UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Population genetics and habitat selection behaviour of Vancouver Island white-tailed ptarmigan Fedy, Bradley Craig 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-267110.pdf [ 7.03MB ]
Metadata
JSON: 831-1.0074963.json
JSON-LD: 831-1.0074963-ld.json
RDF/XML (Pretty): 831-1.0074963-rdf.xml
RDF/JSON: 831-1.0074963-rdf.json
Turtle: 831-1.0074963-turtle.txt
N-Triples: 831-1.0074963-rdf-ntriples.txt
Original Record: 831-1.0074963-source.json
Full Text
831-1.0074963-fulltext.txt
Citation
831-1.0074963.ris

Full Text

POPULATION GENETICS AND HABITAT SELECTION BEHAVIOUR OF VANCOUVER ISLAND WHITE-TAILED PTARMIGAN (Lagopus leucura saxatilis) by Bradley Craig Fedy B.E.S., University of Waterloo, 1999 M.Sc, York University, 2002 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA December 2006 © Bradley Craig Fedy, 2006 ABSTRACT I examined the habitat selection behaviour arid compared regional variation in population performance of a threatened alpine grouse subspecies, the Vancouver Island white-tailed ptarmigan (Lagopus leucura saxatilis). I also examined the genetic population structure and levels of gene flow among 7 populations of white-tailed ptarmigan on Vancouver Island, British Columbia. Population performance was compared between 2 distinct mountain areas of Vancouver Island. The southern region was a more fragmented landscape with smaller isolated alpine patches compared to the more continuous central region of the island. The habitat selection modelling revealed that, unlike other white-tailed ptarmigan subspecies, Vancouver Island white-tailed ptarmigan use a generalist strategy for habitat selection. Individuals preferred areas with a combination of predator cover, food availability and high moisture. The analysis of population performance showed that populations in the central region of the island outperform southern populations. This difference could result from moisture levels which are lower in the southern portion of the island, but likely is not influenced by food abundance during the brood rearing period. I presented data on microsatellite primer optimization and described generalized heterozygote deficiencies and high levels of diversity in Vancouver Island populations. All 7 populations demonstrated high levels of diversity (mean H E = 0.78) combined with high Fis values (0.22) and significant heterozygote deficiencies. The apparent paradox of high diversity combined with high Fis and generalized heterozygote deficiencies are best explained by two scenarios. First, sampling may have captured a snapshot of a group of populations progressing towards severe isolation. This scenario suggests that significant geographic isolation between populations has existed long enough for the increase of inbreeding but not long enough for drift to result in strong population differentiation and a corresponding decrease in diversity. The second scenario evokes a pattern of infrequent dispersal between populations sufficient to maintain high levels of diversity, combined with low densities and" limited mate choice resulting in the relatively quick accumulation of homozygosity levels within populations. I addressed patterns of dispersal using genetic data and direct measures by following the inter-seasonal movements of radio-collared birds. The results showed low, but significant, genetic differentiation between most populations and direct and genetic estimates of dispersal suggested limited gene flow among populations. Analysis of molecular data also demonstrated a generally consistent pattern of isolation-by-distance. However, large areas of unsuitable low elevation habitat might act as barriers to gene flow. The levels of isolation and lack of gene flow imply serious conservation concern for the most southern population. ii T A B L E O F C O N T E N T S A B S T R A C T i i T A B L E O F C O N T E N T S in LIST OF T A B L E S vi L I S T OF F I G U R E S ix A C K N O W L E D G E M E N T S xi C H A P T E R 1: General introduction 1 Resource selection 2 Metapopulation dynamics 3 Study species - white-tailed ptarmigan (Lagopus leucura) 5 Study area - Vancouver Island, British Columbia 6 Study objectives 8 Overview of thesis '. . 8 R E F E R E N C E S 11 C H A P T E R 2: Habitat selection and regional variation in population performance o f white-tailed ptarmigan on Vancouver Island, British Columbia INTRODUCTION 16 Food availability 18 Predator avoidance 19 Thermal regulation , 19 Generalists • 20 Regional variation in reproductive success, age structure and habitat attributes 20 M E T H O D S 21 Study species and study area .' 21 Data collection 22 Fine-scale habitat variables 22 Models 23 Model validation and selection 24 Female reproductive success 26 Age structure and sex ratios 27 Regional differences in key habitat variables 28 R E S U L T S 28 Fine-scale habitat selection 28 Female reproductive success 29 i i i Age structure and sex ratios 29 Regional differences in key habitat variables 30 Predator avoidance • 30 DISCUSSION 31 Habitat selection and key variables 31 Regional variation in population performance 32 Fragmentation and metapopulations 33 Regional variation of key habitat variables , 35 R E F E R E N C E S 47 C H A P T E R 3: Heterozygote deficiencies and high levels of genetic diversity in Vancouver Island white-tailed ptarmigan INTRODUCTION 52 M E T H O D S 55 Study species and study area 55 Primer optimization 55 Population analyses 56 R E S U L T S 58 Primer optimization 58 Within population analyses , 59 Interspecific comparisons ..; 59 Nonrandom association of alleles 60 Private alleles 61 Bottleneck 61 Population structure 62 DISCUSSION 62 Nonrandom association of alleles 63 Bottleneck 65 Heterozygote deficiencies 65 R E F E R E N C E S 78 C H A P T E R 4: Gene flow, patterns of dispersal and metapopulation dynamics of alpine Vancouver Island white-tailed ptarmigan inferred from microsatellite DNA markers. INTRODUCTION 84 M E T H O D S 86 Study species and study area 86 Temporal variation 86 Inter-population analyses 87 Direct observations 88 RESULTS 89 Temporal variation 89 Inter-population analyses 90 Direct movement analysis 91 DISCUSSION 91 Temporal variation and sample populations : 92 Metapopulations • 92 Migration 94 Barriers to gene flow 95 REFERENCES 106 CHAPTER 5 Conclusions 110 Future research directions 113 REFERENCES . 115 v L I S T O F T A B L E S T a b l e 2.1. Habitat variables collected at each sampling plot. F - Food Availability; P r -Predator Avoidance; Th - Thermal Regulation; Gen - Generalist; Gl - Global., p.38 2 T a b l e 2.2. Fine-scale habitat selection sampling effort. Each session is an independent, approximately 3 hr observation period, of a unique individual. Each plot is a 20 m sampling plot p.39 T a b l e 2.3. Coefficient estimates and standard errors for all iterations. Numbers in bold are significant at 10% p.40 T a b l e 2.4. Cross-validated Spearman-rank correlations (rs) between bin ranks and frequencies for individual and average model sets p.41 T a b l e 2.5. Age structure of all ptarmigan in south and central portions of the island throughout the inventory (1995 - 1999) and intensive (2003 - 2004) stages of the study p.42 T a b l e 3.1. Sampling locations for white-tailed ptarmigan DNA samples across Vancouver Island, British Columbia. The number of individuals sampled (N) and the range of years over which birds were sampled for each mountain. Co-ordinates are based on the British Columbia Albers projection p.68 T a b l e 3.2. Primers tested on Lagopus leucura that did not show sufficient variation ( > 4 unique alleles). Primers were first tested on 2% agarose gels. Those primers that amplified product were subsequently tested on 6% polyacrylamide gels. Listed for each primer are the annealing temperature. (Ta), the expected allele size reported in the original papers, whether the primer amplified product that could be visualized on agarose or acrylamide gels and the associated sample sizes of individuals (n). "-" represents primers not tested on acrylamide gels. A = number of alleles p.69 T a b l e 3.3. Gene diversity and amplification results at 10 loci. Fis calculated as Weir and Cockerham (1984). All loci showed significant deviations from Hardy-Weinberg equilibrium (HWE probability test). Hd presents results for tests of heterozygote deficiency (Rousset and Raymond 1995). T a , optimal annealing temperature; A, number of observed alleles; n, number of individuals genotyped at each locus; Ho, mean observed heterozygosity; H E - expected heterozygosity as calculated in ARLEQUIN. rc and ry are estimates of the frequency of null alleles as described by Chakraborty et al. (1992) and Brookfield (1996), respectively. rc = (H E - H 0 ) / (H E + Ho) and r^ = (H E - Ho)/(l + H E). na = no GeneBank accession number provided in the original publication. Brackets beside Wright's fixation indices are p-values.p.70 T a b l e 3.4. Results of matching the sequences of Vancouver Island white-tailed ptarmigan amplification products to GenBank sequences as determined by the ESEE3S software program p.71 T a b l e 3.5. Estimates of within locality genetic diversity for each population, n, number of individuals sampled; A, mean number of alleles; A n , mean allelic richness; Ho, mean observed heterozygosity; H E , mean expected heterozygosity. All mean values vi presented ± Standard error. rc and rb are estimates of the frequency of null alleles as described by Chakraborty et al. (1992) and Brookfield (1996), respectively. rc = (H E - H 0 ) / ( H E + H 0 ) and rb = ( H E - H0)/(l + H E ) . FIS , 95% confidence interval presented in parentheses (Weir and Cockerham 1984). Ris, allele size-based correlation as calculated in genepop. Hd, p-value for test of heterozygote deficiency p.72 Table 3.6. Summary of previous studies using microsatellite DNA markers to investigate the population genetics of grouse (Aves: Tetraoninae). Data presented are averaged over population or loci from the corresponding literature cited. Loci, gives the number of microsatellite loci used in the study; A, number of alleles; A n , allelic ' richness; H 0 , observed heterozygosity; H E , expected heterozygosity; Fis, Wright's inbreeding coefficient; HWE, Hardy-Weinberg equilibrium, n.s. indicates there were no significant departures from Hardy-Weinberg, in other instances the number of significant deviations is reported. There were no heterozygote deficiencies (Hd) reported in any grouse studies except in this study of Vancouver Island white-tailed ptarmigan. If no corresponding measure was found in the original publications na was entered into the cell p.73 Table 3.7. Distribution of private alleles found only in single localities of ptarmigan. T is the total number of alleles of different sizes found in a given population, U is the number of alleles unique to that population with the value in parentheses equal to the number of these alleles that occur at a frequency of >5%, and the percentage is (U/T)*100. p.74 Table 3.8. Significant (p < 0.05; prior to Bonferroni correction) excesses or deficiencies in heterozygosity under each of three models of mutation in each locality and in Lagopus leucura saxatilis, determined via bottleneck (Cornuet and Luikart 1996) simulations. IAM - infinite alleles model, S M M - stepwise mutation model, T P M -two-phase mutation model p.75 Table 3.9. Proportion of membership of each predefined sample population of Vancouver Island white-tailed ptarmigan in each cluster inferred from the Bayesian method of Pritchard et al. (2000) p.76 Table 4.1. Comparison of allelic richness (An) and diversity (HE) in two Vancouver Island white-tailed ptarmigan populations. Values presented are means and values in brackets are the ranges. P-values were determined by Wilcoxon paired-samples tests p.97 Table 4.2. Analysis of Molecular Variance among 7 Vancouver Island white-tailed ptarmigan populations ; p.98 Table 4.3. Pairwise population comparisons of FST and R S T values.^FsT values are presented on the lower diagonal and R S T values above the diagonal. Non-significant values are in bold. Underlined values are significant at the 5% level and all other values at the 1% level. Negative values are due to sampling error and indicate gene flow without limitation.... p.99 vii Table 4.4. Means and 95% confidence intervals of the posterior distributions for migration rates between Vancouver Island white-tailed ptarmigan populations. Values along the diagonal are the proportions of individuals derived from the source population in each generation and are underlined. All standard deviations were < 0.05. Migration rates to other populations > 0.10 are in bold. Migration rates < 0.10 but still with a positive value for the lower 95% confidence interval are in italics p. 100 Table 4 .5. Direct movement and dispersal distances for Vancouver Island white-tailed ptarmigan determined from aerial and ground radio-telemetry data. The diagonal represents movements within a population. The first line presents the number of movements within the population/the total number of movements observed. The number of individual birds contributing to the movement analysis are then presented in brackets. The second line on the diagonal presents the mean distance moved (km) and the range of distances in brackets. The lower diagonal presents the geographic distance between the populations (km). Vancouver Island white-tailed ptarmigan were never observed moving beyond their population boundaries p. 101 viii LIST OF FIGURES Figure 1.1. Vancouver Island, British Columbia, Canada. Dark grey areas represent Alpine Tundra, and lighter grey represents Mountain Hemlock. Habitat classifications are based on British Columbia's Biogeoclimatic Ecosystem Zones. Population boundaries are circled with a dotted line and labelled S - South, SW - South West, B - Beaufort, CS - Central South, C E - Central East, CW - Central West, N - North. See Chapter 3, Table 3.1 for names of mountains included in each population... p. 10 Figure 2.1. a) Frequency of categories of average scores for 'available' data for all models. Frequency values for individual models (n = 5) are depicted with unique symbols, b ) Standard error of categories of average scores for 'available' data for all models. Standard errors for individual models (n = 5) are depicted with unique symbols.p.43 Figure 2.2. Age category comparison for south and central regions. Black bars = adults > 2 years. White bars = yearlings. Comparisons are presented including data on both sexes, females only and males only. The numbers inside the bars represent the respective sample sizes. Data are from both stages of the study : p.44 Figure 2.3. Comparison of sex ratios for south and central regions. Black bars = females of breeding age. White bars = males of breeding age. The numbers inside the bars represent the respective sample sizes. Data are from both stages of the study.... p.45 Figure 2.4. South and central region comparisons of key habitat variables. Values presented are median ± standard error. White bars = south, Black bars = central p.46 Figure 3.1. Neighbour-joining consensus dendrogram showing clustering based on the genetic distance of Nei (1978). The one major break (72%) does not include the CS population which is geographically proximate to the other Central populations. Numbers on branches are percentages of 1000 bootstrap replicates p.77 Figure 4 .1 . Neighbour-joining consensus dendrogram showing population clustering for Vancouver Island white-tailed ptarmigan based on the genetic distance of Nei (1978). The one partition (90%) groups both sampling periods (inventory and intensive) in the South together and those in the Central East together. The number on the branch is percentage of 1000 bootstrap replicates p. 102 Figure 4.2. Isolation by distance pattern for Vancouver Island white-tailed ptarmigan. Regression of genetic differentiation [estimated by FST/(1-FST)] against the logarithm of geographical distances (km) for all pairs of populations. Circled points refer to data causing deviations from a pattern of migration-drift equilibrium and demonstrate restricted gene flow relative to geographic distance p. 103 Figure 4.3. Principal components analysis on multilocus genotypes of Vancouver Island white-tailed ptarmigan. The first three principal components (PC) explained 72% of the total variance p. 104 Figure 4.4. Schematic depicting the results of pairwise F S T comparisons and the between population migration rates (m) inferred by ByesAss for Vancouver Island white-tailed ptarmigan p. 105 x ACKNOWLEDGEMENTS I would first like to thank my advisor Kathy Martin. Her contributions to my thesis, particularly during the writing stage, helped me produce a thesis of which I am proud. Her boundless passion for grouse and alpine areas was always inspiring. Carol Ritland gave me the gifts of infinite patience in the lab and constant encouragement throughout my Ph.D. Her unwavering confidence in my capabilities provided motivation during the inevitable lows that accompany any Ph.D. research. I am grateful to Peter Arcese for his enthusiasm in the final product and for challenging me to find new ways of approaching data. Brian Klinkenberg ensured I always considered the importance of spatial variation and insisted on careful examination of the appropriateness of my methods. Val Lemay nutured my growth in biometrics and I thank her for many hours spent discussing data analyses and her assistance in producing strong quantitative results. Mark Drever thoroughly read every data chapter and provided useful reviews and insights that strengthened the communication of my findings. My research involved climbing many mountains on Vancouver Island, and continuing through all types of weather and real and perceived adversity. Thus the mountains provided a perfect metaphor for working through this thesis and the many, many hours spent hiking through mountains searching for birds provided me with time for reflections on my purpose and my passion. I thank my wife Elissa for everything. Her patience, compassion and kindness throughout the entire process inspired me to climb this mountain and made it possible to do it all with a smile on my face (most of the time). I had one field assistant throughout my entire Ph.D. Mark Wong spent two grueling field seasons following me through mountains, hail, rain and bugs and despite it all, is still a close friend. I admired his commitment to the project and I thank him for our good conversations and for his quiet adaptability when things did not go as planned. An extra 6 hours of hiking up and down mountains or even a torched field vehicle never fazed him. He was an excellent companion and assistant. I think the completion of my Ph.D. represents the end of a long road and a long climb as much for my family as for myself. My graduate career has taken me from tropical low land jungles to high elevation alpine, and although it was difficult for them to follow me, they have always done their best to understand my choices. I cannot possibly thank them enough for their encouragement and for always providing a quiet place of respite from the focus and fury of my research. A NSERC Canadian Graduate Scholarship, and a UBC University Graduate Fellowship supported me throughout my thesis. Field and lab work were supported by NSERC vehicle and research grants to Kathy Martin. Research was also supported by the Society of Canadian Ornithologists' Bailie Award, the Friends of Ecological Reserves and a professional equipment sponsorship with Arc'teryx. xn CHAPTER 1: General introduction One of the primary goals of ecology is to understand the processes that influence the patterns of distribution and abundance in animals (Krebs 1994). These patterns are primarily determined by the spatial and temporal variability of resources which structure vertebrate populations genetically and ecologically. This spatial and temporal variation can be examined at both coarse and fine scale levels. The interactions of populations across a landscape directly influence the coarse scale distribution and abundance of a species. Finer-scale environmental features of the habitat, such as resource availability and distribution, can also affect the reproductive success and survival of individuals. Therefore, animal habitat use should be considered on at least two different scales: coarse scale, landscape-level population interactions and fine scale individual use of specific habitat features. These two levels of investigation are intricately linked because the quality of a local habitat can influence the dynamics of an entire population through its effect on population parameters (Fleishman et al. 2002, Walker et al. 2003). Metapopulation dynamics is the study of interactions between local populations within a larger area, where migration from one local population to another is possible (Hanski and Gilpin 1991). The metapopulation approach is important because it considers the role of spatial structure in forming ecological patterns (Hanski and Simberloff 1997). Most metapopulation studies report that patch area and the degree of isolation between patches explains a sufficient amount of variation in the distribution of individuals to enable prediction of patch occupancy (Thomas and Jones 1993, Hanski 1998, Moilanen and Hanski 1998). However, environmental variables, such as habitat quality, can also have significant influences on the metapopulation dynamics of a species and prediction of habitat use (Fleishman et al. 2002, Matter and Roland 2002, Walker et al. 2003). Determining habitat quality requires demonstration of a correlation between key habitat variables and population performance (e.g., fecundity, survival). Habitat quality can influence social behaviour (Walters et al. 2002), the distribution of individuals (Johnson and Sherry 2001) and reproductive success (Norris et al. 2004). Understanding what elements of the environment influence population performance facilitates predictions regarding the impacts of habitat change at both the landscape and local habitat scales. 1 Understanding how individuals use habitats and the identification of key habitat variables that affect fitness carries particular importance for species of conservation concern. On a population level, comprehending how populations interact across a landscape will enable one to identify factors affecting the distribution of a species. A metapopulation model serves as one potentially effective means of examining population level questions in fragmented populations. Patterns of genetic population structure provide important insights into landscape level population dynamics (Hastings and Harrison 1994, Pannell and Charlesworth 2000). In particular, studies of genetic population structure can identify factors influencing population persistence such as patterns of dispersal (Caizergues et al. 2003), barriers to gene flow (Piertney et al. 1998), and changes in levels of heterozygosity (Hedrick and Gilpin 1997). Empirical data contributing to our understanding of animal metapopulation dynamics and genetic structuring comes primarily from species that are sensitive to environmental stochasticity and have evolved in naturally ephemeral habitat types in which habitat suitability changes temporally (review in Harrison and Taylor 1997, tidal pools - Ebert et al. 2002, Pajunen and Pajunen 2003, meadows - Hanski et al. 1995, fire maintained scrub - Stith et al. 1996). Other empirical studies come from species that have experienced severe anthropogenic fragmentation of their natural habitat (Elmhagen and Angerbjorn 2001). Relatively little work has been done on the metapopulation dynamics of species which have evolved in naturally isolated and stable habitats (but see Moilanen and Hanski 1998, Martin et al. 2000, Elmhagen and Angerbjorn 2001). In my research I explored resource selection, genetic population structure and population dynamics of a species that has evolved in naturally isolated and successionally stable, high-elevation habitats. Resource selection Determining which resources animals select provides fundamental information about how species meet their requirements for survival (Manly et al. 2002). Resource selection studies examine the resources used by a species and compare these to available resources. This approach provides quantitative information regarding the resource requirements of both individuals and populations (Manly et al. 2002). Resource selection behaviour is a spatially hierarchical process including the: 1) geographic range of the species, 2) individual home ranges, 3) use of general features (habitats) within the home range and 4) selection of particular elements (e.g. food, cover). The information obtained from this approach can be related to population density and can also be used in modelling efforts to project the impact of habitat change (Manly et al. 2002, Boyce and McDonald 1999). 