a political issue by the late 1990s. In 11 an effort to address both wildlife conflicts (agricultural depredation, vehicle accidents) and restore the species to other former habitats on the mainland, a multi-year translocation and reintroduction program was begun (Reynolds et al. 2018). The project involved trapping groups of elk living in close proximity to developed areas on the Sechelt Peninsula and Powell River, and translocating them to isolated high quality habitats within the South Coast Region. Simultaneously, a limited elk hunting season was initiated on the Sechelt Peninsula, the original site of the mainland introductions (Quayle and Brunt 2003, Reynolds et al. 2018). Following the recommendations of Komers and Curman (2000), the strategy used for Roosevelt elk reintroductions from 2000 to 2017 was to provide a founding population of at least 20 individuals (min = 4, max = 56), heavily biased towards females. In practice, the majority of elk trapped and moved were cows and calves (84.7%), with lesser numbers of immature bulls. An effort was made to translocate at least one mature bull (4+ years) into each new herd when possible (Reynolds et al. 2018). Through 84 translocations, of 601 elk into 27 new elk population units (EPUs), the mainland population grew from ~400 animals in two EPUs in 2001 to ~2050 animals by 2020. It bears repeating, the vast majority of elk used for mainland translocations were descendants of 27 elk from Vancouver Island. In recent years at least two individuals in the Rainy-Gray EPU were observed with nontypical colouration, putatively the result of genetic mutations (Figure 1.3). Though hypothetical, the individuals were suspected to display partial leucism, wherein a mutation disrupts pigment production pathways (melanogenesis) in some cell types resulting in either uniform discolouration or partial discolouration (piebaldism) of the 12 skin or hair. The melanogenesis pathway is highly complex, involving multiple genes involved in regulation of both qualitative and quantitative expression, some recessive while others are dominant, and identical mutations in specific genes result in different phenotypes in different species (Barsh 1996). While some colour phase animals, such as the Kermode black bear, exhibit recessive alleles that interrupt the production of the black-brown pigment eumelanin (Ritland et al. 2001), it may be that the observed elk have mutations affecting the production of red-yellow pigment pheomelanin in the same MCR1 gene. Further discussion of the genetic causes of pigment disorders are beyond the scope of this thesis, however, it should be noted that such mutations are very infrequently expressed in large, wild populations. As most mutations of the melanogenesis pathway are recessive (Barsh 1996) requiring an individual to have two copies of the mutant allele, and the observed cow does not appear to have reproduced any leucistic offspring, the observation of two affected individuals in the Rainy-Gray EPU raises concern regarding the genetic diversity and effective population size of at least this herd, if not all reintroduced populations. 1.4 Molecular Markers In this study we used both maternally inherited mitochondrial DNA (mtDNA) and neutral, biparentally inherited microsatellite loci to evaluate genetic structure in the Vancouver Island source and mainland reintroduced populations of Roosevelt elk. The D- loop of the mtDNA control region is non-coding, and therefore mutates at a rapid rate relative to other mtDNA regions. Mitochondrial markers can be used to detect distinct lineages within and among demes (Brown et al. 1982) and are particularly sensitive to bottleneck events (Avise 1994). Microsatellites are highly variable due to their short 13 repeated sequences, which are prone to replication errors (Schlötterer 2000, Bhargava and Fuentes 2010). Being biparentally inherited and highly variable, microsatellites provide multi-locus genotypes for individuals, and thus allow the characterization of contemporary population differentiation and comparison of important genetic diversity statistics (Slatkin 1995, Pritchard et al. 2000, Allendorf and Luikart 2007). The combination of both mtDNA and microsatellites has been used in compliment to evaluate various questions about population genetics in wildlife (Graham & Burg, 2012; Larson, Jameson, Bodkin, Staedler, & Bentzen, 2002; Polziehn, Hamr, Mallory, & Strobeck, 1998; Polziehn, Hamr, Mallory, & Strobeck, 2000). Genetic patterns (structure) observed in a species using neutral markers provide evidence of the interactions between gene flow and drift that affect genome wide diversity (Funk et al. 2012). 1.5 Study Design To understand how reintroduction strategy impacted the genetic structure of mainland Roosevelt elk herds, we acquired tissue, blood and faecal samples from biologists with the BC Ministry of Forests, Lands and Natural Resource Operations (FLNRO) across as many EPUs as possible. Faecal samples were collected to enhance the number of samples, and therefore the precision of our analyses, for four mainland populations of particular interest. Where individual EPU sampling was insufficient to provide the desired 25-30 samples to allow adequate characterization of population allele frequencies (Hale et al. 2012), samples were aggregated with adjacent EPUs where no barriers to dispersal were likely. Grouped samples were analysed utilizing various software resources to investigate population diversity and structure to understand changes associated with past bottlenecks, recent translocations, and current connectivity among 14 populations. Redundant analyses were run with alternative software, where available, to confirm calculations and summary statistics. 1.6 Thesis Overview The data chapter of this thesis investigates the legacy of historic bottlenecks associated with overexploitation, and the current state of population genetics within Roosevelt elk in BC. Sequences from mtDNA are used to explore long-term genetic patterns on Vancouver Island, and search for a signal related to recent translocations. Patterns in the genetic data are considered in relation to known historic events, and to current population connectivity. Microsatellite genotypes are then used to further examine population genetics, changes in diversity, population differentiation, bottleneck signals, and effective population size. In the final chapter I summarize and review findings, explore observed genetic patterns to predict causation, and provide context and management options for populations of concern. 15 1.7 Literature Cited Allendorf, F. W., and G. Luikart. 2007. Conservation and the genetics of populations. Blackwell Pub, Oxford, UK. Avise, J. C. 1994. Molecular Markers, Natural History and Evolution. Chapman & Hall, New York, NY. Balkenhol, N., S. A. Cushman, A. T. Storfer, and L. P. Waits. 2016. Landscape genetics: concepts, methods, applications. N. Balkenhol, S. A. Cushman, A. T. Storfer, and L. P. Waits, editors. John Wiley & Sons, Ltd., Chichester, UK. Barsh, G. S. 1996. The genetics of pigmentation: from fancy genes to complex traits. Trends in Genetics 12:299–305. Elsevier Ltd. Barsness, L. 2000. Heads, hides, and horns. TCU Press, Fort Worth, TX. BC Government. 2020. Municipal and sub-provincial areas population, 2011 to 2020. Population Estimates. https://www2.gov.bc.ca/gov/content/data/statistics/people- population-community/population/population-estimates. Begg, M. 2007. Legislating British Columbia: a history of B.C. land law, 1858-1978. University of British Columbia. Betts, M. G., K. J. Gutzwiller, M. J. Smith, W. D. Robinson, and A. S. Hadley. 2015. Improving inferences about functional connectivity from animal translocation experiments. Landscape Ecology 30:585-593. Bhargava, A., and F. F. Fuentes. 2010. Mutational dynamics of microsatellites. Molecular Biotechnology 44:250–266. Humana Press Inc. Brown, W. M., E. M. Prager, A. Wang, A. C. Wilson, and E. M. Prager. 1982. Mitochondrial DNA sequences of primates: tempo and mode of evolution. Journal of Molecular Evolution 18:225–239. Brunt, K. R. 1990. Ecology of Roosevelt elk. Pages 65–98 in J. B. Nyberg and D. W. Janz, editors. Deer and elk habitats in coastal forests of southern British Columbia. Victoria, BC. Charlesworth, D., and J. H. Willis. 2009. The genetics of inbreeding depression. Nature Reviews Genetics 10:783–796. Dratch, P. A. 1986. A marker for red deer-wapiti hybrids. Proceedings of the New Zealand Society of Animal Production 46:179–182. Environment and Climate Change Canada. 2020. 1981-2010 Climate Normals & Averages. https://climate.weather.gc.ca/climate_normals/station_select_1981_2010_e.html?sea rchType=stnProv&lstProvince=BC. Frankham, R. 1995. Effective population size/adult population size ratios in wildlife: a review. Genetical Research 66:95–107. Frankham, R. 2005. Genetics and extinction. Biological Conservation 126:131–140. Frankham, R. 2015. Genetic rescue of small inbred populations: meta-analysis reveals large and consistent benefits of gene flow. Molecular Ecology 24:2610–2618. Frankham, R., C. Bradshaw, and B. W. Brook. 2014. Genetics in conservation management: revised recommendations for the 50/500 rules, Red List criteria and population viability analyses. Biological Conservation 170:56–63. Franklin, A. I. 1980. Evolutionary change in small population. Pages 135–149 in M. E. Soulé and B. A. Wilcox, editors. Conservation Biology, An Evolutionary-Ecological 16 Perspective. Sunderland, MA. Funk, W. C., J. K. McKay, P. A. Hohenlohe, and F. W. Allendorf. 2012. Harnessing genomics for delineating conservation units. Trends in Ecology and Evolution 27:489–496. Giglio, R. M., J. A. Ivy, L. C. Jones, and E. K. Latch. 2018. Pedigree-based genetic management improves bison conservation. The Journal of Wildlife Management 82:766–774. Gilmour, J. S. L., and J. W. Gregor. 1939. Demes: a suggested new terminology. Nature 144:333–334. Graham, B. A., and T. M. Burg. 2012. Molecular markers provide insights into contemporary and historic gene flow for a non-migratory species. Journal of Avian Biology 43:198–214. Gray, G. 1995. The history of wildlife in North America. Page 260 in G. Gray, editor. Wildlife and people: the human dimensions of wildlife ecology. University of Illinois Press, Champagne, Il. Hale, M. L., T. M. Burg, and T. E. Steeves. 2012. Sampling for microsatellite-based population genetic studies: 25 to 30 individuals per population is enough to accurately estimate allele frequencies. PLoS ONE 7:e45170. Hedrick, P. 2005. Large variance in reproductive success and the Ne/N ratio. Evolution 59:1596–1599. Jones, E. 2014. Spatiotemporal analysis of old world diseases in North American, A.D. 1519-1807. American Antiquity 79:487–506. Keller, L. F., and D. M. Waller. 2002. Inbreeding effects in wild populations. Trends in Ecology and Evolution 17:230–241. Kimura, M., and J. F. Crow. 1963. The measurement of effective population number. Evolution 17:279. Koenig, W. D. 1988. On determination of viable population size in birds and mammals. Wildlife Society Bulletin 16:230–234. Komers, P. E., and G. P. Curman. 2000. The effect of demographic characteristics on the success of ungulate re-introductions. Biological Conservation 93:187–193. Larson, S., R. Jameson, J. Bodkin, M. Staedler, and P. Bentzen. 2002. Microsatellite DNA and mitochondrial DNA variation in remnant and translocated sea otter (Enhydra lutris) populations. Journal of Mammalogy 83:893–906. Lorenzini, R., and L. Garofalo. 2015. Insights into the evolutionary history of Cervus (Cervidae, tribe Cervini) based on Bayesian analysis of mitochondrial marker sequences, with first indications for a new species. Journal of Zoological Systematics and Evolutionary Research 53:340–349. Lowe, W. H., and F. W. Allendorf. 2010. What can genetics tell us about population connectivity? Molecular ecology 19:3038–3051. McHugh, T. 1979. Time of the Buffalo. 1st edition. University of Nebraska Press, Lincoln, NE. Mountain Goat Management Team. 2010. British Columbia Management Plan Series Management Plan for the Mountain Goat (Oreamnos americanus) in British Columbia. http://www.env.gov.bc.ca/wld/recoveryplans/rcvry1.htm. Accessed 17 Nov 2018. Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic 17 variability in populations. Evolution 29:1–10. Nei, M., and F. Tajima. 1981. Genetic drift and estimation of effective population size. Genetics 98:625–640. O’Gara, B. W., and R. G. Dundas. 2002. Distribution: past and present. Pages 67–119 in D. E. Toweill and J. W. Thomas, editors. North American elk: ecology and management. Smithsonian Institution Press, Washington, D.C. Palsbøll, P. J., M. Bérubé, and F. W. Allendorf. 2007. Identification of management units using population genetic data. Trends in Ecology and Evolution 22:11–16. Pojar, J., K. Klinka, and D. A. Demarchi. 1991. Coastal Western Hemlock Zone. Pages 95–111 in M. Meidinger and J. Pojar, editors. Ecosystems of British Columbia: Special Report Series #6. Research Branch, BC Ministry of Forests, Victoria, BC. Polziehn, R. O., J. Hamr, F. F. Mallory, and C. Strobeck. 2000. Microsatellite analysis of North American wapiti (Cervus elaphus) populations. Molecular Ecology 9:1561– 1576. Polziehn, R. O., J. Hamr, F. Mallory, and C. Strobeck. 1998. Phylogenetic status of North American wapiti (Cervus elephus) subspecies. Canadian Journal of Zoology 76:998– 1010. Polziehn, R. O., and C. Strobeck. 1998. Phylogeny of wapiti, red deer, sika deer, and other North American Cervids as determined from mitochondrial DNA. Molecular Phylogenetics and Evolution 10:249–258. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945–959. Quayle, J. F., and K. R. Brunt. 2003. Status of Roosevelt elk (Cervus elaphus roosevelti) in British Columbia. Victoria, BC. Ralls, K., J. Ballou, M. Dudash, C. Fenster, M. Eldridge, R. Lacy, P. Sunnucks, and R. Frankham. 2018. Call for a paradigm shift in the genetic management of fragmented populations. Conservation Letters 11: e12412. Reed, D. H., and R. Frankham. 2003. Correlation between fitness and genetic diversity. Conservation Biology 17:230–237. Reynolds, D., J. Kelly, N. Tweddle, and C. Morrison. 2018. Coastal Mainland Roosevelt Elk Recovery and Management Project. Victoria, BC. Ritland, K., C. Newton, and H. D. Marshall. 2001. Inheritance and population structure of the white-phased “Kermode” black bear. Current Biology 11:1468–1472. Schlötterer, C. 2000. Evolutionary dynamics of microsatellite DNA. Chromosoma 109:365–371. Slatkin, M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139:457–462. Soulé, M. E. 1980. Thresholds for survival: maintaining fitness and evolutionary potential. Pages 151–169 in M. E. Soulé and B. A. Wilcox, editors. Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer, Sunderland, MA. Spalding, D. J. 1992. The history of elk in British Columbia. Victoria, BC. Speller, C. F., B. Kooyman, A. T. Rodrigues, E. G. Langemann, R. M. Jobin, and D. Y. Yang. 2014. Assessing prehistoric genetic structure and diversity of North American elk (Cervus elaphus) populations in Alberta, Canada. Canadian Journal of Zoology 92:285–298. Waples, R. S. 2015. Testing for Hardy-Weinberg proportions: have we lost the plot? 18 Journal of Heredity 106:1–19. Wright, S. 1931. Evolution in Mendelian populations. Genetics 16:97–159. Young, T. P. 1994. Natural die-offs of large mammals: implications for conservation. Conservation Biology 8:410–418. 19 Figure 1.1: Stochastic effect of bottleneck events and drift. Theoretical changes in allele frequency for four alleles at a single locus, different alleles represented by different colours, allele frequency at each stage in brackets. Random loss of individuals from a population during a bottleneck may result in substantial changes in allele frequency. 20 Figure 1.2: Historic and current distribution of Roosevelt elk (C. c. roosevelti) in North America; green shading shows species distribution prior to widespread European settlement along the Pacific Coast, orange and white diagonal polygons indicate refugia at the beginning of the 1900s, and green patterned areas show current distribution; map adapted from Spalding 1992, O’Gara and Dundas 2002, Quayle and Brunt 2003, Reynolds et al. 2018. 21 Figure 1.3: Observed Roosevelt elk (C. c. roosevelti) in the reintroduced Rainy-Gray Elk Population Unit on the mainland of British Columbia, with putative genetic mutations of the melanogenesis pathway: (L) mature bull elk exhibiting normal (L) and abnormal (R) colouration, photo by D. Reynolds 2017; (R) a mature cow elk with putative partial leucism, photo by D. Brackett 2018; photos used with permission. 2 2 CHAPTER 2: MULTI-MARKER GENETIC ANALYSIS IDENTIFIES METAPOPULATION STRUCTURE AND BOTTLENECK SIGNATURES IN ROOSEVELT ELK IN BRITISH COLUMBIA 2.1 Introduction Where wildlife populations have been extirpated, or reduced to a relatively small number of individuals (population bottleneck), translocations have been used to augment extant populations or to re-establish species into historic habitats (Fritts and Carbyn 1995, Larson et al. 2002b, Olson et al. 2013, Greenhorn et al. 2018, West et al. 2018). Small reintroduced populations are at increased risk of inbreeding depression and reduced ability to adapt to changing environments; reintroduction strategy, social structure and mating system all potentially contribute to the loss of genetic diversity from even brief population bottleneck events (Keller et al. 2012, Keller and Waller 2002, Olson et al. 2013). Genetic bottlenecks coupled with genetic drift may result in further loss of genetic diversity (Wright 1931, Nei et al. 1975, Stockwell et al. 1996, Larson et al. 2002a, Allendorf and Luikart 2007). The increased impact of drift on small populations is well known (Charlesworth and Willis 2009, Frankham 2005; Keller and Waller 2002, Nei et al. 1975). Small populations experience increased loss of genetic diversity and increased frequency of deleterious alleles (Wright 1931, Nei et al. 1975, Aldridge and Boyce 2007, Frankham 2015). During a bottleneck event, the magnitude of any loss of genetic diversity is a result of the size of the remaining population and the duration of the reduction (Nei et al. 1975). Genetic bottlenecks may increase a population’s vulnerability to inbreeding depression, lowering not only the fitness of the population but potentially 23 that of the entire species, putting it at increased risk of extinction (Franklin 1980, Reed and Frankham 2003). The effect of landscape composition on genetic structure is complicated by the interplay between a species’ evolutionary history (phylogeography), behavioural ecology (habitat niche, mating system, seasonal movement pattern, dispersal capabilities, etc.) and the abiotic environment (climate and geophysical features) in which it occurs (Manel et al. 2003, Cushman and Landguth 2012, Hindley et al. 2018). The relationship between landscape ecology and population genetics has become its own field of study known as landscape genetics (Manel et al. 2003, Storfer et al. 2007). The isolation of populations by barriers to dispersal (physical, ecological or behavioural) can result in further loss of diversity and increased risk of inbreeding depression (Reed and Frankham 2003, Frankham 2005, Ralls et al. 2018). The coastal areas of southern British Columbia (BC) represent extremes of topography and climate. Steep sided fjords reach far inland and elevation can range from sea level to over 2500 m in less than 10 km. This highly heterogenous landscape likely presents substantial barriers to dispersal for numerous organisms. Roosevelt elk (Cervus canadensis roosevelti) are the largest subspecies of North American elk, historically occurring along the west coast of North America from northern California to southern BC (Figure 1.1; Bryant and Maser 1982, Spalding 1992). Roosevelt elk were extirpated from the mainland of the Province by the late 1800s during a time of market hunting and commercial exploitation of wildlife, though a small population at Phillips Arm (Figure 2.1) either survived in isolation or recolonized the area (Spalding 1992). The Vancouver Island population survived this period, experiencing a 24 substantial bottleneck event early in the 20th century with the population possibly reduced to as few as ~375 animals (Spalding 1992, Quayle and Brunt 2003). Roosevelt elk on Vancouver Island are known to have low genetic diversity relative to other elk populations, likely due to small numbers of founding animals, geographic isolation, and at least one known population bottleneck (Polziehn et al. 1998b; Polziehn et al. 2000; Spalding, 1992). In BC Roosevelt elk are a provincially Blue Listed species; a species of special concern due to the degradation and loss of habitat and poaching (BC Conservation Data Centre, 2020). During the 1980s, biologists and conservation groups became interested in reintroducing the species to former habitat on the BC mainland. Initial reintroductions (1987-89) consisted of 22 animals translocated from Vancouver Island to the Sechelt Peninsula (Spalding 1992): 13 individuals (10 females, three males) from the Campbell River area, and nine individuals (eight females, one male) from the Qualicum area, with an additional five elk (unknown sex and age) moved onto the mainland near Powell River in 1994 from a site near Comox (Figure 2.1; Spalding 1992, Quayle and Brunt 2003). A translocation and reintroduction program was initiated in 2001 with the goal of re-establishing viable populations throughout much of the subspecies’ historic range in BC (Figure 1.1). During the 17 years the program was active, more than 600 elk were moved through 84 translocation events within the South Coast Region (Appendix I; Reynolds, Kelly, Tweddle, & Morrison, 2018). These translocations (Figure 2.2) helped the mainland Roosevelt elk population increase from approximately 400 individuals occurring in two elk population units (EPUs) in 2001 to more than 2050 in 27 EPUs by 2019 (Figure 2.2; D. Reynolds FLNRO personal communication, 2020). 25 Despite the overall success of this reintroduction program, and that of earlier Wapiti reintroductions across North America, concern has been raised about the potential for reduced genetic diversity in polygynous ungulate species generally (Stephen et al. 2005, Ortego et al. 2011, Olson et al. 2013, Hopken et al. 2015, Bérénos et al. 2016, Sattler et al. 2017, Giglio et al. 2018), and in reintroduced elk specifically (Eberhardt 1996, Hicks et al. 2007, Conard et al. 2010, Hundertmark and Van Daele 2010, Frankham et al. 2014, Muller et al. 2018). All elk used for South Coast Region (SCR) reintroductions from 2001-2017 were descendants of the initial 27 elk translocated to the South Coast Region from Vancouver Island between 1987-1996. Additional translocations occurred from Vancouver Island to the mainland in the West Coast Region in 2017 (Figure 2.1). As mitochondrial markers represent an effective population size (Ne) one-quarter that of nuclear markers due to mitochondrial DNA being haploid and maternally inherited, they are more sensitive to stochastic events such as bottlenecks and specific lineages are more likely to be lost (Avise 1994). Nuclear microsatellite loci have been widely used to investigate contemporary population structure and relatedness in various taxa, from plants to vertebrates, due to their relatively high rates of mutation (Slatkin 1995, Schlötterer 2000, Balloux and Lugon-Moulin 2002). To discern how past bottlenecks, sequential translocations, and reintroduction strategy may have affected genetic diversity in British Columbia’s Roosevelt elk, we evaluated both mitochondrial and nuclear diversity in extant Vancouver Island and reintroduced mainland populations. 26 Specifically, we: 1. evaluated the genetic diversity and structure of the source populations on Vancouver Island and reintroduced populations on the mainland, 2. compared reintroduced populations to determine if a relationship exists between the number of source populations, number of founding individuals, or years since founding and observed genetic diversity, 3. studied the role sequential population bottlenecks play in the genetic diversity of polygynous species, and 4. identified populations that may be vulnerable to inbreeding depression associated with low genetic diversity. Understanding the genetic consequences and potential implications of past events is important for securing the future survival of sensitive species. This research was undertaken in partnership with the BC Ministry of Forests, Lands and Natural Resource Operations (FLNRO) to provide baseline genetic data important for informed stewardship of Roosevelt elk within the Province (Stockwell et al. 1996, Reynolds et al. 2018, West et al. 2018). 2.2 Methods 2.2.1 Sample Acquisition Roosevelt elk samples (n = 357), collected between 2012 and 2019 from 48 EPUs in BC were acquired from FLNRO and analyzed in this study: 90 samples were from faeces, 22 from blood, and 245 from tissue (Appendix II). Of the 245 tissue samples, 243 were from incisor teeth collected for aging data during compulsory inspection of resident and indigenous hunter harvests. Tissue was cut or scraped from the tooth root using a 27 flame sterilized blade and collected in 1.5 ml microcentrifuge tubes and stored at ambient temperature until DNA extraction. Two additional tissue samples were provided by the BC Conservation Officer Service from euthanized animals. Faecal samples were opportunistically collected during census surveys similar to Ramón-Laca, Soriano, Gleeson, & Godoy (2015). A sterile flocked swab (Puritan PurFlock Ultra 25-3606-U) moistened with ASL buffer (stool lysis buffer - Qiagen GmbH cat.no. 1014755) was swabbed around pellets, focussing on areas not exposed to direct sunlight and areas with observable mucus. The swab tip was broken off into a 1.5 ml microcentrifuge tube containing 1.5 ml of ASL buffer, and stored at ambient temperature until DNA extraction (Hajkova et al. 2006). 2.2.2 DNA Extraction and Amplification DNA from tissue and blood samples was extracted using a modified Chelex (Bio- Rad Chelex 100 resin) protocol (Walsh et al. 1991, Burg and Croxall 2001). Faecal DNA was extracted using a QIAamp® DNA Stool Mini Kit (Qiagen GmbH cat.no. 51504) following the published extraction protocol modified for the use of faecal swabs, where the microcentrifuge tube containing the swab and 1.5 mL ASL buffer was vortexed for 30 seconds, then a 1.2 mL aliquot was transferred to a new tube, and the standard extraction protocol followed. A subset of samples was selected for mtDNA analysis (Appendix II); Vancouver Island samples (n=28) were selected to represent as many EPUs as possible, while mainland samples (total n=31) focussed on two of the oldest reintroduced populations, the Sechelt Peninsula (n=11) and Powell River (n=7), Pitt River (n=5) represented a rapidly growing population, and Squamish (n=5) was considered highly connected to 28 other populations. Remaining mainland samples were from peripheral Roosevelt herds (n=3) that may be connected to Rocky Mountain elk (C. c. nelsoni) populations occurring to the east and south. A 567 bp portion of the mitochondrial D-loop known to be variable in Roosevelt elk (Polziehn et al. 1998) was amplified in 25 µL polymerase chain reactions (PCR). D- loop PCR reactions contained 5x Green GoTaq® Flexi 5x buffer (Promega), 0.2 mM dNTP, 2.5 mM MgCl2, 0.4 μM primers EK-F23 and EK-R663 (Speller et al., 2014; Appendix III), 1 U GoTaq® Flexi polymerase, and DNA template. PCR conditions were as follows: one cycle of denaturing at 94°C for 120 seconds (s), annealing at 52°C for 45 s, extension at 72°C for 60 s; 35 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 45 s, followed by a final extension step of 72°C for 300 s. To confirm successful amplification, 3 µL of PCR product was run on a 0.8% agarose gel. Samples that produced a clean band of approximately 600 bp were subsequently Sanger sequenced using the forward primer (EK-F23) at NanuQ, Genome Quebec (McGill University, Montreal, QC, Canada). To further evaluate contemporary population structure and differentiation, 23 microsatellite loci (Appendix III) known to be variable in elk were individually screened using 10 μL PCR reactions containing GoTaq® Flexi 5x buffer (Promega), 0.2 mM dNTP, 2.0 mM MgCl2 (see Appendix III), 0.2 μM forward primer, 0.4 μM reverse primer, 0.2 ng/μL BSA, 0.5 U GoTaq® Flexi polymerase, and DNA template. Forward primers were synthesized with a 5’ M13 tag sequence (CAC GAC GTT GTA AAA CGA C) to incorporate a fluorescent tag during amplification allowing PCR products to be visualized on a 6% acrylamide gel with a Li-COR 4300 DNA Analyzer (Li-COR Inc., Lincoln, NE, USA). Li-COR runs were conducted with three positive control samples, to 29 ensure consistent amplification and scoring, and one negative control. The positive controls provided a reference across multiple gels allowing consistent scoring of alleles. Of the 23 loci screened, fourteen loci successfully amplified, and a subset of samples (n=12) was screened for variation; one locus appeared monomorphic (RT5) and one showed low levels of heterozygosity in the samples tested (RT27) and both were subsequently dropped from further analyses. Twelve polymorphic loci were retained for genotyping (BL42, BM203, BM888, BM4107, BM4513, BM6506, BMC1009, CSSM041, INRA107, RT7, RT13, OarFCB193; Appendix III). Loci RT13 and OarFCB193 later appeared highly sensitive to template DNA concentration in faecal samples and were subsequently removed from further genotyping. Primers for loci BM4107 and BM4513, originally developed for cattle (Bos taurus), produced PCR products with excessive stutter (additional bands) that made accurate scoring impossible; these primers were redesigned (BM4107RD and BM4513RD; Appendix III) while retaining the variable repeat motif. Final PCR amplification conditions were as follows: one cycle of denaturing at 94°C for 120 s, annealing at 55°C for 45 s, extension at 72°C for 60 s; 7 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 45 s, followed by 25 cycles of 89°C for 30 s, 57°C for 30 s, and 72°C for 45 s; and a final extension step of 72°C for 300 s. Four loci required different annealing temperatures: BM203 and CSSM041 at 52/54°C, BMC1009 at 48/50°C, and BM6506 at 60/62°C. When loci showed poor amplification for faecal samples, an additional three cycles were added at the higher annealing temperature (BM203, BM888, BM6506 and BMC1009). Alleles were scored manually, cross checked multiple times, and scored by a second experienced individual. Samples with faint bands, substantial stutter, or 30 anomalous elements were reamplified until a clear genotype could be identified or were discarded from further analyses. A subset of samples from all gels was run together on a single load as an additional check to ensure consistent allele calls across runs. 2.2.3 Sequence Analyses Study sequences were initially aligned to 1211 bp D-loop sequences for two Roosevelt and one tule elk (GenBank accession no. AF016970, AF016971 and AF016977 respectively) in MEGAX v10.1.7 (Kumar et al. 2018) to confirm subspecies and known polymorphic sites (Polziehn et al. 1998). All variable sites were confirmed or rejected through visual review of sequence chromatograms. Sequences were evaluated for standard population genetic measures (haplotype diversity HD, nucleotide diversity - π) in DnaSP v6.12.03 (Rozas et al. 2017) to allow comparison between source and reintroduced populations. To evaluate population differentiation, pairwise genetic distances (Tajima and Nei 1984) were calculated in MEGAX with 1000 bootstraps to determine variance, and for FST (ΦST) in ARLEQUIN v.3.5.2 (Excoffier & Lischer, 2010). 2.2.4 Microsatellite Analyses Previous studies have reported that a sample size of 25-30 individuals from a population is required to evaluate the genetic structure of a population using variable microsatellite loci (Hale et al. 2012), with some suggesting population differentiation detectable with smaller sample sizes (Landguth et al. 2012). To achieve a reasonable sample size for population analyses, EPUs with small sample size (<8) were grouped with adjacent units where no putative barrier to dispersal was suspected (Table 2.3). Where an EPU’s sample size was low (8-12) and connectivity to adjacent herds was 31 unlikely, samples were analyzed as a separate group. Significance values for multiple tests were adjusted for false discovery rate (FDR;α= 0.05) using the procedure detailed in Benjamini & Hochberg (1995) to reduce type 1 errors without greatly reducing power to detect potential differences. To check for multiple sampling of individuals, important with non-invasive samples, a procedure to evaluate samples (Paetkau 2003) was undertaken for genotype matches with up to two mismatched loci identified using CERVUS v.3.0.7 (Marshall et al. 1998). Populations and loci were checked for deviation from Hardy-Weinberg equilibrium (HWE) in GENODIVE v.3.04 (Meirmans and Van Tienderen 2004) with 10,000 permutations via analysis of molecular variance (AMOVA). Deviations from linkage equilibrium (LD) between loci were checked with GENEPOP v.4.7.5 (Raymond and Rousset 1995, Rousset 2008). Standard genetic measures were calculated for loci and populations, including number of alleles per locus (Na), effective number of alleles (Neff), Shannon Information Index (I), observed heterozygosity (Ho), expected heterozygosity (He), unbiased expected heterozygosity (uHe), and fixation index (F) were calculated in GenAlEx v6.5 (Peakall and Smouse 2012), allelic richness (Ar) was calculated with hierfstat (Goudet 2005) implemented in R (R Core Team 2013), and the inbreeding coefficient Gis, Nei’s (1987) analogue of Wright’s Fis, was calculated in GENODIVE. To evaluate the effects of translocation on population diversity, changes in allele frequency from the primary source population(s) for each locus within each population and genetic diversity statistics were compared against population demographic history. Comparisons included: number of founding individuals, source population(s), initial male to female ratio, and years since founding (Appendix 1). 