2 Determining which features of the environment species prefer identifies environmental attributes contributing to the quality of the habitat, particularly if these features are correlated with levels of reproductive success and survival (population performance). Furthermore, as mentioned above, habitat quality can also affect the dynamics of the entire metapopulation by influencing rates of immigration and emigration (Fleishman et al. 2002, Matter and Roland 2002, Walker et al. 2003). Metapopulation dynamics The metapopulation approach to population biology is based on two key premises: 1) populations are spatially structured into assemblages of breeding subpopulations and 2) migration among local populations has some effect on local dynamics, including the possibility of population rescue (Hanski and Simberloff 1997). Dispersal between populations is a key aspect of metapopulation biology and has the potential to affect the persistence of a species. Habitat patch size and the degree of isolation between populations influence the probability and effectiveness of dispersal, and can also influence the vulnerability of local populations to extinction. Isolated populations may become vulnerable to extinction if emigration from an occupied patch is not compensated by immigration (Stacey and Taper 1992). Migration is of particular concern for populations with low fecundity. In populations suffering from low fecundity or survival, dispersal from neighbouring populations may help prevent local extinction (Brown and Kodric-Brown 1977). For white-tailed ptarmigan (Lagopus leucura altipetens) in Colorado, immigration of individuals from separate subpopulations played an important role in demographic rescue of declining populations (Martin et al. 2000). Frequent dispersal among populations and demographic rescue can result in panmictic genetic population structure and, in these instances, patchy populations are more accurately considered one large population. Limited dispersal can result in low levels of gene flow and may lead to the genetic isolation of certain subpopulations resulting in a decrease in heterozygosity (Hedrick and Gilpin 1997). Thus, metapopulation dynamics can influence the demographic and genetic characteristics of populations. The landscape matrix surrounding local populations can present significant barriers to gene ^ flow, and these barriers can have genetic and evolutionary consequences for local populations (Segelbacher et al. 2003). Mountain ranges rising above the preferred habitat of black grouse (Tetrao tetrix) and capercaillie have been suggested as natural barriers to gene flow (Segelbacher and Storch 2002, Caizergues et al. 2003). In red grouse (Lagopus lagopus scoticus), a river system and the surrounding unsuitable habitat was revealed as a natural barrier 3 to gene flow (Piertney et al. 1998). Metapopulation studies often only consider patch area and isolation when examining gene flow. However, the permeability of the surrounding habitat matrix can have important influence on the 'effective isolation' of habitat patches (Ricketts 2001). A potential consequence of limited gene flow is higher relatedness between individuals in small populations due to restricted mate choice. Inbreeding results from matings between closely related individuals that can produce offspring with reduced fitness due to the expression of partially recessive deleterious mutations in homozygous form (Darwin 1876, 1877). This process is referred to as inbreeding depression and can affect the persistence of populations (Crnokrak and Roff 1999, Hedrick 2001) and small populations are particularly at risk (Keller and Waller 2002). Inbreeding depression has been related to a reduction in hatching rates and annual survival within natural avian populations (Keller 1998, Daniels and Walters 2000). Furthermore, inbreeding depression can increase an individual's susceptibility to environmentally-induced mortality (Hedrick and Kalinowski 2000). Inbred soay sheep (Ovis aries) are more susceptible to the negative effects of parasites than less inbred individuals (Coltman et al. 1999), and inbred song sparrows (Melospiza melodia) are less likely to survive severe weather conditions than outbred individuals (Keller et al. 1994). In Darwin's finches (Keller et al. 2002), inbreeding depression reduced the probability of recruitment and was much more severe in years with low food availability and high competition. Harsh environmental conditions can potentially magnify inbreeding depression, thus putting small populations exposed to both inbreeding and heightened stress at particular risk (Hedrick and Kalinowski 2000, Keller et al. 2002). A number of factors influence patterns of genetic variability in small populations. Highly inbred populations may have low genetic load because inbreeding potentially exposes recessive deleterious alleles to purging by natural selection (Keller et al. 1994). This effect may be particularly pronounced in species exposed to harsh environments (Keller et al. 1994, Keller et al. 2002). Predictions regarding levels of heterozygosity in small populations are further confounded by the effects of genetic bottlenecks, which, if frequent, can increase homozygosity in populations (Keller et al. 1994, Bouzat et al. 1998). Moreover, in small populations, a limited amount of immigration can have disproportionate effects on the vigour and persistence of local populations if heterozygous individuals have higher fitness (Ingvarsson and Whitlock 2000, Whitlock et al. 2000). 4 Maintenance of genetic variation within populations helps retain deleterious recessive mutations in a heterozygous state and may provide adaptive potential in a changing environment (Amos and Balmford 2001). Genetic variability and adaptability of populations may be important in the ability of animals to respond to extreme environmental change, such as global warming (Lande 1999). In red-cockaded woodpecker (Picoides borealis) populations, inbred birds did not adjust to climate change by laying earlier as other birds have, and inbred individuals suffered reproductive costs as a result (Schiegg et al. 2002). Models of metapopulation dynamics are used to predict patch occupancy, extinction and colonization of local populations (Hanski and Gilpin 1991). The majority of these models focus primarily on the influences of patch area and isolation (Hanski 1998). Theory suggests the probability of local extinction increases with decreasing patch size and the probability of colonization increases with decreasing isolation, and empirical studies (e.g., Thomas and Jones 1993, Moilanen and Hanski 1998) support this pattern. However, other studies found measures of habitat quality explained significant variance in population extinction and colonization rates, by showing that environmental variables such as food, topography and vegetation structure can also affect the dynamics of metapopulations (Dennis and Eales 1999). In several cases, habitat quality superseded patch geometry as an explanation for metapopulation dynamics (Fleishman et al. 2002, Walker et al. 2003), including patterns of dispersal (Matter and Roland 2002). Furthermore, Fleishman et al. (2002) suggested that habitat quality is particularly important in highly variable environments where metapopulation dynamics may be closely related to multiple aspects of habitat quality that are not static in their relative importance. Study species - white-tailed ptarmigan (Lagopus leucura) White-tailed ptarmigan are a small grouse species (375 - 400g) and live in alpine areas at or above treeline throughout western mountain Cordilleras (Braun et al. 1993). White-tailed ptarmigan are restricted to North America, have an almost continuous distribution from Alaska to Colorado, and can be found in isolated populations as far south as New Mexico (Braun et al. 1993). This species feeds on most available plant parts in the alpine, including stems, leaves, seeds, buds, flowers, and fruits. Chicks and hens that are laying eggs will opportunistically feed on insects. Clutch initiation is often dependent on environmental variables such as winter snowfall and the timing of snow melt in extreme years (Clarke and Johnson 1992, Martin and Wiebe 2004), but on average, begins in early June with an incubation period of 23 - 24 days (Wiebe and Martin 2000). Average first clutch size (± standard deviation) for white-tailed ptarmigan in Colorado is 6.2 ± 0.7 for adult birds (> 2yr) and 5.8 ± 0.9 for yearlings (Braun et 5 al. 1993). First clutch size appears similar in Vancouver Island white-tailed ptarmigan including birds of all ages (n — 10, mean = 6.0 ± 0.7, B. Fedy, unpubl. data), although demographic data are limited for L. I. saxatilis, particularly for early in the breeding period. White-tailed ptarmigan produce one brood of precocial young per season, and will re-nest if the clutch is lost early in the season or before the first two weeks of incubation (Braun et al. 1993). Females raise chicks to independence without male assistance. Males and unsuccessful females form mixed flocks later in the summer. The species also tends to migrate altitudinally each year, breeding at mid-high elevations, moving to mountain tops in late summer, and wintering across a range of mid-high elevations, using snow free ridges for forage and soft snow for burrows. Sandercock et al. (2005a) reported age-specific variation in ptarmigan demography, with clutch size and the probability of re-nesting increasing with age. Compared to other ptarmigan, white-tailed ptarmigan populations have higher survival and typically older stable age distribution appearing to follow a survivor or bet-hedging life history strategy (Sandercock et al. 2005a,b). There are five subspecies of white-tailed ptarmigan (Braun et al. 1993). The focus of my research was the subspecies Lagopus leucura saxatilis, restricted to Vancouver Island. The L. I. saxatilis subspecies designation was based on 12 specimens and differences in plumage colouration and morphology (larger body size and bill) than nominate subspecies L. I. alpitens, (McTaggart Cowan 1938). L. I. saxatilis also has heavier body mass and shorter wings relative to Colorado (L. I. alpitens) populations (Martin and Forbes 2001). White-tailed ptarmigan on Vancouver Island may also use different habitat types, likely at lower elevation, than mainland conspecifics (Martin and Elliot 1996, see Chapter 2). Furthermore, L.l. saxatilis is a 'threatened,' Blue-listed subspecies in British Columbia, which means it may be vulnerable to declines from human activities and/or natural events (Fraser et al. 1999). Baseline inventory data (presence/absence, abundance, etc.) were collected from 1995-2000 in response to the lack of ecological data on this subspecies (Martin and Elliot 1996, Martin and Commons 1997, Martin et al. 2004). The intensive period of field data collection for this study ran from May - August 2003 and 2004. Martin et al. (2004) studied the historic and current distribution of Vancouver Island white-tailed ptarmigan and reported observations on 81 mountains from 1905 to 2004 using a combination of public survey records and 253 captured and marked birds. They concluded white-tailed ptarmigan occupied most of the suitable habitat on Vancouver Island and found no evidence of range contraction over the 25 years examined. Martin et al. (2004) also suggested that population stability was lower in the southern region of the island than in the central region. 6 Study area - Vancouver Island, British Columbia Vancouver Island, located off the west coast of mainland British Columbia, Canada is the largest island on the North American west coast. It is ~ 460 km long (north - south) and 50 - 80 km wide (east to west) with an area of 31,284 km 2 (Figure 1.1). Vancouver Island has extensive alpine areas and the highest mountain is Golden Hinde at 2,682m a.b.s.l. Compared to alpine habitats on the mainland, Vancouver Island alpine habitats are lower elevation and possibly lower quality due to their small size and fragmented distribution. Alpine areas on Vancouver Island are naturally divided into a continuous group of mountains in the central region of the island and a more fragmented collection of lower elevation mountains in the southern region of the island. Much of the central portion of the island falls within Strathcona Provincial Park (approx. 2,500 km ) and has likely experienced less alteration of the low-elevation landscape matrices surrounding the alpine areas. I examined seven populations of Vancouver Island white-tailed ptarmigan (Figure 1.1). The southern region includes the South, South West and Beaufort populations. The central region includes the Central South, Central West and Central East populations. I surveyed only one population in the northern region of the island. All populations represent a collection of mountains divided by natural barriers such as lakes and low-elevation areas. The southern populations are characterized by limited alpine habitat, a higher proportion of subalpine and upper montane (mountain hemlock parkland) and greater isolation than the more continuous central region populations. White-tailed ptarmigan reside year-round and require all necessary habitat types for summer and winter within close proximity. On the mainland, white-tailed ptarmigan have measurable habitat preferences throughout the year which are described in detail in Chapter 2. Vancouver Island white-tailed ptarmigan are found in mountain ranges of various altitudes (822 m - 2682 m; mainland white-tailed ptarmigan in B.C. 1,280 m - 2,650 m Campbell et al. 1990; in Colorado, 3,350 m-4,250 m Braun et al. 1993). Alpine regions are some of the most rapidly warming habitats in response to global climate change, particularly through the winter (Beniston et al. 1997). If climate change is going to detrimentally affect any species,it is likely that alpine species will show this impact first (the proverbial 'canary in the coal mine'; Martin 2001, Beniston 2003). Ecological responses to climate change such as the altitudinal shift of plant communities could lead to greater separation of populations and a decrease in available habitat (Grabherr et al. 1994). Populations may respond to these changes in habitat by a decrease in population size or perhaps a change in habitats used. As available alpine habitat 7 decreases, individuals may move into less preferred, lower elevation habitat types. Drier climates and advancing treelines are two predicted effects of climate change and will likely impact white-tailed ptarmigan i f individuals select for moist alpine habitats. Study objectives Ecology is the scientific study of the interactions that determine the distribution and abundance of organisms from genetic to landscape scales (Krebs 1994). Studying metapopulation dynamics can provide a greater understanding of how populations interact across a landscape and identify potential threats associated with limited population interaction such as inbreeding depression. Quantifying how organisms select habitats on a finer scale and their resource selection behaviour can provide additional detail explaining distribution and abundance. Vancouver Island white-tailed ptarmigan are ideal for this type of analysis because of their isolated populations, restricted habitat requirements and potential sensitivity to human impacts. Objective 1: To examine resource selection behaviour of white-tailed ptarmigan on Vancouver Island, to determine the characteristics of preferred habitats, and to examine regional variation in measures of population performance (Chapter 2). Objective 2: To examine the genetic population structure and levels of dispersal and gene flow among different white-tailed ptarmigan populations on Vancouver Island (Chapters 3 and "4). Overview of thesis In Chapter 1,1 described the importance of understanding resource selection behaviour and discussed its potential influence on measures of population performance, particularly in species of conservation interest. I explained metapopulation dynamics and the potential influences on the demographic and genetic characteristics of populations. Finally, I outlined the basic ecology and life-history attributes of white-tailed ptarmigan and the physical characteristics of their habitats on Vancouver Island. In Chapter 2,1 investigated the habitat selection behaviour and regional variation in population performance of Vancouver Island white-tailed ptarmigan. I examined habitat selection data using a generalized linear mixed model approach which compared competing habitat selection models and identified key habitat variables. I analyzed variation in population performance using data on female reproductive success, female age structure and sex ratio. Population performance was compared between the south and central regions of Vancouver Island. I determined that, unlike other white-tailed ptarmigan subspecies, Vancouver Island 8 white-tailed ptarmigan use a generalist strategy to habitat selection and 1 identified several key habitat variables. I discussed the implications of regional variation in population performance and considered the role of regional variation in key habitat variables. In Chapter 2,1 also-documented unique habitat selection behaviour and discussed the implications of the results for the conservation of this threatened subspecies in light of .the predicted impacts of climate change. In Chapter 3,1 described the population genetics of Vancouver Island white-tailed ptarmigan. I presented data on microsatellite primer optimization, and described generalized heterozygote deficiencies and high levels of diversity. A brief analysis of population subdivision revealed weak population structure among the seven populations. I discussed the potential influence of null alleles and examined several scenarios to explain the observed genetic patterns of high inbreeding and generalized heterozygote deficiencies. In Chapter 4,1 used direct and indirect measures of dispersal to investigate the patterns of dispersal and identify barriers to gene flow for Vancouver Island white-tailed ptarmigan. I examined dispersal patterns and population connectivity among seven populations using 10 microsatellite loci. Patterns of dispersal were also addressed using direct measures by following the inter-seasonal movements of radio-collared birds. I presented results on patterns of genetic differentiation between populations, discussed the most appropriate model of dispersal and the potential impacts of genetic isolation. In Chapter 5,1 summarized the major conclusions of each study within this thesis. I provided an overview of the entire research programme and discussed future directions for studies of white-tailed ptarmigan and grouse population genetics. Chapters 2, 3, and 4 were written with the intention of publishing them as stand alone papers after my thesis defence. 9 Figure 1.1. Vancouver Island, British Columbia, Canada. Dark grey areas represent Alpine Tundra, and lighter grey represents Mountain Hemlock. Habitat classifications are based on British Columbia's Biogeoclimatic Ecosystem Zones. Population boundaries are circled with a dotted line and labelled S - South, S W - South West, B - Beaufort, C S - Central South, C E - Central East, C W - Central West, N - North. See Chapter 3, Table 3.1 for names of mountains included in each population. o R E F E R E N C E S Amos, W. and A. Balmford. 2001. When does conservation genetics matter? Heredity 87: 257-265. Beniston, M . , H. F. Diaz and R. S. Bradley. 1997. Climatic change at high elevation sites: an overview. Climate Change 36: 233-251. Beniston, M . 2003. Climate change in mountain regions: a review of possible impacts. Climate Change 59: 5-31. Bouzat, J. L. , H. H. Cheng, H. A. Lewin, R. L. Westemeier, J. D. Brawn and K. N. Paige. 1998. Genetic evaluation of a demographic bottleneck in the greater prairie-chicken. Conservation Biology 12: 836-843. Boyce, M . S. and L. L. McDonald. 1999. Relating populations to habitats using resource selection functions. Trends in Ecology & Evolution 14: 268-272. Braun, C. E. , K. Martin and L. A. Robb. 1993. White-tailed ptarmigan (Lagopus leucurus). In: F. Gill and A. Poole. The Birds of North America, 68. Academy of Natural Sciences, Philadelphia, PA, and American Ornithologists' Union, Washington, D.C.: 22. Brown, J. H. and A. Kodric-Brown. 1977. Turnover rates in insular biogeography: effect of immigration on extinction. Ecology 58: 445-449. Caizergues, A., O. Ratti, P. Helle, L. Rotelli, L. Ellison and J.-Y. Rasplus. 2003. Population genetic structure of male black grouse (Tetrao tetrix L.) in fragmented vs. continuous landscapes. Molecular Ecology 12: 2297-2305. Campbell, R. W., N. K. Dawe, I. McTaggart Cowan, J. M . Cooper, G. W. Kaiser and M . C. E. McNall. 1990. The birds of British Columbia. Royal British Columbia Museum, Victoria. Clarke, J. A. and R. E. Johnson. 1992. The influence of spring snowpack on white-tailed ptarmigan (Lagopus leucurus) breeding success, spacing and movement in the Sierra Nevada. Condor 94: 622-627. Coltman, D. W., J. G. Pilkington, J. A. Smith and J. M . Pemberton. 1999. Parasite-mediated selection against inbred soay sheep in a free-living, island population. Evolution 53: 1259-1267. Crnokrak, P. and D. A. Roff. 1999. Inbreeding depression in the wild. Heredity 83: 260-270. Daniels, S. J. and J. R. Walters. 2000. Inbreeding depression and its effects on natal dispersal in red-cockaded woodpeckers. The Condor 102: 482-491. Darwin, C. R. 1876. The effects of cross and self fertilization in the vegetable kingdom. London: John Murray. 11 Darwin, C. R. 1877. The different forms of flowers on plants of the same species. London: John Murray. Dennis, R. L. H. and H. T. Eales. 1999. Probability of site occupancy in the large heath butterfly Coenonympha tullia determined from geographical and ecological data. Biological Conservation 87: 295-301. Ebert, D., C. Haag, M . Kirkpatrick, M . Riek, J. W. Hottinger and V. I. Pajunen. 2002. A selective advantage to immigrant genes in a Daphnia metapopulation. Science 295: 485-488. Elmhagen, B. and A. Angerbjorn. 2001. The applicability of metapopulation theory to large mammals. Oikos 94: 89-100. Fleishman, E. , C. Ray, P. Sjogren-Gluve, C. L. Boggs and D. D. Murphy. 2002. Assessing the roles of patch quality, area, and isolation in predicting metapopulation dynamics. Conservation Biology 16: 706-716. Fraser, D. F., W. L. Harper, S. G. Cannings and J. M . Cooper. 1999. Rare birds of British Columbia. Wildlife Branch and Resource Inventory Branch, British Columbia Ministry of Environment, Lands and Parks,. 60 - 61. http://srmapps.gov.bc.ca/apps/eswp/ Grabherr, G. M . , M . Gottfried and H. Pauli. 1994. Climate effects on mountain plants. Nature 369: 448. Hanski, I. and M . Gilpin. 1991. Metapopulation dynamics; brief history and conceptual domain. Biological Journal of Linnean Society 42: 3-16. Hanski, I., T. Pakkala, M . Kuussaari and G. C. Lei. 1995. Metapopulation persistence of an endangered butterfly in a fragmented landscape. Oikos 72: 21-28. Hanski, I. 1998. Metapopulation dynamics. Nature 396: 41-49. Hanski, I. A. and D. Simberloff. 1997. The metapopulation approach, its history, conceptual domain, and application to conservation. In: I. A. Hanksi and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 5-26. Harrison, S. and D. A. Taylor. 1997. Empirical evidence for metapopulation dynamics. In: I. A. Hanksi and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 27-42. Hastings, A. and S. Harrison. 1994. Metapopulation dynamics and genetics. Annual Review of Ecology and Systematics 25: 167-188. Hedrick, P. W. and M . E. Gilpin. 1997. Genetic effective size of a metapopulation. In: I. A. Hanski and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 165-181. 12 Hedrick, P. W. and S. T. Kalinowski. 2000. Inbreeding depression in conservation biology. Annual Review of Ecology and Systematics 31:139-162. Hedrick, P. W. 2001. Conservation genetics: where are we now? Trends in Ecology and Evolution 16: 629-636. Ingvarsson, P. K. and M . C. Whitlock. 2000. Heterosis increases the effective migration rate. Proceedings of the Royal Society of London Series B 267: 1321-1326. Johnson, M . D. and T. W. Sherry. 2001. Effects of food availability on the distribution of migratory warblers among habitats in Jamaica. Journal of Animal Ecology 70: 546-560. Keller, L. F., P. Arcese, J. N. M . Smith, W. M . Hochachka and S. C. Stearns. 1994. Selection against inbred song sparrows during a natural population bottleneck. Nature 372: 356-357. Keller, L. F. 1998. Inbreeding and its fitness effects in an insular population of song sparrows (Melospiza melodia). Evolution 52: 240-250. Keller, L. F., P. R. Grant, B. R. Grant and K. Petren. 2002. Environmental conditions affect the magnitude of inbreeding depression in survival of Darwin's finches. Evolution 56: 1229-1239. Keller, L. F. and D. M . Waller. 2002. Inbreeding effects in wild populations. Trends in Ecology and Evolution 17: 230-241. Krebs, C. J. 1994. Ecology: the experimental analysis of distribution and abundance. HarperCollins College Publishers, New York; Lande, R. 1999. Extinction risks from anthropogenic, ecological, and genetic factors. In: L. F. Landweber and A. P. Dobson. Genetics and the Extinction of Species, Princeton University Press: 1-22. Manly, B. F. J., L. L. McDonald, D. L. Thomas, T. L. McDonald and W. P. Erickson. 2002. Resource Selection by Animals: statistical design and analysis for field studies. Kluwer Academic Publishers, Dordecht, The Netherlands. Martin, K. and L. Elliot. 1996. Vancouver Island white-tailed ptarmigan inventory progress report (1995-1996). Centre for Alpine Studies, Forest Sciences, University of British Columbia, http://www.forestry.ubc.ca/alpine/docs/wtpvi-l.pdf Martin, K. and M . L. Commons. 1997. Vancouver Island white-tailed ptarmigan inventory project: progress report. 1997 surveys. Centre for Alpine Studies, Forest Sciences, University of British Columbia, http://www.forestry.ubc.ca/alpine/docs/wtpvi-3.pdf Martin, K., P. B. Stacey and C. E. Braun. 2000. Recruitment, dispersal, and demographic rescue in spatially-structured white-tailed ptarmigan populations. Condor 102: 503-516. Martin, K. and L. Forbes. 2001. Vancouver Island white-tailed ptarmigan: species information. British Columbia Ministry of the Environment, Lands and Parks. 10. 13 Martin, K., G. A. Brown and J. R. Young. 2004. The historic and current distribution of the Vancouver Island white-tailed ptarmigan (Lagopus leucurus saxatilis). Journal of Field Ornithology 75: 239-256. Martin, K. and K. L. Wiebe. 2004. Coping mechanisms of alpine and arctic breeding birds: Extreme weather and limitations to reproductive resilience. Integrative and Comparative Biology 44: 177-185. Martin, K. M . 2001. Wildlife in alpine and sub-alpine habitats. In: D. H. Johnson and T. A. O'Neil. Wildlife-Habitat Relationships in Oregon and Washington., Oregon State University Press: 285-310. Matter, S. F. and J. Roland. 2002. An experimental examination of the effects of habitat quality on the dispersal and local abundance of the butterfly Parnassius smintheus. Ecological Entomology 27: 308-316. McTaggart Cowan, I. 1938. The white-tailed ptarmigan of Vancouver Island. Condor 41: 82-83. Moilanen, A. and I. Hanski. 1998. Metapopulation dynamics: effects of habitat quality and landscape structure. Ecology 79: 2503-2515. Norris, D. R., P. P. Marra, T. K. Kyser, T. W. Sherry and L. M . Ratcliffe. 2004! Tropical winter habitat limits reproductive success on the temperate breeding grounds in a migratory bird. Proceedings of the Royal Society of London Series B 271: 59-64. Pajunen, V. I. and I. Pajunen. 2003. Long-term dynamics in rock pool Daphnia metapopulations. Ecography 26: 731-738. Pannell, J. R. and B. Charlesworth. 2000. Effects of metapopulation processes on measures of genetic diversity. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 355: 1851-1864. Piertney, S. B., A. D. C. MacColl, P. J. Bacon and J. F. Dallas. 1998. Local genetic structure in red grouse (Lagopus lagopus scoticus): evidence from microsatellite DNA markers. Molecular Ecology 7: 1645-1654. Ricketts, T. H. 2001. The matrix matters: effective isolation in fragmented landscapes. The American Naturalist 158: 87-99. Sandercock, B. K., K. Martin and S. J. Hannon. 2005a. Demographic consequences of age-structure in extreme environments: population models for arctic and alpine ptarmigan. Oecologia 146: 13-24. Sandercock, B. K., K. Martin and S. J. Hannon. 2005b. Life history strategies in extreme environments: comparative demography of arctic and alpine ptarmigan. Ecology 86: 2176-2186. 14 Schiegg, K., G. Pasinelli, J. R. Walters and S. J. Daniels. 2002. Inbreeding and experience affect response to climate change by endangered woodpeckers. Proceedings of the Royal Society of London Series B 269: 1153-1159. Segelbacher, G. and I. Storch. 2002. Capercaillie in the Alps: genetic evidence of metapopulation structure and population decline. Molecular Ecology 11:1669-1677. Segelbacher, G., J. Hoglund and I. Storch. 2003. From connectivity to isolation: genetic consequences of population fragmentation in capercaillie across Europe. Molecular Ecology 12: 1773-1780. Stacey, P. B. and M . Taper. 1992. Environmental variation and the persistence of small populations. Ecological Applications 2: 18-29. Stith, B. M . , J. W. Fitzpatrick, G. E. Woolfenden and B. Pranty. 1996. Classification and conservation of metapopulations: a case study of the Florida scrub jay. In: D. R. McCullough. Metapopulations and Wildlife Conservation, Island Press: 187-214. Thomas, C. D. and T. M . Jones. 1993. Partial recovery of a skipper butterfly (Hesperia-Comma) from population refuges - lessons for conservation in a fragmented landscape. Journal of Animal Ecology 62: 472-481. Walker, R. S., A. J. Novaro and L. C. Branch. 2003. Effects of patch attributes, barriers, and distance between patches on the distribution of a rock-dwelling rodent (Lagidium viscacid). Landscape Ecology 18: 187-194. Walters, J. R., S. J. Daniels, J. H. Carter and P. D. Doerr. 2002. Defining quality of red-cockaded woodpecker foraging habitat based on habitat use and fitness. Journal of Wildlife Management 66: 1064-1082. Whitlock, M . C , P. K. Ingvarsson and T. Hatfield. 