32 Pairwise population differentiation was calculated for both Fst (AMOVA; Excoffier, Smouse, & Quattro, 1992; Meirmans, 2006) and Jost’s D in GENODIVE, with 5,000 permutations to determine significance. Jost’s D provides a measure of differentiation that is independent of within population diversity (Jost 2008). To further understand genetic differentiation between populations, genetic data were analyzed using the software STRUCTURE v.2.9.4 (Pritchard et al. 2000). STRUCTURE is a model-based Bayesian clustering software that infers ancestry of individuals based on their multi-locus genotypes. To elucidate population differences on Vancouver Island and in the South Coast Region, a hierarchical analysis was conducted using the admixture model with locprior setting. Settings were chosen assuming limited dispersal between populations, with some correlation of allele frequencies by location (Porras-Hurtado et al. 2013). The locprior setting has been shown to identify genetic structure in populations undergoing contemporary change (Hubisz et al. 2009, Porras- Hurtado et al. 2013). Five iterations for each genetic cluster value (K=1-5), with individual STRUCTURE runs consisting of 50,000 step burn-in and 200,000 Monte Carlo Markov Chain (MCMC) steps, were conducted. To determine the most likely number of genetic clusters the statistics ∆K (Evanno et al. 2005) and Prob(K) (Pritchard et al. 2000) were calculated in the online software CLUMPAK (Kopelman et al. 2015), and histograms were visually checked at each value of K. As the source populations on Vancouver Island are known to have experienced at least one bottleneck event at the beginning of the 20th century, and reintroduced populations may have been subject to sequential bottlenecks associated with reintroduction, populations with a minimum of 20 samples were evaluated using the 33 software BOTTLENECK v.1.2.02 (Piry et al. 1999) for deviations from mutation / drift equilibrium and mode shift. BOTTLENECK implements four tests: sign test, standard difference test, Wilcoxon sign rank test, and mode shift test; a mode shift is indicated when there is a deviation from an L-shaped distribution and is suggested by the authors to be a reliable indicator of past bottleneck events. As the standard difference test requires at least 20 microsatellite loci to reliably detect a bottleneck, this test was not done. Both the sign and Wilcoxon tests were run for three mutation models: Infinite Alleles Model (IAM), Stepwise Mutation Model (SMM), and Two Phase Mutation model (TPM). The TPM model allows for adjustment of the ratio of strict single step mutation, as well as variance in size of multiple step mutations. As microsatellite loci rarely evolve under a strict SMM, the proportion of single step mutations was set at 0.90, with variance set at 0.12 (Garza and Williamson 2001, Hundertmark and Van Daele 2010). To further understand how reintroduction strategy has affected genetic diversity in these recently established populations, the genetic effective population size (Ne) was calculated with the software NeESTIMATOR v.2.1 (Do et al. 2014) using the linkage disequilibrium model for random mating; this option provides an estimate for a single time point. Multiple tests were run for two thresholds for rare alleles (0.05, 0.01) to allow comparison of the model using different run parameters. As the Ne for each population is estimated independently, all populations were included. 34 2.3 Results 2.3.1 Sampling Samples for 356 individuals (Figure 2.3) were successfully amplified and genotyped at seven or more microsatellite loci (Appendix II), with one additional sample that amplified with mtDNA primers but failed to amplify with most microsatellite loci. Genotype comparisons in CERVUS showed a total of 119 genotype matches with one or two mismatched microsatellite loci, and no matches with zero mismatched loci. Each genotype pair was reviewed to determine its potential for being the same individual sampled more than once. Out of 119 potential identity matches, all were rejected but one, as they were either tissue samples from different deceased individuals, tissue samples from dead elk and faecal samples collected at a later date, or Vancouver Island samples matched with mainland samples. One matched pair consisted of a faecal sample collected in spring and a tissue sample from an elk killed later that year at a nearby location. As a true match could not be excluded, the faecal sample was removed from the data set. All of the following analyses, except for mtDNA sequencing, were completed using the remaining 355 samples. 2.3.2 Sequencing A subset of 58 samples were successfully sequenced and analyzed; all individuals (Vancouver Island n = 28; mainland n = 30) were determined to have Roosevelt elk mtDNA, presence of an A to G transition at site 493 (Figure 2.4b). We identified three variable sites and four Roosevelt elk haplotypes in our analysis (Table 2.2). Of the four haplotypes observed (Figure 2.4a), two haplotypes occurred exclusively south of Alberni Inlet on Vancouver Island (herein referred to as VIS; Hap-C and Hap-D) and two other 35 haplotypes to the north (VIN; Hap-A and Hap-B). Overall haplotype diversity (HD) was 0.494 and overall nucleotide diversity (π) was 0.0011 (Table 2.2). VIN had the highest haplotype and nucleotide diversities (n = 18, Hd = 0.523, π = 0.0009) compared to VIS (n = 10, HD = 0.356, π = 0.0006). Among mainland populations all 30 individuals shared a single haplotype (HD = 0, π = 0). This haplotype (Hap-A) was also found on VIN (Figure 2.4). Genetic distance among populations (Table 2.2) was highest between VIS and VIN (d = 0.00292, ±0.00201) and lowest between VIN and the mainland (d = 0.00079, ±0.00081), with intermediate values between VIS and mainland (d = 0.00213, ±0.00183). Pairwise FST difference was greatest between VIS and mainland (FST = 0.925, P <0.001), lowest between VIN and mainland (FST = 0.494, P <0.001), and intermediate between VIN and VIS (FST = 0.721, P <0.001). 2.3.3 Microsatellites The 355 samples were aggregated into 13 populations in further analyses (Table 2.3, Figure 2.3): Vancouver Island South (VIS, n = 41), Vancouver Island North (VIN, n = 55), Sechelt Peninsula (SP, n = 48), Central Coast (CC, n = 35), Rainy Gray (RG, n = 53), Squamish – Indian River (SQ, n = 16), Pitt River (PITT, n = 29), Stave River – Tipella (STV, n = 16), Brittain – Skwawka (BRIT, n = 10), Powell River South (PRS, n = 17), Powell River North (PRN, n = 16), Toba River (TOBA, n = 11), and Homathko – Southgate – Orford (HO, n = 8). Deviations from HWE were observed for two loci and four populations (Appendix IV), however, none remained significant after correcting for FDR (130 pairwise; adjusted P=0.0004). Two pair of loci showed significant LD (Appendix V) after controlling for FDR (45 pairwise; adjusted P = 0.0033): BM888 and Inra107 (χ2 ≥116.77, 36 df = 26, P <1.81x10-13; LD in VIN, SP and PITT populations, full population by locus- pair data not shown) and BM4107RD & Inra107 (χ2 ≥53.49, df = 26, P <0.0012; LD in four populations not significant after FDR adjustment). Bottlenecked populations often show deviations from HWE, as their characteristics often violate HWE assumptions (Nei et al. 1975, Avise 1994) and as such, all loci and populations were retained for further analyses. Genetic diversity statistics were calculated to evaluate general patterns among populations (Table 2.4). The average number of alleles (Na) showed the highest values in the primary source population VIN and the reintroduced (tertiary) population SQ (Na = 3.0), the next highest value in the main secondary source population SP and tertiary RG, PITT and STV (Na = 2.9), followed by CC (Na = 2.8), primary source VIS, tertiary BRIT, PRS and TOBA (Na = 2.7), and PRN and HO (Na = 2.6). The effective number of alleles (Neff) showed a similar pattern except STV had the highest value (Neff = 2.29) followed by VIN (Neff = 2.20), TOBA, SP, PITT, RG, and PRN (Neff = 2.09-2.07), HO and CC (Neff = 2.05-2.03), with BRIT, VIS, SQ, PRS with the lowest values (Neff = 1.99-1.84). Allelic richness (Ar) was calculated for all populations, though it should be noted that this statistic is sensitive to small sample sizes and rare alleles. Ar showed a mixed pattern wherein four of the ten reintroduced populations (STV, SQ, PITT, TOBA) had equal or greater richness as VIN (Ar = 2.635), and an additional two (BRIT and HO) greater than secondary source SP (Ar = 2.577). The remaining four tertiary populations (CC, PRN, RG, PRS) had lower Ar values than SP, but greater richness than VIS (Ar = 2.229). The information index (I) showed greatest diversity in STV and VIN (I = 0.866 and 0.843 respectively), followed by SP (I = 0.816) and tertiary populations (in 37 descending order TOBA, PITT, RG, PRN, CC, HO, SQ, BRIT (I = 0.807-0.763). VIS and PRS had the lowest values (I = 0.710 and 0.694 respectively). Observed (Ho), expected (He) and unbiased heterozygosity (uHe) showed mixed patterns, however, four populations showed significant reductions in Ho as indicated by Gis (Table 2.5): VIS, RG, HO and TOBA (Gis = 0.111, 0.121, 0.227, 0.231, respectively). The fixation index (F) showed the same pattern as Gis, as they are both based on differences between observed and expected heterozygosity. Diversity statistics for microsatellite loci (Table 2.5) showed a pattern of lower observed heterozygosity than expected for all loci but two, BL42 and INRA107. Comparisons of founding population demographics to allele frequency (Appendix VI) and diversity statistics (Table 2.5) failed to show any clear associations. Allele frequency changes ± 50% were common among the reintroduced populations when compared to VIN, the primary source population. All reintroduced populations lost at least one low frequency allele present in VIN, with PRN, TOBA and HO each losing four alleles. Four populations (RG, SQ, PITT and TOBA) had a unique allele at locus BM888 not observed on Vancouver Island or SP. Of note, VIS was missing three alleles, compared to VIN. Pairwise population differentiation was similar with both Fst and Jost’s D, with one exception (PRS x PRN significantly different with Fst only; not shown), therefore only Fst values are discussed (Table 2.7). After FDR correction (78 pairwise tests, adjusted P <0.0179), VIS was significantly different from all other populations except HO; primary source VIN was different from all populations except PRN, TOBA and HO. Among the reintroduced mainland populations, PRS showed significant differentiation 38 from five populations (SP, CC, PITT, STV and PRN), PITT was significantly different from three (SP, PRS and RG) as was SP (SQ, PITT and PRS). The only other significant difference was SQ and SP. STRUCTURE analyses of microsatellite genotypes showed clear genetic differentiation between VIS and all other populations (Figure 2.5a). Analyses of all 13 populations showed ∆K = 2, as did Prob(K). A closer evaluation of the different ancestry plots, specifically K = 4, suggested further hierarchical structure may be present. To ascertain if there was further hierarchical structure, a series of STRUCTURE runs were conducted following the recommendations in Wang (2017), where the parameters for the model’s prior assumptions were adjusted, namely allowing independent alpha values for each population, and lower initial values of alpha (a = 0.1-0.5). Adjusting these values allows STRUCTURE to more precisely assign individual ancestry contributions (Wang 2017). Using an initial value for alpha of 0.25, and allowing individual priors for each population, the estimate for ∆K increased to four, while Prob(K) remained at two (Figure 2.5a). To evaluate the less developed genetic structure observed among the mainland populations further runs were conducted without VIS (Figure 2.5b). Including populations that are highly different from others can reduce the ability of the model to identify substructure. An initial run with admixture, loc prior and default settings resulted in ∆K = 3 and Prob(K) = 1 (not shown). Visual examination of ancestry plots indicated some population differentiation was evident at K = 3. In an attempt to reconcile the discrepancy between ∆K and Prob(K), parameters were set to allow individual alpha values and an initial alpha value of 0.25. At those settings both ∆K and Prob(K) 39 supported three clusters (Figure 2.5b). VIN is clearly differentiated, while the mainland populations show individuals assigned to two different ancestral populations (mostly orange, mostly light blue), regardless of their sampled population. Populations run through BOTTLENECK (sample size >20) included VIS, VIN, SP, RG, PITT, and PR which included samples from three EPUs 2-12A, 2-12E and 2- 12D with no barriers to dispersal between them other than distance. BOTTLENECK results (Table 2.7) showed significant deviations from expected heterozygosity at mutation / drift equilibrium (Hex) for the IAM in all populations for both the sign test (P ≤0.006) and Wilcoxon sign rank test (P ≤0.003). In contrast, the strict SMM was only significant for two populations in the Wilcoxon test (VIN and SP, P = 0.009 and 0.012 respectively). For the TPM model both VIS and VIN were significant in the sign test (P = 0.018 and 0.017 respectively) with nine loci showing heterozygosity excess, and all populations except PITT were significant in the Wilcoxon test (P = 0.003-0.042). A mode shift was observed for SP, PITT and PR. Estimates of effective population size were calculated for all 13 populations (Table 2.8). The Ne estimates for VIS and VIN were very similar at 58.5-59.7. Ne values for reintroduced populations ranged from a low of Ne = 7 (minimum allele frequency 0.05) for PITT to a maximum Ne=230 (minimum allele frequency 0.01) for RG, with an average Ne value of 72. Seven of the mainland populations (SP, RG, SQ, BRIT, PRS, TOBA and HO) had Ne estimates greater than their sample size, while four were lower (CC, PITT, STV and PRN). Confidence intervals (95%) for most Ne values were large, with maximum values of infinity in seven populations and a ten-fold difference between 40 minimum and maximum values. Only two populations (CC and PITT) had 95% CI values that differed roughly five-fold (5.3-24.5 and 2.9-13.8 respectively). 2.4 Discussion Translocations have been used widely to reintroduce populations of wildlife to former habitat with varied measures of success (Lubow 1996, Wolf et al. 1996, Stephen et al. 2005). The reintroduction program undertaken by the British Columbia government between 2001 and 2017 has resulted in the reestablishment of Roosevelt elk to historic habitats, with a population increase of over 500% and species’ range expansion of >30,000 km2 from prior to initiation (Reynolds et al. 2018). While the program’s overall success cannot be overstated, concern about genetic diversity loss associated with reintroductions, sequential translocations, and specific species life history requires careful evaluation of such projects, especially for species at risk (Stockwell et al. 1996, Larson et al. 2002a, West et al. 2018, White et al. 2020). 2.4.1 Population Structure The results of this study suggest that population bottlenecks have reduced mitochondrial genetic diversity in both extant and reintroduced populations of Roosevelt elk in British Columbia. The initial reintroduction of 22 elk to the mainland occurred on the Sechelt Peninsula with eight females and one male from the Qualicum area on southern Vancouver Island; VIS has two unique haplotypes not shared with VIN. Our results suggest that either: 1) the translocated Qualicum area females shared the A haplotype common on VIN, or 2) the C and / or D haplotype was lost due to drift on the mainland within 25 years of translocation. Unfortunately, neither archival samples from the founding animals nor contemporary DNA samples from the Qualicum area were 41 available for analysis. Reviewing species distribution and density maps from near the time of translocation (Brunt 1990) suggest elk occurring near Qualicum were likely more connected to populations further south than those to the north (Billy Wilton FLNRO, personal communication 2020). Regardless of which haplotypes were present at the time of reintroduction, at best it appears mainland populations are lacking the mitochondrial diversity present on Vancouver Island, and at worst, there has been a substantial loss of diversity in mainland populations, despite a female bias which should have helped conserve mtDNA diversity. Three of the haplotypes found in this study were previously identified on Vancouver Island in Polziehn et al. (1998). One haplotype from that study (Roosevelt 33), with an A to G transition at site 541, was not observed in our samples. Unfortunately, the source location of the Roosevelt 33 sample is unknown (R. Polzeihn, University of Alberta, personal communication 2020). The D haplotype observed in the VI South population, with a T to C transition at site 441, has not been previously reported. Multi-locus genotype analyses supported the mitochondrial pattern seen on Vancouver Island, showing the same differentiation between VIS and VIN associated with Alberni Inlet, both in pairwise differentiation (FST) and Bayesian clustering (STRUCTURE) analyses. This suggests that the Alberni Inlet represents a substantial barrier to both male and female dispersal on Vancouver Island. The Alberni Inlet runs southwest to northeast from Barkley Sound on the west coast of Vancouver Island to Port Alberni some 62 km inland. The inlet creates a barrier two-thirds of the island’s width, with areas from Port Alberni east to the Salish Sea comprised of mostly poor to very poor quality elk winter habitat (Quayle and Brunt 2003). Degradation of habitat quality is 42 likely associated with extensive loss of old growth forest, important in high snow years, transportation corridors, and development. Winter habitat appears to be the critical element of Roosevelt elk persistence in an area (Brunt 1990). Analyses of multi-locus genotypes showed clear population differentiation between VIS and VIN, corresponding to the barrier represented by Alberni Inlet. Both pairwise Fst and Bayesian analyses show a marked difference between these two populations, suggesting very little gene flow is occurring. As the Alberni Inlet represents a barrier to dispersal on Vancouver Island, it is likely that the extensive inlets and the steep, rocky terrain of much of the mainland coast also represent significant barriers to Roosevelt elk dispersal. While the Alberni Inlet appears to be a somewhat obvious barrier to dispersal, the reasons underlying its status are not as obvious. Elk are good swimmers, as evidenced by occasional sightings on islands. A group of elk were translocated to the Chehalis EPU on the west side of Harrison Lake in 2013. A year and a half after release, a GPS collared female elk swam east across the lake, a distance of more than 3.2 km. Alberni Inlet is less than a kilometre wide for much of its length, and less than 700 m in a few spots, raising the question: is the inlet a substantive barrier to dispersal, or is the lack of gene flow more related to low population density in adjacent areas? 