2000. Local drift load and the heterosis of interconnected populations. Heredity 84: 452-457. Wiebe, K. L. and K. Martin. 2000. The use of incubation behavior to adjust avian reproductive costs after egg laying. Behavioral Ecology and Sociobiology 48: 463-470. 15 CHAPTER 2 Habitat selection and regional variation in population performance of white-tailed ptarmigan on Vancouver Island, British Columbia INTRODUCTION Determining how animals select habitat features is a fundamental question in ecology. Indeed, the primary goal of ecology is to understand the processes that influence the patterns of distribution and abundance in animals (Krebs 1994). The patterns of animal distribution and abundance are often determined by the spatial and temporal variability of resources which can structure vertebrate populations genetically and ecologically. Animals' selection of habitats can reflect the variability and importance of resources and takes place at multiple spatial scales (Orians and Wittenberger 1991, Johnson et al. 2004, Anderson et al. 2005). Fine-scale selection is examined at the level of immediate influence and refers to the decisions an animal makes as it moves throughout the landscape, selecting specific objects. These fine-scale habitat selection decisions can influence the survival (Patten et al. 2005) and reproductive success (Walters et al. 2002) of an individual and may affect the performance of the population as indicated by population persistence and productivity. Habitat selection may also incorporate tradeoffs among the resources required for survival and reproduction, including food availability and shelter, within the constraints of predation, competition, and intraspecific attraction. For example, the habitat characteristics that provide the best cover from predators may not provide good forage. Given the large number of factors that influence habitat selection decisions made by animals, analysis of habitat selection behaviour is best addressed using a multivariate approach. Many studies have examined the specific habitat characteristics of areas used by species and compared those habitats with areas not used by the species (Jones 2001, Manly et al. 2002). An 16 alternate, and more informative, approach to examining habitat selection involves comparing areas that are used by an animal with those that are available to the animal (Johnson 1980). A used/available comparison provides key insights into the fine-scale choices animals make, which, as noted above, can influence population performance. A used/available comparison also allows researchers to make inferences about choice, whereas information on the quality of used versus unused habitat is only informative if the unused habitat is actually available to the animal of interest (Jones 2001). The classification of available habitats for a species must be done carefully to ensure a biologically relevant (i.e., suitable habitat for the species in that time period) and detailed comparison of decision behaviour within existing local resource variation. Animals with flexibility in their resource selection behaviour should adjust their behaviour based on the abundance and availability of habitat components. Fine-scale comparisons take this flexibility into account and ensure the comparisons of used sites relate directly to the habitat characteristics available to individuals. Regional variation in population performance can be caused by many factors such as resource availability, predation and competition. An important step to understanding the influence of habitat structure on population performance involves the identification of key habitat variables. After identifying those variables examination of corresponding regional variation in key habitat variables may elucidate which components of the habitat have the greatest effect on individual and population health. This allows us to scale up from individual-level habitat selection choices and fine-scale habitat features to population-level processes. I studied the fine-scale habitat selection behaviour of a subspecies of white-tailed ptarmigan restricted to Vancouver Island (Lagopus leucura saxatilis). White-tailed ptarmigan are an obligate alpine grouse species that spend the entire year at or above the treeline in mountain areas. Vancouver Island (British Columbia) is a large island (31,284 km2), and the alpine areas divide into two general regions. The central portion of the island has extensive continuous 17 alpine habitat inhabited by white-tailed ptarmigan. The southern portion also contains inhabited alpine areas, but relative to the central region of Vancouver Island, they are smaller and more isolated, at lower elevation, and experience more human disturbance in the lowland areas between mountains. Furthermore, the densities of white-tailed ptarmigan are lower in the southern portion than in the central region of the island (Martin et al. 2004). Studies of white-tailed ptarmigan on the mainland (Colorado, Alberta) document the importance of food and moisture availability and predation to the life history and habitat selection behaviour of this species (e.g. Herzog 1977, Giesen and Braun 1992, Sandercock et al. 2005b). White-tailed ptarmigan are well adapted to cold environments and need to cool during the heat of the summer (Johnson 1968). Given the importance of the aforementioned factors, I used generalized linear mixed models to compare 4 hypotheses concerning the habitat selection behaviour of white-tailed ptarmigan. The models I tested were: 1) food availability, 2) predator avoidance, 3) thermal regulation, 4) generalists (food, predation and thermal regulation). I then compared the performance of all these models to a separate global model that included all uncorrected habitat variables that were related to the above models and may have significantly influenced the habitat selection behaviour of white-tailed ptarmigan. Food availability White-tailed ptarmigan feed primarily on vegetation (Braun et al. 1993) and have high energy and nutrient requirements due to a higher than expected metabolic rate and approximately 8 months of molt annually (Johnson 1968). Consequently, they must maximize their use of plant productivity. White-tailed ptarmigan are herbivores with extensive foraging time (25 - 30% of daily activity, Artiss and Martin 1995) and processing time. Their crop contents account for approximately 1 - 5% of body weight during the day, and up to 25% nearing dusk (Braun et al. 1993). Willow (Salix sp.) constitutes an important and nutritious component of white-tailed ptarmigan diets (Weeden 1967, May and Braun 1972). All previous 18 studies of white-tailed ptarmigan habitat selection behaviour have found that availability of willow is the principal factor determining fine-scale ptarmigan distribution (Choate 1963, Herzog 1977, 1980, Frederick and Gutierrez 1992, Giesen and Braun 1992). While white-tailed ptarmigan will feed on a range of herbaceous and woody plants, particularly during brood rearing, willow remains the principal food source throughout the year (Allen and Clarke 2005). Vancouver Island does not have a sufficient abundance of willow for it to contribute significantly to white-tailed ptarmigan diet (B. Fedy, unpubl. data). Vancouver Island white-tailed ptarmigan feed primarily on the flowers of the ericaceous shrubs white mountain-heather (Cassiope mertensiana) and pink mountain-heather (Phyllodoce empetriformis) during the brood rearing period. They also feed regularly on the seeds of alpine grasses and sedges and the flowers of other forbs (e.g. partridge foot Luetkea pectinata; B. Fedy, unpubl. data). Predator avoidance Predation risk alters the habitat selection behaviour of many prey species (Hik 1995, Sih and Christensen 2001, Verdolin 2006). White-tailed ptarmigan are prey for an array of predators, with raptors representing the primary predators of adult birds (Braun et al. 1993). Ptarmigan chicks are precocial and experience high levels of predation (approx. 35%) during their hatch year (Harmon and Martin 2006). Sandercock et al. (2005a, 2005b) compared the life history strategies of congeneric ptarmigan and found that white-tailed ptarmigan follow a survivor, or possibly bet-hedging, life history strategy. Therefore, given that adult survival plays a dominant role in the population dynamics of species with these life histories, predator avoidance could significantly influence the habitat selection behaviour of white-tailed ptarmigan. Thermal regulation Summer temperatures in the alpine on Vancouver Island frequently reach 30°C. White-tailed ptarmigan have low evaporative efficiency (Johnson 1968) and high insulative value of feathers (Veghte and Herreid 1965, Johnson 1968), making them poorly adapted to high 19 ambient temperatures. Thus, they require habitat characteristics which reduce thermoregulatory energy demands. Birds will crouch in shallow puddles of water, snow bathe and gular flutter when ambient temperature is higher than 21°C (Bradbury 1915, Johnson 1968). I collected habitat selection data during the warmest months of the year. Therefore, individual white-tailed ptarmigan may have selected shaded areas with lower levels of solar incidence to assist with thermoregulation. Generalists Previous studies of habitat use by white-tailed ptarmigan suggest they are specialists in their selection of habitat, choosing moist areas of predominantly willow cover (Choate 1963, Weeden 1967, May and Braun 1972, Herzog 1977, Herzog 1980, Frederick and Guitierrez 1992, Giesen and Braun 1992, Allen and Clarke 2005). Similarly, other species in the subfamily Tetraoninae also seem quite specialized in their habitat selection behaviour (Storch 1993, Baines et al. 1996, Summers et al. 2004). However, given the presumed tradeoffs between foraging, predator avoidance and thermal regulation involved in habitat selection, more than one variable may influence the habitat selection behaviour of individuals. The generalist model combines the most important variable from each of the other models, thus addressing a combination of food availability, predator avoidance and thermal regulation requirements. Regional variation in reproductive success, age structure and habitat attributes Vancouver Island mountain areas are divided into a larger, more continuous central region and a smaller, more fragmented, southern region. The smaller, more isolated, populations may have a higher probability of local extinction (Hanski 1998). This can be a product of low reproductive success and/or survival. I predicted that population performance, as measured by reproduction and persistence, would be lower in the southern region of the island than in the central region. Young breeding females have lower survival and reproductive success compared to older birds (Wiebe and Martin 1998). Therefore, populations with an age structure biased 20 towards older birds have higher reproductive output and future population growth potential. 1 also predicted a greater proportion of older birds in the central region of the island because age structure skewed towards older birds is indicative of higher population performance, both as an indirect indication of survival and because older birds have higher reproductive output than younger birds. I also predicted differing sex ratios between the two regions as another indication of regional variation in population performance. Most monogamous birds have male biased sex ratios (Wittenberger and Tilson 1980, Black 1996), and when populations of monogamous species decline, they typically become even more male biased (Hannon and Martin 1992). White-tailed ptarmigan are a monogamous species and populations are generally characterized by an excess of males (Braun et al. 1993). Therefore, I predicted greater male-biased sex ratios in the more fragmented southern region compared to the central region. Reproductive success, age structure, and sex ratios combine to give an accurate indication of overall population performance, as these variables capture juvenile production, recruitment, and potential for population level production. Two objectives of this study were to identify key habitat features preferred by white-tailed ptarmigan, and to determine if population performance (reproductive success, age structure, sex ratios) varied regionally on Vancouver Island. I correlated population performance with regional variation in habitat features selected by ptarmigan, in order to achieve the objective of scaling up from individual-level choices and fine-scale habitat features to population level processes. METHODS Study species and study area In Chapter 1 I provided detailed descriptions of the study species and Vancouver Island. 21 Data collection Birds were located using playbacks of male territorial calls during the early breeding season (May - late June), and by playbacks of chick distress calls in later season (late June - October). Individuals were captured using a noose pole (Zwickel and Bendell 1967), and outfitted with a necklace radio collar (RI-2D/2B, 18 month battery life, 9g, Holohil Systems Ltd., Carp, Ontario) and a unique combination of coloured leg bands. Individuals were aged by comparing the shape of the 8 t h and 9 t h primary feathers, and by pigmentation on the 9 t h and 10th primary coverts, which distinguishes first-year birds from adults (Bergerud et al. 1963). Previous to my study, birds were captured as part of an inventory project from June to October 1995 - 1999 (Martin and Elliot 1996, Martin and Commons 1997, Martin et al. 2004). My subsequent intensive habitat selection study captured birds from May to August 2002 - 2004. Both stages (inventory and intensive) combined resulted in a total of 9 years of banding. Researchers in both stages also noted species and abundance of all potential predators of adults, chicks, and nests while monitoring white-tailed ptarmigan in the field. The most abundant predators were primarily avian and included ravens (Corvus corax), eagles, and hawks (B. Fedy and K. Martin, unpubl. data) and also likely included wolves (Canis lupus) and cougars (Felis concolor; Martin and Elliot 1996, Martin and Commons 1997). Fine-scale habitat variables Fine-scale resource selection data were collected on radio-collared birds during the chick rearing period from July to August 2003 and 2004. Researchers maintained sufficient distance (approximately 35m) from the focal birds to ensure minimal disturbance, with some variation in distance based on local topography. Observers used binoculars to watch birds and bird locations were visually marked every half hour over intensive observation sessions, each of which was approximately 3 hours. Each session then resulted in 7 observations per bird (at 0, 30, 60, 90, 120, 150, 180 min.). The duration of these sessions allowed time for birds to engage in a 22 number of behaviours (feeding, vigilance, etc.), and allowed them to choose multiple sites from the habitat in the immediate area. Occasionally weather and/or terrain forced us to shorten the session. After a 3 hr observation session ended, habitat data were collected for a 20 m 2 plot (approximately 2.5 m radius) with the observed bird locations as the centre points. These locations represented used habitat. An individual bird was only observed for 1 sampling session each year (2 birds were observed in both years). The selection of available habitat is often not assessed in a manner relevant to the study (Jones 2001). I ensured that sites were actually available to individuals by constraining the selection of available sites to within an individual's territory. For the selection of available habitat, I selected 2 random compass bearings and chose distances in 5 m intervals from 30 m -50 m from the centre of the known used sites. The end point of this transect was the centre point for a 20 m habitat sampling plot, following the same protocol as at the used sites. Generally, I measured 2 samples of available habitat for every used habitat sample. Used plots were coded as 1 and available plots as 0. The randomly selected available plots had to meet two criteria for inclusion in sampling: 1) The site had to be accessible. Random sites were occasionally inaccessible given the steep and rugged nature of the habitat. 2) The site had to have less than 75% snow cover. High snow cover areas are seldom used by white-tailed ptarmigan during late summer (pers. obs) and therefore do not represent available habitat. I collected a number of distance measures for each plot (e.g. distance to water). Cases in which I could not detect water within a reasonable distance of the plot, I entered the maximum distance I could accurately measure which was always greater than 200m, but ranged up to 640m. Models I developed a priori hypotheses concerning fine-scale habitat selection behaviour and built models to compare these hypotheses. Each model included the habitat attributes I considered most likely to represent a specific hypothesis. Aspect was sine-transformed so that aspects with 23 higher levels of solar incidence were close to 1, and aspects with lower solar incidence close to -1. Since the importance of aspect is slope dependent, aspects with slopes < 5 0 were assigned a neutral value of 0 (Whittington et al. 2005). Percent cover variables were arcsine transformed prior to analysis for a closer approximation of a normal distribution (Zar 1999). I used principal components analysis (PCA) for variable reduction and removed any variables with r > 0.65. When two variables were highly correlated, I retained the most biologically relevant variable. Table 2.1 presents all variables retained after the PCA. Habitat selection data are commonly analyzed using logistic regression in which predictor variables are combined in an effort to predict the probability of use (probability y = 1; Manly et al. 2002, Keating and Cherry 2004). This type of analysis allows for inclusion of continuous and categorical variables and identifies the relative influence of specific habitat features. However, inferences from a logistic regression are not valid unless all model assumptions are met. My habitat sampling observation sessions violated a key assumption of logistic regression. Because I had multiple measures from individuals which were correlated in both space and time, my data did not satisfy the logistic regression assumption of independence of errors. Therefore, I used a mixed model approach to determine the characteristics of preferred habitats and test models of habitat selection behaviour (GLIMMIX procedure, SAS Institute, 2005, Gillies et al. 2006). Individual identity was entered into the models as a random factor to account for correlations within individuals. I specified a first-order autoregressive covariance structure to account for the effect of repeated measurements within a particular subject and ensure normal distribution of the random effects residuals. Model validation and selection Models were validated by withholding a fraction of the data using a &-fold partitioning of the original samples, where k represents the number of partitions ranging from 2 to N- 1 (Fielding and Bell 1997). Using Huberty's (1994) rule of thumb for determining the ratio of training to 24 testing cases, I used 75% of the data for my training set. Huberty's heuristic suggests a ratio of [ 1 + (p - 1) / 2 ]"' where p is the number of predictors (Huberty 1994). I used 4-fold partitioning in which the data were randomly divided into 4 groups of approximately equal size. Each model was then built using a training set containing approximately 75% of the data, and then was tested against the remaining data. Each individual quarter of the data was iteratively removed from the training set and used as an independent test set to analyze the reliability of the model (built on the remaining three quarters of the data) and to provide a means for comparison across models. The GLIMMIX procedure fits generalized linear mixed models based on linearizations that employ Taylor series expansions to approximate the model by a model based on pseudo-data with fewer nonlinear components (GLIMMIX procedure documentation, SAS Institute, 2005 http://support.sas.com/rnd/app/papers/glimmix.pdf). I specified the use of restricted pseudo-likelihood estimation for each model (Wolfinger and O'Connell 1993). This technique does not result in a true objective function for the overall optimization process, and therefore is not appropriate for comparison among models. The development of a single objective function value for model comparison is an active area of research, not yet resolved within the statistical community. Therefore, to assess model performance and compare models, I used an adaptation of a method proposed by Boyce et al. (2002). This technique is based on the patterns of predicted scores obtained from the test data sets in the &-fold process. Each of the four test data sets produced a set of predicted values for all observations. I partitioned the test data to only include the predicted values from used sites (coded as l's). These values were then separated into categories of predicted values (histogram bins). Frequencies of predicted values were adjusted (divided) by the proportion of 'used' observations in the original data (area adjusted frequencies) and were scaled for unequal test sample sizes. A model with good predictive performance has a strong positive correlation between bin number and area-adjusted frequency, 25 as more used locations should fall within higher bins (Boyce et al. 2002). Bin size and number were chosen following Boyce et al. (2002), and resulted in 8 bins of approximately equal sample size. This method merges rare predicted values at the tails of the distribution, and bins that do not have validation points with the nearest bin. A Spearman-rank correlation between area-adjusted frequency of cross-validation points within individual bins and the bin rank was calculated for each cross-validated model. A significant value suggests a positive correlation between bins and predicted values indicating the model does well at predicting use from the suite of available sites. When the data are partitioned to include only the available sites (coded as 0's), a Spearman-rank correlation will show a negative correlation for models with good predictive capabilities. Female reproductive success White-tailed ptarmigan chicks remain with the hen until they reach independence in the fall of their hatch year (approximately 60 days of age, Hannon and Martin 2006). Chicks are capable of sustained flights at approximately 25 days after hatch. By convention, and because juvenile grouse experience the highest mortality during their first two weeks, a female was considered successful if her young survived to 25 days (Hannon and Martin 2006). Data on female reproductive success were collected throughout 8 years of the study during both the inventory and intensive stages (1995 - 1999 and 2002 - 2004) by following radio-tagged hens until reproductive outcome could be determined. I used a generalized linear mixed model (GLIMMIX) to examine differences in female reproductive success (0-7, 15 chicks/female) between the southern and central portions of the island. I included one observation per year per female observed between mid-July and late September. The distribution of observation dates of fledging success between these months was similar for both regions (B. Fedy and K. Martin, unpubl. data). Females observed during the post-hatch period without young that exhibited behaviours of an unsuccessful hen (e.g. flocking 26 with other adult birds) were assumed unsuccessful. The response variable was the number of chicks and had a Poisson distribution (0 values account for 21% of observations). The model included reproductive success data from 98 individual females (south n = 21, central n = 77) over 8 years (1995 - 1999, 2002 - 2004). Because some individual females were observed in multiple years, the total number of female-year observations contributing to the analysis of reproductive success was 125 (south n = 29, central n = 96). Individual identity and year were entered into the model as random factors to isolate the effect of region on female reproductive success. Age structure and sex ratios Levels of survival and reproduction in birds are often age-dependent, with older birds generally exhibiting higher survival and reproduction (Martin 1995). I compared the age structure of birds from the south portion of the island to those from the central. Birds were classified as either adults (>1 year) or yearlings (< 1 year, i.e. born the previous season). Data were combined across the 8 study years. About 31% of birds (57 of 182) were observed over more than one year, and each year was considered an independent data point for this analysis. Females are the limited sex in white-tailed ptarmigan populations. Surveys of Vancouver Island white-tailed ptarmigan throughout the inventory and intensive stages of the project did not have an equal probability of encountering both sexes. For example, females are easier to find during the chick rearing stage (if they have chicks) than males. Conversely, unmarked males are easier to locate when they are territorial during the early season. However, since field work was conducted from May to October, both groups were encountered. Nevertheless, the following analysis of sex ratios may not represent true population sex ratios. Since the same techniques were used in both the south and central regions, any biases introduced by the techniques should be consistent across regions and therefore allow for reliable comparison. 