2.4.2 Reintroduction Effects To understand the effect that reintroduction strategy has had on genetic diversity, we searched for patterns within the genetic data; changes in heterozygosity, allele frequency and richness were examined, and pairwise comparisons between populations were conducted (Tables 2.4, 2.6 and Appendix IV). Overall, rare alleles were lost in many populations, and allele frequencies varied substantially between them. Changes in 43 heterozygosity (He – Ho) between source and reintroduced populations often exceeded 20%, though changes were somewhat more likely to represent a loss of heterozygosity (~59%) than an increase. No clear pattern could be ascertained related to single vs. multiple source populations, number of founding individuals, or years since reintroduction, suggesting that the observed changes were likely the result of drift. Pairwise differences were highly significant between both VIS, VIN and most mainland populations (Table 2.6). Differences among a handful of mainland populations were significant, with PRS showing five pairwise differences, and PITT and SP showing three each. Interestingly SP and to a lesser amount PRS, were the source populations for all of the other mainland populations sampled. Differentiation between the reintroduced populations and their source(s) is likely due to drift associated with small numbers of founding individuals, and / or the capture and translocation of closely related individuals. Unfortunately, little work has been conducted on the genetic relatedness of winter herds in elk, and none in Roosevelt elk. High home range fidelity and social influences on herd interactions (or lack thereof) suggests that elk social dynamics may contribute to genetic changes observed in reintroduced populations (Franklin et al. 1975, Larkin et al. 2004, Muller et al. 2018). The PITT represents a fast growing reintroduced population, which should have helped to minimize diversity loss and drift (Nei et al. 1975). Bayesian analyses showed increasing differences between VIS, VIN and the mainland populations. Within the reintroduced groups, individual ancestry assignment appeared unrelated to its sampled population (Figure 2.5b), which may be indicative of founder effect, or could possibly be related to different levels of drift between overlapping generations. Roosevelt elk are long lived, and sampling may have captured 44 multiple generations. Also, some individuals with allele frequencies similar to source populations may have been sampled, or later cohorts may be more affected by inbreeding and drift making it appear as though there were two different source populations. Overlapping generations, unequal fecundity (as would be expected in polygynous mating systems), and variations in population size are all known to affect allele frequencies and population measures like Ne (Nunney 1993, Frankham 1995, Waples et al. 2014). The underlying model used in STRUCTURE assumes populations are in HWE, and loci are not linked (Wang et al. 2016). While no populations were significantly out of HWE and only two pairs of loci showed LD after adjusting for FDR, numerous loci in individual populations exhibited low probability values, just not enough to be significant. To see if specific loci were implicated, data sets were analyzed with the highly significant loci pair in LD (BM888 and INRA107), and a third locus possibly out of HWE (RT7) were removed one at a time (data not shown) and reanalyzed. The results of both pairwise differentiation and Bayesian analysis remained unchanged. STRUCTURE is known to start ‘splitting’ assignments (assigning individuals to multiple populations) in groups when K values are higher than the likely number of populations (Lawson et al. 2018). While this is observed in many mainland individuals, at K=4 in Figure 2.5a and at K=3 in Figure 2.5b, the VIN population becomes differentiated from the mainland. Comparing STRUCTURE results to pairwise differences (Table 2.6) supported the conclusion that mainland populations are differentiated from VIN. The question of why different individuals in reintroduced herds sourced from a single population appear to be assigned to two, remains unanswered. 45 These results, taken as a whole, suggest that some mainland populations of Roosevelt elk are differentiating due to genetic drift, likely from a combination of factors including founder effect, small population size, and limited or complete isolation from other herds. The lack of reduced diversity, or even the increase in some diversity measures, in reintroduced populations is not unheard of. It appears that for species that have experienced substantial bottleneck events, the initial reduction of low frequency alleles results in only common alleles remaining, and as such they are more likely to be carried to new populations by founders (Clegg et al. 2002, Larson et al. 2002a, b). The Wilcoxon test implemented in BOTTLENECK is reported to have the highest power of the program’s test for detecting recent bottleneck events when using a moderate number of loci (Piry et al. 1999), and the TPM model is most representative of microsatellite mutation processes (DiRienzo et al. 1994). The results of the bottleneck analyses suggest that all Roosevelt elk populations in BC show some evidence of recent bottleneck events. Only the PITT herd was not significant, though close, in the Wilcoxon – TPM analysis (P=0.065), yet it still showed deviation from an L-shaped allele distribution (mode shift), suggesting that this population should also be considered bottlenecked. The authors of the program report that testing for bottleneck signatures is likely only detectable for 4Ne generations after the event; therefore, the greater the magnitude of the genetic bottleneck, the sooner it will become undetectable. The populations on Vancouver Island appear to still show the signature of a substantial bottleneck event after more than 100 years, which may be the result of a few hundred individuals surviving. In contrast, an insular population of Roosevelt elk introduced to an island in the Kodiak Archipelago, Alaska founded by five females and three males in 46 1929 showed no heterozygosity excess after 79 years when evaluated with various bottleneck detection software (Hundertmark and Van Daele 2010). The authors used genetic population models to show that the signature of the bottleneck may have been lost in the population within eight years. This surprising result was suggested to be due to a rapid loss of heterozygosity in the first year, slowing to virtually no change by year 10. The authors also reported that while other methods were also able to detect bottlenecks in the subspecies, they were unable to detect sequential bottlenecks. The estimates effective population sizes calculated for Vancouver Island raises a number of important management considerations. Study data for Vancouver Island were less likely to be affected by the sampling of close relatives or family groups, and may be a complication in EPUs where non-invasive samples were included, because hunted samples were widely dispersed and could be considered relatively random. Both areas, VIS and VIN, had good numbers and quality of samples, yet Ne estimates were barely above the IUCN Red List criteria (Ne >50) for critically endangered. It should be noted that linkage disequilibrium estimators of Ne can become downwardly biased when sampling includes overlapping generations (Nunney 1993, Waples et al. 2014). While it is undoubtable study samples represented various age class animals, the extent to which this affected our Ne estimates is unknown as age class information was not available from the compulsory inspection data for teeth, which was the sole source of Vancouver Island samples. The unfortunate reality is that Roosevelt elk on Vancouver Island appear to have lost significant genetic diversity due to the bottleneck of the late 1800s, and moderate population growth in the decades that followed. Ne is also known to be adversely affected by fluctuations in population size and variance in individual reproductive success (Nei 47 and Murata 1966, Nunney 1993). Extreme weather events, specifically harsh and long winters, are known to have caused many Vancouver Island herds to have experienced substantial fluctuations in population size (Brunt 1990, Quayle and Brunt 2003). The strategy undertaken in the reintroduction of Roosevelt elk to the mainland of BC resulted in the creation of numerous new herds in areas of high quality habitat, through translocation of a minimum of 20 individuals (Reynolds et al. 2018). The reintroduction of at least 20 individuals has been shown to result in increased population growth and increased reintroduction success in various artiodactyls (Komers and Curman 2000). From a demographic perspective, the reintroduction program can only be described as exemplary with all 27 mainland EPUs, except for one, showing healthy growth rates in the initial years after establishment (Reynolds et al. 2018). From a genetic diversity viewpoint, our results suggest that there is reason for concern. It is important to note that roughly 70% of all elk translocated within the region were captured on the Sechelt Peninsula, including 20 of the 25 founding individuals for the Powell River herd, the other source for mainland translocations, with the exception of the Phillips and Heydon EPUs discussed in the introduction. While Komers & Curman (2000) acknowledged that reintroduced species with polygynous mating systems could result in increased rates of inbreeding, only maintenance of heterozygosity was mentioned and dismissed, as “loss of heterozygocity [sic.] in particular can be virtually prevented if the founding population is allowed to increase rapidly”. More recent research has shown that even within reintroduced populations where the founding individuals were not closely related, and with roughly three times as many founders, differential reproductive success such as that expected with one or two males 48 dominating all breeding in a small population, could lead to the rapid loss of genetic diversity within the first few years of establishment (Wilson et al. 2005). Other researchers have found similar results in reintroduced, polygynous ungulates (Fitzsimmons et al. 1997, Slate et al. 2000, Zachos et al. 2007, Olson et al. 2013). In a recent study evaluating different approaches on maintaining genetic diversity in bison (Bison bison), a severely bottlenecked species, Giglio et al. (2018) found increased inbreeding if a male was a dominant breeder for two years. As bison and elk share a similar breeding system, the potential for increased loss of genetic diversity in reintroduced elk populations is likely exacerbated by strategies where only one or two mature males are present during initial herd establishment. Furthermore, as most male elk translocated with captured herds were predominantly yearlings or calves, any mature male present early in herd development was likely to dominate all breeding for multiple years, with subsequent increased levels of inbreeding and decreased levels of diversity. It is also possible, if not likely, that yearling males are closely related to at least some of the females in a captured herd (Clutton-Brock et al. 1988, Geist 2002), thus further compounding the potential for loss of diversity within these populations. While historic bottleneck events may have reduced genetic diversity in British Columbia’s Roosevelt elk, it is critical that the remaining genetic diversity of the population is maintained. The mainland populations appear to be undergoing genetic differentiation from their Vancouver Island source. Mainland Roosevelt elk show a clear reduction in mitochondrial diversity and loss of multiple low frequency alleles in multiple herds. Bottlenecked populations often show a heterozygosity increase, however this effect is ephemeral and there is typically an eventual loss of heterozygosity (Cornuet and 49 Luikart 1996, Luikart and Cornuet 1998). Low heterozygosity is associated with reduced fitness of individuals, as well as populations and species (Aldridge & Boyce 2007, Bérénos et al. 2016, Keller et al. 2012, O’Brien & Evermann, 1988, Reed & Frankham, 2003, Slate et al. 2000; however, see Britten 1996). While the loss of microsatellite alleles on their own is likely inconsequential, as they are putatively neutral, it may be indicative of the loss of irreplaceable adaptive potential in the form of rare alleles at non- neutral loci. Roosevelt elk have low genetic diversity when compared to other elk subspecies (Polziehn et al. 1998, 2000) For maternally inherited markers, rapid population growth is the critical element to reduce the effect of drift on mitochondrial diversity (Avise 1994). Between the initial reintroduction in 1987, and initiation of the mainland reintroduction project in 2000, the mainland population grew from 22 to over 400, likely representing maximum population growth for this subspecies. Despite this, it appears only a single mitochondrial haplotype occurs in mainland populations of Roosevelt elk. While mitochondrial sequencing occurred for only a subset of mainland elk, the sample size of 30 provided a reasonable chance of observing low frequency haplotypes, though it is possible that rare haplotypes are also present. In practice, the majority of captures for Roosevelt elk reintroductions took place during winter months as attracting elk into traps is easier when food resources are scarce, and elk are more likely to congregate in larger groups (Reynolds et al. 2018). Relatedness within winter herds has not been studied for Roosevelt elk. Some work on other subspecies of elk showed herds are matrilineal (Clutton-Brock et al. 1988, Nussey et al. 2005), while others show lack of stability in family groups (Vander Wal et al. 2012). The 50 potential for loss of genetic diversity in elk may be complicated in reintroductions as genetic structuring has been shown to persist in herds with no physical barriers to dispersal for up to 13 years after release, at least partly due to social segregation among herds (Muller et al. 2018). Female elk are highly philopatric to a home range, and Roosevelt elk appear to be no exception. In 2001, two females were captured on the Sechelt Peninsula wearing VHF collars deployed during translocation from Vancouver Island between 1987-1989. Their capture occurred only a few hundred metres from their release location 12-14 years earlier, suggesting limited dispersal of established females in quality habitats (I. Gazeley personal observation, 2001). 2.5 Conclusions This study used nuclear and mitochondrial molecular markers to reveal metapopulation structure of Roosevelt elk throughout much of their current range in British Columbia. Historic and contemporary bottlenecks have likely resulted in substantial losses of genetic diversity in the study populations, though the eventual extent of loss in the mainland populations will not be known for many years. To better understand genetic diversity changes in populations of Roosevelt elk, as well as other polygynous species, further monitoring and research is needed. Of particular interest, the Alberni Inlet and developed areas to the east on Vancouver Island represent a complete barrier to dispersal. Greater understanding of the role this landscape feature is playing in isolating elk on southern Vancouver Island is required. It should be recognized that Alberni Inlet was previously identified as a putative barrier dividing the Vancouver Island metapopulation into two main subpopulations (Brunt 1990, Quayle and Brunt 2003), however, this study represents the first genetic confirmation of that hypothesis. 51 As gene flow between populations is integral to maintaining genetic diversity, and therefore population fitness and persistence, an understanding of connectivity, in the context of the landscape, is required. While landscape genetics is effective and appropriate for understanding dispersal processes at the metapopulation level, in the case of mainland Roosevelt elk it is probably too soon for patterns to have become detectable. As individual elk herds have only been established from 3-23 years, it is unlikely that dispersal can be detected among populations that are mostly sourced from the same secondary population, with molecular markers similar to those used in this study. New molecular methods involving genome wide sequencing may be able to provide more power to identify population origin of suspected migrants. The analyses presented here suggest that some populations of Roosevelt elk may require intervention. Elk in the southern part of Vancouver Island have very low genetic diversity compared to most other BC Roosevelt populations, including sequentially bottlenecked reintroduced herds on the mainland. On the mainland we suggest that the Rainy-Gray, Pitt, Stave, and Powell River North and South (Haslam, Lois, Eldred, Powell-Daniels, Theodosia EPUs) groups represent population units of concern due to loss of allelic diversity, increased inbreeding and fixation metrics, and / or low effective population sizes. Considering the samples sizes from Homathko-Southgate-Orford EPUs, and a single point sample (n=8) from the Toba EPU, we suggest additional efforts should be made to improve our knowledge of the genetic health of these herds. While the high inbreeding values in these EPUs are of concern, the small sample sizes increase uncertainty around these measures. 52 To make the most of scarce financial and human resources, any future translocations to augment genetically depauperate populations should consider a ‘genetically informed’ strategy (Wilson et al. 2005, Pemberton 2008, West et al. 2018, White et al. 2020) where source populations or candidate individuals are evaluated and targeted to maximize genetic differences from the receiving population. While human- mediated dispersal should be considered a last resort, and some projects have failed due to genetic swamping of locally adaptive alleles when source populations are poorly chosen, in the case of Roosevelt elk this is probably not a concern. It is unlikely that different populations of the subspecies in British Columbia express highly adaptive alleles or have evolved potential fitness reducing differences, such as major histocompatibility complex incompatibilities. In contrast, increased heterozygosity and allelic diversity have been shown to contribute to improved fitness in most animals (Lacy 1997, Reed and Frankham 2003, Johnson et al. 2010). It is our opinion that a strategy of translocating a few individuals per year, from areas on Vancouver Island with high frequency of other mitochondrial haplotypes (haplotypes B, C or D), would provide important additional genetic diversity to some, if not all, mainland populations. 53 2.6 Literature Cited Aldridge, C. L., and M. S. Boyce. 2007. Linking occurrence and fitness to persistence: habitat based approach for endangered sage-grouse. Ecological Applications 17:508–526. Allendorf, F. W., and G. Luikart. 2007. Conservation and the genetics of populations. Blackwell Pub, Oxford, UK. Avise, J. C. 1994. Molecular Markers, Natural History and Evolution. Chapman & Hall, New York, NY. Balloux, F., and N. Lugon-Moulin. 2002. The estimation of population differentiation with microsatellite markers. Molecular Ecology 11:155–165. BC Conservation Data Centre. 2020. Conservation Status Report - Roosevelt Elk. BC Species and Ecosystems Explorer. B.C. Minist. of Environ. Victoria, B.C. Available: http://a100.gov.bc.ca/pub/eswp/ (accessed Feb 24, 2020). Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological) 57:289–300. Bérénos, C., P. A. Ellis, J. G. Pilkington, and J. M. Pemberton. 2016. Genomic analysis reveals depression due to both individual and maternal inbreeding in a free-living mammal population. Molecular Ecology 25:3152–3168. Britten, H. B. 1996. Meta-analyses of the association between multilocus heterozygosity and fitness. Evolution 50:2158–2164. Brunt, K. R. 1990. Ecology of Roosevelt elk. Pages 65–98 in J. B. Nyberg and D. W. Janz, editors. Deer and Elk habitats in coastal forests of southern British Columbia. Victoria, BC. Bryant, L. D., and C. Maser. 1982. Classification and distribution. Pages 1–59 in J. W. Thomas and D. E. Toweill, editors. Elk of North America. First Ed. Stackpole Books, Harrisburg, PA. Burg, T. M., and J. P. Croxall. 2001. Global relationships amongst black-browed and grey-headed albatrosses: analysis of population structure using mitochondrial DNA and microsatellites. Molecular Ecology 10:2647–2660. Charlesworth, D., and J. H. Willis. 2009. The genetics of inbreeding depression. Nature Reviews Genetics 10:783–796. Clegg, S. M., S. M. Degnan, J. Kikkawa, C. Moritz, A. Estoup, and I. P. F. Owens. 2002. Genetic consequences of sequential founder events by an island-colonizing bird. Proceedings of the National Academy of Sciences of the United States of America 99:8127–8132. Clutton-Brock, T. H., S. D. Albon, and F. E. Guinness. 1988. Reproductive success in male and female red deer. Pages 325–343 in T. H. Clutton-Brock, editor. Reproductive success: studies of individual variation in contrasting breeding systems. University of Chicago Press, Chicago, Illinois. Conard, J. M., M. J. Statham, P. S. Gipson, and S. M. Wisely. 2010. The influence of translocation strategy and management practices on the genetic variability of a reestablished elk (Cervus elaphus) population. Restoration Ecology 18:85–93. Cornuet, J. M., and G. Luikart. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele fequency data. Genetics 54 144:2001-2014. Cushman, S. A., and E. L. Landguth. 2012. Multi-taxa population connectivity in the northern Rocky Mountains. Ecological Modelling 231:101–112. DiRienzo, A., A. C. Peterson, J. C. Garzat, A. M. Valdes, M. Slatkin, and N. B. Freimer. 1994. Mutational processes of simple-sequence repeat loci in human populations. Proceedings of the National Academy of Sciences of the United States of America 91:3166–3170. Do, C., R. S. Waples, D. Peel, G. M. Macbeth, B. J. Tillett, and J. R. Ovenden. 2014. NeEstimator v2: re-implementation of software for the estimation of contemporary effective population size (Ne) from genetic data. Molecular Ecology Resources 14:209–214. Eberhardt, L. E. 1996. Growth of an isolated elk population. The Journal of Wildlife Management 60:369–373. Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14:2611–2620. Excoffier, L., and H. Lischer. 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10:564–567. Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131:479–491. Fitzsimmons, N. N., S. W. Buskirk, and M. H. Smith. 1997. Genetic changes in reintroduced Rocky Mountain bighorn sheep populations. The Journal of Wildlife Management 61:863–872. Frankham, R. 1995. Effective population size/adult population size ratios in wildlife: a review. Genetical Research 66:95–107. Frankham, R. 2005. Genetics and extinction. Biological Conservation 126:131–140. Frankham, R. 2015. Genetic rescue of small inbred populations: meta-analysis reveals large and consistent benefits of gene flow. Molecular Ecology 24:2610–2618. Frankham, R., C. Bradshaw, and B. W. Brook. 2014. Genetics in conservation management: revised recommendations for the 50/500 rules, Red List criteria and population viability analyses. Biological Conservation 170:56–63. Franklin, A. I. 1980. Evolutionary change in small population. Pages 135–149 in M. E. Soulé and B. A. Wilcox, editors. Conservation Biology, An Evolutionary-Ecological Perspective. Sunderland, MA. Franklin, W. L., A. S. Mossmna, and M. Dole. 1975. Social organization and home range of Roosevelt elk. Journal of Mammalogy 56:102–118. Fritts, S. H., and L. N. Carbyn. 1995. Population viability, nature reserves, and the outlook for gray wolf conservation in North America. Restoration Ecology 3:26–38. Garza, J. C., and E. G. Williamson. 2001. Detection of reduction in population size using data from microsatellite loci. Molecular Ecology 10:305–318. Geist, V. 2002. Adaptive behavioral strategies. Pages 389–483 in D. E. Toweill and J. W. Thomas, editors. North American elk: ecology and management. Second Ed. Smithsonian Institution Press, Washington, D.C. Giglio, R. M., J. A. Ivy, L. C. Jones, and E. K. Latch. 2018. Pedigree-based genetic 55 management improves bison conservation. The Journal of Wildlife Management 82:766–774. Goudet, J. 2005. HIERFSTAT, a package for R to compute and test hierarchical F- statistics. Molecular Ecology Notes 5:184–186. B Greenhorn, J. E., J. Bowman, and P. J. Wilson. 2018. Genetic monitoring suggests increasing structure following recolonization by fishers. The Journal of Wildlife Management 82:1403–1416. Hajkova, P., B. Zemanova, B. J., B. Hajek, K. Roche, E. Tkadlek, and J. Zima. 2006. Factors affecting success of PCR amplification of microsatellite loci from otter faeces. Molecular Ecology Notes 6:559–562. Hale, M. L., T. M. Burg, and T. E. Steeves. 2012. Sampling for microsatellite-based population genetic studies: 25 to 30 individuals per population is enough to accurately estimate allele frequencies. PLoS ONE 7:e45170. Hicks, J. F., J. L. Rachlow, O. E. Rhodes, C. L. Williams, L. P. Waits, M. D., N. R., and Larkin J. 2007. Reintroduction and genetic structure: Rocky Mountain elk in Yellowstone and the western states. Journal of Mammalogy 88:129–138. Hindley, J. A., B. A. Graham, P. C. Pulgarin-R, and T. M. Burg. 2018. The influence of latitude, geographic distance, and habitat discontinuities on genetic variation in a high latitude montane species. Scientific Reports 8:1–11. Hopken, M. W., T. M. Lum, P. M. Meyers, and A. J. Piaggio. 2015. Molecular assessment of translocation and management of an endangered subspecies of white- tailed deer (Odocoileus virginianus). Conservation Genetics 16:635–647. Hubisz, M. J., D. Falush, M. Stephens, and J. K. Pritchard. 2009. Inferring weak population structure with the assistance of sample group information. Molecular Ecology Resources 9:1322–1332. Hundertmark, K. J., and L. J. Van Daele. 2010. Founder effect and bottleneck signatures in an introduced, insular population of elk. Conservation Genetics 11:139–147. Johnson, W. E., D. P. Onorato, M. E. Roelke, E. D. Land, M. Cunningham, R. C. Belden, R. McBride, D. Jansen, M. Lotz, D. Shindle, J. G. Howard, D. E. Wildt, L. M. Penfold, J. A. Hostetler, M. K. Oli, and S. J. O’Brien. 2010. Genetic restoration of the Florida panther. Science 329:1641–1645. Jost, L. 2008. Gst and its relatives do not measure differentiation. Molecular Ecology 17:4015–4026. Keller, L. F., I. Biebach, S. R. Ewing, and P. E. A. Hoek. 2012. The genetics of reintroductions: inbreeding and genetic drift. J. G. Ewen, D. P. Armstrong, K. P. Parker, and P. J. Seddon, editors. Reintroduction biology : integrating science and management. Wiley-Blackwell, Hoboken, NJ. Keller, L. F., and D. M. Waller. 2002. Inbreeding effects in wild populations. Trends in Ecology and Evolution 17:230–241. Komers, P. E., and G. P. Curman. 2000. The effect of demographic characteristics on the success of ungulate re-introductions. Biological Conservation 93:187–193. Kopelman, N. M., J. Mayzel, M. Jakobsson, N. A. Rosenberg, and I. Mayrose. 2015. Clumpak: a program for identifying clustering modes and packaging population structure inferences across K. Molecular Ecology Resources 15:1179–1191. Kumar, S., G. Stecher, M. Li, C. Knyaz, and K. Tamura. 2018. MEGA X: Molecular evolutionary genetics analysis across computing platforms. Molecular Biology and 56 Evolution 35:1547–1549. Lacy, R. C. 1997. Importance of genetic variation to the viability of mammalian populations. Journal of Mammalogy 78:320–335. Landguth, E. L., B. C. Fedy, S. J. Oyler-McCance, A. L. Garey, S. L. Emel, M. Mumma, H. H. Wagner, M.-J. Fortin, and S. A. Cushman. 2012. Effects of sample size, number of markers, and allelic richness on the detection of spatial genetic pattern. Molecular Ecology Resources 12:276–284. Larkin, J. L., J. J. Cox, M. W. Wichrowski, M. R. Dzialak, and D. S. Maehr. 2004. Influences on release-site fidelity of translocated elk. Restoration Ecology 12:97– 105. Larson, S., R. Jameson, J. Bodkin, M. Staedler, and P. Bentzen. 2002a. Microsatellite DNA and mitochondrial DNA variation in remnant and translocated sea otter (Enhydra lutris) populations. Journal of Mammalogy 83:893–906. Larson, S., R. Jameson, M. Etnier, M. Fleming, and P. Bentzen. 2002b. Loss of genetic diversity in sea otters (Enhydra lutris) associated with the fur trade of the 18th and 19th centuries. Molecular Ecology 11:1899–1903. Lawson, D. J., L. van Dorp, and D. Falush. 2018. A tutorial on how not to over-interpret STRUCTURE and ADMIXTURE bar plots. Nature Communications 9:3258. Lubow, B. C. 1996. Optimal translocation strategies for enhancing stochastic metapopulation viability. Ecological Applications 6:1268–1280. Luikart, G., and J.-M. Cornuet. 1998. Empirical evaluation of a test for identifying recently bottlenecked populations from allele frequency data. Conservation Biology 12:228–237. Manel, S., M. K. Schwartz, G. Luikart, and P. Taberlet. 2003. Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology & Evolution 18:189–197. Marshall, T. C., J. Slate, L. E. B. Kruuk, and J. M. Pemberton. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology 7:639–655. Meirmans, P. G. 2006. Using the AMOVA framework to estimate a standardized genetic differentiation measure. Evolution 60:2399–2402. Meirmans, P. G., and P. H. Van Tienderen. 2004. genotype and genodive: two programs for the analysis of genetic diversity of asexual organisms. Molecular Ecology Notes 4:792–794. Muller, L. I., J. L. Murrow, J. L. Lupardus, J. D. Clark, J. G. Yarkovich, W. H. Stiver, E. K. Delozier, B. L. Slabach, J. J. Cox, and B. F. Miller. 2018. Genetic structure in elk persists after translocation. The Journal of Wildlife Management 82:1124–1134. Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic variability in populations. Evolution 29:1–10. Nei, M., and M. Murata. 1966. Effective population size when fertility is inherited. Genetical Research 8:257–260. Nunney, L. 1993. The influence of mating system and overlapping generations on effective population size. Evolution 47:1329–1341. Nussey, D. H., D. W. Coltman, T. Coulson, L. E. B. Kruuk, A. Donald, S. J. Morris, T. H. Clutton-Brock, and J. Pemberton. 2005. Rapidly declining fine-scale spatial genetic structure in female red deer. Molecular Ecology 14:3395–3405. 57 O’Brien, S. J., and J. F. Evermann. 1988. Interactive influence of infectious disease and genetic diversity in natural populations. Trends in Ecology and Evolution 3:254– 259. Olson, Z. H., D. G. Whittaker, and O. E. Rhodes. 2013. Translocation history and genetic diversity in reintroduced bighorn sheep. The Journal of Wildlife Management 77:1553–1563. Ortego, J., G. Yannic, A. B. A. Shafer, J. Mainguy, M. Festa-Bianchet, D. W. Coltman, and S. D. Côté. 2011. Temporal dynamics of genetic variability in a mountain goat (Oreamnos americanus) population. Molecular Ecology 20:1601–1611. Paetkau, D. 2003. An empirical exploration of data quality in DNA-based population inventories. Molecular Ecology 12:1375–1387. Peakall, P. E., and R. Smouse. 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research–an update. Bioinformatics 28:2537– 2539. Pemberton, J. M. 2008. Wild pedigrees: The way forward. Proceedings of the Royal Society B: Biological Sciences 275:613–621. Piry, S., G. Luikart, and J.-M. Cornuet. 1999. BOTTLENECK: a computer program for detecting recent reductions in the effective population size using allele frequency data. Journal of Heredity 90:502–503. Polziehn, R. O., J. Hamr, F. F. Mallory, and C. Strobeck. 2000. Microsatellite analysis of North American wapiti (Cervus elaphus) populations. Molecular Ecology 9:1561– 1576. Polziehn, R. O., J. Hamr, F. Mallory, and C. Strobeck. 1998. Phylogenetic status of North American wapiti (Cervus elephus) subspecies. Canadian Journal of Zoology 76:998– 1010. Porras-Hurtado, L., Y. Ruiz, C. Santos, C. Phillips, A. Carracedo, and M. V. Lareu. 2013. An overview of STRUCTURE: applications, parameter settings, and supporting software. Frontiers in genetics 4:1–13. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945–959. Quayle, J. F., and K. R. Brunt. 2003. Status of Roosevelt elk (Cervus elaphus roosevelti) in British Columbia. Victoria, BC. R Core Team. 2013. R: A language and environment for statistical computing. Vienna, Austria. Ralls, K., J. Ballou, M. Dudash, C. Fenster, M. Eldridge, R. Lacy, P. Sunnucks, and R. Frankham. 2018. Call for a paradigm shift in the genetic management of fragmented populations. Conservation Letters 11:1–6. Ramón-Laca, A., L. Soriano, D. Gleeson, and J. A. Godoy. 2015. A simple and effective method for obtaining mammal DNA from faeces. Wildlife Biology 21:195–203. Raymond, M., and F. Rousset. 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity 86:248–249. Reed, D. H., and R. Frankham. 2003. Correlation between fitness and genetic diversity. Conservation Biology 17:230–237. Reynolds, D., J. Kelly, N. Tweddle, and C. Morrison. 2018. Coastal Mainland Roosevelt Elk Recovery and Management Project. Victoria, BC. Rousset, F. 2008. genepop’007: a complete re-implementation of the genepop software 58 for Windows and Linux. Molecular Ecology Resources 8:103–106. Rozas, J., A. Ferrer-Mata, J. C. Sanchez-DelBarrio, S. Guirao-Rico, P. Librado, S. E. Ramos-Onsins, and A. Sanchez-Gracia. 2017. DnaSP 6: DNA sequence polymorphism analysis of large data sets. Molecular Biology and Evolution 34:3299–3302. Sattler, R. L., J. R. Willoughby, and B. J. Swanson. 2017. Decline of heterozygosity in a large but isolated population: a 45-year examination of moose genetic diversity on Isle Royale. PeerJ 5:e3584 Schlötterer, C. 2000. Evolutionary dynamics of microsatellite DNA. Chromosoma 109:365–371. Slate, J., L. E. B. Kruuk, T. C. Marshall, J. M. Pemberton, and T. H. Clutton-Brock. 2000. Inbreeding depression influences lifetime breeding success in a wild population of red deer (Cervus elaphus). Proceedings of the Royal Society B: Biological Sciences 267:1657-1662. Slatkin, M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139:457–462. Spalding, D. J. 1992. The history of elk in British Columbia. Victoria, BC. Speller, C. F., B. Kooyman, A. T. Rodrigues, E. G. Langemann, R. M. Jobin, and D. Y. Yang. 2014. Assessing prehistoric genetic structure and diversity of North American elk (Cervus elaphus) populations in Alberta, Canada. Canadian Journal of Zoology 92:285–298. Stephen, C. L., D. G. Whittaker, D. Gillis, L. L. Cox, and O. E. Rhodes. 2005. Genetic consequences of reintroductions: an example from Oregon pronghorn antelope (Antilocapra americana). Journal of Wildlife Management 69:1463–1474. Stockwell, C. A., M. Mulvey, and G. L. Vinyard. 1996. Translocations and the preservation of allelic diversity. Conservation Biology 10:1133–1141. Storfer, A., M. A. Murphy, J. S. Evans, C. S. Goldberg, S. Robinson, S. F. Spear, R. Dezzani, E. Delmelle, L. Vierling, and L. P. Waits. 2007. Putting the ‘landscape’ in landscape genetics. Heredity 98:128–142. Tajima, F., and M. Nei. 1984. Estimation of evolutionary distance between nucleotide sequences. Molecular Biology and Evolution 1:269–285. Vander Wal, E., P. C. Paquet, and J. A. Andrés. 2012. Influence of landscape and social interactions on transmission of disease in a social cervid. Molecular Ecology 21:1271–1282. Walsh, P. S., D. A. Metzger, and R. Higuchi. 1991. Chelex® 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. BioTechniques 10:506–513. Wang, J. 2017. The computer program STRUCTURE for assigning individuals to populations: easy to use but easier to misuse. Molecular Ecology Resources 17:981– 990. Wang, J., E. Santiago, and A. Caballero. 2016. Prediction and estimation of effective population size. Heredity 117:193-206. Waples, R. S., T. Antao, and G. Luikart. 2014. Effects of overlapping generations on linkage disequilibrium estimates of effective population size. Genetics 197:769–780. West, R., S. Potter, D. Taggart, and M. D. B. Eldridge. 