27 Only breeding age birds (yearling and older) were considered in the analysis of sex ratios. All frequency data were analyzed using Chi-square tests (FREQ procedure, SAS Institute). Regional differences in key habitat variables Key habitat variables were identified from the modelling procedure by examining the parameter estimates (P - values) and their standard errors for all model iterations and the overall average. The most important variables are those with high p coefficients (significant at p = 0.10) in more than one iteration of the model. I used data from the 500 habitat selection sampling sites to compare the distribution of these key habitat variables in the southern and central regions. All data were analyzed using SAS ver. 9.1 or SPSS ver. 11.5. R E S U L T S Fine-scale habitat selection I collected a total of 500 (used n = 176, available n = 324) samples contributing to the habitat selection models over 2003 (n = 11 individuals) and 2004 (n = 18 individuals), with the majority of the samples coming from the central region of Vancouver Island (Table 2.2). Plots of the area-adjusted frequency for each model showed that all models have upward tendencies indicating reasonable to good predictive capabilities (Figure 2.1a). The food availability, predator avoidance and generalist models had the highest number of predicted values falling in bins 7 and 8. The food availability, predator avoidance and global models showed the lowest standard error in bins 3 - 8 (Figure 2.1b). The thermal model generally performed poorly and with high standard error. The global model also performed fairly poorly, but with low standard error. The key predictor variables associated positively with percent subshrub cover, percent boulder cover and distance to water. There is also weaker evidence for a negative relationship with sine-transformed aspect suggesting an avoidance of areas with higher levels of solar incidence (Table 2.3). All of the models, except for the thermal regulation model, showed a 28 significant positive relationship in predictive capabilities as determined by a Spearman's rank correlation (Table 2.4). Assessments of model performance need to consider the error associated with the predictions. Smaller errors indicate more consistent and better performing models. All the above results contributed to the assessment of model performance and identification of key habitat variables. Female reproductive success Region strongly influenced the breeding success of female white-tailed ptarmigan, with consistently higher reproductive success in the central (mean number of chicks = 3.7 ± 0.27 per female n = 77 females) region of the island than in the south (mean number of chicks = 2.0 ± 0.52 per female n = 21 females, GLIMMIX p = 0.56, s.e. = 0.19, df = 20, t-value = 2.97 p=0.01). The southern region also experienced almost complete reproductive failure in 1995 and 2004. Age structure and sex ratios The southern region of the island had a significantly higher proportion of yearling birds than the central region (south = 31/87 = 36% yearlings, central = 41/164 = 25% yearlings, n = 251, Fisher's exact test one-sided x 2 = 3.14, p = 0.05). Amongst females the difference in age structure was not as pronounced (n = 127, Fisher's exact test one-sided x 2 = 1.10, p = 0.20, Figure 2.2, Table 2.5). I also analyzed a subset of individuals (n=142) of known age (>lyear). I used GLIMMIX to examine the effect of region on the age of known age birds of both sexes. The response variable (age) had a poisson distribution and I entered individual identity and year as random factors. I specified a compound-symmetry covariance structure which had constant variance and covariance suitable to the correlation of observations of age over years. The specification of the covariance structure resulted in the appropriate normal distribution of residuals. This mixed model approach allowed for the isolation of the effect of region on age structure. In general, the 29 central region tended to have older birds than the southern region. However, the influence of region on the age of birds > 1 year was not statistically significant (range 1 - 5 years, (3 = 0.14, s.e. = 0.06, t-value = 2.22, p = 0.16). The ratio of females to males differed dramatically between the two regions (Figure 2.3). In the central region, females made up 58% of the marked population, whereas in the southern region, they made up 37% of the marked population (x2 = 10.17, p = 0.01). Regional differences in key habitat variables Habitat sampling plots were used to analyze differences in key habitat variables between the central (n=420 plots) and the southern (n=80 plots) regions of the island. Boulder cover did not differ between the two regions (median: south = 21.9%, central = 28.1%, Mann-Whitney U-test Z = -1.060, p = 0.29). Subshrub cover differed between the two areas with the southern portion having greater subshrub cover than the central area (median: south = 18.8%, central = 3.75%, Mann-Whitney U-test Z = -4.102, p < 0.001). The distance to water also varied significantly between the two areas with the central sampling plots consistently closer to water than the southern plots (median: south = 112 m, central = 38.5 m, Mann-Whitney U-test Z = -6.036, p < 0.001, Figure 2.4). Predator abundance Predator abundance (raptors, ravens Corvus cor ax, bears Ursus americanus, Canid sp.) did not differ significantly between the southern and central regions of the island. The southern region had an average of 2.5 predators per day based on a 14 hr day (307 predator detections/1755 person hours). In the central region of the island, there were 273 predator sightings over 1,968 person hours which results in an average of 2 predators per 14 hour day ((273/1968) *14hr). 30 DISCUSSION Habitat selection and key variables All models, except thermal regulation, performed well with no clearly superior model of fine-scale habitat selection. The relatively equal performance of all models provides strong evidence for a generalist approach to fine-scale habitat selection in Vancouver Island white-tailed ptarmigan and suggests the importance of tradeoffs in microhabitat choices. The examination of the P coefficients and associated standard errors revealed several key habitat variables (Table 2.3). In particular, I found strong support for the importance of three key habitat variables: percent boulder cover within a site, percent subshrub cover within a site, and the distance to water. There was weaker support for the importance of aspect in the generalist model, but this predictor did not contribute much to either the thermal regulation or global models. Resource selection data were collected during brood rearing which is an important period for the survival and training of juvenile birds (Allen and Clarke 2005, Hannon and Martin 2006). The selection for boulder cover represents a predator avoidance strategy for adult and juvenile birds. In addition to providing cover from predators, boulder cover could also provide shade and lower levels of solar incidence, thus aiding in the thermoregulatory demands of birds. Aspect also showed a consistently negative coefficient further supporting an avoidance of areas with high solar incidence. Alpine habitats have an abundance of microhabitat structural heterogeneity due to high levels of erosion. In all sampling plots (used and available, n = 500) the median boulder cover was approximately 25% (used median = 35%, available median = 20%). The preference for boulder cover, supports the importance of predator avoidance. White-tailed ptarmigan preferred moist areas close to water. Because white-tailed ptarmigan do not typically get the moisture they need for survival from surface water (Braun et al. 1993), distance to water may be a proxy measure for other habitat characteristics, such as insects. 31 Alpine snow pack remaining late into the summer provides the majority of moisture in white-tailed ptarmigan habitats. Moisture levels influence alpine vegetation communities and could influence the abundance and diversity of available food items. However, the preferred food species during the brood rearing period were the flowers of white and pink mountain heather, which are quite tolerant of dry conditions (Pojar and MacKinnon 1994). Subshrub cover was an important component in the habitat selection behaviour of white-tailed ptarmigan as birds showed a preference for sites with greater subshrub cover. On Vancouver Island, pink and white mountain heather begin to bloom at the end of June and the flowers, a preferred food, are still available through August. Moisture levels may have influenced the production of inflorescences on the ericaceous shrubs, however, I did not test for differences in the abundance of flowers. The fruits of other subshrub species may replace this food source later in the season (e.g., crowberry Empetrum nigrum, bear berry Arctostaphylos uva-ursi). There could be specialization on ericaceous shrubs throughout the year, but it is likely not specific to one species as in areas where willow is abundant. The difference in food selection habitats of Vancouver Island white-tailed ptarmigan, compared to mainland birds, highlights the importance of considering regional variation (Island - Mainland), and the inherent flexibility of fine-scale resource selection behaviour of this species. In this study, I demonstrated a behavioural gradient for white-tailed ptarmigan from specialist to generalist in white-tailed ptarmigan. Regional variation in population performance The regional differences I found in female reproductive success, age structure and sex ratios indicate much lower population performance in the southern region than in the central region of Vancouver Island. Female white-tailed ptarmigan exhibit strong age-dependent effects (Sandercock et al. 2005a). First-year female birds have significantly higher annual mortality, later laying dates, lower spring body condition, smaller clutches, and less re-nesting (Wiebe and 32 Martin 1998, Sandercock et al. 2005a). Given the influence of these factors on population performance, a trend of this magnitude likely represents a biologically meaningful pattern. The analysis of known age birds also showed a trend towards a higher mean age in the central region of the island than in the south. These trends, combined with the results of the age structure analysis including all birds, suggest a biologically meaningful difference exists in the age structures of white-tailed ptarmigan, suggesting overall lower population performance in the southern region. Further indication of lowered population performance in the southern region lies in the much lower ratio of females to males in the southern region. The difference in age structure could be an important influence on the reproductive success of females. It may be this difference that is responsible for the difference in reproductive success between the two regions. / Fragmentation and metapopulations The lower population performance in the southern region is consistent with predictions from metapopulation theory. Alpine areas on Vancouver Island are habitat patches for white-tailed ptarmigan surrounded by a matrix of unsuitable, low elevation habitat. Most metapopulation studies report that coarse scale measurement of patch area and the degree of isolation between patches is sufficient to predict patch occupancy (Thomas and Jones 1993, Hanski 1998, Moilanen and Hanski 1998). The probability of patch occupancy is related to, and a clear indication of, population performance. Alpine areas in the southern region of the island are more isolated and have less available habitat than the central region. However, explaining lowered population performance as a result of patch area and isolation does not provide sufficient insight into the mechanisms influencing the observed variation. White-tailed ptarmigan have a unique life history compared to the majority of study species contributing to metapopulation theory. Empirical data contributing to our understanding of metapopulation dynamics come primarily from species that, unlike white-tailed ptarmigan, are sensitive to environmental stochasticity and have evolved in naturally ephemeral habitat types in 33 which habitat suitability changes temporally (review in Harrison and Taylor 1997, tidal pools -Ebert et al. 2002, Pajunen and Pajunen 2003, meadows - Hanski et al. 1995, fire maintained scrub - Stith et al. 1996). Other empirical studies examine species that have experienced severe anthropogenic fragmentation of their natural habitat (e.g. Elmhagen and Angerbjorn 2001). Relatively little work has been done on the metapopulation dynamics of species which have evolved in naturally isolated and stable habitats (Moilanen and Hanski 1998, Martin et al. 2000, Elmhagen and Angerbjorn 2001). The lower population performance in the southern region suggests that patch area and isolation are: effective indicators of performance in a species which has evolved in naturally patchy environments that are still relatively undisturbed by human impacts. This is further evidence to highlight the importance and relevance of metapopulation theory to the conservation of white-tailed ptarmigan on Vancouver Island. The Pacific Northwest, including Vancouver Island, is experiencing significant increases in temperature as a result of global climate change (Service 2004). In many mountain areas, an increase in mean annual temperature has reduced snow packs and shifted low elevation plant and animal communities to higher elevations (Beniston 2003). Treelines are advancing to higher altitudes (Wardle and Coleman 1992, Meshinev et al. 2000, Kullman 2001), alpine plant ranges are shifting upwards in elevation from l-4m per decade (Grabherr et al. 1994, Parmesan 1996), and lowland birds are extending their distribution from lower mountain slopes to higher areas (Pounds et al. 1999). These will result in a decrease in the amount of alpine habitat. The rising tide of low elevation communities and treelines may result in smaller habitat patches for white-tailed ptarmigan and greater distances between populations. Predictions for southern Vancouver Island based on a mean annual temperature increase of ~3°C show an upward elevational shift in biogeoclimatic zones and a significant decrease of available alpine habitat (Hebda 1998). This loss of alpine habitat could result in greater separation of the south and central portions of the island, with the central area progressing, structurally and functionally, to 34 the current patterns observed in the south island. Given the findings of this study, greater isolation and smaller patches may result in a decrease in population performance that will have direct affects on the persistence of this unique subspecies of white-tailed ptarmigan. Regional variation of key habitat variables The distribution of key habitat variables may contribute to the observed regional variation in population performance. Predation influences the reproductive success of females as juveniles typically suffer high predation rates. If predation was greater in the southern region it could also influence the age structure, if birds had a low probability of survival to adult. However, the amount of available boulder cover did not differ between the two regions. Furthermore, predation risk was not significantly higher in the southern region. Thus predation of young or adult birds is unlikely to influence the difference in population performance. Food availability may influence population performance. The brood rearing period is a key stage for the reproductive success of white-tailed ptarmigan (Hannon and Martin 2006). The preferred food sources on Vancouver Island for white-tailed ptarmigan during the brood rearing period are the flowers of white and pink mountain heather. A greater availability of subshrub cover in the southern region suggests a greater availability of this key food source in the southern region; however, this finding was counter to the prediction that food availability during brood rearing is important to population success. Food availability may not be a limiting factor at this life-history stage. However, more experienced birds in the central area manage to find and select for this food source. The greater experience of the central birds (higher proportion of adults) in finding quality forage may offset any potential regional variation in population performance due to food availability. Furthermore, the resource selection data were collected during only one stage of the ptarmigan's annual cycle. Influence of food availability may be more pronounced or critical at other times of the year. 35 The distance to water was the only key habitat variable that correlated with population performance in the predicted direction. The southern region had a significantly greater distance to water than the central region. The summer of 2004 was particularly dry on Vancouver Island and white-tailed ptarmigan in the southern populations experienced complete reproductive failure. This further highlights the correlation of water and reproductive success. Moisture is at its annual low in alpine areas during the late summer when I conducted the intensive habitat selection study. Most available water comes from snow pack during this time and high summer temperatures and early snow melt can leave alpine areas desiccated in late summer. The southern region mountains are at lower elevation than central region mountains and the snow pack melts earlier in the season. Other studies have documented a preference in white-tailed ptarmigan for moist areas during this life stage (Choate 1963, Herzog 1977, Frederick and Gutierrez 1992). Perhaps the moist areas provide greater insect availability, which may be important for juvenile birds (Hudson 1986). Moist areas may also be cooler than drier areas and thus assist with thermoregulation. Vancouver Island white-tailed ptarmigan showed pronounced regional variation in population performance. This variation was consistent with predictions from metapopulation theory, reinforcing the importance of patch area and isolation. In this study, I also demonstrated the relevance of these factors to long-lived species that have evolved in naturally fragmented populations. Studies using molecular markers to examine the population genetic structure of Vancouver Island white-tailed ptarmigan populations will provide further insights into the persistence and population dynamics of this species. In this current study I also highlighted the behavioural flexibility of white-tailed ptarmigan. This species employs a generalist approach to fine-scale habitat selection on Vancouver Island in contrast to mainland populations which demonstrate a strong preference for willow throughout the year. Finally, I stressed cautions concerning the impact of climate change on resident alpine species. Climate change is raising 36 the elevation limit of flora and fauna which will result in smaller alpine habitat patches and greater isolation. Furthermore, moisture levels seem key to the brood rearing stage of ptarmigan. Warmer temperatures will decrease the snow pack and leave many alpine areas desiccated in late summer, a crucial reproductive stage for white-tailed ptarmigan. 37 Table 2.1. Habitat variables collected at each sampling plot. F - Food Availability; Pr - Predator Avoidance; Th - Thermal Regulation; Gen - Generalist; Gl - Global. Variable Description Range Model Topographic Variables slope aspect Slope angle measured in degrees with a clinometre 0 - 66° Orientation of steepest slope angle measured with a compass 0 - 360° Th, Gen, Gl Distance Variables distance to water Distance to nearest detectable surface water of any depth distance to cover Distance to nearest object greater than 30cm height that could provide cover to adult birds 0 - 637.m 0-24m Th, Gl Pr, Gl Ground Cover Variables boulder cover snow cover water cover shrub cover subshrub cover herb cover turf cover Percent of plot area covered by bedrock or boulders greater 0 - 100 % Pr, Gl than 30cm in diameter Percent of plot area covered by snow 0 -'75 % Percent of plot area covered by surface water of any depth 0-29 % Percent of plot area covered by woody plants greater than 0-100 % Pr, Th, Gl 30cm tall. Common species: Tsuga mertensiana, Pinus contorta Percent of plot area covered by ericaceous shrubs less than or 0 - 100 % F, Gl equal to 30cm tall. Common species: Cassiope mertensiana, Phyllodoce empetriformis Percent of plot covered by herbaceous plants. Common species: 0-43 % F, Gl Luetkea pectinata, Silene acaulis Percent of plot covered by grasses and sedges. Common species: 0-55% F, Gl Festuca occidentalis, Carex nardina oo Table 2 . 2 . Fine-scale habitat selection sampling effort. Each session is an independent, approximately 3 hr observation period, of a unique individual. Each plot is a 20 m sampling plot. Plots Region Sessions Used Available Total Central 23 148 272 420 South 6 28 52 80 Total 29 176 324 500 Table 2 . 3 . Coefficient estimates and standard errors for all iterations. Numbers in bold are significant at 10%. Iteration 1 Iteration 2 Iteration 3 Iteration 4 Average Food Availability Variable Estimate Error Estimate Error Estimate Error Estimate Error Estimate Error intercept -0.499 0.300 -0.500 0.294 -0.421 0.312 -0.451 0.338 -0.468 0.311 subshrub cover 0.316 0.132 0.355 0.139 0.254 0.150 0.504 0.151 0.357 0.143 turf cover 0.371 0.316 0.316 0.311 0.416 0.310 -0.149 0.378 0.238 0.329 herb cover 0.151 0.315 0.114 0.339 -0.445 0.346 0.008 0.344 -0.043 0.336 Predator Avoidance intercept -0.404 0.343 -0.387 0.346 -0.438 0.457 -0.405 0.377 -0.408 0.381 cover present y/n -0.099 0.193 -0.058 0.204 -0.220 0.350 -0.075 0.209 -0.113 0.239 distance to cover -0.003 0.026 0.000 0.026 -0.016 0.088 0.004 0.026 -0.004 0.041 boulder cover 0.284 0.170 0.162 0.167 0.488 0.173 0.196 0.191 0.282 0.175 shrub cover -0.023 0.210 0.171 0.210 0.383 0.223 0.302 0.210 0.208 0.213 Thermal Regulation intercept -0.510 0.323 -0.495 0.312 -0.570 0.337 -0.486 0.359 -0.515 0.333 aspect -0.518 0.412 -0.657 0.392 -0.358 0.438 0.016 0.505 -0.379 0.437 aspect*time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 distance to water 0.002 0.001 0.002 0.001 0.003 0.001 0.001 0.001 0.002 0.001 shrub cover -0.131 0.184 0.077 0.180 0.102 0.204 0.197 0.176 0.061 0.186 Generalists intercept -0.988 0.343 -0.852 0.334 -1.078 0.359 -0.848 0.377 -0.942 0.353 aspect -0.105 0.058 -0.096 0.060 -0.173 0.064 -0.054 0.063 -0.107 0.061 distance to water 0.002 0.001 0.002 0.001 0.003 0.001 0.001 0.001 0.002 0.001 boulder cover 0.478 0.157 0.311 0.158 0.563 0.173 0.386 0.172 0.434 0.165 subshrub cover 0.466 0.143 0.468 0.154 0.418 0.163 0.667 0.164 0.505 0.156 Global Model intercept -0.876 0.584 -0.670 0.563 -0.911 0.563 -0.819 0.570 chicks present? (y/n) -0.276 0.636 -0.364 0.621 -0.364 0.660 -0.335 0.639 aspect -0.460 0.420 -0.656 0.402 0.031 0.526 -0.362 0.449 aspect*time 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 distance to water ' 0.002 0.001 0.002 0.001 0.001 0.001 0.002 0.001 distance to cover 0.004 0.026 0.002 0.026 0.018 0.026 0.008 0.026 cover present y/n -0.101 0.194 -0.175 0.208 -0.017 0.219 -0.098 0.207 boulder cover 0.578 0.189 0.472 0.192 0.600 0.218 0.550 0.199 shrub cover 0.090 0.216 0.295 0.221 0.426 0.222 0.270 0.219 subshrub cover 0.523 0.148 0.537 0.160 0.706 0.176 0.589 0.161 turf cover 0.344 0.329 0.258 0.330 -0.113 0.402 0.163 0.354 herb cover 0.269 0.329 0.232 0.354 0.383 0.381 0.294 0.355 40 Table 2.4. Cross-validated Spearman-rank correlations (rs) between bin ranks and frequencies for individual and average model sets. Model Iteration Food Generalist Predation Thermal Global r s P r 5 P r s P r s P r s P 1 0.927 0.001 0.964 0.000 0.819 0.013 0.873 0.005 0.756 0.030 2 0.736 0.038 0.635 0.091 0.970 0.000 0.434 0.282 0.765 0.027 3 0.964 0.000 0.507 0.200 0.830 0.011 0.683 0.062 0.898 0.002 4 0.439 0.276 0.982 0.000 0.880 0.004 -0.366 0.373 na na Average 0.833 0.010 0.976 0.000 0.976 0.000 0.714 0.470 0.976 0.000 Table 2.5. Age structure of all ptarmigan in south and central portions of the island throughout the inventory (1995 - 1999) and intensive (2003 - 2004) stages of the study. Year South Central Total Adult Yearling Total Adult Yearling Total 1995 7 1 8 14 2 16 24 1996 12 6 18 20 6 26 44 1997 12 2 14 35 10 45 59 1998 9 7 16 24 7 31 47 1999 8 9 17 1 - 1 18 2003 5 5 10 10 4 14 24 2004 3 1 4 19 12 31 35 Total 56 31 87 123 41 164 251 42 a) 35 30 o a 3 cr T3 3 -c? b ) o b C —^» r/3 25 20 15 10 16 14 12 io H 8 6 4 2. 0 Food Predator • Thermal • Generalist •Global 3 4 5 6 Binned predicted scores Food • Predator • Thermal • Generalist •Global 3 4 5 6 Binned predicted scores Figure 2 .1 . a) Frequency of categories o f average scores for 'available' data for all models. Frequency values for individual models (n = 5) are depicted with unique symbols, b ) Standard error o f categories of average scores for 'available' data for all models. Standard errors for individual models (n = 5) are depicted with unique symbols. 43 100% 80% a o H—» 60% — 3 OH O 40% OH o 20% 0% south I central Both sexes south I central Females only south I central Males only Figure 2.2. Age category comparison for south and central regions. Black bars = adults > 2 years. White bars = yearlings. Comparisons are presented including data on both sexes, females only and males only. The numbers inside the bars represent the respective sample sizes. Data are from both stages of the study. 44 100% c -4—» a P. o P. t«—i o 80% 6 0 % 4 0 % 1 2 0 % 0 % 4-1 69 55 S o u t h Centra l Figure 2.3. C o m p a r i s o n o f sex ratios for south and central regions. B l a c k bars - females o f breeding age. W h i t e bars = males o f breeding age. T h e numbers ins ide the bars represent the respective sample sizes. D a t a are from both stages o f the study. 45 Figure 2.4. South and central region comparisons of key habitat variables. Values presented are median ± standard error. White bars = south, black bars = central. 4 6 R E F E R E N C E S Allen, T. and J. A. Clarke. 2005. Social learning of food preferences by white-tailed ptarmigan chicks. Animal Behaviour 70: 305-310. Anderson, D. P., M . G. Turner, J. D. Forester, J. Zhu, M . S. Boyce, H. Beyer and L. Stowell. 2005. Scale-dependent summer resource selection by reintroduced elk in Wisconsin, USA. Journal of Wildlife Management 69: 298-310. Artiss, T. and K. Martin. 1995. Male vigilance in white-tailed ptarmigan, Lagopus leucurus -mate guarding or predator detection. Animal Behaviour 49: 1249-1258. Baines, D., I. A. Wilson and G. Beeley. 1996. Timing of breeding in black grouse Tetrao tetrix and capercaillie Tetrao urogallus and distribution of insect food for the chicks. Ibis 138: 181-187. Beniston, M . 2003. Climate change in mountain regions: a review of possible impacts. Climate Change 59: 5-31. Bergerud, A. T., S. S. Peters and R. McGrath. 1963. Determining sex and age of willow ptarmigan in New Foundland. Journal of Wildlife Management 27: 700 -711. Black, J. M . 1996. Partnerships in birds: the study of monogamy. Oxford Ornithology Series. Oxford University Press, New York. Boyce, M . S., P. R. Vernier, S. E. Nielson and F. K. A. Schmiegelow. 2002. Evaluating resource selection functions. Ecological Modelling 157: 281-300. Bradbury, W. C. 1915. Notes on the nesting of the white-tailed ptarmigan in Colorado. Condor 17:214-222. Braun, C. E. , K. Martin and L. A. Robb. 1993. White-tailed ptarmigan (Lagopus leucurus). In: F. Gill and A. Poole. The Birds of North America, 68. Academy of Natural Sciences, Philadelphia, PA, and American Ornithologists' Union, Washington, D.C.: 22. Choate, T. S. 1963. Habitat and population dynamics of white-tailed ptarmigan in Montana. Journal of Wildlife Management 27: 684-699. Ebert, D., C. Haag, M . Kirkpatrick, M . Riek, J. W. Hottinger and V. I. Pajunen. 2002. A selective advantage to immigrant genes in a Daphnia metapopulation. Science 295: 485-488. Elmhagen, B. and A. Angerbjorn. 2001. The applicability of metapopulation theory to large mammals. Oikos 94: 89-100. Fielding, A. H. and J. F. Bell. 1997. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation 24: 38-49. 47 Frederick, G. P. and R. J. Gutierrez. 1992. Habitat use and population characteristics of the white-tailed ptarmigan in the Sierra Nevada, California. The Condor 94: 889-902. Giesen, K. M . and C. E. Braun. 1992. Winter home range and habitat characteristics of white-tailed ptarmigan in Colorado. Wilson Bulletin 104: 263-272. Gillies, C. S., M . Hebblewhite, S. E. Nielsen, M . A. Krawchuk, C. L. Aldridge, J. L. Frair, D. J. Saher, C. E. Stevens and C. L. Jerde. 2006. Application of random effects to the study of resource selection by animals. Journal of Animal Ecology 75: 887-898. Grabherr, G. M . , M . Gottfried and H. Pauli. 1994. Climate effects on mountain plants. Nature 369:448. Hannon, S. J. and K. Martin. 1992. Monogamy in willow ptarmigan - Is male vigilance important for reproductive success and survival of females. Animal Behaviour 43: 747-757. Hannon, S. J. and K. Martin. 2006. Ecology of juvenile grouse during the transition to adulthood. Journal of Zoology 269: 422-433. Hanski, I., T. Pakkala, M . Kuussaari and G. C. Lei. 1995. Metapopulation persistence of an endangered butterfly in a fragmented landscape. Oikos 72: 21-28. Hanski, I. 1998. Metapopulation dynamics. Nature 396: 41-49. Harrison, S. and D. A. Taylor. 1997. Empirical evidence for metapopulation dynamics. In: I. A. Hanksi and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 27-42. Hebda, R. 1998. Atmoshperic change, forests and biodiversity. Environmental Monitoring and Assessment 49: 195-212. Herzog, P. W. 1977. Summer habitat use by white-tailed ptarmigan in southwestern Alberta. Canadian Field-Naturalist 91: 367-371. Herzog, P. W. 1980. Winter habitat use by white-tailed ptarmigan (Lagopus leucurus) in southwestern Alberta, Canada. Canadian Field-Naturalist 94: 159-162. Hik, D. S. 1995. Does risk of predation influence population-dynamics - evidence from the cyclic decline of snowshoe hares. Wildlife Research 22: 115-129. Huberty, C. J. 1994. Applied Discriminant Analysis. Wiley Interscience, New York. Hudson, P. 1986. Red Grouse: the biology and management of a wild gamebird. The Game Conservancy Trust, Fordingbridge. Johnson, C. J., D. R. Seip and M . S. Boyce. 2004. A quantitative approach to conservation planning: using resource selection functions to map the distribution of mountain caribou at multiple spatial scales. Journal of Applied Ecology 41: 238-251. 48 Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65-71. Johnson, R. E. 1968. Temperature regulation in the white-tailed ptarmigan, Lagopus leucurus. Comparative Biochemistry and Physiology 24: 1003-1014. Jones, J. 2001. Habitat selection studies in avian ecology: a critical review. The Auk 118: 557-562. Keating, K. A. and S. Cherry. 2004. Use and interpretation of logistic regression in habitat selection studies. Journal of Wildlife Management 68: 774-789. Krebs, C. J. 1994. Ecology: the experimental analysis of distribution and abundance. HarperCollins College Publishers, New York. Kullman, L. 2001. 20th century climate warming and tree-limit rise in the southern Scandes of Sweden. Ambio 30: 72-80. Manly, B. F. J., L. L. McDonald, D. L. Thomas, T. L. McDonald and W. P. Erickson. 2002. Resource Selection by Animals: statistical design and analysis for field studies. Kluwer Academic Publishers, Dordecht, The Netherlands. Martin, K. 1995. Patterns and mechanisms for age-dependent reproduction and survival in birds. American Zoologist 35: 340-348. Martin, K. and L. Elliot. 1996. Vancouver Island white-tailed ptarmigan inventory progress report (1995-1996). Centre for Alpine Studies, Forest Sciences, University of British Columbia, http://www.forestry.ubc.ca/alpine/docs/wtpvi-l.pdf Martin, K. and M . L. Commons. 1997. Vancouver Island white-tailed ptarmigan inventory project: progress report. 1997 surveys. Centre for Alpine Studies, Forest Sciences, University of British Columbia, http://www.forestry.ubc.ca/alpine/docs/wtpvi-3.pdf Martin, K., P. B. Stacey and C. E. Braun. 2000. Recruitment, dispersal, and demographic rescue in spatially-structured white-tailed ptarmigan populations. Condor 102: 503-516. Martin, K., G. A. Brown and J. R. Young. 2004. The historic and current distribution of the Vancouver Island white-tailed ptarmigan (Lagopus leucurus saxatilis). Journal of Field Ornithology 75: 239-256. May, T. A. and C. E. Braun. 1972. Seasonal foods of adult white-tailed ptarmigan in Colorado. Journal of Wildlife Management 36: 1180-1186^ Meshinev, T., I. Apostolova and E. Koleva. 2000. Influence of warming on timberline rising: a case study on Pinuspeuce Griseb. in Bulgaria. Phytocoenologia 30: 431-438. Moilanen, A. and I. Hanski. 1998. Metapopulation dynamics: effects of habitat quality and landscape structure. Ecology 79: 2503-2515. 49 c Orians, G. H. and J. F. Wittenberger. 1991. Spatial and temporal scales in habitat selection. American Naturalist 137: S29-S49. Pajunen, V. I. and I. Pajunen. 2003. Long-term dynamics in rock pool Daphnia metapopulations. Ecography 26: 731-738. Parmesan, C. 1996. Climate and species' range. Nature 382: 765-766. Patten, M . A., D. H. Wolfe, E. Shochat and S. K. Sherrod. 2005. Effects of microhabitat and microclimate selection on adult survivorship of the lesser prairie-chicken. Journal of Wildlife Management 69: 1270-1278. Pojar, J. and A. MacKinnon. 1994. Plants of Coastal British Columbia: including Washington, Oregon and Alaska. Lone Pine Publishing, Vancouver. Pounds, J. A., M . P. L. Fogden and J. H. Campbell. 1999. Biological response to climate change on a tropical mountain. Nature 398: 611-615. Sandercock, B. K., K. Martin and S. J. Hannon. 2005a. Demographic consequences of age-structure in extreme environments: population models for arctic and alpine ptarmigan. Oecologia 146: 13-24. Sandercock, B. K., K. Martin and S. J. Hannon. 2005b. Life history strategies in extreme environments: comparative demography of arctic and alpine ptarmigan. Ecology 86: 2176-2186. Service, R. F. 2004. Water resources: as the west goes dry. Science 303: 1124-1127. Sih, A. and B. Christensen. 2001. Optimal diet theory: when does it work, and when and why does it fail? Animal Behaviour 61: 379-390. Stith, B. M . , J. W. Fitzpatrick, G. E. Woolfenden and B. Pranty. 1996. Classification and conservation of metapopulations: a case study of the Florida scrub jay. In: D. R. McCullough. Metapopulations and Wildlife Conservation, Island Press: 187-214. Storch, I. 1993. Habitat selection by capercaillie in summer and autumn: is bilberry important? Oecologia 95: 257-265. Summers, R. W., R. Proctor, M . Thorton and G. Avey. 2004. Habitat selection and diet of the capercaillie Tetrao urogallus in Abernethy Forest, Strathspey, Scotland. Bird Study 51: 58-68. Thomas, C. D. and T. M . Jones. 1993. Partial recovery of a skipper butterfly (Hesperia-Comma) from population refuges - lessons for conservation in a fragmented landscape. Journal of Animal Ecology 62: 472-481. Veghte, J. H. and C. F. Herreid. 1965. Radiometric determination of feather insulation and metabolism of arctic birds. Physiological Zoology 38: 267-275. 50 Verdolin, J. L. 2006. Meta-analysis of foraging and predation risk trade-offs in terrestrial systems. Behavioral Ecology and Sociobiology 60: 457-464. Walters, J. R., S. J. Daniels, J. H. Carter and P. D. Doerr. 2002. Defining quality of red-cockaded woodpecker foraging habitat based on habitat use and fitness. Journal of Wildlife Management 66: 1064-1082. Wardle, P. and M . C. Coleman. 1992. Evidence for rising upper limits of 4 native New-Zealand forest trees. New Zealand Journal of Botany 30: 303-314. Weeden, R. B. 1967. Seasonal and geographic variation in the foods of adult white-tailed ptarmigan. Condor 69: 303-309. Whittington, J., C. C. St Clair and G. Mercer. 2005. Spatial responses of wolves to roads and trails in mountain valleys. Ecological Applications 15: 543-553. Wiebe, K. L. and K. Martin. 1998. Age-specific patterns of reproduction in white-tailed and willow ptarmigan Lagopus leucurus andZ. lagopus. Ibis 140: 14-24. Wittenberger, J. F. and R. L. Tilson. 1980. The evolution of monogamy - hypotheses and evidence. Annual Review of Ecology and Systematics 11: 197-232. Wolfinger, R. and M . O'Connell. 1993. Generalized liners mixed models: a pseudo-likelihood approach. Journal of Statistical Computation and Simulation 4: 233-243. Zar, J. H. 1999. Biostatistical Analysis. Simon & Schuster, New Jersey. Zwickel, F. C. and J. F. Bendell. 1967. A snare for capturing blue grouse. Journal of Wildlife Management 31: 202-204. 51 CHAPTER 3 Heterozygote deficiencies and high levels of genetic diversity in Vancouver Island white-tailed ptarmigan INTRODUCTION Patterns of genetic variation within and between populations are often the results of geographic isolation and barriers to gene flow. In addition, local genetic structure is influenced by effective population size, gene flow, rates of genetic drift, population turnover, and levels of inbreeding (Harrison and Hastings 1996). All of these factors interact at relatively fine spatial and temporal scales to determine population structure and are reflected in allele frequencies at a given location. Patterns of genetic structure are also influenced by rates of drift and variation among individuals in reproductive success (Scribner and Chesser 1993), population bottlenecks or founder events following local extinction (Wade and McCauley 1988), and levels of inbreeding and assortative mating. Processes that affect genetic diversity and structure often operate at different time scales, influencing the patterns of genetic structure at different rates. For example, in small, isolated populations, drift may lead to the genotypic separation of populations over relatively short time periods. However, inbreeding in small populations or populations with positive assortative mating can reduce the observed heterozygosity of small populations in an even shorter time frame. Small, isolated populations are expected to have higher levels of relatedness between individuals than larger populations due to restricted mate choice. Inbreeding can lead to the production of offspring with reduced fitness due to the expression of partially recessive deleterious mutations in homozygous form (Keller and Waller 2002). Inbreeding depression reflects the negative consequences of increased homozygosity at genes that affect fitness and can influence the persistence of populations (Crnokrak and Roff 1999, Hedrick and Kalinowski 2000). Inbreeding depression has been related to a reduction in hatching rates, annual survival, and decreased male fitness within natural avian populations (Keller 1998, Daniels and Walters 2000, Hoglund et al. 2002). Furthermore, harsh environmental conditions can potentially magnify inbreeding depression, thus putting small populations exposed to both inbreeding and heightened stress at particular risk (Hedrick and Kalinowski 2000, Keller et al. 2002). In addition to affecting current fitness related traits, the loss of genetic variation can potentially hamper adaptation to new selective regimes (e.g. climate change). The benefits of maintaining genetic variation within populations include the retention of deleterious recessive 52 mutations in a heterozygous state and may provide adaptive potential in a changing environment (Amos and Balmford 2001). Genetic variability and adaptability of populations may be important in the ability of animals to respond to extreme environmental change, such as global warming (Lande 1999). Indeed, inbred endangered red-cockaded woodpeckers (Picoides borealis) did not adjust to climate change by laying earlier as other birds have and these inbred individuals have suffered reproductive costs as a result (Schiegg et al. 2002). Genetic load is the relative chance that an average individual will die before reproducing because of deleterious genes it possesses. Highly inbred populations may actually have low genetic load if inbreeding exposes recessive deleterious alleles to purging by natural selection (Keller et al. 1994), thus allowing the population to inbreed with minimal impact. This effect may be particularly pronounced in species which are exposed to harsh environments (Keller et al. 1994, Keller et al. 2002). However, a general pattern of purging has not been found in all species studied (Brewer et al. 1990) and may not be strong enough to eliminate inbreeding depression in animals (Ballou 1997). Predictions regarding levels of heterozygosity in small populations can be confounded by the effects of genetic bottlenecks, which if frequent, can lead to a reduction in the number of alleles and an increase in homozygosity in populations (Keller et al. 1994, Bouzat et al. 1998). Moreover, in small populations, a limited amount of immigration can have disproportionate effects on the vigour and persistence of local populations if heterozygous individuals have higher fitness (heterosis; Ingvarsson and Whitlock 2000, Whitlock et al. 2000). Heterosis has led to increased population performance and genetic rescue of adders (Vipera berus; Madsen et al. 1999) and big horn sheep (Ovis canadensis; Hogg et al. 2006). However, heterosis detected in Fi hybrids is often lost in subsequent generations leading to outbreeding depression, in which the hybrid progeny in .subsequent generations have lower performance or fitness than either parent (Allendorf and Luikart 2007). Outbreeding depression may result in reduced fitness in hybrids because of loss of local adaptation by ecologically mediated selection (Sage et al. 1986). General processes that shape genetic diversity and structure (e.g. drift, inbreeding) affect neutral and quantitative loci in similar ways, suggesting molecular markers can be used as indicators of overall genetic change (Reed and Frankham 2001, Gilligan et al, 2005). Thus, differentiation in neutral markers influenced by gene flow, drift, effective population size, and breeding patterns, plays a major role in determining population structure and reflects genetic processes. Molecular markers are also effective in detecting dispersal and population information on hard to access, low-density populations and provide otherwise unavailable 53 insights on demographic parameters and patterns of genetic variation partitioning within and among populations. Vancouver Island represents a more fragmented distribution of alpine habitats compared to much of the range of white-tailed ptarmigan {Lagopus leucura) throughout the western mountain Cordilleras of North America. The fragmented nature of these populations and narrow habitat requirements could have significant influence on population genetic structure by limiting dispersal and access to mates (Boone and Rhodes 1996). Compared to mainland and arctic populations of white-tailed ptarmigan, ptarmigan on Vancouver Island have relatively little continuous alpine area, all of it at relatively low elevation. Furthermore, these areas lack willow (Salix sp.), an important food source which is abundant throughout the rest of white-tailed ptarmigan distribution (see Chapter 2). Long-term radio-tracking studies have not documented long-distance dispersal of Vancouver Island white-tailed ptarmigan, suggesting a strong tendency for birds to remain in their natal patch (see Chapter 4). Limited dispersal could lead to a larger than expected influence of inbreeding or drift in Vancouver Island white-tailed ptarmigan given the low density, low effective population size and apparent lack of long-distance dispersal. In this study, I describe the optimization of heterologous microsatellite markers in white-tailed ptarmigan and the patterns of population genetics in the Vancouver Island white-tailed ptarmigan subspecies (L. I. saxatilis). I document one of the first empirical examples of extreme heterozygote deficiencies in unexploited, wild populations of a terrestrial vertebrate that is not likely a product of the Wahlund effect (Martinez et al. 2002, Ruiz-Garcia et al. 2005). Furthermore, these study populations have high levels of genetic diversity despite the consistent deficit of heterozygous individuals. Studies of white-tailed ptarmigan in Colorado have revealed a capacity for external recruitment that permits the persistence of small populations which experience stochastic conditions for breeding and survival (Martin et al. 2000); recruitment patterns were highly variable and recruits tended to be yearling birds. Similar patterns have been observed in white-tailed ptarmigan populations in Rocky Mountain National Park (Giesen and Braun 1993). White-tailed ptarmigan in Colorado have high dispersal capabilities with estimated demographic exchange likely occurring between populations within 5-10 km for males and 20-30 km for females (Martin et al. 2000). In addition to reporting extreme heterozygote deficiencies and high levels of diversity, I explore the potential mechanisms underlying this unique pattern of population genetics and present preliminary evidence of low levels of population structure in Vancouver Island white-tailed ptarmigan. 54 M E T H O D S Study species and study area White-tailed ptarmigan were located using playbacks of male territorial calls during the early breeding season (May - late June), and chick distress calls later in the season (late June -October). Individuals were captured using a noose pole (Zwickel and Bendell 1967) and sampling took place during two time periods: inventory and intensive. The inventory stage of the research was carried out from 1995 - 1999 and was followed by an intensive sampling stage, May to August 2003 - 2004. Samples were collected from 15 different mountains across Vancouver Island (Table 3.1). These mountains were grouped into seven different populations based on geographical features that represent potential population boundaries (Chapter 1, Figure 1.1). I then grouped these seven populations into three major regions on Vancouver Island; 1) the southern region, which includes the South, South West and Beaufort populations, 2) the central region, which encompasses the three central populations and 3) the northern region (Table 3.1). See Chapter 1 for more details on study species and area. Blood collection procedures varied slightly between the inventory and intensive stages of research. During the inventory stage, blood samples were collected from the brachial vein and placed in a vacuum tube containing EDTA. During the intensive stage of research, blood samples (approx. 0.1 - 0.5 ml) were collected into an EDTA coated syringe and then were placed into a microtube containing 80% ethanol and mixed gently. During both stages, blood samples were kept at ambient temperature in the field and stored at -20°C once they arrived at the lab. Only adult individuals were genotyped to prevent the sampling of families that could potentially bias population allele frequencies by over-representation of individual lineages. Primer optimization I tested the effectiveness of 34 primers from various genera (Lagopus sp., Centrocercus sp., Tetrao sp.) that could potentially cross-amplify to white-tailed ptarmigan based on observed patterns of genetic variation from other studies (Piertney and Dallas 1997, Piertney et al. 1998, Piertney and Hoglund 2001, Caizergues et al. 2001, Taylor et al. 2003; Tables 3.2 and 3.3). DNA was isolated from the inventory samples using Puregene DNA isolation kits for whole blood following the manufacturer instructions (Gentra Systems Inc., Minneapolis, MN). DNA from the intensive stage of research was extracted from samples using a phenol-chloroform extraction procedure (Carter 2000). Polymerase chain reactions (PCR) were performed on all samples in a total volume of 10 pl using an MJ Research PTC-100 thermal cycler (Bio-Rad Inc., Hercules, CA). Each reaction mix contained 10 ng template DNA, IX Buffer (Roche 55 Diagnostics, Mississauga, ON), 0.2 mM dNTP, 0.5 pmol primer (forward and reverse), 0.3 pmol of M l 3-29 Infrared Label Primer and 1.5 units Taq (Roche Diagnostics, Mississauga, ON; Oetting et al. 1995). PCR profiles followed those outlined in the respective papers (Table 3.2 and 3.3). PCR products were first visualized on 2% agarose gels. Those primers that amplified product (n = 24 of 34 tested) were subsequently run on 6% polyacrylamide gels on a LiCor 4200 automated sequencer (LiCor Inc., Lincoln, NE) and alleles were sized by comparison with size standards using the RFLP software program (LiCor Inc. Lincoln, NE). Observed and expected heterozygosity levels were calculated using ARLEQUIN (ver. 3.01; Excoffier et al. 2005), following Guo and Thompson (1992). The unbiased estimator of Wright's inbreeding coefficient Fis was calculated (Weir and Cockerham 1984) and significance was tested by permutation (FSTAT ver. 2.9.3.2; Goudet 1995). Amplification products from each of the 10 microsatellite markers used to assess population genetic structure were sequenced from white-tailed ptarmigan (n = 30 individuals) to verify the cross-species amplification of microsatellite fragments (Table 3.4). Homozygous PCR products were purified for each primer using Qiaquick PCR amplification kits (QIAGEN, Mississauga, ON) and sequenced using SequiTherm E X C E L II Long-Read DNA sequencing Kits-LC (Epicentre Technologies) on a LiCor 4200 automated sequencer (LiCor Inc. Lincoln, NE). Nucleotide sequences were aligned using ESEE3S software (ver. 2.00a; Cabot and Beckenbach 1989) and then edited manually. Sequences of each locus were compared to the original sequences from GenBank to determine similarity in flanking regions and repeat motifs. Alignments were determined by minimizing the number of mismatches and gaps assumed in the sequences and were compared using the 'match' option in the ESEE3S program. In two cases (BG4 and BG6), the original sequences were not available. In these instances I confirmed the presence of the correct repeat motif and then randomly assigned one of my sample sequences as the baseline and compared the match of the remaining samples to the baseline sample. Population analyses Departures from Hardy-Weinberg equilibrium were addressed with 2 tests implemented in GENEPOP (web ver. 3.4; Raymond and Rousset 1995). The first test accounted for both heterozygote deficit and heterozygote excess. The second test examined the specific alternative hypothesis of a heterozygote deficiency (Hd) using the score (U) test of Rousset and Raymond (1995) available in GENEPOP. Significance values were computed for each locus by unbiased estimates of Fisher's exact test using the Markov chain method through 1000 iterations (Guo and Thompson 1992). Linkage disequilibrium was examined using GENEPOP in which each 56 locus pair is compared across populations using Fisher's method. P-values from multiple tests were assessed for significance using sequential Bonferroni correction (Rice 1989) adjusting by the number of populations and loci. Average number of alleles (A), allelic richness corrected for the minimum sample size of individuals (An), and allele frequencies were calculated using FSTAT (ver. 2.9.3.2; Goudet 1995). These values were averaged for Vancouver Island white-tailed ptarmigan and compared to those reported for other microsatellite studies involving Tetraonids (Piertney and Dallas 1997, Piertney et al. 1998, Oyler-McCance et al. 1999, Caizergues et al. 2003a, Caizergues et al. 2003b, Larsson et al. 2003, Johnson et al. 2004). Populations that have recently experienced a decline in effective population size (N )^ will show reduced numbers of alleles (A) and, more slowly, a decrease in heterozygosity (HE) at polymorphic loci (Piry et al. 1999). To make inferences regarding population decline, I followed the method of Cornuet and Luikart (1996). Since A is reduced quicker than H E , in a population that has recently experienced a bottleneck, the observed H E will be higher in populations that recently experienced a bottleneck than the mutation-drift equilibrium heterozygosity (H e q) expected from the observed number of alleles. Populations without a recent change in N e will be in mutation-drift equilibrium where the H e q will be equal to the Hardy-Weinberg heterozygosity (HE). This expected heterozygosity (H e q) was calculated in the ( BOTTLENECK program (ver. 1.2.02; Piry et al. 1999) through simulation under three different mutation models: 1) the infinite-alleles model (IAM); 2) the stepwise mutation model (SMM) and 3) the two-phase model (TPM). The IAM assumes that every mutation creates a new allele never before present in the population. The S M M tries to better represent microsatellite mutations wherein single repeat units are added and deleted with near equal frequency (Valdes et al. 1993). However, the actual mechanisms of microsatellite evolution are much more complicated than the simple S M M (Li et al. 2002). The TPM is a variant of S M M that allows a certain proportion of mutations to involve a greater number of repeat units. For the TPM, I assumed 70% stepwise mutations and 30% multistep mutations. All models were run with 1000 simulation iterations. A "recent" bottleneck is defined as within approximately the past 2N e -4N e generations, depending on several factors such as the mutation rate and severity of the bottleneck (Cornuet and Luikart 1996). I used the Wilcoxon sign-rank test as suggested by Piry et al. (1999) to test the null hypothesis of no heterozygosity excess on average across all loci. I investigated geographic genetic population structure using a Bayesian clustering approach implemented in the program STRUCTURE (ver. 2.1; Pritchard et al. 2000). This method estimates the most appropriate number (k) of populations needed for interpreting the observed genotypes. 