2018. Looking back to go forward: genetics informs future management of captive and reintroduced 59 populations of the black-footed rock-wallaby Petrogale lateralis. Conservation Genetics 19:235–247. White, K. S., T. Levi, J. Breen, M. Britt, J. Meröndun, D. Martchenko, Y. N. Shakeri, B. Porter, and A. B. A. Shafer. 2020. Integrating genetic data and demographic modeling to facilitate conservation of small, isolated mountain goat populations. The Journal of Wildlife Management 85:271–282. Wilson, G. A., J. S. Nishi, B. T. Elkin, and C. Strobeck. 2005. Effects of a recent founding event and intrinsic population dynamics on genetic diversity in an ungulate population. Conservation Genetics 6:905–916. Wolf, C. M., B. Griffith, C. Reed, and S. A. Temple. 1996. Avian and mammalian translocations: update and reanalysis of 1987 survey data. Conservation Biology 10:1142–1154. Wright, S. 1931. Evolution in Mendelian populations. Genetics 16:97–159. Zachos, F. E., C. Althoff, Y. V. Steynitz, I. Eckert, and G. B. Hartl. 2007. Genetic analysis of an isolated red deer (Cervus elaphus) population showing signs of inbreeding depression. European Journal of Wildlife Research 53:61–67. 60 Table 2.1 Mitochondrial haplotypes identified in 58 British Columbia Roosevelt elk (Cervus canadensis roosevelti); nucleotide positions aligned to consensus sequence (OG) for North American Wapiti from Polziehn et al. (1998); † = diagnostic site for Roosevelt subspecies. Haplotype Nucleotide position 441 450 476 493† OG T G A A A T G A G B T G G G C T A A G D C A A G Table 2.2 Summary of mitochondrial diversity for 58 Roosevelt elk D-loop sequences (C. c. roosevelti) and pairwise Fst values; *** P <0.001. Population Number of Haplotype Nucleotide Number of Pairwise FST individuals diversity diversity haplotypes VIN VIS (n) (HD) (π) observed Vancouver 18 0.523 0.0009 A (n = 10) - - Island B (n = 8) North Vancouver 10 0.356 0.0006 C (n = 8) 0.721*** - Island D (n = 2) South Mainland 30 0.000 0.0000 A (n = 30) 0.494*** 0.925*** Total 58 0.494 0.0011 - - - 61 Table 2.3 Population groupings and associated management units for Roosevelt elk (C. c. roosevelti) study samples (n=355) used in microsatellite and mitochondrial analyses. Grouping WMU / EPU Number of MSAT success D-loop samples (min 7/10 loci) Sequence VIS 1-03A-B 9 104A-C 26 1-05A-C 9 Total 44 41 10 VIN 1-06A-B/D 9 1-09A-D 12 1-10A-H 20 1-11A-B 9 1-12C-E 5 Total 55 55 18 SP 2-05A 49 48 11 CC 2-05D 8 2-05E-H 27 Total 35 35 - RG 2-05I 55 53 0 SQ 2-06A/2-08A 17 16 5 PITT 2-08B 29 29 5 STV 2-08C/2-09A 16 16 2 BRIT 2-12E/Q 10 10 - PRS 2-12A/S 18 17 7 PRN 2-12D/F 17 16 - TOBA 2-13A 17 11 - HO 2-14A-B/2-15A 10 8 - TOTAL 372 355 58 62 Table 2.4 Genetic diversity statistics at 10 microsatellite loci for 13 populations of Roosevelt elk (Cervus c. roosevelti) in British Columbia, including number of alleles per locus (Na), effective number of alleles (Neff), Shannon Index (I), observed heterozygosity (Ho), expected heterozygosity (He), unbiased expected heterozygosity (uHe), fixation index (F), allelic richness (Ar), and the inbreeding coefficient Gis, * indicates significance at a = 0.05, adjusted for FDR (P ≤0.015). Pop n Na Neff Ar I Ho He uHe F Gis VIS 41 2.7 1.94 2.229 0.710 0.423 0.469 0.475 0.101 *0.111 VIN 55 3.0 2.20 2.635 0.843 0.538 0.519 0.524 -0.020 -0.027 SP 48 2.9 2.08 2.577 0.816 0.500 0.507 0.512 0.008 0.024 CC 35 2.8 2.03 2.549 0.787 0.481 0.484 0.492 0.013 0.021 RG 53 2.9 2.07 2.489 0.792 0.444 0.499 0.504 0.125 *0.121 SQ 16 3.0 1.96 2.689 0.765 0.471 0.455 0.470 -0.005 -0.003 PITT 29 2.9 2.08 2.635 0.804 0.491 0.485 0.493 -0.012 0.004 STV 16 2.9 2.29 2.729 0.866 0.504 0.535 0.554 0.074 0.093 BRIT 10 2.7 1.99 2.633 0.763 0.480 0.468 0.493 -0.007 0.027 PRS 17 2.7 1.84 2.487 0.694 0.448 0.423 0.436 -0.019 -0.028 PRN 16 2.6 2.07 2.512 0.788 0.497 0.498 0.514 0.005 0.035 TOBA 11 2.7 2.09 2.640 0.807 0.409 0.499 0.525 0.199 *0.231 HO 8 2.6 2.05 2.600 0.782 0.413 0.493 0.526 0.159 *0.227 Mean 27.1 2.8 2.05 2.570 0.786 0.469 0.487 0.501 0.048 0.066 63 Table 2.5 Averaged diversity statistics for 10 microsatellite loci in 13 populations of Roosevelt elk (C. c. roosevelti) in British Columbia, including number of alleles per locus (Na), effective number of alleles (Neff), observed heterozygosity (Ho), expected heterozygosity (He), corrected expected heterozygosity (Ht), and inbreeding coefficient Gis. Locus Na Neff Ho He Ht Gis BL42 3 2.437 0.628 0.606 0.621 -0.037 BM203 3 1.842 0.448 0.471 0.504 0.048 BM4107RD 4 2.098 0.527 0.539 0.558 0.022 BM4513RD 4 2.027 0.472 0.523 0.532 0.097 BM6506 2 1.891 0.437 0.487 0.483 0.101 BM888 3 1.549 0.318 0.366 0.367 0.132 BMC1009 3 1.789 0.413 0.455 0.469 0.092 CSSM041 2 1.849 0.449 0.473 0.470 0.050 Inra107 5 2.831 0.682 0.665 0.681 -0.026 RT7 2 1.734 0.317 0.439 0.467 0.278 Overall 3.1 2.005 0.469 0.502 0.515 0.066 64 Table 2.6 Pairwise genetic difference for 13 populations of Roosevelt elk (C. c. roosevelti) in British Columbia; Fst values below diagonal (significant values in bold), P values above diagonal (FDR adjusted a = 0.05, P = 0.0179). POP VIS VIN SP CC RG SQ PITT STV BRIT PRS PRN TOBA HO VIS -- <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 0.043 VIN 0.067 -- <0.001 <0.001 <0.001 <0.001 <0.001 0.016 0.007 <0.001 0.365 0.018 0.485 SP 0.074 0.020 -- 0.216 0.056 0.001 0.010 0.649 0.233 <0.001 0.085 0.212 0.326 CC 0.086 0.037 0.004 -- 0.071 0.035 0.873 0.777 0.601 0.002 0.282 0.114 0.102 RG 0.065 0.028 0.008 0.009 -- 0.040 0.007 0.582 0.543 0.027 0.063 0.926 0.210 SQ 0.103 0.054 0.038 0.020 0.019 -- 0.132 0.037 0.699 0.078 0.030 0.051 0.108 PITT 0.105 0.049 0.016 -0.007 0.018 0.011 -- 0.294 0.887 0.003 0.134 0.096 0.019 STV 0.067 0.023 -0.005 -0.008 -0.004 0.029 0.004 -- 0.408 0.007 0.331 0.505 0.493 BRIT 0.092 0.035 0.008 -0.005 -0.003 -0.01 -0.015 0.001 -- 0.076 0.542 0.397 0.100 PRS 0.090 0.041 0.040 0.035 0.021 0.019 0.041 0.044 0.025 -- 0.017 0.087 0.124 PRN 0.071 0.001 0.012 0.004 0.015 0.031 0.011 0.004 -0.004 0.036 -- 0.163 0.405 TOBA 0.052 0.028 0.008 0.016 -0.017 0.035 0.019 -0.003 0.002 0.024 0.017 -- 0.311 HO 0.035 -0.001 0.005 0.022 0.012 0.031 0.046 -0.004 0.037 0.023 0.002 0.010 -- 65 Table 2.7 Bottleneck analysis results for six populations of Roosevelt elk (C. c. roosevelti) in British Columbia for the Sign test, Wilcoxon Sign Rank test, and deviation from an L-shaped allele frequency distribution for three different mutation models: Infinite Allele model (IAM), Two Phase Mutation model (TPM; strict stepwise mutation=90%, variance=0.12), and Stepwise Mutation model (SMM) for populations with sample size >20: Hex = expected number of loci with heterozygosity excess at mutation-drift equilibrium, He = observed number of loci with heterozygosity excess, L = expected “L” shaped allele frequency distribution, * denotes significance at a=0.05. Test VIS n=41 VIN n=55 SP n=48 RG n=53 PITT n=29 PR n=26 Sign Hex\He P Hex \ He P Hex \ He P Hex \ He P Hex \ He P Hex \ He P IAM 4.67 \ 9 *0.006 4.77 \ 10 *0.001 4.78 \ 10 *<0.001 4.35 \ 10 *<0.001 4.94 \ 10 *0.001 4.84 \ 10 *<0.001 TPM 5.38 \ 9 *0.018 5.23 \ 9 *0.017 5.19 \ 8 0.067 5.33 \ 8 0.080 5.56 \ 7 0.088 5.35 \ 7 0.234 SMM 5.45 \ 8 0.091 5.40 \ 8 0.087 5.20 \ 7 0.205 5.41 \ 8 0.088 5.61 \ 6 0.528 5.41 \ 7 0.246 Wilcoxon IAM *0.003 *<0.001 *<0.001 *<0.001 * <0.001 *<0.001 TPM *0.042 *0.005 *0.003 *0.007 0.065 *0.016 SMM 0.053 *0.009 *0.012 0.053 0.080 0.065 Dist. Shape L L Shifted L Shifted Shifted 66 Table 2.8 Effective population size (Ne) for 13 populations of Roosevelt elk (C. c. roosevelti) in British Columbia, linkage disequilibrium model for a single time-point with two minimum allele frequency cut-off values (0.05, 0.01) with 95% confidence intervals, with N values calculated from census estimates (Wilson 2015, Reynolds et al. 2018). Pop’n Sample Crit. Weighted Ne CI (95%) Ne:N Ne:N Name Size Value Mean Min Max Values Ratio VIS 41 0.05 40.9 58.5 18.9 Infinite 0.01 44.4 17.0 1302.3 58.5:1400 0.04 VIN 55 0.05 53.7 59.7 26.4 389.6 0.01 68.9 28.7 1187.2 68.9:4870 0.01 SP 48 0.05 46.9 52.9 21.2 935.7 0.01 55.4 22.7 887.5 55.4:300 0.18 CC 35 0.05 34.5 11.9 5.3 24.8 0.01 18.8 8.8 49.3 18.8:356 0.05 RG 53 0.05 47.6 143.0 34.4 Infinite 0.01 229.7 40.2 Infinite 40.2:101 0.40 SQ 16 0.05 15.5 161.5 12.2 Infinite 0.01 51.7 9.5 Infinite 51.7:170 0.30 PITT 30 0.05 28.1 7.0 2.9 13.8 0.01 9.0 3.6 18.4 9:79 0.11 STV 15 0.05 14.6 10.9 2.9 90.9 0.01 12.0 3.1 164.7 12:120 0.10 BRIT 10 0.05 10 24.1 2.7 Infinite 0.01 24.1 2.7 Infinite 24.1:130 0.18 PRS 17 0.05 16 22.4 5.2 Infinite 0.01 35.7 7.0 Infinite 35.7:120 0.30 PRN 16 0.05 15.5 10.7 2.7 178.3 0.01 11.4 2.7 368 11.4:123 0.09 TOBA 11 0.05 9.6 131 3.6 Infinite 0.01 99.2 3.3 Infinite - - HO 8 0.05 8 18.6 1.9 Infinite 0.01 18.6 1.9 Infinite 18.6:130 0.14 67 Figure 2.1: Study area and translocations of Roosevelt elk (Cervus canadensis roosevelti) from Vancouver Island to the mainland of British Columbia between 1986 and 2017; red triangles = primary source populations, yellow diamond = Sechelt Peninsula secondary source population, orange triangle = Powell River (Haslam) tertiary source population; translocations to Phillips and Heydon EPUs (green circles) occurred in 2017. 68 Figure 2.2: Translocations of Roosevelt elk (C. c. roosevelti) in the South Coast Region of British Columbia from 2001 – 2017. 69 Figure 2.3: Elk population groupings used for analyses of Roosevelt elk (C. c. roosevelti) genetic analyses, and locations of samples (n = 355) collected between 2012 and 2019 in British Columbia. 70 Fig. 2.4: A) Observed haplotypes and associated sampling locations for Roosevelt elk (C. c. roosevelti) in British Columbia, based on 566 bp mitochondrial D-loop sequences (n = 58). B) Phylogenetic tree and haplotype network for D-loop sequences observed in Roosevelt elk (C. c. roosevelti) in British Columbia; consensus sequence for C. c. nelsoni (Polziehn et al., 1998) used as the outgroup; numbers are transition sites with nucleotide transition in brackets, letters on branch tips are observed haplotypes. 71 Figure 2.5 STRUCTURE ancestry plots for 355 Roosevelt elk (C. c. roosevelti) in British Columbia: A) for 13 populations ∆K = 2 and Prob(K) = 4; B) for 12 populations ∆K = 3 and Prob(K) = 3. STRUCTURE settings = 50,000 burn in and 200,000 MCMC steps, with admixture and locprior model, infer a, independent values for a, initial a = 0.25; plots and Best K values calculated in CLUMPAK (Kopelman et al., 2015). 72 CHAPTER 3: GENERAL DISCUSSION The reintroduction of species into the landscapes they have been extirpated from is a challenging undertaking. The science behind reintroductions, and the augmentation of extirpated and / or endangered populations, has shifted towards understanding the underlying factors that influence the establishment and long term success of those species, including the important role of genetics (Franklin 1980, Frankham 1995, Ralls et al. 2018). It is to this body of knowledge I hope this study makes a contribution. For many at risk species former habitats have become unsuitable (Fritts et al. 1997), while for others the underlying cause of their extirpation, such as direct exploitation or isolation from conspecifics, is still a concern (Wolf et al. 1996). In a review of 180 reintroduction and translocation projects, Fischer & Lindenmayer (2000) identified three main aims of successful reintroduction and translocation projects: to address human – wildlife conflicts, to restock game species, and to conserve species. The reintroduction of Roosevelt elk (Cervus canadensis roosevelti) from Vancouver Island to their former habitat on the mainland was initiated to address conservation concerns; low species diversity, struggling populations and ongoing illegal hunting. When the subspecies population grew rapidly and began to find conflict with the area’s human residents, a long term reintroduction and translocation project was initiated to mediate conflict, create sustainable harvest opportunities for resident and indigenous hunters, solidify the return of Roosevelt elk, and restore ecosystem function to the coastal mainland of BC (Reynolds et al. 2018). However, despite a ~2.5x increase in the provincial population since 2001, and range expansion of over 30,000 km2 (Wilson 2015, 73 Reynolds et al. 2018), there remains concern about the future of this economically and culturally important, charismatic animal. 3.1 Management Implications Genetic evaluation and monitoring of reintroduced populations of Roosevelt elk have been identified as areas of need, as Recommendation 6 in Reynolds et al. (2018). Our assessment has identified numerous reintroduced populations with reduced diversity compared to the core population on northern Vancouver Island. Mitochondrial analyses showed that mainland populations have half the diversity of northern Vancouver Island elk, and only one quarter the diversity of Roosevelt elk across the island metapopulation, with a complete division of haplotypes north and south of the Alberni Inlet (Table 2.2, Figure 2.4a). This finding, on its own, should be enough to initiate a review of past management strategies and careful consideration of potential mitigations. Further analyses with multi-locus nuclear genotypes identified more cause for concern; as seen with mtDNA, microsatellite genotypes revealed Roosevelt elk populations are highly structured between the northern and southern areas of Vancouver Island (Figure 2.5a), with low effective population sizes (Table 2.8). The southern island population exhibits less diversity, loss of low frequency alleles, and substantial shifts in allele frequencies compared to northern Vancouver Island (Appendix VI). The mainland population is genetically isolated from Vancouver Island populations, and without mitigation, this is likely to continue. Loss of alleles and changes in allele frequencies among the mainland populations is likely the combined result of pre-existing low diversity, polygynous breeding system, and translocation of limited numbers of founding individuals (Nunney 1993, Polziehn et al. 2000). 74 Currently the threshold for a Roosevelt elk population unit to be considered ‘recovered’ is a population of >50 animals, with a bull to cow ratio 20:100 or greater, with 30% of bulls branch antlered (Wilson 2015), which is usually seen in two year-old and older bulls. In practice, most populations have a more even bull:cow ratio by the time the population has reached 50. Once designated as recovered, hunting opportunities are allocated first to indigenous communities, and then through a controlled limited draw system for resident hunters (Wilson 2015). The low effective population (Ne) values observed in many sampled populations, while probably downward biased somewhat due to sampling multiple years and potentially multiple generations (Waples et al. 2014), suggest that the census threshold for recovered status may currently be too low. However, the observed Ne estimates for Vancouver Island are also very low, equating to Ne:N ratios of approximately 0.04 and 0.01 for VIS and VIN, respectively (Table 2.8). These values are extremely concerning and require further evaluation. The observed overall low genetic diversity, low Ne, reduced allelic diversity, rapid differentiation due to drift, and increased levels of inbreeding, serve to underline the status of Roosevelt elk as a “species of special concern”. Individually these signals are problematic, in combination they require a strategy to mitigate against further genetic degradation. 3.2 Future Directions The reintroduction of Roosevelt elk to the British Columbia mainland should be heralded as a success in the restoration of species at risk and ecosystem function. The increased population size and range expansion represent significant steps towards securing the future of this species. The identification of populations of concern should be 75 used as an opportunity to reinvigorate continued management of Roosevelt elk, and conservation of their critical habitats. Research is currently underway to identify ecological drivers of important habitats for elk survival. This critical work is being done to assist in identification and designation of ungulate winter range on Vancouver Island, and aid in forestry management planning (A. Ford, University of British Columbia, personal communication 2021). An important element in the reintroduction strategy for Roosevelt elk has been an assumption of gene-flow among mainland populations. Long-distance dispersal along the coast by male elk, most likely from the Haslam herd, was anecdotally supported in the late 1990s and early 2000s. Likewise, a single female translocated into the Chehalis EPU travelled over 120 km, including a 3.2 km swim across a lake, before being killed in a motor vehicle collision (D. Reynolds, FLNRO, personal communication 2021). These events support the assumption that gene-flow among many mainland groups is possible, if at low rates. Low rates of dispersal between populations can mitigate drift, preventing or reducing differentiation between populations (Wright 1951, Balloux and Lugon- Moulin 2002, Allendorf and Luikart 2007). It is therefore important to understand the landscape level connectivity in the South Coast Region. While current genetic diversity and close relatedness of mainland populations is not likely conducive to identification of migrants, it may be possible to use genetic studies of other species in the region to predict important multi-species corridors for dispersal (Cushman and Landguth 2012, Brennan et al. 2020). Currently unpublished genetic connectivity analyses have been previously been conducted in the South Coast Region for wolverine (Gulo gulo) and grizzly bears (C. Neivelt and S. Rochetta, FLNRO 76 personal communication 2019). Additional species that may be suitable as proxies for modelling population connectivity could include black bear, black-tailed deer, grey wolf, or even ruffed-grouse (Bonasa umbellus). The provincial government has set a goal of removing Roosevelt elk from the Blue List of species at risk by 2025 (Wilson 2015, Reynolds et al. 2018) through the following objectives: “1. Maintain self-sustaining populations of Roosevelt elk throughout their current range in the West Coast and South Coast Regions, 2. Re-establish Roosevelt elk in their historic range where ecological conditions are suitable, 3. Maintain or restore the contribution of Roosevelt elk to natural biodiversity and ecosystem function.” To this list I suggest adding a fourth objective in support of the first: to maintain and / or improve genetic diversity in Roosevelt elk within individual population units, and for the subspecies as a whole. The maintenance of genetic diversity in reintroduced populations has been identified as a critical element in their long term survival and adaption to environmental change (IUCN/SSC 2013, Frankham et al. 2014). As such, it is recommended that a mitigation strategy be developed to improve the genetic diversity of Roosevelt elk within mainland populations, through the use of translocation of individuals from populations of high or different diversity (Hogg et al. 2006, Frankham 2015, Poirier et al. 2019). Occasional translocations of elk from herds in conflict with residents on Vancouver Island could be partially or wholly diverted to reinforce genetic diversity among mainland populations. Of considerable genetic value would be individuals with mitochondrial haplotypes not found on the mainland, such as those that 77 occur on northern Vancouver Island in the vicinity of the Adam, Eve, Tsitika, Nimpkish, White and Salmon Rivers, near Sayward. Risks associated with further translocations between Vancouver Island and the mainland, such as outbreeding or disease, are far outweighed by the potential benefits for the subspecies (Frankham 2015). 