57 The most likely value of k is assessed by comparing the likelihood of the data for different values of k. I assessed P(K |X ) for a maximum of 9 populations. The model with the number of populations with the highest posterior probability was identified and the proportional membership of the genome of each ptarmigan was used to determine similarity or distinctiveness among sampling locations. This software performed well at low levels of population differentiation and worked well for inferring the number of population clusters with FST levels around 0.02 - 0.03 (Latch et al. 2006). The program is started from a random configuration and follows a series of steps through parameter space. Each step depends on the parameter values at the previous step. This procedure results in correlations between the state of the Markov Chain at different points during the run; however, by running the simulation long enough, the correlations become negligible. The burning length refers to how long the simulation is run before collecting data and is set long enough to minimize the effect of the starting configuration. Iset a burning period of 50,000 iterations to ensure all key summary statistics had converged and I obtained probability estimates using 50,000 Markov Chain Monte Carlo repetitions. A dendrogram of relationships among populations was constructed from Nei's D matrices. Calculations for the dendrogram were carried out in the program P H Y L I P (ver. 3.6; Felsenstein J http://evolution.geneticas.washington.edu/phylip.html). Dendrograms were joined using the neighbour-joining algorithm (Saitou and Nei 1987) as implemented by P H Y L I P . S E Q B O O T was used to bootstrap allele frequencies 1000 times (Felsenstein 1985). This procedure creates a new data set by sampling N characters randomly without replacement, so the resulting data set has the same size as the original, but some characters have been left out and others duplicated. The random variation of results from analyzing the boostrapped data sets are statistically typical of the variation observed by collecting new data sets. G E N D I S T was used to create multiple Nei's D matrices, and neighbour-joining trees (1000) were produced from lower triangulation matrices of D using N E I G H B O R . A consensus tree was built with C O N S E N S U S and visualized using T R E E V I E W . R E S U L T S Primer optimization Twenty-four of the 34 primers tested for their use in white-tailed ptarmigan showed little or no polymorphism (Table 3.2). The remaining 10 showed heterozygosity ( H E ) levels ranging from 0.687 to 0.936 (mean = 0.789) with a range of 7 to 32 alleles at each locus (A; mean = 13.9; Table 3.3). All 10 loci showed significant departures from HWE (p < 0.005), and all but 58 three loci (LLST7, BG16, BG18) demonstrated a significant heterozygote deficit (HJ). Six of the 10 loci also had Fis values significantly greater than zero (p < 0.005). Across 10 loci Fis = 0.20 and the lower limit of the boostrapped 95% confidence interval was greater than zero indicating high levels of Fis over all loci. Linkage disequilibrium probability tests calculated based on contingency tables for all 10 loci pairs across all populations revealed only 1 locus-pair (BG16 and BG18) of 45 possible comparisons that was significant at the 5% level after sequential Bonferroni corrections. Of the 315 possible comparisons of locus pairs on a population level there were no locus-pairs that exhibited significant linkage disequilibrium. Given these observations, it is reasonable to assume independence among loci in this data set. All loci showed high levels of correspondence when matched with GenBank sequences (63% -98% match in ESEE3S software) confirming that all primers were amplifying their target regions of the genome (Table 3.4). Furthermore, the observed range of allele sizes in Vancouver Island white-tailed ptarmigan falls within the expected sizes reported in the original primer optimization studies for all but two primers (LLSD10 and BG14; Table 3.3). The Vancouver Island white-tailed ptarmigan microsatellite data set also produced similar numbers of unique alleles as reported in the original optimization studies in all but 2 primers (LLST7 and SGCA6; Table 3.3). Within population analyses Among populations the average number of alleles, corrected for sample size (An), ranged from 4.5 to 5.1 (Table 3.5). Observed heterozygosity (Ho) in populations ranged from 0.56 -0.67 and Fis values ranged from 0.14 (South) to 0.27 (South-West). Fis values were significantly greater than zero in five of seven populations. All populations showed significant departures from the expectations of Hardy-Weinberg equilibrium and significant heterozygote deficiencies (HJ). Interspecific comparisons The average number of alleles (A) and allelic richness (An) in Vancouver Island white-tailed ptarmigan were similar to the average of the corresponding values reported for other grouse species (Table 3.6). The observed levels of heterozygosity (Ho) in Vancouver Island white-tailed ptarmigan (0.60) were similar to the average Ho reported for other grouse (0.64). Vancouver Island white-tailed ptarmigan differed from other grouse in three important genetic measures. First, Vancouver Island white-tailed ptarmigan had higher than average levels of expected heterozygosity ( H E ; 0.78) compared to other grouse (mean = 0.67, range 0.48 to 0.87). Second, the mean Fis value of 0.21 was much higher than those reported for other grouse 59 species (mean = -0.002, range -0.06 to 0.11). Finally, there are very few cases of significant departures from HWE for other grouse populations, whereas all Vancouver Island white-tailed ptarmigan populations showed significant departures from HWE and heterozygote deficiencies. Nonrandom association of alleles The comparisons of observed and expected numbers of heterozygotes at each locus under Hardy-Weinberg expectations showed that heterozygote deficiencies occurred frequently in Vancouver Island white-tailed ptarmigan. In 55 of 70 (79%) locus-population comparisons, the observed number of heterozygotes was less than expected. I found significant heterozygote deficits in all 7 populations when all loci were combined (p < 0.005). I measured the extent of nonrandom mating within populations using the inbreeding coefficient (Fis) for each locus-population combination. The majority (55 of 70 = 79%) of F i S values for all populations across all loci were greater than zero. Overall, Fis values (combined across loci) for each population ranged from 0.14 to 0.27 (mean = 0.21) and 5 were significantly different from zero based on 95% confidence intervals that were greater than zero (Table 3.5). This demonstrates that nonrandom associations of alleles frequently occur in the different populations. Assuming the entire heterozygote deficiency is due to null alleles, I used two formulae to estimate the frequency of null alleles in each population (rc - Chakraborty et al. 1992 and n, -Brookfield 1996). The rc estimator assumes that null homozygotes are not present in the sample whereas rb allows such genotypes to occur. Values for rc averaged across loci for each population range from 0.06 (South) to 0.13 (South West) with mean 0.10 ± 0.01. Corresponding values for rb for each population range from 0.07 (South) to 0.17 (South West) with mean 0.13 ± 0.01. These estimates suggest that if heterozygote deficiencies are due entirely to null alleles then such alleles are present at relatively high frequencies (approximately 11 - 12%) in Vancouver Island white-tailed ptarmigan populations. However, as emphasized by Brookfield (1996), this assumes that all of the heterozygote deficiency is due to null alleles; which seems unlikely (see below). Amplification failed for a small number of individuals at 1-3 loci despite repeated attempts using new sample dilutions, whereas they were successfully genotyped at the other loci. Of the 52 (of 1330 possible) individual-locus amplifications that failed all but one (98%) were samples from the earlier inventory stage of the research and had been in storage up to seven years (as DNA) before attempts were made to genotype the samples. Three loci LLSD10, LLST7, and 60 BG14 accounted for 75% of the failed amplifications. Two of these 3 loci, LLST7 and BG14, did not show significant Fis values. Sample sizes were not large enough to test the presence of fine-scale (within-population) genetic structure in all populations in order to address the possibility of the Wahlund effect on observed levels of Fi S . However, fine scale subpopulation structure did not appear to exist in the largest sampled populations. I compared the samples from 2 mountain peaks separated by approximately 5 km of continuous alpine habitat in the Central East population for 2003 (Mt. Albert Edward, n = 8 individuals and Mt. Jutland, n = 5). Sampling within only one year (2003) eliminates the effect of year on genetic variation. There were no significant differences between these two mountains in allelic (x2 = 16.91, p = 0.66) or genotypic distributions (%2 = 13.99, p = 0.83 F S T = -0.02). I also pooled the samples from Mt. Albert-Edward and Mt. Jutland to test for annual differences by comparing genotypes of birds captured in 2003 (n = 13) to those in 2004 (n = 17). These two years showed significant differences in both allelic (x2 = 51.39, p < 0.001) and genotypic distributions (x = 39.00, p = 0.007 F S T = 0.02), suggesting that, in this region, observed heterozygote deficiencies and high levels of Fis may be the result of temporal variation. However, independent analyses of both 2003 and 2004 samples also demonstrated significant heterozygote deficiencies across all loci (2003 p < 0.001, 2004 p < 0.001) and high Fis values (2003 Fis= 0.16, 2004 F i S = 0.19). I also examined annual variation in distribution of alleles and genotypes on Mt. Arrowsmith, comparing 1997 (n = 8), 1998 (n = 7) and 2003 (n = 6), and I found significant differences in both allelic and genotypic distributions between the years. Private alleles Each population sampled also contained a number of private alleles. Between 1.6% and 10.1% (mean = 6.1%) of all alleles detected within population samples were population specific (Table 3.7). In total, 33 population specific alleles were identified. Of these, 12 (36%) occurred at frequencies of > 5%. The number of individuals sampled per population was not related to the proportion of population specific alleles detected (Spearman's rank correlation, p = 0.13). Thus, the populations contained a distinct subset of the overall genetic variation resolved by the microsatellite markers used. Bottleneck The B O T T L E N E C K results under the IAM indicated an excess of heterozygosity at 3 of the 7 populations in Vancouver Island white-tailed ptarmigan compared to that expected at mutation drift equilibrium (Table 3.8). Potentially, this could suggest the populations showing significant 61 excess recently went through a population bottleneck. However, under S M M and TPM, the Wilcoxon sign-rank test detected different patterns. S M M showed a deficit in 2 sample populations and TPM demonstrated excess heterozygosity in only 1 population and a deficit in the Vancouver Island white-tailed ptarmigan test. Taken together, these results are equivocal regarding the occurrence of past changes in population size, but rather point to the likelihood that mutation-genetic drift equilibrium exists in these populations. Population structure According to the Bayesian clustering analysis, the most probable number of populations, or clusters, for interpreting the observed genotypes was 6 (i.e. the Ln probability of the data was maximum with 6 clusters). However, the assignment of individuals to clusters did not coincide well with their population of origin because the proportion of membership of each predefined sample population in each cluster was fairly evenly spread across the clusters (Table 3.9). I carried out a similar analysis using each individual's mountain of origin as the predefined locality and tested k = 1 - 15. In this case, the analysis suggested the most probable number of clusters for interpreting the observed genotypes was 5. This analysis also resulted in fairly even distribution of the proportion of membership of each predefined mountain across the 5 clusters (data not presented). Therefore, the most probable number of clusters based on observed genotypes is 5 or 6. However, the clusters cannot be used to reliably group the predefined sampling populations. The neighbour-joining tree based on Nei's D resolved several partitions of the populations (Figure 3.1). One group includes the Central West, Central East and Northern populations. The Central South joins this group next. This combined group is separated from all populations in the southern region. This pattern has moderate support with the strongest break separating the Central West, Central East and North populations from the remaining populations. Overall, this suggests fairly weak population structuring between the 7 populations. DISCUSSION All 10 primers selected for the analysis of the population genetics of white-tailed ptarmigan showed sufficient genetic variation for the detection of population genetic patterns. Also, the sequences matched using ESEE3S provided clear evidence that the microsatellite primers used amplified the specified target regions of the genome. Furthermore, the primers also showed good congruence to the values reported in the original publications in number and size range of alleles. The sequence matches and similarities in number of alleles and allele size allowed confidence in the analysis and interpretation of the observed genotypic polymorphisms. In this 62 study, I confirmed the cross-species applicability suggested by several of the original primer developers (Caizergues et al. 2001, Taylor et al. 2003). I provided the first description of the population genetics of white-tailed ptarmigan and revealed high levels of diversity and low levels of genetic isolation between populations of Vancouver Island white-tailed ptarmigan (see Chapter 4). Populations also showed generalized heterozygote deficiencies. This combination of high levels of diversity, limited population structure, and heterozygote deficiencies is a unique pattern of population genetic structure in published reports of wild terrestrial vertebrates. Nonrandom association of alleles Three possible explanations exist for positive Fis values, aside from inbreeding. They are: 1) the presence of non-amplifying (null) alleles, 2) short allele dominance or 3) undetected genetic structure within populations which results in heterozygote deficiencies in samples taken from such populations. Explanations 1 and 2 address the potential artefactual sources of heterozygote deficits. The presence of alleles that are not amplified because of nonmatching primer sequences (null alleles) can be suspected for microsatellite loci, particularly when using heterologous primers (Pemberton et al. 1995). The third explanation is known as the Wahlund effect, which causes reduced heterozygosity in sample populations due to undetected subpopulation structure. If two or more subpopulations with different allele frequencies are combined then overall heterozygosity will be reduced and Fis will increase. Although I could not test for the presence of null alleles using parent-offspring comparisons, several lines of evidence suggested the observed heterozygote deficiency was not due to null alleles. Individual samples that fail to amplify at a specific locus, despite repeated attempts, suggest the presence of a null allele. The majority of Vancouver Island white-tailed ptarmigan samples (all but one) that did not amplify at one or more loci were from the inventory stage of research. This suggests that failure to amplify is more likely due to experimental difficulties such as DNA degradation.' Furthermore, the 2 loci that accounted for over 40% of the non-amplified samples did not have F] S values significantly greater than zero. If high Fis was the result of null alleles, these 2 loci should also show high Fis values. Another, more putative, source of nondetection of microsatellite alleles is short allele dominance. In heterozygote individuals that present a short and a long allele it is possible that only the short allele is detected and scored (Wattier et al. 1998). In all heterozygote Vancouver Island white-tailed ptarmigan individuals that showed a 'long-size' allele and a 'short-size' allele, both were present with similar band intensity. I tested the reliability of this conclusion to ensure that long alleles were not suppressed during PCR and thus not scored by examining allele 63 intensity levels at each locus. Two types of individuals were amplified at each locus: heterozygotes for two short alleles and heterozygotes for two long alleles. Each combination was amplified in the same PCR reaction and analyzed on 6% acrylamide gel. I did not observe a decrease in band intensity for the alleles of the 'long' heterozygotes compared with the alleles of the 'short' heterozygotes in any locus-allele combination. Thus, there is no evidence of the suppression of long alleles at any of the 10 loci used to investigate the population genetics of white-tailed ptarmigan. Individuals within a sample population often came from different mountains. It is possible, but unlikely, that fine-scale structuring between mountains could be the cause of the observed heterozygote deficiencies (Wahlund effect). Analyses of within sample populations did not detect fine-scale spatial subpopulation structure. Although sample size was limited in these analyses, Fjs values were likely not the result of undetected spatial population structure. However, if there is high population turnover between years, it is possible for the Wahlund effect to apply on a temporal scale if samples come from multiple years. Heterozygote deficiency could be observed because of undetected genetic variation between years. The tests between years within two mountain areas suggested significant differences in allelic and genotypic distributions. However, F S T values are very low between years arguing for little substructure between years. Examination of each year independently revealed Fis levels significantly greater than zero and significant heterozygote deficiencies. Overall, these results provide strong evidence the observed heterozygote deficiencies are not due to the Wahlund effect on either a spatial or temporal scale. Final evidence that,high F ) S and heterozygote deficiencies are real population genetic trends in Vancouver Island white-tailed ptarmigan comes from the consistency of the pattern. All loci but 2 (BG16 and BG18) showed positive Fis values with 6 of these significantly greater than zero, and an overall average Fis, with.95% confidence intervals greater than zero. Furthermore, all 7 populations also showed positive Fis values, with 5 of these significantly greater than zero. Finally, all 7 populations have a significant deficit of heterozygotes. This provides unequivocal evidence of a real and consistent heterozygote deficiency in Vancouver Island white-tailed ptarmigan. Heterozygote deficiencies are fairly common in molecular studies of fish (e.g. Castric et al. 2002). However, heterozygote deficiencies are quite rare in microsatellite studies of terrestrial vertebrates and are often explained as a product of the Wahlund effect or the deficits are observed in highly managed and exploited populations. For example, the Wahlund effect is the 64 most likely explanation of heterozygote deficiencies in microsatellite studies of Massasauga rattlesnakes (Sistrurns c. catenatus; Gibbs et al. 1997), alpine marmots {Marmota marmota; Goossens et al. 2001) and spectacled bears (Tremarctos ornatus; Ruiz-Garcia et al. 2005). Significant heterozygote deficiencies have been observed in highly managed and exploited populations of red deer (Cervus elaphus; Martinez et al. 2002). Finally, heterozygote deficiencies and high Fis have been documented in 2 species of hyrax (Heterohyrax brucei, Procavia Johnston; Gerlach and Hoeck 2001); however, both hyrax species had very low levels of genetic diversity. Furthermore, 6 of 8 microsatellite loci used to infer patterns of population genetics in the hyrax study (Gerlach and Hoek 2001) had very few alleles (1 - 4) and, given the small number of alleles, are questionable in their ability to resolve patterns of genetic structure. Bottleneck The results from Cornuet and Luikart's (1996) tests did not support the hypothesis of a recent bottleneck. The interpretations of results from bottleneck tests using microsatellite polymorphism data are dependent on the assumed models (infinite alleles model - IAM, stepwise mutation model - S M M or two phase model - TPM). Many microsatellite loci appear to fit a variant of the S M M (Dirienzo et al. 1994). Thus, S M M seems to best describe the evolutionary dynamics of most microsatellite loci (Shriver et al. 1993) and S M M or TPM analyses should be more accurate than those assuming the IAM. However, there is no evidence of a recent bottleneck in Vancouver Island white-tailed ptarmigan regardless of the model employed and allele frequency distributions do not suggest S M M is the most applicable model for this study (B. Fedy, unpubl. data). Heterozygote deficiencies An apparent paradox exists in the population genetic structure of Vancouver Island white-tailed ptarmigan. This subspecies has an extreme deficiency of heterozygotes, however, diversity levels are higher than average reported for other species of Tetraoninae. Small population size alone cannot account for heterozygote deficiencies because random mating, even in a small isolated population, would not lead to lowered individual multilocus heterozygosity compared to random expectations. Population isolation and lack of gene flow can lead to increased levels of Fis, and small populations are particularly susceptible to these influences. Severe barriers to gene flow will result in a decrease in heterozygosity and, ultimately, a decrease in diversity. Two lines of evidence show that population isolation is only one process influencing the heterozygote deficiencies in Vancouver Island white-tailed ptarmigan. First, the brief analysis of population structure suggests insufficient isolation for this 65 to be the primary cause of the heterozygote deficiency. Second, in isolated populations drift should result in decreased levels of diversity. The high levels of diversity observed across all populations further suggests that strong geographic isolation is not the only factor influencing the levels of observed heterozygosity. The processes of inbreeding (increasing Fis) and drift (decreasing diversity) are likely working on different time scales to influence the genetic population structure of Vancouver Island, white-tailed ptarmigan. The loss of diversity due to drift is a slower process than inbreeding and works on an evolutionary time scale to reduce heterozygosity and genetically discriminate populations (Lande 1985). Alternatively, inbreeding works on an ecological time scale and can accumulate in small populations within a few generations. My sampling may have captured a snapshot of a group of populations that were in the process of becoming more isolated than in the recent evolutionary past. Significant geographic isolation may have existed long enough between populations to result in an increase in inbreeding but not long enough for drift to result in strong population differentiation and a decrease in diversity. This scenario predicts further isolation of Vancouver Island white-tailed ptarmigan populations in the future and an even greater increase in inbreeding. This could eventually threaten the existence of these populations through potential detrimental effects of inbreeding depression and decreased gene flow. Positive assortative mating refers to a situation in which similar phenotypes mate more commonly than expected by chance and can lead to a progressive increase in the number of homozygotes. Inbreeding is positive assortative mating in its most extreme form. Positive assortative mating is common in birds (Delestrade 2001, Jawor et al. 2003, Tryjanowski and Simek 2005), and white-tailed ptarmigan in Colorado demonstrate positive assortative mating by age (Hannon and Martin 1996). Modelling efforts have shown that positive assortative mating can increase inbreeding while still conserving genetic diversity (Rosvall and Mullin 2003) and could potentially explain the observed genetic patterns in Vancouver Island white-tailed ptarmigan. However, if positive assortative mating exists in this subspecies, and is the primary factor influencing low levels of observed heterozygosity and high diversity, then similar patterns should exist in other white-tailed ptarmigan populations and/or related grouse. This pattern did not exist in the Yukon samples of white-tailed ptarmigan, though sample size limited the level of inference. To my knowledge, this pattern of high inbreeding and heterozygote deficiency combined with high levels of diversity has not been reported for any other monogamous grouse species or documented in other bird species with significant positive 66 assortative mat ing. L i m i t e d access to mates is a more pars imonious explanat ion than posi t ive assortative mat ing because it o n l y impl i e s a l imi ta t ion to dispersal and mat ing opportunities, rather than a change i n patterns o f mate choice. V a n c o u v e r Island white- ta i led ptarmigan differ f rom other populat ions o f white- tai led ptarmigan p r i m a r i l y i n their density and landscape dis tr ibut ion. V a n c o u v e r Island white- ta i led ptarmigan are at l ower densities than other populat ions o f white- ta i led ptarmigan (Freder ick and Gut ier rez 1992). These l o w densities, when coupled w i t h l imi ted dispersal , restrict mate choice and a l l o w inbreeding to accumulate re la t ive ly q u i c k l y . In a more continuous group o f populat ions i n Co lo rado , dispersal was found to be per iodic and almost episodic i n nature ( M a r t i n et a l . 2000). Occas iona l and infrequent recruitment to populat ions across V a n c o u v e r Island cou ld act to main ta in the h igh levels o f diversi ty , mi t iga t ing the effects o f drift and l i m i t i n g strong popula t ion structure. T h e h igh levels o f Fis and heterozygote deficiencies are most l i k e l y mainta ined b y a process o f per iodic and l imi t ed successful recruitment sufficient to main ta in h igh divers i ty but not frequent enough to decrease inbreeding, w h i c h accumulates q u i c k l y o n an eco log ica l t ime scale. 67 Table 3.1. Sampling locations for white-tailed ptarmigan D N A samples across Vancouver Island, British Columbia. The number of individuals sampled (n) and the range of years over which birds were sampled for each mountain. Co-ordinates are based on the British Columbia Albers projection. B C Albers Region Population Mountain Name n Easting Northing Years South Central North s El Capitan 5 1130544' 439340 1995-1998 s Mt. Moriarty 7 1113012 459101 1997-2004 s Mt. McQuillan 4 1102028 455744 1995-1997 s Mt. Arrowsmith 21 1102643 468268 1997-2003 SW 5040 Peak 6 1052355 464217 1996-1997 SW Mt. Klitsa 3 1056039 471040 1997-1998 B Mt. Chief Frank 4 1068795 500320 1997-1998 B Mt. Joan 3 1078380 489416 2003-2004 CS Mt. Tom Taylor 13 1026339 495947 1997-1998 CW Marble Meadows 11 1023137 517702 1996-1998 CW King's Peak 1 1011750 532772 1996 CE Mt. Albert Edward 38 1041022 518073 1995-2004 CE Mt. Jutland 8 1042255 520797 2003-2004 N Mt. Cain 5 976943 579286 1997-2004 N Mt. Hkusam 4 1011426 590938 2004 68 Table 3.2. Primers tested on Lagopus leucura that did not show sufficient variation ( > 4 unique alleles). Primers were first tested on 2% agarose gels. Those primers that amplified product were subsequently tested on 6% polyacrylamide gels. Listed for each primer are the annealing temperature (Ta), the expected allele size reported in the original papers, whether the primer amplified product that could be visualized on agarose or acrylamide gels and the associated sample sizes of individuals (n). "-" represents primers not tested on acrylamide gels. A = number of alleles. Suggested Tested Expected Agarose Acrylamide Primer Focal species T a (°C) T a (°C) Size (bp) n product? n product? Results LLSTl' Lagopus lagopus scoticus 53.1,53.6 54 180 6 y 10 y monomorphic LLSD21 Lagopus lagopus scoticus 53.8,53.9 54 110 6 y 10 y monomorphic LLSD41 Lagopus lagopus scoticus 56.2,58.7 56 200 6 y 10 y monomorphic LLSD61 Lagopus lagopus scoticus 56.8,61.3 58 116 6 y 10 y monomorphic LLSD71 Lagopus lagopus scoticus 58.1,59.3 58 135 6 y 10 y A = 4 LLSD81 Lagopus lagopus scoticus 55.7,56.9 56 135 6 y 10 y monomorphic LLSD91 Lagopus lagopus scoticus 61.5,64.3 60 145 • 6 y 10 y monomorphic LLST22 Lagopus lagopus scoticus 60 - 50(TD) 60 - 50(TD) 210 6 n - - -LLST42 Lagopus lagopus scoticus 60 - 50(TD) 60 -50(TD) 339 6 y 10 n no product LLST52 Lagopus lagopus scoticus 60 - 50(TD) 60 - 50(TD) 355 6 n - - -LLST62 Lagopus lagopus scoticus 60 - 50(TD) 60 - 50(TD) 286 6 y 10 y monomorphic SGCA53 Centrocercus urophasianus 56 54 257-279 4 n - -SGCA93 Centrocercus urophasianus 54 54 319-363 4 y 10 y monomorphic SGCA113 Centrocercus urophasianus 62 54 363-383 4 y 10 y monomorphic SGCATAT13 Centrocercus urophasianus 66-56 54 90-130 4 y 10 y monomorphic BG104 Tetrao tetrix 55 55 225 4 y 10 n no product BG124 Tetrao tetrix 55 55 232 4 n - - -BG194 Tetrao tetrix 59 55 178 4 n - - -BG204 Tetrao tetrix 57 55 147 4 n - - -TTD35 Tetrao tetrix 65 - 55 (TD) 65 - 55 (TD) 138 4 y 10 y monomorphic TTD45 Tetrao tetrix 65 - 55 (TD) 65 - 55 (TD) 248 4 y 10 y A = 3 TTD55 Tetrao tetrix 65 - 55 (TD) 65 - 55 (TD) 224 4 n - - -TTT15 Tetrao tetrix 65 - 55 (TD) 65 - 55 (TD) 247 4 n - -TTT25 Tetrao tetrix 65 - 55 (TD) 65 - 55 (TD) 179 4 n - - -1 Piertney and Dallas 19972 Piertney et al. 1998 3 Taylor et al. 2003 4 Piertney and Hoglund 2001 5 Caizergues et al. 2001 Table 3.3. Gene diversity and amplification results at 10 loci. F I S calculated as Weir and Cockerham (1984). All loci showed significant deviations from Hardy-Weinberg equilibrium (HWE probability test). Rd presents results for tests of heterozygote deficiency (Rousset and Raymond 1995). T a , optimal annealing temperature; A, number of observed alleles; n, number of individuals genotyped at each locus; H 0 , mean observed heterozygosity; H E - expected heterozygosity as calculated in ARLEQUIN. rc and rb are estimates of the frequency of null alleles as described by Chakraborty et al. (1992) and Brookfield (1996), respectively. rc = (H E - H 0)/(H E + H 0 ) and rb = (H E - H 0)/(l + HE). na = no GeneBank accession number provided in the original publication. Brackets beside Wright's fixation indices are p-values. A HWE Rd Allele Size (bp) Primer T a(°C) Exp Obs n H 0 H E P P rb rc FIT FST Expected Observed P Accession LLSD10' . 56 14 11 116 0.41 0.80 * * 0.33 0.22 0.47* 0.05(0.04) 0.50* 200 314-118 0.00 X99060 LLST3 2 55(TD) 15 20 129 0.45 0.78 * * 0.27 0.19 0.42* 0.01(0.77) 0.43* 260 168-236 0.00 Y16827 LLST7 2 55(TD) 14 32 119 0.87 0.94 * 0.16 0.04 0.04 0.07(0.01) 0.01(0.75) 0.08* 346 307-449 0.01 Y16831 SGCA6 3 62 33 13 133 0.59 0.88 * * 0.20 0.15 0.32* 0.02(0.35) 0.33* 300-368 298-324 0.00 AY 190931 TTD64 65(TD) 11 10 131 0.47 0.74 * * 0.23 0.16 0.32* 0.10* 0.38* 131 106-132 0.00 AF303097 BG4 5 55(TD) 7 10 131 0.59 0.74 * * 0.11 0.09 0.19* 0.02(0.44) 0.21* 178 183-197 0.10 na BG6 5 57 7 13 131 0.47 0.71 * * 0.20 0.14 0.33* 0.00(0.97) 0.33* 185 187-219 0.00 na BG145 54 8 13 123 0.72 0.79 * * 0.05 0.04 0.06(0.11) 0.03(0.06) 0.09(0.03) 259 178-218 0.10 AF381548 BG165 54 14 10 132 0.86 0.83 * 0.07 -0.02 -0.02 -0.05(0.91) 0.02(0.33) -0.03(0.85) 161 136-161 0.04 AF381550 BG185 53 10 7 133 0.77 0.69 * 0.35 -0.06 -0.05 -0.15(0.99) 0.03(0.02) -0.12(0.99) 168 152-174 0.90 AF38-1551 Overall 13.9 127.8 0.62 0.79 0.20 0.02 0.03 SE 2.3 2.0 0.06 0.02 (0.10;0.30)a (0.12; 0.33)a (0.02; 0.05)a a Lower and upper limits of boostrap 95% confidence intervals for the fixation indices * p < 0.005 ' Piertney and Dallas 1997 2 Piertney et al. 1998 3 Taylor et al. 2003 4 Caizerques et al. 2001 5 Piertney and Hoglund 2001 o Table 3.4. Results of matching the sequences of Vancouver Island white-tailed ptarmigan amplification products to GenBank sequences as determined by the ESEE3S software program. Repeat Average Primer Motif n % Match LLSD10 (CA) (A) 4 91 LLST3 C A A A 3 81 LLST7 G A A A 2 82 SGCA6 (CA) 19 4 77 TTD6 (CA)17 1 63 BG4 (GATA)15 4 96 BG6 (GATA)15 3 98 BG14 (CTAT)10(CCAT)9 2 87 (CATA)7 BG16 (CTAT)15 3 78 BG18 (CTAT)17 4 76 Table 3 . 5 . Estimates of within locality genetic diversity for each population, n, number of individuals sampled; A, mean number of alleles; A n , mean allelic richness; Ho, mean observed heterozygosity; H E , mean expected heterozygosity. A l l mean values presented ± Standard error. rc and rb are estimates of the frequency of null alleles as described by Chakraborty et al. (1992) and Brookfield (1996), respectively. rc = (HE - H0)/(HE + H0) and rb = (HE - H0)/(l + HE). Fis, 95% confidence interval presented in parentheses (Weir and Cockerham 1984). Ris, allele size-based correlation as calculated in genepop. Hd, p-value for test of heterozygote deficiency. Population n A - A n H 0 H E rc rb Fis Ris Hd p - value South 37 9.3 ± 1.1 4.8 ±0.3 0.67 ±0.05 0.77 ±0.02 0.06 0.07 0.137 (0.005; 0.132) 0.02 < 0.005 South West 9 6.3 ±0.7 5.1 ±0.5 0.60 ±0.09 0.84 ±0.02 0.13 0.17 0.267 (0.071; 0.196) 0.08 < 0.005 Beaufort 7 5.2 ±0.5 4.6 ±0.4 0.57± 0.07 0.79± 0.03 0.12 0.16 0.261 (0.049; 0.212) 0.10 < 0.005 Central South 13 6.9 ± 1.1 4.8 ±0.5 0.59 ± 0.09 0.78 ±0.03 0.11 0.14 0.226 (-0.008; 0.234) 0.11 < 0.005 Central West 12 6.2 ±0.9 4.5 ±0.5 0.56 ± 0.09 0.74 ± 0.04 0.10 0.14 0.225 (0.032; 0.193) 0.02 < 0.005 Central East 46 10 ± 1.5 5.0 ±0.4 0.61 ±0.05 0.78 ± 0.03 0.10 0.12 0.21 (0.129; 0.081) 0.19 < 0.005 North 9 5.8 ±0.7 4.7 ±0.5 0.62 ± 0.09 0.77 ±0.04 0.08 0.11 0.169 (-0.042; 0.211) 0.27 < 0.005 ) T a b l e 3.6. Summary of previous studies using microsatellite D N A markers to investigate the population genetics of grouse (Aves: Tetraoninae). Data presented are averaged over population or loci from the corresponding literature cited. Loc i , gives the number of microsatellite loci used in the study; A , number of alleles; A n , allelic richness; Ho, observed heterozygosity; H E , expected heterozygosity; Fis, Wright's inbreeding coefficient; H W E , Hardy-Weinberg equilibrium, n.s. indicates there were no significant departures from Hardy-Weinberg, in other instances the number of significant deviations is reported. There were no heterozygote deficiencies (Hd) reported in any grouse studies except in this study of Vancouver Island white-tailed ptarmigan. If no corresponding measure was found in the original publications na was entered into the cell. Average Species Loci A A n H 0 H E Fis HWE Reference Red grouse (Lagopus lagopus scoticus) 7 9.6 na 0.79 0.87 -0.02 n.s. Piertney et al. 1997/Piertney et al. Sage grouse (Centrocercus urophasianus) 4 4.3 na 0.49 0.48 na 6% locus/pop. Oyler-Mcance et al. 1999 Capercaillie (Tetrao urogallus) Isolated populations 10 4.35 2.3 0.59 0.54 -0.06 9 of 18 pop. Larsson et al. 2003* ' Continuous populations 10 4.9 2.4 0.60 0.62 0.11 Lesser prairie chicken (Tympanuchus pallidicinctus) 5 3.5 na 0.42 0.66 na n.s. Van Den Bussche et al. 2003 Rock ptarmigan (Lagopus muta) 6 10.9 9.4 0.81 0.83 na n.s. Caizergues et al. 2003 Black grouse (Tetrao tetrix) Alps 14 11.7 5.4 0.74 0.75 na n.s. Caizergues et al. 2003 Finland 14 12.5 6.3 0.76 0.78 na n.s. Chinese grouse (Bonasa sewerzowi) 5 4.2 3.6 0.57 0.52 -0.04 n.s. Larsson et al. 2003 Greater prairie chicken (Tympanuchus cupido) 6 na 7.1 0.64 na na n.s. Johnson et al. 2004. Average 8.1 7.3 5.2 0.64 0.67 -0.002 White-tailed ptarmigan (Lagopus leucura) Yukon n=6 9. 5.1 5.1 0.72 0.75 0.04 n.s. B. Fedy, unpublished data Vancouver Island White-tailed ptarmigan (Lagopus leucura saxatilis) 10 7.1 4.8' 0.60 0.78 0.21 all pop. sign Hd * Data from Segelbacher 2002, Segelbacher et al. 2003 presented in Larsson et al. 2003, H W E data from Segelbacher and Storch2002 Table 3.7. Distribution of private alleles found only in single localities of ptarmigan. T is the total number of alleles of different sizes found in a given population, U is the number of alleles unique to that population with the value in parentheses equal to the number of these alleles that occur at a frequency of >5%, and the percentage is (U/T)*100. Localities South SouthWest Beaufort Central South Central West Central East North Locus T U % T U % T U % T U % T U % T U % T U % LLSD10 9 0 0.0 5 K D 20.0 5 0 0.0 5 . 0 0.0 4 0 .0.0 8 1(0) 12.5 5 0 0.0 LLST3 12 3(0) 25.0 5 1(1) 20.0 5 0 0.0 6 1(1) 16.7 6 1(0) 16.7 12 4(1) 33.3 6 0 0.0 LLST7 1.7 2(1) 11.8 11 2(2) 18.2 9 1(1) 11.1 15 3(3) 20.0 14 0 0.0 21 0_ 0.0 11 1(1) 9.1 SGCA6 8 . 0 0.0 8 0 0.0 6 0 0.0 9 0 0.0 7 0 0.0 13 2(0) 15.4 7 0 0,0 TTD6 9 1(0) 11.1 5 0 0.0 5 0 0.0 6 0 0.0 4 0 0.0 7 0 0.0 5 0 0.0 BG4 8 0 0.0 7 0 0.0 5 0 0.0 8 1 12.5 5 0 0.0 6 0 0.0 3 0 0.0 BG6 6 0 0.0 7 1 14.3 4 1 25.0 .4 0 0.0 5 0 0.0 9 2 22.2 4 0 0.0 BG14 11 1(0) 9.1 6 0 0.0 7 0 0.0 8 0 0.0 6 0 0.0 8 0 0.0 5 0 0.0 BG16 8 1(0) 12.5 5 0 0.0 4 0 0.0 4 0 0.0 6 0 0.0 9 b 0.0 7 0 0.0 BG18 5 0 0.0 4 0 0.0 3 0 0.0 4 0 0.0 5 0 0.0 6 1(0) 16.7 5 1 20.0 Total 93 8(1) 8.6 63 5(4) 7.9 53 2(1) 3.8 69 5(4) 7.2 62 1(0) 1.6 99 10(1) 10.1 58 2(1) 3.4 Table 3.8. Significant (p < 0.05; prior to Bonferroni correction) excesses or deficiencies in heterozygosity under each of three models of mutation in each locality and in Lagopus leucura saxatilis, determined via bottleneck (Cornuet and Luikart 1996) simulations. 1AM - infinite alleles model, S M M - stepwise mutation model, TPM - two-phase mutation model. Locality I A M P S M M P T P M P South excess 0.042 deficit 0.012 none South West excess 0.003 none excess 0.012 Beaufort none none none Central South none none none Central West none deficit 0.009 none Central East excess 0.001 none none North none none none L. 1. saxatilis excess 0.012 none deficit 0.0005 75 Table 3.9. Proportion of membership o f each predefined sample population o f Vancouver Island white-tailed ptarmigan in each cluster inferred from the Bayesian method o f Pritchard et al. (2000). Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 South 0.16 0.14 0.18 0.11 0.20 0.20 South west 0.17 0.21 0.16 0.16 0.13 0.17 Beaufort 0.15 0.20 0.15 0.15 0.20 0.15 Central south 0.21 0.18 0.19 0.12 0.14 0.16 Central west 0.21 0.14 0.15 0.19 0.13 0.17 Central east 0.16 0.15 0.16 0.21 0.15 0.17 North 0.12 0.21 0.14 0.20 0.20 0.13 Central West South West South Figure 3.1. Neighbour-joining consensus dendrogram showing clustering based on the genetic distance of Nei (1978). The one major break (72%) does not include the CS population which is geographically proximate to the other Central populations. Numbers on branches are percentages of 1000 bootstrap replicates. 77 REFERENCES Allendorf, F. W. and G. Luikart. 2007. Conservation and the Genetics of Populations. Blackwell Publishing, Oxford. Amos, W. and A. Balmford. 2001. When does conservation genetics matter? Heredity 87: 257-265. Ballou, J. D. 1997. Ancestral inbreeding only minimally affects inbreeding depression in mammalian populations. Journal of Heredity 88: 169-178. Boone, M . D. and O. E. Rhodes. 1996. Genetic structure among subpopulations of the eastern wild turkey (Meleagris gallopavo silvestris). American Midland Naturalist 135: 168-171. Bouzat, J. L., H. H. Cheng, H. A. Lewin, R. L. Westemeier, J. D. Brawn and K. N. Paige. 1998. Genetic evaluation of a demographic bottleneck in the greater prairie-chicken. Conservation Biology 12: 836-843. Brewer, B. A., R. C. Lacy, M . L. Foster and G. Alaks. 1990. Inbreeding depression in insular and central populations of peromyscus mice. Journal of Heredity 81: 257-266. Brookfield, J. F. Y. 1996. A simple new method for estimating null allele frequency from heterozygote deficiency. Molecular Ecology 5: 453-455. Cabot, E. L. and A. T. Beckenbach. 1989. Simultaneous editing of multiple nucleic-acid and protein sequences with Esee. Computer Applications in the Biosciences 5: 233-234. Caizergues, A., S. Dubois, A. Loiseau, G. Mondor and J.-Y. Rasplus. 2001. Isolation and characterization of microsatellite loci in black grouse .{Tetrao tetrix). Molecular Ecology Notes 1: 36-38. Caizergues, A., A. Bernard-Laurent, J.-F. Brenot, L. Ellison and J.-Y. Rasplus. 2003a. Population genetic structure of rock ptarmigan Lagopus mutus in Northern and Western Europe. Molecular Ecology 12: 2267-2274. Caizergues, A., O. Ratti, P. Helle, L. Rotelli, L. Ellison and J.-Y. Rasplus. 2003b. Population genetic structure of male black grouse (Tetrao tetrix L.) in fragmented vs. continuous landscapes. Molecular Ecology 12: 2297-2305. Carter, R. E. 2000. General molecular biology. In: A. J. Baker. Molecular Methods in Ecology, Blackwell Science: 7-49. Castric, V. , L. Bernatchez, K. Belkhir and F. Bonhomme. 2002. Heterozygote deficiencies in small lacustrine populations of brook charr Salvelinus Fontinalis Mitchill (Pisces, Salmonidae): a test of alternative hypotheses. Heredity 89: 27-35. 78 Chakraborty, R., M . Deandrade, S. P. Daiger and B. Budowle. 1992. Apparent heterozygote deficiencies observed in DNA typing data and their implications in forensic applications. Annals of Human Genetics 56: 45-57. Cornuet, J. M . and G. Luikart. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014. Crnokrak, P. and D. A. Roff. 1999. Inbreeding depression in the wild. Heredity 83: 260-270. Daniels, S. J. and J. R. Walters. 2000. Inbreeding depression and its effects on natal dispersal in red-cockaded woodpeckers. The Condor 102: 482-491. Delestrade, A. 2001. Sexual size dimorphism and positive assortative mating in alpine choughs (Pyrrhocorax graculus). Auk 118: 553-556. Dirienzo, A., A. C. Peterson, J. C. Garza, 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. Excoffier, L., G. Laval and S. Schneider. 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1: 47-50. Felsenstein, J. 1985. Confidence-limits on phylogenies - an approach using the bootstrap. Evolution 39: 783-791. Frederick, G. P. and R. J. Gutierrez. 1992. Habitat use and population characteristics of the white-tailed ptarmigan in the Sierra Nevada, California. The Condor 94: 889-902. Gerlach, G. and H. N. Hoeck. 2001. Islands on the plains: metapopulation dynamics and female biased dispersal in hyraxes (Hyracoidea) in the Serengeti National Park. Molecular Ecology 10: 2307-2317. Gibbs, H. L. , K. A. Prior, P. J. Weatherhead and G. Johnson. 1997. Genetic structure of populations of the threatened eastern Massasauga rattlesnake, Sistrurus c. catendtus: evidence from microsatellite DNA markers. Molecular Ecology 6: 1123-1132. Giesen, K. M . and C. E. Braun. 1993. Natal dispersal and recruitment of juvenile white-tailed ptarmigan in Colorado. Journal of Wildlife Management 57: 72-77. Gilligan, D. M . , D. A. Briscoe and R. Frankham. 2005. Comparative losses of quantitative and molecular genetic variation in finite populations of Drosophila melanogaster. Genetical . Research 85: 47-55. Goossens, B., L. Chikhi, P. Taberlet, L. P. Waits and D. Allaine. 2001. Microsatellite analysis of genetic variation among and within alpine marmot populations in the French Alps. Molecular Ecology 10: 41-52. 79 Goudet, J. 1995. FSTAT (Version 1.2): A computer program to calculate F-statistics. Journal of Heredity 86: 485-486. Guo, S. W. and E. A. Thompson. 1992. Performing the exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics 48: 361-372. . Hannon, S. J. and K. Martin. 1996. Mate fidelity and divorce in ptarmigan: polygyny avoidance on the tundra. In: J. M . Black. Partnerships in Birds: the study of monogamy, 69616. Oxford University Press: 192-210. Harrison, S. and A. Hastings. 1996. Genetic and evolutionary consequences of metapopulation structure. Trends in Ecology & Evolution 11: 180-183. Hedrick, P. W. and S. T. Kalinowski. 2000. Inbreeding depression in conservation biology. Annual Review of Ecology and Systematics 31: 139-162. 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 Series B 273: 1491-1499. Hoglund, J., S. B. Piertney, R. V. Alatalo, J. Lindell, A. Lundberg and P. T. Rintamaki. 2002. Inbreeding depression and male fitness in black grouse. Proceedings of the Royal Society of London Series B 269: 711-715. Ingvarsson, P. K. and M . C. Whitlock. 2000. Heterosis increases the effective migration rate. Proceedings of the Royal Society of London Series B 267: 1321-1326. Jawor, J. M . , S. U. Linville, S. M . Beall and R. Breitwisch. 2003. Assortative mating by multiple ornaments in northern cardinals (Cardinalis cardinalis). Behavioral Ecology 14:515-520. Johnson, J. A., M . R. Bellinger, J. E. Toepfer and P. Dunn. 2004. Temporal changes in allele frequencies and low effective population size in greater prairie-chickens. Molecular Ecology 13:2617-2630. Keller, L. F., P. Arcese, J. N. M . Smith, W. M . Hochachka and S. C. Stearns. 1994. Selection against inbred song sparrows during a natural population bottleneck. Nature 372: 356-357. Keller, L. F. 1998. Inbreeding and its fitness effects in an insular population of song sparrows (Melospiza melodid). Evolution 52: 240-250. Keller, L. F., P. R. Grant, B. R. Grant and K. Petren. 2002. Environmental conditions affect the magnitude of inbreeding depression in survival of Darwin's finches. Evolution 56: 1229-1239. Keller, L. F. and D. M . Waller. 2002. Inbreeding effects in wild populations. Trends in Ecology and Evolution 17: 230-241. 80 Lande, R. 1985. Expected time for random genetic drift of a population between stable phenotypic states. Proceedings of the National Academy of Sciences of the United States of America 82: 7641-7645. Lande, R. 1999. Extinction risks from anthropogenic, ecological, and genetic factors. In: L. F. Landweber and A. P. Dobson. Genetics and the Extinction of Species, Princeton University Press: 1-22. Larsson, J. K., Y. H. Sun, Y. Fang, G. Segelbacher and J. Hoglund. 2003. Microsatellite variation in a Chinese grouse Bonasa sewerzowi population: signs of genetic impoverishment? Wildlife Biology 9: 261-266. Latch, E. K., G. Dharmarajan, J. C. Glaubitz and O. E. Rhodes. 2006. Relative performance of Bayesian clustering software for inferring population substructure and individual assignment at low levels of population differentiation. Conservation Genetics 7: 295-302. Li , Y. C , A. B. Korol, T. Fahima, A. Beiles and E. Nevo. 2002. Microsatellites: genomic distribution, putative functions and mutational mechanisms: a review. Molecular Ecology 11: 2453-2465. Madsen, T., R. Shine, M . Olsson and H. Wittzell. 1999. Conservation biology - restoration of an inbred adder population. Nature 402: 34-35. Martin, K., P. B. Stacey and C. E. Braun. 2000. Recruitment, dispersal, and demographic rescue in spatially-structured white-tailed ptarmigan populations. Condor 102: 503-516. Martinez, J. G., J. Carranza, J. L. Fernandez-Garcia and C. B. Sanchez-Prieto. 2002. Genetic variation of red deer populations under hunting exploitation in southwestern Spain. Journal of Wildlife Management 66: 1273-1282. Oetting, W. S., H. K. Lee, D. J. Flanders, G. L. Wiesner, T. A. Sellers and R. A. King. 1995. Linkage analysis with multiplexed short tandem repeat polymorphisms using infrared fluorescence and M l 3 tailed primers. Genomics 30: 450-458. Oyler-McCance, S. J., N. W. Kahn, K. P. Burnham, C. E. Braun and T. W. Quinn. 1999. A population genetic comparison of large- and small-bodied sage grouse in Colorado using microsatellite and mitochondrial DNA markers. Molecular Ecology 8: 1457-1465. Pemberton, J. M . , J. Slate, D. R. Bancroft and J. A. Barrett. 1995. Nonamplifying alleles at microsatellite loci - a caution for parentage and population studies. Molecular Ecology 4: 249-252. Piertney, S. B. and J. F. Dallas. 1997. Isolation and characterization of hypervariable microsatellites in the red grouse Lagopus lagopus scoticus. Molecular Ecology 6: 93-95. Piertney, S. B., A. D. C. MacColl, P. J. Bacon and J. F. Dallas. 1998. Local genetic structure in red grouse (Lagopus lagopus scoticus): evidence from microsatellite DNA markers. Molecular Ecology 7: 1645-1654. 81 Piertney, S. B. and J. Hoglund. 2001. Polymorphic microsatellite DNA markers in black grouse (Tetrao tetrix). Molecular Ecology Notes 1: 303-304. 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. Pritchard, J. K., M . Stephens and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945-959. 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. 2001. How closely correlated are molecular and quantitative measures of genetic variation? A meta-analysis. Evolution 55: 1095-1103. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Rosvall, O. and T. J. Mullin. 2003. Positive assortative mating with selection restrictions on group coancestry enhances gain while conserving genetic diversity in long-term forest tree breeding. Theoretical and Applied Genetics 107: 629-642. Rousset, F. and M . Raymond. 1995. Testing heterozygote excess and deficiency. Genetics 140: 1413-1419. Ruiz-Garcia, M . , P. Orozco-terWengel, A. Castellanos and L. Arias. 2005. Microsatellite analysis of the spectacled bear {Tremarctos ornatus) across its range distribution. Genes & Genetic Systems 80: 57-69. Sage, R. D., D. Heyneman, K. C. Lim and A. C. Wilson. 1986. Wormy mice in a hybrid zone. Nature 324: 60-63. Saitou, N. and M . Nei. 1987. The Neighbor-Joining Method - a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4: 406-425. Schiegg, K., G. Pasinelli, J. R. Walters and S. J. Daniels. 2002. Inbreeding and experience affect response to climate change by endangered woodpeckers. Proceedings of the Royal Society of London Series B 269: 1153-1159. Scribner, K. T. and R. K. Chesser. 1993. Environmental and demographic correlates of spatial and seasonal genetic-structure in the Eastern Cottontail (Sylvilagus Floridanus). Journal of Mammalogy 74: 1026-1044. Segelbacher, G. and I. Storch. 2002. Capercaillie in the Alps: genetic evidence of metapopulation structure and population decline. Molecular Ecology 11: 1669-1677. 82 Segelbacher, G., J. Hoglund and I . Storch. 2003. From connectivity to isolation: genetic consequences of population fragmentation in capercaillie across Europe. Molecular Ecology 12: 1773-1780. Shriver, M . D., L. Jin, R. Chakraborty and E. Boerwinkle. 1993. Vntr allele frequency-distributions under the Stepwise Mutation Model - a computer-simulation approach. Genetics 134: 983-993. Taylor, S. E. , S. J. Oyler-McCance and T. W. Quinn. 2003. Isolation and characterization of microsatellite loci in greater sage-grouse (Centrocercus urophasianus). Molecular Ecology Notes 3: 262-264. Tryjanowski, P. and J. Simek. 2005. Sexual size dimorphism and positive assortative mating in red-backed shrike Lanius collurio: an adaptive value? Journal of Ethology 23: 161-165. Valdes, A. M . , M . Slatkin and N. B. Freimer. 1993. Allele frequencies at microsatellite loci - the Stepwise Mutation Model revisited. Genetics 133: 737-749. Wade, M . J. and D. E. McCauley. 1988. Extinction and recolonization - their effects on the genetic differentiation of local-populations. Evolution 42: 995-1005. Wattier, R., C. R. Engel, P. Saumitou-Laprade and M . Valero. 1998. Short allele dominance as a source of heterozygote deficiency at microsatellite loci: experimental evidence at the dinucleotide locus G v l C T in Gracilaria gracilis (Rhodophyta). Molecular Ecology 7: 1569-1573. Weir, B. S. and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 3 8: 13 5 8-13 70. Whitlock, M . C., P. K. Ingvarsson and T. Hatfield. 2000. Local drift load and the heterosis of interconnected populations. Heredity 84: 452-457. Zwickel, F. C. and J. F. Bendell. 1967. A snare for capturing blue grouse. Journal of Wildlife Management 31: 202-204. 83 CHAPTER 4 Gene flow, patterns of dispersal, and metapopulation dynamics of alpine Vancouver Island white-tailed ptarmigan inferred from microsatellite DNA markers INTRODUCTION Dispersal is an important process affecting the evolution of animal populations in terms of demographic and adaptive trajectories. For animals in patchily distributed habitats dispersal can be fundamental to the persistence of local populations. Metapopulation dynamics is the study of the interactions between local populations within a larger area, where dispersal from one population to another is possible (Hanski and Gilpin 1991). Metapopulations exist in heterogenous landscapes in which habitat patches suitable for population existence are located within a matrix of unsuitable habitat. Thus, dispersal is fundamental to understanding the metapopulation dynamics of a species. Accurate information on dispersal patterns is difficult to collect for many species at relevant spatial scales for populations and two methodological approaches exist to assess dispersal. Direct methods of measuring dispersal involve marking individuals and following their movement patterns over time. These methods are very accurate but tend to underestimate long-distance dispersal and overlook the effects of important 'outlier' years, unless they occur during data collection (Crochet 1996). Indirect methods of studying dispersal analyze allele frequencies to estimate levels of gene flow, population differentiation, and patterns of distribution (Slatkin 1987). Indirect methods provide patterns of genetic population structure that can lead to important insights into landscape-level population dynamics (Hastings and Harrison 1994, Pannell and Charlesworth 2000). Of particular importance for species of conservation concern, studies of genetic population structure can identify factors influencing population persistence such as patterns of dispersal (Caizergues et al. 2003) and barriers to gene flow (Piertney et al. 1998). Finally, indirect methods can also identify the influence of landscape structure and fragmentation on population genetic structure (Giles and Goudet 1995). Empirical data contributing to our understanding of animal metapopulation dynamics comes primarily from species that are sensitive to environmental stochasticity and have evolved in naturally ephemeral habitats that change in suitability over short time scales (review in Harrison and Taylor 1997, Chapter 2). These characteristics have important influence on the patterns of extinction and recolonization of populations. For example, the ephemeral nature of habitat 84 suitability influences the probability of extinction in local populations. For some amphibians, lowered precipitation prevents the formation of some bodies of water and results in complete extirpation of local populations (Bradford et al. 2003). Similar species life history characteristics are found in many studies that used indirect genetic techniques to examine metapopulation dynamics and genetic structure (Ebert et al. 2002, Harper et al. 2003). Furthermore, many genetic studies of metapopulation dynamics examine species that have experienced severe anthropogenic fragmentation of their natural habitat resulting in an evolutionarily novel habitat distribution (Galbusera et al. 2000). Few studies have been conducted on the metapopulation dynamics of species which have evolved in naturally isolated and stable habitats using either direct or indirect methods (Moilanen et al. 1998, Martin et al. 2000, Elmhagen and Angerbjorn 2001, Gerlach and Hoeck 2001, Goossens et al. 2001, Franken and Hik 2004). Animal populations that experience a patchy distribution of suitable habitats usually have significant levels of genetic differentiation between populations. This type of population genetic structure is typical of metapopulations, reflects the patterns of limited dispersal between populations, and often results in deviations from panmixia (Newman and Squire 2001, Gerlach and Hoeck 2001, Goossens et al. 2001). Furthermore, the landscape matrix surrounding local populations can present barriers to gene flow that can have significant genetic and evolutionary consequences for local populations (Segelbacher et al. 2003), influencing the 'effective' isolation of habitat patches (Ricketts 2001) and resulting in deviations from a continuous pattern of isolation by distance. Landscape features can act as barriers to gene flow in some ground-dwelling birds with restricted flight capacity or narrow habitat requirements (Boone and Rhodes 1996). Mountain ranges rising above the preferred habitat of black grouse {Tetrao tetrix) and capercaillie (Tetrao urogallus) function as natural barriers to gene flow (Segelbacher and Storch 2002, Caizergues et al. 2003). In red grouse (Lagopus lagopus scoticus), a river system and the surrounding unsuitable habitat were revealed as barriers to gene flow (Piertney et al. 1998). These studies, and others, have addressed fragmentation and isolation impacts on population genetics in Tetraonids (Oyler-McCance et al. 1999), however, these species primarily inhabit anthropogenically altered ecosystems such as forests and prairies. Alpine ecosytems are beginning to show the impacts of climate change (Krajick 2004), accumulation of foreign compounds (Blais et al. 1998), and impacts from increasing recreational use (Watson and Moss 2004). However, alpine ecosytems are one of the few ecosystem types that remain relatively undisturbed by human activities (Martin 2001). 