3.3 Conclusions A growing number of studies on the genetics of reintroduced polygynous ungulates have shown potential negative consequences due to drift, isolation and subsequent inbreeding depression (Stephen et al. 2005, Olson et al. 2013, Bérénos et al. 2016, Sattler et al. 2017). The results of this study add to the current body of knowledge around genetic changes associated with species reintroductions; identifying the potential for bottlenecks, founder effects, isolation, and source population characteristics to contribute to increased drift and population differentiation. Populations of at-risk species benefit from our understanding of the fundamental biodiversity that population genetic studies provide. The continued persistence and successful expansion of many species depends on maintaining their capacity to adapt to changing environments, and without knowing which species or populations are vulnerable, it is impossible to know where wildlife managers should prioritize their resources. This author proposes that Roosevelt elk should be considered high priority for additional management, through ongoing translocations from Vancouver Island to the mainland, to prevent continued loss of genetic diversity. 78 3.4 Literature Cited Allendorf, F. W., and G. Luikart. 2007. Conservation and the genetics of populations. Blackwell Pub, Oxford, UK. Balloux, F., and N. Lugon-Moulin. 2002. The estimation of population differentiation with microsatellite markers. Molecular Ecology 11:155–165. Bérénos, C., P. A. Ellis, J. G. Pilkington, and J. M. Pemberton. 2016. Genomic analysis reveals depression due to both individual and maternal inbreeding in a free-living mammal population. Molecular Ecology 25:3152–3168. Brennan, A., P. Beytell, O. Aschenborn, P. Du Preez, P. J. Funston, L. Hanssen, J. W. Kilian, G. Stuart‐Hill, R. D. Taylor, and R. Naidoo. 2020. Characterizing multispecies connectivity across a transfrontier conservation landscape. Journal of Applied Ecology 57:1700–1710. Cushman, S. A., and E. L. Landguth. 2012. Multi-taxa population connectivity in the northern Rocky Mountains. Ecological Modelling 231:101–112. Fischer, J., and D. B. Lindenmayer. 2000. An assessment of the published results of animal relocations. Biological Conservation. 96:1-11. Frankham, R. 1995. Effective population size/adult population size ratios in wildlife: a review. Genetical Research 66:95–107. Frankham, R. 2015. Genetic rescue of small inbred populations: meta-analysis reveals large and consistent benefits of gene flow. Molecular Ecology 24:2610–2618. Frankham, R., C. Bradshaw, and B. W. Brook. 2014. Genetics in conservation management: revised recommendations for the 50/500 rules, Red List criteria and population viability analyses. Biological Conservation 170:56–63. Franklin, A. I. 1980. Evolutionary change in small population. Pages 135–149 in M. E. Soulé and B. A. Wilcox, editors. Conservation Biology, An Evolutionary-Ecological Perspective. Sunderland, MA. Fritts, S. H., E. E. Bangs, J. A. Fontaine, M. R. Johnson, M. K. Phillips, E. D. Koch, and J. R. Gunson. 1997. Planning and implementing a reintroduction of wolves to Yellowstone National Park and central Idaho. Restoration Ecology 5:7–27. Hogg, J. T., S. H. Forbes, B. M. Steele, and G. Luikart. 2006. Genetic rescue of an insular population of large mammals. Proceedings of the Royal Society B: Biological Sciences 273:1491–1499. IUCN/SSC. 2013. Guidelines for Reintroductions and Other Conservation Translocations. Gland, CH. Nunney, L. 1993. The influence of mating system and overlapping generations on effective population size. Evolution 47:1329–1341. Olson, Z. H., D. G. Whittaker, and O. E. Rhodes. 2013. Translocation history and genetic diversity in reintroduced bighorn sheep. The Journal of Wildlife Management 77:1553–1563. Poirier, M., D. W. Coltman, F. Pelletier, J. Jorgenson, and M. Festa‐Bianchet. 2019. Genetic decline, restoration and rescue of an isolated ungulate population. Evolutionary Applications 12:1318–1328. Polziehn, R. O., J. Hamr, F. F. Mallory, and C. Strobeck. 2000. Microsatellite analysis of North American wapiti (Cervus elaphus) populations. Molecular Ecology 9:1561– 1576. 79 Ralls, K., J. Ballou, M. Dudash, C. Fenster, M. Eldridge, R. Lacy, P. Sunnucks, and R. Frankham. 2018. Call for a paradigm shift in the genetic management of fragmented populations. Conservation Letters 11:1–6. Reynolds, D., J. Kelly, N. Tweddle, and C. Morrison. 2018. Coastal Mainland Roosevelt Elk Recovery and Management Project. Victoria, BC. Sattler, R. L., J. R. Willoughby, and B. J. Swanson. 2017. Decline of heterozygosity in a large but isolated population: a 45-year examination of moose genetic diversity on Isle Royale. PeerJ 5:e3584 Stephen, C. L., D. G. Whittaker, D. Gillis, L. L. Cox, and O. E. Rhodes. 2005. Genetic consequences of reintroductions: an example from Oregon pronghorn antelope (Antilocapra americana). Journal of Wildlife Management 69:1463–1474. Waples, R. S., T. Antao, and G. Luikart. 2014. Effects of overlapping generations on linkage disequilibrium estimates of effective population size. Genetics 197:769–780. Wilson, S. 2015. A Management Plan for Roosevelt Elk in British Columbia. Victoria, BC. Wolf, C. M., B. Griffith, C. Reed, and S. A. Temple. 1996. Avian and mammalian translocations: update and reanalysis of 1987 survey data. Conservation Biology 10:1142–1154. Wright, S. 1951. The genetical structure of natural populations. Annals of Eugenics 15:323–354. 80 Appendix I: Translocation history of Roosevelt elk (Cervus canadensis roosevelti) for mainland reintroductions between 1987-2017 (Reynolds et al. 2018); Sechelt Peninsula served as the main translocation source from 2000-2017 (72.3% of individuals), followed by Haslam (near Powell River, 19.3%), Vancouver Island (2017 translocations into Heydon and Phillips EPUs, 5.3%) and Rainy-Gray EPU (3.1%); * this value is reported in Spalding (1992) and Quayle and Brunt (2003), however, it is later reported as 11 in Wilson (2015) and Reynolds et al. (2018). Year Source Location Release EPU Released (n) 1987 Vancouver Island (Campbell R) Sechelt Peninsula 7 1988 Vancouver Island (Campbell R) Sechelt Peninsula 6 1989 Vancouver Island (Qualicum) Sechelt Peninsula *9 1994 Vancouver Island (Fanny Bay) Haslam 5 1996 Sechelt Peninsula Haslam 9 1996 Sechelt Peninsula Haslam 5 1996 Sechelt Peninsula Haslam 6 2000 Sechelt Peninsula McNab 25 2001 Sechelt Peninsula Skwawka 12 2001 Sechelt Peninsula Rainy-Gray 6 2002 Haslam Narrows 7 2002 Sechelt Peninsula McNab 1 2002 Sechelt Peninsula Skwawka 7 2002 Sechelt Peninsula Narrows 4 2003 Haslam Narrows 9 2003 Haslam Rainy-Gray 3 2003 Sechelt Peninsula Clowhom 5 2003 Sechelt Peninsula Narrows 2 2003 Sechelt Peninsula Rainy-Gray 2 2004 Haslam Clowhom 7 2004 Sechelt Peninsula Clowhom 8 2004 Sechelt Peninsula Rainy-Gray 2 2004 Sechelt Peninsula Deserted 13 2005 Sechelt Peninsula Brittain 20 2005 Sechelt Peninsula Deserted 7 2005 Sechelt Peninsula Pitt 23 2006 Haslam McNab 8 2006 Haslam Vancouver 11 2006 Sechelt Peninsula Indian 20 2006 Sechelt Peninsula Vancouver 10 2007 Haslam Squamish 26 2007 Haslam Quatum 12 2007 Sechelt Peninsula Stave 19 2008 Haslam Quatum 6 2008 Sechelt Peninsula Powell-Daniels 17 2008 Sechelt Peninsula Quatum 1 2008 Sechelt Peninsula Stave 1 2008 Sechelt Peninsula Theo 3 81 2009 Haslam Homathko 2 2009 Haslam Orford 19 2009 Sechelt Peninsula Brem 14 2009 Sechelt Peninsula Homathko 18 2009 Sechelt Peninsula Orford 1 2009 Sechelt Peninsula Toba 10 2010 Haslam Theo 3 2010 Sechelt Peninsula Powell-Daniels 7 2010 Sechelt Peninsula Toba 10 2011 Sechelt Peninsula Mamquam 8 2011 Sechelt Peninsula Rainy-Gray 5 2011 Sechelt Peninsula Southgate 20 2011 Sechelt Peninsula Theo 13 2012 Sechelt Peninsula Brem 10 2012 Sechelt Peninsula Rainy-Gray 8 2013 Sechelt Peninsula Chehalis 5 2013 Sechelt Peninsula Lower Lillooet 14 2014 Sechelt Peninsula Chehalis 24 2015 Sechelt Peninsula Chehalis 5 2015 Sechelt Peninsula Lower Lillooet 14 2015 Sechelt Peninsula Chehalis 10 2016 Rainy-Gray Chehalis 12 2016 Rainy-Gray Sechelt Peninsula 4 2016 Sechelt Peninsula Eldred 14 2017 Sechelt Peninsula Phillips 11 2017 Vancouver Island (Campbell R) Phillips 11 2017 Vancouver Island (Lower Salmon) Heydon 18 2017 Vancouver Island (Campbell R) Heydon 4 82 Appendix II: Summary of 356† samples from Roosevelt elk (C. c. roosevelti) in British Columbia; BC WID = wildlife identification number for compulsory inspection of resident, alien and indigenous hunter kills (6 digits) or wildlife health identification number for samples from elk captured for global positioning system collar deployment (7 digits), or Sample (ID) for faecal samples (3 digits); one sample (SP465) was successfully sequenced, but failed to amplify at 7+ microsatellite loci and was excluded from further analyses. Sample BC WID / Sample ID WHID / ID EPU Group Latitude dd Longitude dd Type Haplotype VIS120 127569 103A VIS 48.593907 -124.124541 T C VIS410 140604 103A VIS T VIS414 144174 103A VIS T VIS421 137991 103A VIS T VIS422 141443 103A VIS T VIS427 144182 103A VIS T VIS420 144159 103B VIS T C VIS429 144184 103B VIS T VIS121 127554 104A VIS 48.763174 -123.748715 T C VIS122 127587 104A VIS 48.709152 -123.624961 T VIS123 131907 104A VIS 48.767542 -123.718139 T VIS124 135905 104A VIS 48.767544 -123.718140 T VIS125 143882 104A VIS 48.767546 -123.718141 T VIS126 142430 104A VIS 48.768191 -123.811036 T VIS127 149892 104A VIS 48.767548 -123.718143 T VIS413 147259 104A VIS T VIS415 144172 104A VIS T VIS416 144230 104A VIS T VIS419 144162 104A VIS T VIS451 144101 104A VIS T VIS452 142465 104A VIS T VIS456 142449 104A VIS T VIS128 127411 104B VIS 49.049792 -123.956777 T VIS129 143915 104B VIS 48.904039 -124.212684 T VIS130 143919 104B VIS 48.914796 -124.302660 T VIS409 144238 104B VIS T VIS423 142471 104B VIS T VIS176 137396 104C VIS 48.907486 -124.572274 T D VIS177 139516 104C VIS 48.907475 -124.572284 T VIS426 144207 104C VIS T VIS428 144183 104C VIS T VIS449 144161 104C VIS T VIS181 127575 105C VIS 48.933008 -124.103835 T C VIS182 142439 105C VIS 48.940110 -124.166757 T VIS412 144235 105C VIS T C VIS417 144171 105C VIS T C VIS180 127585 105B VIS 49.064286 -124.296341 T C VIS455 144144 105B VIS T C VIS178 127584 105A VIS 49.031167 -124.367789 T VIS179 140843 105A VIS 49.109949 -124.437192 T D VIS457 144143 105A VIS T VIC186 140979 106B VIN 49.535297 -124.980065 T A VIC458 144204 106B VIN T A 83 VIC187 121597 106D VIN 49.579056 -125.212720 T VIC453 144121 106D VIN T VIC183 140969 106A VIN 49.879228 -125.470202 T VIC184 140974 106A VIN 49.936597 -125.294321 T VIC185 143880 106A VIN 49.838666 -125.347393 T A VIC408 144222 106A VIN T A VIC454 144092 106A VIN T VIN213 121588 110F VIN 50.054299 -125.639279 T VIN214 121594 110F VIN 49.978404 -125.630855 T VIN215 137389 110F VIN 49.971325 -125.601293 T VIN211 133683 110E VIN 50.097263 -125.379139 T VIN212 137056 110E VIN 50.090736 -125.403515 T VIN206 121582 110C VIN 50.234533 -125.724426 T VIN207 127413 110C VIN 50.157658 -125.719591 T B VIN208 121581 110D VIN 50.264731 -125.835128 T VIN209 127414 110D VIN 50.212338 -125.802853 T B VIN210 127421 110D VIN 50.282934 -125.899883 T A VIN218 127573 110H VIN 50.367497 -125.914568 T VIN219 140810 110H VIN 50.363295 -125.941925 T B VIN203 121589 110B VIN 50.060980 -126.157265 T B VIN204 121592 110B VIN 50.083334 -126.114549 T VIN205 140850 110B VIN 50.070755 -126.022860 T B VIN200 121587 110A VIN 50.362152 -126.035592 T VIN201 121591 110A VIN 50.293592 -126.080989 T VIN202 121600 110A VIN 50.247060 -126.049118 T B VIN216 127417 110G VIN 50.370232 -126.223748 T B VIN217 127571 110G VIN 50.349684 -126.329504 T B VIN220 127406 111A VIN 50.322964 -126.753457 T VIN221 137381 111A VIN 50.445098 -126.798917 T VIN222 121580 111B VIN 50.174741 -126.437482 T VIN223 127403 111B VIN 50.150557 -126.526540 T VIN224 127407 111B VIN 50.132278 -126.580187 T VIN225 127408 111B VIN 50.164133 -126.319784 T VIN226 127574 111B VIN 50.053837 -126.530734 T VIN227 127599 111B VIN 50.125509 -126.251962 T A VIN228 132025 111B VIN 50.174750 -126.437490 T VIN229 140973 112C VIN 49.807220 -126.359405 T A VIN230 121590 112D VIN 49.824137 -126.531014 T VIN231 127578 112D VIN 49.857999 -126.569837 T VIN232 127596 112D VIN 49.797168 -126.575105 T A VIN233 127422 112E VIN 49.699694 -126.205696 T VIC196 121578 109D VIN 49.943950 -126.236673 T VIC197 127409 109D VIN 49.960230 -126.255307 T VIC198 127416 109D VIN 49.867130 -126.206848 T VIC199 140799 109D VIN 49.765476 -126.028782 T A VIC188 121579 109A VIN 49.838596 -126.084910 T A VIC189 131928 109A VIN 50.005403 -126.145032 T VIC190 133682 109A VIN 49.997462 -126.068508 T VIC191 140849 109A VIN 49.840044 -126.098668 T VIC192 121598 109B VIN 49.597055 -125.953028 T VIC193 127420 109B VIN 49.715655 -126.035082 T VIC194 127410 109C VIN 49.860887 -126.043541 T VIC195 127418 109C VIN 49.860950 -126.043543 T SP031 031 205A SP 49.656194 -123.957273 F SP032 032 205A SP 49.656195 -123.957274 F SP033 033 205A SP 49.656196 -123.957275 F 84 SP034 034 205A SP 49.656197 -123.957276 F SP035 035 205A SP 49.656198 -123.957277 F SP036 036 205A SP 49.656199 -123.957278 F SP037 037 205A SP 49.656200 -123.957279 F SP038 038 205A SP 49.656201 -123.957280 F SP039 039 205A SP 49.656202 -123.957281 F SP040 040 205A SP 49.656203 -123.957282 F SP041 041 205A SP 49.656204 -123.957283 F SP042 042 205A SP 49.656205 -123.957284 F SP043 120689 205A SP 49.569727 -123.883227 T SP044 120680 205A SP 49.559306 -123.932889 T SP047 120688 205A SP 49.535257 -123.884633 T SP048 120679 205A SP 49.656206 -123.957285 T SP050 120691 205A SP 49.656207 -123.957286 T SP054 120690 205A SP 49.656208 -123.957287 T SP057 120701 205A SP 49.744814 -123.994676 T SP058 120702 205A SP 49.656209 -123.957288 T SP061 120706 205A SP 49.525291 -123.892384 T SP062 120707 205A SP 49.525350 -123.892401 T SP063 120708 205A SP 49.613849 -123.989777 T SP065 120710 205A SP 49.656210 -123.957289 T SP066 120711 205A SP 49.656211 -123.957290 T SP067 120712 205A SP 49.656213 -123.957292 T SP068 120713 205A SP 49.620565 -123.963517 T SP069 120714 205A SP 49.530660 -123.803154 T SP070 120715 205A SP 49.530665 -123.803180 T SP074 135953 205A SP 49.483573 -123.820467 T SP075 135955 205A SP 49.618141 -123.966165 T A SP076 135956 205A SP 49.656212 -123.957291 T A SP079 135960 205A SP 49.618143 -123.966167 T SP080 135961 205A SP 49.744820 -123.994686 T SP081 135962 205A SP 49.618145 -123.966169 T SP082 135963 205A SP 49.626565 -123.953517 T A SP085 135966 205A SP 49.525295 -123.892400 T A SP090 135971 205A SP 49.618149 -123.966173 T SP094 100851 205A SP 49.514189 -123.805327 T A SP095 101622 205A SP 49.525310 -123.892320 T A SP098 101625 205A SP 49.514195 -123.805330 T SP099 113911 205A SP 49.525313 -123.892323 T A SP111 120428 205A SP 49.618147 -123.966171 T A SP249 135945 205A SP 49.933975 -123.296253 T SP281 135994 205A SP 49.620785 -123.966293 T A SP283 135996 205A SP 49.526821 -123.903543 T A SP288 143754 205A SP 49.637941 -123.898290 T SP290 1813835 205A SP 49.759472 -123.963353 B SP465† 140610 205A SP T A JIE045 120686 205E CC 49.793618 -123.404113 T JIE071 120720 205E CC 49.793625 -123.404121 T JIE100 113913 205E CC 49.793618 -123.404113 T JIE108 113925 205E CC 49.793610 -123.404111 T JIE287 143752 205E CC 49.795274 -123.404427 T JIE010 1811031 205D CC 49.801059 -123.724938 B JIE078 135959 205D CC 49.708212 -123.771019 T JIE097 101624 205D CC 49.831717 -123.705270 T JIE102 113915 205D CC 49.708220 -123.771010 T JIE103 113917 205D CC 49.708225 -123.771015 T 85 JIE234 135993 205D CC 49.752015 -123.714669 T JIE279 1813509 205D CC 49.834638 -123.700405 B JIE002 1811023 205H CC 49.919399 -123.813037 B JIE003 1811024 205H CC 49.942349 -123.749369 B JIE060 120705 205H CC 49.934669 -123.805042 T JIE092 135973 205H CC 49.525289 -123.805000 T JIE433 147706 205H CC T JIE072 120721 205F CC 49.674948 -123.465045 T JIE091 135972 205F CC 49.592937 -123.410079 T 000093 no JIE093 WID recorded 205F CC 49.573578 -123.381571 T JIE112 120429 205F CC 49.573573 -123.381575 T JIE114 120431 205F CC 49.573583 -123.381577 T JIE115 120432 205F CC 49.573593 -123.381579 T JIE116 120433 205F CC 49.573603 -123.381561 T JIE235 99635 205F CC 49.572259 -123.381265 T JIE236 135903 205F CC 49.595792 -123.393722 T JIE237 135937 205F CC 49.601914 -123.413880 T JIE432 135917 205F CC T JIE096 101623 205G CC 50.065680 -123.750360 T JIE059 120703 205G CC 50.065675 -123.750378 T JIE101 113914 205G CC 50.065670 -123.750380 T JIE110 120427 205G CC 50.065680 -123.750372 T JIE238 137390 205G CC 50.107232 -123.705816 T JIE280 1813510 205G CC 50.096438 -123.723998 B JIE284 135997 205G CC 50.021158 -123.708499 T RG011 011 205I RG 49.549977 -123.594104 F RG012 012 205I RG 49.549978 -123.594105 F RG013 013 205I RG 49.549979 -123.594106 F RG014 014 205I RG 49.549980 -123.594107 F RG015 015 205I RG 49.549981 -123.594108 F RG016 016 205I RG 49.549982 -123.594109 F RG017 017 205I RG 49.549983 -123.594110 F RG018 018 205I RG 49.549984 -123.594111 F RG020 020 205I RG 49.533129 -123.498200 F RG021 021 205I RG 49.533130 -123.498201 F RG022 022 205I RG 49.533131 -123.498202 F RG023 023 205I RG 49.533132 -123.498203 F RG024 024 205I RG 49.533133 -123.498204 F RG025 025 205I RG 49.533134 -123.498205 F RG026 026 205I RG 49.548348 -123.496457 F RG027 027 205I RG 49.548349 -123.496458 F RG028 028 205I RG 49.548350 -123.496459 F RG029 029 205I RG 49.548351 -123.496460 F RG030 030 205I RG 49.548352 -123.496461 F RG077 135957 205I RG 49.548355 -123.496451 T RG084 135965 205I RG 49.548360 -123.496465 T RG107 113924 205I RG 49.533150 -123.498195 T RG117 117 205I RG 49.511031 -123.503915 F RG118 118 205I RG 49.511032 -123.503916 F RG119 119 205I RG 49.511033 -123.503917 F RG239 121593 205I RG 49.567830 -123.510077 T RG240 135944 205I RG 49.538420 -123.475611 T RG241 136418 205I RG 49.521037 -123.499105 T RG242 136451 205I RG 49.531113 -123.485156 T RG278 1813508 205I RG 49.511076 -123.504265 B 86 RG286 146928 205I RG 49.518362 -123.511083 T RG303 163 205I RG 49.443980 -123.503430 F RG304 164 205I RG 49.443990 -123.503440 F RG305 165 205I RG 49.444000 -123.503450 F RG320 163 205I RG 49.443985 -123.503451 F RG321 164 205I RG 49.444010 -123.503460 F RG322 165 205I RG 49.444015 -123.503462 F RG323 166 205I RG 49.444020 -123.503464 F RG324 167 205I RG 49.444025 -123.503466 F RG325 168 205I RG 49.444030 -123.503468 F RG326 169 205I RG 49.444035 -123.503470 F RG327 170 205I RG 49.444040 -123.503472 F RG328 171 205I RG 49.444045 -123.503474 F RG329 172 205I RG 49.444050 -123.503476 F RG330 173 205I RG 49.508335 -123.504050 F RG331 174 205I RG 49.508336 -123.504052 F RG332 175 205I RG 49.508337 -123.504054 F RG333 333 205I RG 49.508338 -123.504056 F RG334 334 205I RG 49.508339 -123.504058 F RG335 335 205I RG 49.508340 -123.504060 F RG336 336 205I RG 49.508341 -123.504062 F RG337 337 205I RG 49.508342 -123.504064 F RG338 338 205I RG 49.508343 -123.504066 F SQ001 1811022 206A SQ 50.016654 -123.350003 B SQ073 135951 206A SQ 49.843409 -123.223693 T A SQ244 135902 206A SQ 50.179613 -123.387268 T SQ245 135904 206A SQ 50.318761 -123.564271 T SQ246 135936 206A SQ 50.368372 -123.567841 T SQ247 135940 206A SQ 50.373019 -123.574470 T SQ248 135943 206A SQ 50.139770 -123.494396 T SQ250 135947 206A SQ 49.983129 -123.316231 T A SQ251 135949 206A SQ 50.183489 -123.373858 T SQ401 401 206A SQ 50.529136 -122.157748 F SQ405 1912074 206A SQ 50.416898 -122.907282 T A SQ430 146927 206A SQ T A SQ448 135907 206A SQ T A SQ460 140625 206A SQ T A SQ253 135946 208A SQ 49.509039 -122.900124 T SQ466 135923 208A SQ T FVN434 141478 208B PITT T A FVN437 146983 208B PITT T A FVN438 135922 208B PITT T FVN439 135918 208B PITT T FVN440 147255 208B PITT T A FVN441 143858 208B PITT T A FVN442 143618 208B PITT T FVN443 143619 208B PITT T FVN447 143666 208B PITT T A FVN462 140609 208B PITT T FVN463 140601 208B PITT T FVN467 143743 208B PITT T FVN252 140698 208B PITT 49.615312 -122.642281 T FVN254 143564 208B PITT 49.710629 -122.711023 T FVN291 1813836 208B PITT 49.550601 -122.610351 B FVN292 1813837 208B PITT 49.594265 -122.643159 B FVN295 141 208B PITT 49.550602 -122.610352 F 87 FVN296 142 208B PITT 49.550603 -122.610353 F FVN297 143 208B PITT 49.550604 -122.610354 F FVN298 144 208B PITT 49.550605 -122.610355 F FVN299 145 208B PITT 49.550606 -122.610356 F FVN300 147 208B PITT 49.594267 -122.643160 F FVN301 148 208B PITT 49.594500 -122.643100 F FVN306 149 208B PITT 49.594502 -122.643105 F FVN307 150 208B PITT 49.594505 -122.643106 F FVN308 151 208B PITT 49.594508 -122.643107 F FVN309 152 208B PITT 49.594511 -122.643108 F FVN310 153 208B PITT 49.594514 -122.643109 F FVN311 154 208B PITT 49.594517 -122.643110 F FVN009 1811030 208C STV 49.518087 -122.279600 B FVN255 140840 208C STV 49.489570 -122.201137 T FVN256 143562 208C STV 49.497676 -122.258562 T FVN257 143557 208C STV 49.540640 -122.172087 T FVN403 1911923 208C STV 49.087048 -122.167567 T A TI008 1811029 208C STV 49.419160 -121.851985 B TI302 161 208C STV 49.746017 -122.152777 F TI312 155 208C STV 49.746018 -122.152770 F TI313 156 208C STV 49.746020 -122.152780 F TI314 157 208C STV 49.746022 -122.152790 F TI315 158 208C STV 49.746024 -122.152800 F TI316 159 208C STV 49.746026 -122.152810 F A TI317 160 208C STV 49.746028 -122.152820 F TI318 161 208C STV 49.746030 -122.152830 F TI319 162 208C STV 49.746032 -122.152840 F FVN461 140897 208C STV T PRT004 1811025 212Q BRIT 50.300607 -124.042384 B PRT243 135984 212Q BRIT 50.226043 -123.963411 T PRT267 140828 212Q BRIT 50.260762 -124.033982 T PRT431 146926 212Q BRIT T PRT046 120681 212E BRIT 50.012996 -124.028309 T PRT264 135978 212E BRIT 50.023873 -124.042931 T PRT289 1813834 212E BRIT 50.005834 -124.020573 B PRT293 137 212E BRIT 50.023885 -124.042932 F PRT294 138 212E BRIT 50.023890 -124.042934 F PRT444 144219 212E BRIT T PRT051 120693 212S PRS 50.057557 -124.230664 T PRT055 120695 212S PRS 49.814152 -124.394257 T A PRT056 120700 212S PRS 49.951556 -124.397890 T PRT087 135968 212S PRS 49.994525 -124.411924 T PRT088 135969 212A PRS 49.994300 -124.420000 T PRT089 135970 212S PRS 49.994450 -124.430000 T PRT104 113919 212S PRS 49.994455 -124.432500 T PRT105 113920 212S PRS 49.804198 -124.457136 T A PRT106 113922 212S PRS 49.961139 -124.366110 T PRT109 120426 212S PRS 49.833414 -124.302408 T PRT113 120430 212S PRS 49.833410 -124.302418 T A PRT258 135991 212S PRS 49.829618 -124.286573 T PRT259 135992 212S PRS 49.856334 -124.295400 T A PRT260 135990 212S PRS 49.846546 -124.444193 T A PRT268 135986 212S PRS 49.908105 -124.500042 T A PRT269 135988 212S PRS 49.892162 -124.468615 T A PRT273 1813828 212S PRS 49.880079 -124.420377 B PRT049 120696 212D PRN 50.266832 -124.401644 T 88 PRT086 135967 212D PRN 50.266835 -124.401650 T PRT261 121584 212D PRN 50.015356 -124.339855 T PRT263 135989 212D PRN 50.323950 -124.439176 T PRT274 1813829 212D PRN 50.251394 -124.377361 B PRT275 1813830 212D PRN 50.337601 -124.427449 B PRT285 143751 212D PRN 50.335193 -124.426162 T PRT407 144120 212D PRN T PRT445 144226 212D PRN T PRT052 120698 212F PRN 50.093910 -124.609228 T PRT053 120699 212F PRN 50.093919 -124.609235 T PRT064 120709 212F PRN 50.093950 -124.609350 T PRT265 127565 212F PRN 50.094328 -124.627012 T PRT266 135987 212F PRN 50.099185 -124.613787 T PRT276 1813832 212F PRN 50.090680 -124.592228 B PRT406 124764 212F PRN T PRT005 1811026 213A TOBA 50.512057 -124.206327 B PRT006 1811027 213A TOBA 50.595810 -124.183924 B PRT339 339 213A TOBA 50.567324 -124.092812 F PRT340 340 213A TOBA 50.567327 -124.092814 F PRT341 341 213A TOBA 50.567330 -124.092816 F PRT342 342 213A TOBA 50.567333 -124.092818 F PRT344 344 213A TOBA 50.567339 -124.092822 F PRT345 345 213A TOBA 50.567342 -124.092824 F PRT346 346 213A TOBA 50.567345 -124.092826 F PRT348 348 213A TOBA 50.567351 -124.092830 F PRT349 349 213A TOBA 50.567354 -124.092832 F HO270 127415 214A HO 50.668139 -124.771313 T HO271 127419 214A HO 50.650106 -124.798452 T HO277 1813833 214A HO 50.392028 -124.862724 B HO436 144225 214B HO T HO272 121595 215A HO 50.946868 -124.912139 T HO424 144123 215A HO 50.946875 -124.912140 T HO425 144124 215A HO T HO007 1811028 215A HO B 89 Appendix III: Mitochondrial and microsatellite primers used in this study; all microsatellite forward primers contained an M13 primer sequence (CAC GAC GTT GTA AAA CGA C) added to allow incorporation of a fluorescent M13 tag during polymerase chain reaction amplification; allele size (in bp), number of alleles observed (#A), expected heterozygosity (He), observed 1 2 o heterozygosity (Ho), accession number in GenBank (if known), PCR temperature = 8 cycles_25 cycles at temp _temp C, and magnesium chloride in millimolar. # PCR Primer ID sequence 5' to 3' source size A He Ho Rpt Acces. # Temp MgCl2 EK-F23 TAC CAA TCA CCA GCA CAAT CG Speller et 54 EK-R663 CGG GTT GCT GGT TTC ACG al 2014 BL42 F CAA GGT CAA GTC CAA ATG CC Bishop et 269-279 3 0.620 0.628 Di G18455 55_57 2.5mM BL42 R GCA TTT TTG TGT TAA TTT CAT GC al 1994 BM203 F GGG TGT GAC ATT TTG TTC CC Bishop et 246-250 3 0.502 0.448 Di G18500 52_54 2.0mM BM203 R CTG CTC GCC ACT AGT CCT TC al 1994 BM4107 F AGC CCC TGC TAT TGT GTG AG Bishop et 4 NA Di G18519 55_57 2.0mM BM4107 R ATA GGC TTT GCA TTG TTC AGG al 1994 BM4107RD F GCATTGTTCAGGGTTCTCTA 176-190 4 0.557 0.527 Di 55_57 2.0mM BM4107RD R GCTATTGTGTGAGGCAATTC Redesign BM4513 F GCG CAA GTT TCC TCA TGC Bishop et 4 NA Di G18507 55_57 2.0mM BM4513 R TCA GCA ATT CAG TAC ATC ACC C al 1994 BM4513RD F CTCATGCACTTTTCCTTCTG 143-149 4 0.531 0.472 Di 55_57 2.0mM BM4513RD R GCTTATTCAAGTGGTGTAGGC Redesign BMC1009 F GCA CCA GCA GAG AGG ACA TT Bishop et 302-306 3 0.468 0.413 Di 48_50 2.0mM BMC1009 R ACC GGC TAT TGT CCA TCT TG al 1994 BM888 F AGG CCA TAT AGG AGG CAA GCT T Bishop et 199-209 3 0.367 0.318 Di G18484 55_57 2.0mM BM888 R CTC GGT GAG CTC AAA ACG AG al 1994 BM6506 F GCA CGT GGT AAA GAG ATG GC Bishop et 226-234 2 0.483 0.437 Di G18455 60_62 2.0mM BM6506 R AGC AAC TTG AGC ATG GCA C al 1994 CSSM041 F AAT TTC AAA GAA CCG TTA CAC AGC Moore et 144-148 2 0.47 0.449 Tet U03816 52_54 2.0mM CSSM041 R AAG GGA CTT GCA GGG ACT AAA ACA al 1994 INRA107 F TCC CAG ATA CAG ATG CAA CAG Vaiman et 177-193 5 0.679 0.682 Di X71577 55_57 2.0mM INRA107 R GGA GAG CCG AGG GCT TCA G al 1994 90 Rt7 F CCT GTT CTA CTC TTC TTC TC Wilson et 234-235 2 0.465 0.317 Di U90740 55_57 2.0mM Rt7 R ACT TTT CAC GGG CAC TGG TT al 1997 211/205 AML-1 F CAG CCA AAC CTC CCT CTG C Pajeres et or 255 2 0.5 NA Sex X/Y Failed AML-1 R CCC GCT TGG TCT TGT CTG TTG C al 2007 BM415 F GCT ACA GCC CTT CTG GTT TG Bishop et X NA Failed BM415 R GAG CTA ATC ACC AAC AGC AAG al 1994 BM848 F TGG TTG GAA GGA AAA CTT GG Bishop et X NA Failed BM848 R CCT CTG CTC CTC AAG ACA C al 1994 BM1225 F TTT CTC AAC AGA GGT GTC CAC Bishop et X NA Failed BM1225 R ACC CCT ATC ACC ATG CTC TG al 1994 Rt1 F TGC CTT CTT TCA TCC AAC Wilson et X NA Di U90737 Failed Rt1 R CAT CTT CCC ATC CTC TTT AC al 1997 Wilson et RT5 F CAG CAT AAT TCT GAC AAG TG al 1997 Failed RT5 R AAT TCC ATG AAC AGA GGA G Rt13 F GCC CAG TGT TAG GAA AGA AG Wilson et X NA Di U90743 55_57 Rt13 R CAT CCC AGA ACA GGA GTG AG al 1997 OvirH F AAG TCT ACA ATC CAT GGG CTT GC DeWoody 133-143 2 NA Failed OvirH R GTT CTT TAC CAC CTG CAC CA et al 1995 OarFCB193 F TTCATCTCAGACTGGGATTCAGAAAGGC Buchanan 114 NA Di LO1533 58_60 Crawford OarFCB193 R GCTTGGAAATAACCCTCCTGCATCCC 1993 C32 F CATCACCTCCACTAGCTTTG Meredith NA Tet Failed C32 R ATCTGAGCCACTAGGGAAAC et al 2005 C36 F TATGGTGGAGATGTAGGTG Meredith NA Tet Failed C36 R CCATTATGTGTAACCCTCCA et al 2005 ELK_T115 F TGGTTATCTGGGTCATGAAG Meredith NA Tet Failed ElK_T115 R TTGCTATTGAGCCATAGG et al 2005 91 Appendix IV: Deviations from Hardy-Weinberg Equilibrium for 13 populations of Roosevelt elk (C. c. roosevelti) at 10 microsatellite loci, AMOVA with 10000 permutations to estimate significance adjusted for FDR (P=0.015). Significant values in bold. BM4107 BM4513 Pop’n BL42 BM203 RD RD BM6506 BM888 BMC1009 CSSM041 Inra107 Rt7 Multi-locus VIS 0.190 0.141 -0.040 0.192 0.264 0.149 0.032 -0.115 0.071 0.250 *0.111 VIN -0.142 -0.008 0.029 0.094 -0.095 0.294 0.114 -0.174 -0.189 -0.032 -0.027 SP 0.133 -0.035 0.009 -0.073 0.020 -0.025 0.276 -0.177 0.103 -0.04 0.024 CC -0.192 0.120 0.102 0.112 0.200 -0.142 -0.091 0.005 -0.040 0.202 0.021 RG 0.172 0.200 0.124 0.029 0.138 *0.395 0.196 -0.224 -0.009 0.324 *0.121 SQ -0.389 -0.161 0.178 -0.091 -0.037 0.038 0.236 0.400 -0.115 0.200 -0.003 PITT -0.107 -0.096 0.147 -0.222 0.164 0.293 0.066 0.016 -0.033 -0.174 0.004 STV -0.217 0.346 -0.045 -0.011 0.101 0.200 -0.077 -0.077 -0.021 *0.860 0.093 BRIT 0.027 0.153 -0.019 0.385 -0.370 -0.059 -0.301 0.043 -0.086 0.640 0.027 PRS 0.083 -0.435 -0.129 0.255 0.050 0.458 -0.091 0.125 -0.258 0.050 -0.028 PRN -0.007 -0.206 0.011 0.336 0.084 0.118 0.352 -0.132 -0.250 0.063 0.035 TOBA 0.196 0.137 0.125 0.161 0.226 0.302 0.429 0.386 0.200 0.310 *0.231 HO -0.135 0.276 -0.296 0.282 0.548 -0.273 0.164 0.517 0.300 0.774 *0.227 Overall -0.020 0.040 0.018 0.033 0.058 *0.149 0.090 -0.070 -0.062 *0.128 0.028 92 Appendix V: Linkage disequilibrium for 10 microsatellite loci in 13 populations of Roosevelt elk (C. c. roosevelti) in British Columbia, significance adjusted for FDR (P=0.002), * a = 0.05, *** a = 0.001. Locus pair Chi2 df P-Value -------------------- -------- --- -------- BL42 & BM203 29.814155 26 0.275382 BL42 & BM4107RD 21.586875 26 0.711074 BM203 & BM4107RD 21.805299 26 0.699301 BL42 & BM4513RD 28.455439 26 0.336419 BM203 & BM4513RD 27.194096 26 0.399211 BM4107RD & BM4513RD 20.984569 26 0.74276 BL42 & BM6506 26.958661 26 0.411504 BM203 & BM6506 33.842002 26 0.139049 BM4107RD & BM6506 26.73083 26 0.423552 BM4513RD & BM6506 34.011405 26 0.134737 BL42 & BM888 17.5218 26 0.892473 BM203 & BM888 19.74874 26 0.803312 BM4107RD & BM888 23.199969 26 0.621611 BM4513RD & BM888 33.185286 26 0.15679 BM6506 & BM888 25.356648 26 0.498866 BL42 & BMC1009 27.094049 26 0.404414 BM203 & BMC1009 29.85085 26 0.273836 BM4107RD & BMC1009 22.38148 26 0.667655 BM4513RD & BMC1009 34.650226 26 0.119422 BM6506 & BMC1009 34.483764 26 0.123272 BM888 & BMC1009 23.064513 26 0.629295 BL42 & CSSM041 27.860356 26 0.365364 BM203 & CSSM041 32.333618 26 0.18232 BM4107RD & CSSM041 16.454449 26 0.924641 BM4513RD & CSSM041 21.796943 26 0.699754 BM6506 & CSSM041 37.17449 26 0.072091 BM888 & CSSM041 45.274708 26 0.01098 BMC1009 & CSSM041 25.368371 26 0.498208 BL42 & Inra107 27.573406 26 0.379764 BM203 & Inra107 15.80768 26 0.94072 BM4107RD & Inra107 53.492337 26 *0.001176 BM4513RD & Inra107 20.941219 26 0.744992 BM6506 & Inra107 36.703372 26 0.079472 BM888 & Inra107 >116.7698 26 ***<1.81e-13 BMC1009 & Inra107 27.610948 26 0.377864 CSSM041 & Inra107 35.995345 26 0.091755 BL42 & Rt7 18.788961 26 0.845215 BM203 & Rt7 15.746589 26 0.942109 BM4107RD & Rt7 12.069678 26 0.990773 BM4513RD & Rt7 33.899137 26 0.137583 BM6506 & Rt7 23.185709 26 0.622421 BM888 & Rt7 25.987624 26 0.463785 BMC1009 & Rt7 17.996824 26 0.875889 CSSM041 & Rt7 25.574413 26 0.486679 Inra107 & Rt7 35.685261 26 0.097611 93 Appendix VI: Allele frequency for 13 loci in 10 populations of Roosevelt elk (C. c. roosevelti) in British Columbia; primary source population VIN and VIS, secondary source populations SP and PRS; shaded cells indicate missing alleles or frequency reduced by ~40% or more from VIN. Locus Allele VIS VIN SP CC RG SQ PITT STV BRIT PRS PRN TOBA HO BL42 n 41 55 48 35 52 16 28 16 10 17 16 10 8 269 0.463 0.509 0.271 0.329 0.433 0.375 0.268 0.344 0.300 0.500 0.594 0.450 0.500 277 0.037 0.145 0.125 0.143 0.087 0.344 0.214 0.125 0.150 0.206 0.094 0.100 0.250 279 0.500 0.345 0.604 0.529 0.481 0.281 0.518 0.531 0.550 0.294 0.313 0.450 0.250 BM203 n 41 55 48 35 51 16 29 16 10 16 16 11 8 246 0.171 0.273 0.219 0.171 0.382 0.156 0.224 0.313 0.250 0.313 0.156 0.500 0.188 248 0.341 0.691 0.698 0.786 0.608 0.813 0.741 0.656 0.700 0.656 0.750 0.364 0.688 250 0.488 0.036 0.083 0.043 0.010 0.031 0.034 0.031 0.050 0.031 0.094 0.136 0.125 BM4107RD n 41 55 47 35 52 16 28 15 10 17 15 8 8 176 0.524 0.391 0.638 0.629 0.692 0.719 0.536 0.667 0.600 0.824 0.400 0.625 0.688 186 0.463 0.255 0.181 0.229 0.250 0.188 0.214 0.200 0.200 0.118 0.233 0.250 0.250 188 0.000 0.091 0.011 0.000 0.010 0.094 0.000 0.033 0.000 0.000 0.000 0.000 0.063 190 0.012 0.264 0.170 0.143 0.048 0.000 0.250 0.100 0.200 0.059 0.367 0.125 0.000 BM4513RD n 41 55 48 35 49 15 28 16 10 16 16 11 8 143 0.110 0.100 0.240 0.257 0.265 0.233 0.214 0.281 0.350 0.094 0.313 0.227 0.125 145 0.012 0.173 0.115 0.143 0.143 0.100 0.143 0.188 0.150 0.094 0.094 0.182 0.063 147 0.012 0.018 0.021 0.014 0.061 0.033 0.054 0.031 0.000 0.000 0.000 0.045 0.000 149 0.866 0.709 0.625 0.586 0.531 0.633 0.589 0.500 0.500 0.813 0.594 0.545 0.813 BM6506 n 41 53 47 33 51 15 28 16 10 17 15 9 8 226 0.573 0.604 0.617 0.591 0.618 0.533 0.589 0.500 0.550 0.765 0.567 0.722 0.500 234 0.427 0.396 0.383 0.409 0.382 0.467 0.411 0.500 0.450 0.235 0.433 0.278 0.500 BM888 n 41 55 47 35 50 16 29 16 10 17 16 11 8 199 0.000 0.000 0.000 0.000 0.010 0.031 0.017 0.000 0.000 0.000 0.000 0.045 0.000 207 0.695 0.791 0.734 0.700 0.770 0.813 0.707 0.625 0.900 0.882 0.781 0.773 0.750 209 0.305 0.209 0.266 0.300 0.220 0.156 0.276 0.375 0.100 0.118 0.219 0.182 0.250 94 BMC1009 n 41 54 48 35 52 16 25 16 10 17 15 11 8 302 0.012 0.019 0.000 0.014 0.000 0.031 0.020 0.000 0.050 0.000 0.000 0.000 0.063 304 0.634 0.472 0.594 0.686 0.721 0.813 0.760 0.625 0.700 0.618 0.567 0.818 0.438 306 0.354 0.509 0.406 0.300 0.279 0.156 0.220 0.375 0.250 0.382 0.433 0.182 0.500 CSSM041 n 41 55 48 35 51 16 29 16 10 16 16 10 8 144 0.610 0.564 0.615 0.757 0.608 0.469 0.672 0.719 0.550 0.594 0.688 0.650 0.625 148 0.390 0.436 0.385 0.243 0.392 0.531 0.328 0.281 0.450 0.406 0.313 0.350 0.375 Inra107 n 41 54 47 35 48 16 29 16 10 16 16 11 8 177 0.683 0.306 0.213 0.286 0.354 0.375 0.259 0.281 0.350 0.469 0.281 0.273 0.375 179 0.012 0.009 0.021 0.071 0.000 0.063 0.103 0.156 0.000 0.031 0.000 0.000 0.000 185 0.000 0.148 0.287 0.200 0.167 0.094 0.224 0.156 0.100 0.063 0.188 0.273 0.375 191 0.305 0.463 0.436 0.443 0.438 0.438 0.414 0.344 0.500 0.375 0.500 0.455 0.250 193 0.000 0.074 0.043 0.000 0.042 0.031 0.000 0.063 0.050 0.063 0.031 0.000 0.000 Rt7 n 40 50 46 33 48 15 27 15 10 17 16 11 8 243 0.563 0.520 0.457 0.288 0.417 0.200 0.167 0.367 0.150 0.235 0.344 0.455 0.563 245 0.438 0.480 0.543 0.712 0.583 0.800 0.833 0.633 0.850 0.765 0.656 0.545 0.438 95