85 In this study, I described the patterns of dispersal in Vancouver Island white-tailed ptarmigan (Lagopus leucura saxatilis). This species has evolved in naturally isolated and successionally stable, high-elevation habitats. White-tailed ptarmigan are a ground-dwelling grouse species that, compared to other birds, have somewhat limited dispersal distances even in relatively continuous alpine habitat. Telemetry studies of white-tailed ptarmigan in Colorado suggested that demographic exchange occurs between populations within 5-10 km for males and 20 - 30 km for females (Martin et al. 2000). Martin et al. (2000) also noted a capacity for external recruitment that suggested a metapopulation structure for populations of white-tailed ptarmigan. Studies of animal population genetics commonly combine samples from a number of years (e.g. 1997 - 2000: Segelbacher and Storch 2002, 1994 - 2000: Ciofi et al. 2006). Temporal variation in genetic structure is typically of less biological interest than spatial structure and, therefore, is less often addressed (but see Berthier et al. 2006). The pattern of sampling in this study of Vancouver Island white-tailed ptarmigan involved two discrete sampling periods (inventory: 1995 - 1998 and intensive: 2003 - 2004), providing an opportunity to examine temporal variation in the population genetic structure of Vancouver Island white-tailed ptarmigan. Furthermore, it seems prudent to first ensure that temporal variation will not mask the patterns of spatial variation before examining spatial genetic structure with samples from several time periods. In this study, I examined whether habitat fragmentation equals population fragmentation in a species that evolved in naturally fragmented and stable habitats. I described the patterns of dispersal and migration across Vancouver Island, identified potential barriers to gene flow, and addressed temporal variation between two discrete time periods. METHODS Study species and study area This study focused on a subspecies of white-tailed ptarmigan restricted to Vancouver Island, British Columbia. See Chapter 1 for detailed description of the study species and area. Blood sample collection, DNA isolation, and genotyping for 10 loci, followed procedures outlined in Chapter 3. Temporal variation To examine temporal variation in Vancouver Island white-tailed ptarmigan, I divided genotypic data into inventory (1995 - 1998) and intensive (2003 - 2004) stages. Sufficient data existed to compare the genetic structure between the 2 time periods in the South and Central 86 East populations. Locus LLSD10 was not included in comparisons between the sampling periods because of a paucity of data in some populations. For example, the inventory Central East population was missing allele sizes for 9 of 16 individuals at locus D10. Population differences in allelic and genotypic distribution between the 2 time periods within the South and Central East populations were tested following procedures outlined in Chapter 3. These analyses test the null hypotheses of no differences in allelic and genotypic distribution between the inventory and intensive stages of sample collection. Average number of alleles (A), allelic richness for the minimum sample size (An) and diversity ( HE ) were calculated using FST A T (ver. 2.9.3.2 Goudet 1995). I tested for differences between time periods within populations in A n and H E using Wilcoxon signed ranks test, which paired the data by locus. Dendrograms of relationships among localities and time periods were constructed from Nei's D matrices calculated using the software program P H Y L I P ver. 3.6 (Felsenstein J http://evolution.geneticas.washington.edu/phylip.html) using bootstrapped allele frequencies as described in Chapter 3. Inter-population analyses I examined the partitioning of genetic variation using an Analysis of Molecular Variation (AMOVA, A R L E Q U I N ver. 3.01 Excoffier et al. 2005). Significance was assessed through 10,000 random permutations (Excoffier et al. 1992). A M O V A can be applied to microsatellite data to estimate F-statistics (Weir and Cockerham 1984) or to obtain ®ST, an analogue of R S T (Slatkin 1995) . The main difference between the two statistics is that allele size is taken into consideration in R S T , whereas in Weir and Cockerham's (1984) analogue to FST , only allele identity is considered (Slatkin 1995). To facilitate comparisons of estimates of population differentiation with other studies, I used FST and R S T analogue ®ST, (Michalakis and Excoffier 1996) . Isolation by distance over Vancouver Island was assessed by testing the correlation between genetic and geographical distance considering all population pairs. I regressed FST / (1-FST) estimates on logarithm of distance for populations (Rousset 1997). For geographical distances, I calculated straight-line distances between all pairs of populations. Measures were taken from the approximate centre of each region. Mantel's test was used to determine significance of isolation by distance and all calculations were performed with G E N E P O P (3.4 Raymond and Rousset 1995). Population pairwise F S T and R S T values were calculated and significance was tested by permutation of individual haplotypes between populations ( A R L E Q U I N ver. 3.01 Excoffier etal. 2005). 87 Sampling was extensive across Vancouver Island; however, some individuals may have emigrated from unsampled populations or subpopulations. I used the exclusion procedure (Bayesian method) ofGENECLASS (ver. 1.0.02 Cornuet et al. 1999) to estimate the likelihood of an individual being identified as an immigrant from areas other than the 7 populations sampled. GENECLASS calculates a likelihood value which relates each individual to the 7 populations. If the value was below a threshold probability (0.05), the individual was excluded from that population, i.e. did not emigrate or originate from that location. If all 7 populations were excluded for a given individual, that individual may have emigrated from an unsampled population. Allele frequency data were subjected to PCA using PCAGEN (http://www.unil.ch/popgen/softwares/pcagen.htm) to resolve similarities among localities. The genetic structure of the 7 populations was then illustrated using the results of the PCAGEN analysis. Evidence of recent migration (within the last few generations) events across Vancouver Island was assessed using a Bayesian multilocus genotyping procedure implemented with Markov Chain Monte Carlo (MCMC) methods in BayesAss (ver 1.3 Wilson and Rannala 2003). In this analysis, migration refers to the movement of individuals among populations. This method assumes loci are in linkage equilibrium but does not require populations to be in either migration drift or Hardy-Weinberg equilibrium. To examine the strength of the information in the Vancouver Island white-tailed ptarmigan data set, 95% confidence intervals were determined for migration rates and compared to a scenario where all proposed changes throughout the Markov Chain are accepted. Thus, simulating the event in which any information that may exist in the data is insufficient to affect the posterior distribution of the migration rates. If there is sufficient information in the data supplied to the program, the reported confidence intervals should differ from those that occurred in the scenario involving insufficient information in the data. The M C M C ran for 3.0 x 106iterations, with the first 106 iterations discarded as burning to allow the chain to reach stationary state. Samples were collected every 2000 iterations to infer posterior probability distributions for migration rate. Migration rate from a population into the same population was defined as the proportion of individuals in each generation that are not migrants. Direct observations Individual birds were located, captured, and radio-collared as outlined in Chapter 2. In addition to ground-based telemetry, bird locations were also determined through helicopter aerial telemetry during the inventory stage. For the analysis of direct movements, observations 88 were classified into either summer (June - November) or winter season observations. Analysis of direct movements includes all birds with two or more observations that spanned from one season to the next. The locations of dead birds were not included in the calculation of dispersal distances as it could not be determined whether the bird moved to that location or was transported by a predator. RESULTS Temporal variation I examined genetic variation between the inventory and intensive stages of research in the South population (inventory: n = 28, intensive: n = 9) and the Central East population (inventory 1995 - 1998: n = 16, intensive 2003 - 2004: n = 30). The allelic distribution within populations differed between these two periods (South: %2 = 52.34, p < 0.001, Central East: £ = 62.02, p < 0.001). The test for similar genotypic distribution between populations (G-test; Goudet et al. 1996) also showed significant differences between the two stages of data collection (South: x 2 = 39.24 p < 0.001, Central East: x 2 = 53.31, p < 0.001). Table 4.1 contains a summary of the comparisons of allelic richness and diversity between the two time periods. Allelic richness (An) did not differ in the South population between the inventory and intensive stages (Wilcoxon paired samples test Z = -0.059, p = 0.953). In the Central East population, allelic richness differed between the two time periods with an increase in A n detected from the inventory to the intensive stage (Z = -2.31, p = 0.021). When all inventory samples from both populations were compared with all intensive samples from both populations, the intensive samples showed significantly higher A n than the inventory samples (Z — -2.025, p = 0.043). Diversity (HE) levels did not differ between the two time periods in the South (Z - -0.415, p = 0.678). The Central East population showed a significant increase in diversity between the inventory and the intensive stages (Z = -2.547, p = 0.011). Comparison of all inventory samples from both populations to all intensive samples from both populations showed significantly higher diversity in the intensive samples (Z = -1.938, p = 0.053). The neighbour-joining tree based on Nei's D resolved one major partition of the different sampling stages (Figure 4.1). Despite differences detected within populations between sampling stages (outlined above), one group contained both South sampling stages and the other group contained both Central East sampling stages. This partitioning was resolved in 90% of the trees. This result indicates that observed temporal variation within these populations is insufficient to mask the importance of spatial variation in allele frequency distributions. 89 Inter-population analyses The A M O V A hierarchical analysis of variance revealed that between 91 and 96% of microsatellite variance resides within populations (Table 4.2). A significant proportion (3.4 -8.6%) is distributed among populations. Vancouver Island white-tailed ptarmigan allele distributions are generally not continuous within loci (B. Fedy, unpubl. data); therefore, the F S T values likely represent the most accurate description of genetic partitioning. A significant positive relationship existed between pairwise F S T (or F S T/(1 - F S T ) ; Rousset 1997) and geographical distance (rs = 0.60, p = 0.03, Figure 4.2). F S T over all loci and populations was fairly low (mean = 0.03, range = 0.01 - 0.06). R S T calculations resulted in higher average values than F S T ( R S T mean = 0.09, range -0.03 - 0.23) however, as mentioned above, F S T likely represents a more accurate picture of differentiation as the 10 microsatellite loci used do not appear to follow the stepwise mutation model in Vancouver Island white-tailed ptarmigan. Pairwise F S T and R S T values were calculated for all pairs of populations and most comparisons were significant at either the 5 or 1% level (Table 4.3). The South population had significant pairwise F S T values for all comparisons. The Central South had the highest number of non-significant pairwise F S T values and demonstrated a lack of differentiation from the South West, Beaufort and Central West populations. These similarities are likely due to location, as the Central South is geographically the most central of the 7 populations. Two other pairwise comparisons were not significant: the South West - Beaufort and Central West - Central East. The exclusion procedure in G E N E C L A S S was not able to assign populations at a threshold level of 0.05 for 11% (14) of the 133 individuals sampled. Possibly these individuals emigrated from unsampled populations. The 'unassigned' individuals were evenly spread between the two regions (South Region = 7, Central Region = 7). The low percentage of unassigned individuals suggests sampling captured the majority of putative source populations. The first 3 Principal Components (PC) explained 72% of the variation in the microsatellite data (PCI = 42%, PC2 = 18%, PC3 = 12% Figure 4.3). PCI showed the South population as markedly distinct and highlighted a pattern of similarity among the Central and North populations. The 2 n d component suggested similarity among the South West, Beaufort and Central South populations. The 3 r d component distinguished the North population from all other populations. The analysis software BayesAss simulates the effect of no information in the data from which to estimate migration rates by accepting all proposed changes in the Markov Chain. This analysis resulted in nonmigration rates of 0.83 with a 95% confidence interval of (0.68, 0.99) 90 and migration rates of 0.03 with a 95% confidence interval of (0, 0.14). All standard deviations for the posterior probabilities generated from the Vancouver Island white-tailed ptarmigan data set were < 0.05. Confidence intervals recovered from the data set were also considerably smaller and migration rates were occasionally greater than those obtained from the simulated scenario of no information. These results suggest the microsatellite data set contained an appreciable amount of information from which to estimate migration rates. The majority of populations had a very low proportion of migrants (Table 4.4). Several populations had migration rates > 0.10. The Central East contributed migrants to the South West (m = 0.20) and North (m = 0.12) populations. The Central West contributed migrants to the Central East (m = 0.13) population, and the Beaufort also contributed a portion of migrants to the Central South (m = 0.16) population. A positive value for the lower 95% confidence interval was also recorded for the migration between other populations despite the low proportion of migrants (m = 0.06 -0.08). However, a pattern of note is the positive lower bound of the 95% confidence interval from the Southern population to 3 other populations (Beaufort m = 0.07, Central South m = 0.08, Central West m = 0.06) that may represent a meaningful pattern for this population. Figure 4.4 gives a schematic representation and summary of the estimation of migration rates and pairwise FST comparisons. Direct movement analysis I analyzed the direct movement data of 118 individuals which resulted in 348 movement observations. These direct observations of radio-tagged birds revealed no movement among any populations. Average distances moved within each of the populations were typically around Hon (Table 4.5). Maximum distances observed were approximately 10 km. The shortest distance between 2 populations was 22 km (Central South - Central East), which is twice the maximum observed dispersal distance. This suggests that inter-population movement is rare on Vancouver Island. D I S C U S S I O N Analysis of Vancouver Island white-tailed ptarmigan allele frequencies revealed significant isolation by distance, significant (but low) population differentiation, and low movement rates among populations. Several lines of evidence support a stepping-stone model of migration. An extended area of low elevation habitat separating the South population appears to act as a barrier to gene flow, suggesting the existence of a threshold distance of unsuitable habitat for Vancouver Island white-tailed ptarmigan. 91 Temporal variation and sample populations Allelic richness and diversity increased between the inventory and intensive sampling periods in the Central East population. Given the data presented thus far, it is difficult to provide a clear explanation for the temporal variation in the Central East population. I analyzed temporal variation for this study to ensure temporal variation did not mask the genetic signature of spatial variation. The dendrogram comparing spatial and temporal variation demonstrated a clear grouping of populations by geographic (not temporal) variation. These findings provide support that temporal variation in genetic structure did not confound the spatial variation of interest. One of my primary goals in this study was to estimate the rates and patterns of migration among populations. Non-sampling of putative populations may influence estimates of migration rates. The Bayesian approach used to assess migration rates was appropriate because it does not have the limiting assumptions of either Hardy-Weinberg equilibrium or mutation-drift equilibrium (Wilson and Rannala 2003). However, the analysis assumes the sample populations represent all populations in the estimation of migration rates. The non-sampling of contributing populations may lead to a false impression of the accuracy of migration estimates (Beerli 2004). This concern is unlikely in the analysis of Vancouver Island white-tailed ptarmigan because the exclusion analysis suggested that sampling captured the majority of contributing populations. Thus, the probability that unsampled populations are contributing individuals and confounding the estimation of migration rates is low. Metapopulations Dispersal among populations is a fundamental concept of metapopulation biology. If local extinction of populations does not occur due to frequent dispersal among populations, the concept of a metapopulation may not apply. In these situations the regional population would be more accurately thought of as a single spatially (albeit patchily) distributed population (panmixia). Naturally patchy distributions of a species do not necessarily result in a metapopulation structure. Metapopulation models can be described in three ways: 1) equilibrium, 2) non-equilibrium, and 3) patchy population. In an equilibrium metapopulation model, patch area and isolation are key factors influencing the probability of population extinction, and extinctions are generally balanced by recolonization (Harrison and Taylor 1997). In a non-equilibrium metapopulation model, recolonization is insufficient to balance extinction and predicts region-wide declines for the species. Under a non-equilibrium model, most suitable habitat patches would not be occupied (Harrison 1991). In the patchy population 92 model, dispersal among patches is sufficiently frequent so that local extinctions virtually never occur, and the system effectively consists of a single large population occupying many habitat patches or a complex of several such populations (Harrison and Taylor 1997). A patchy population dynamic has been observed in several species that exploit patchy habitats and this model predicts frequent dispersal among populations (Harrison 1991, Doak and Mills 1994, Martin et al. 2000, Bradford et al. 2003). Vancouver Island white-tailed ptarmigan do not appear to fit the patchy population model (panmixia). The low estimated migration rates for Vancouver Island white-tailed ptarmigan suggest that dispersal among populations is too rare to be considered a single large population. Furthermore, pairwise comparisons of populations reveal significant differentiation between populations. Very few reproductively successful migrants are required per generation to eliminate genetic differences between two populations (e.g. Haiti and Clark 1997 p. 194). Thus Vancouver Island white-tailed ptarmigan are unlikely to represent one large population of patchily distributed subpopulations, and habitat fragmentation seems to equal population fragmentation. This conclusion is further supported by the absence of any direct observations of dispersal between populations. It is also unlikely that Vancouver Island white-tailed ptarmigan represent a non-equilibrium metapopulation. The analysis of migration rates suggests that, although infrequent, migration does occur between populations. Furthermore, several populations were not genetically differentiated from each other. The non-equilibrium metapopulation model also predicts that most suitable habitat patches will not be occupied. Surveys of suitable mountain habitat on Vancouver Island revealed the majority of alpine habitat was occupied by white-tailed ptarmigan (Martin et al. 2004). The most appropriate model of metapopulation dynamics for Vancouver Island white-tailed ptarmigan is potentially the equilibrium metapopulation model. The observed genetic differentiation suggests populations are not a demographically single unit since most populations are genetically differentiated from each other and dispersal between populations does occur, albeit infrequently. The regional variation in reproductive success and age and sex ratios described in Chapter 2 also support this conclusion. The patterns of genetic structure of Vancouver Island white-tailed ptarmigan support a metapopulation structure which is in agreement with the conclusions from telemetry studies of white-tailed ptarmigan in Colorado (Martin et al. 2000). 93 One of the key assumptions of the equilibrium model is a pattern of extinctions and recolonizations. Populations of Vancouver Island white-tailed ptarmigan probably do not experience regular extinctions and recolonizations. Instead, populations likely represent a pattern of stochastic and infrequent demographic rescue. Thus, Vancouver Island white-tailed ptarmigan do not fit particularly well into any model of metapopulation dynamics. The concept of metapopulations and their relevance to real populations has been recently addressed in the literature (Baguette 2004, Armstrong 2005, Smith and Green 2005). This study was not intended to further explore the meaning and relevance of the metapopulation paradigm. However, it seems clear that recognizing the importance of spatial structure in forming patterns of dispersal and genetic structure lends insights into the barriers to gene flow and patterns of dispersal in Vancouver Island white-tailed ptarmigan. Frequent extinction-recolonization dynamics between populations would lead to less differentiation between populations and higher estimated migration rates than those observed. However, surveys throughout the duration of the project of individual mountains within populations found some mountains have birds in some years but not others (Martin et al. 2004, B. Fedy, unpubl. data). This variation suggests a fine-scale equilibrium metapopulation structure among mountains within populations, although data were insufficient to test this hypothesis explicitly. When a mountain extirpated of white-tailed ptarmigan is recolonized, the founder groups that recolonize local mountains likely originate from within the population area. Extinction-recolonization events may be relatively frequent among mountains within populations. Migration Patterns of genetic isolation by distance result from the changing relative influence of gene flow and drift as populations become more geographically isolated (Hutchison and Templeton 1999). Vancouver Island white-tailed ptarmigan populations demonstrated a more or less consistent pattern of isolation by distance. This pattern suggests the stepping-stone model (Kimura and Weiss 1964) is the most likely explanation of gene flow patterns in Vancouver Island white-tailed ptarmigan. Mean dispersal distances within populations, determined from radio-telemetry, ranged from 0.8 to 2.8 km. Despite telemetry observations of 348 inter-seasonal movements, no radio-tagged individuals moved between populations. Dispersing individuals are likely moving relatively short distances to the nearest neighbour patches and direct dispersal events between populations appear to be rare. 94 Nevertheless, direct gene flow between populations is not necessary to explain gene flow over greater distances than those observed using direct methods in dispersing individuals. Genes dispersed by the 'stepping-stone' model can result in the spread of genes over larger distances than the dispersal ability of the animals might predict. Compared to other bird species, white-tailed ptarmigan are perceived as a relatively sedentary species and, if movements were severely restricted, we may expect no gene flow and very high levels of differentiation between populations. However, Vancouver Island white-tailed ptarmigan populations did not show high levels of differentiation between populations, suggesting some connectivity. The geographic distances among populations ranged from 18 to 168 km, generally greater than previously inferred for white-tailed ptarmigan dispersal (Martin et al. 2000). Such discrepancies between direct and indirect estimates have been reported in other presumably sedentary species and are likely the product of stepping-stone dispersal (Kvist et al. 1998, Baker et al. 2001, Segelbacher and Storch 2002). Absence of substantial past and recent white-tailed ptarmigan movements among populations was inferred by a Bayesian method for estimating recent migration rates from microsatellite allele frequencies. Estimated migration rates were low but I found evidence of exchange, in one direction or another, among all populations except the South. The lack of significant migration either to or from the South population suggests it is the most genetically isolated of the 7 populations. Barriers to gene flow Much of the alpine in the Central Region of Vancouver Island is within Strathcona Provincial Park (approx. 2,500 km2). This relatively continuous mountain area showed a greater number of non-significant pairwise FST values than other populations (Figure 4.4). Only two populations in the central region showed significant differentiation, the Central East and Central South. The Central South population showed the fewest significant pairwise differences and given its geographically central location, likely functions as a stepping stone for movements among populations. Buttle Lake is approximately 18 km long (north - south) and 1.5 km wide and extends through the middle of the central populations, ending with the Central South at its southern end. Piertney et al. (1998) identified a river system as an important barrier to gene flow for red grouse. In contrast, Buttle Lake does not appear to be a barrier to gene flow for white-tailed ptarmigan, possibly because it is surrounded on 3 sides by suitable habitat and birds are able to disperse around or, potentially with a short flight, over the lake. 95 Under the isolation by distance model, the frequency of alleles, and hence PCI scores, are expected to change gradually across two-dimensional space and any discontinuities can be viewed as barriers to gene flow (Piertney et al. 1998). In this study, the discontinuity showed correspondence with areas of unsuitable low elevation habitat separating the South population. The size of this barrier appeared to have a disproportionate effect on limiting allelic movement. The South population also showed consistent genetic differentiation from all populations. The four upper points of the isolation by distance graph showed greater genetic distance given geographic distance and suggested the low elevation areas act as barriers to gene flow to and from the South population. Low elevation habitat per se does not block gene flow since areas of unsuitable habitat separate many of the other populations. However, the low elevation areas separating the South population are the largest continuous expanses of unsuitable habitat separating populations of Vancouver Island white-tailed ptarmigan. The size of this unsuitable habitat is likely acting as a barrier and this finding suggests that Vancouver Island white-tailed ptarmigan have a threshold of unsuitable habitat past which dispersal is extremely limited. In Chapter 2 I demonstrated lower population performance in the entire southern region (including South West and Beaufort populations). The South population could be in danger of population decline with decreased population performance and little dispersal between the South and neighbouring populations. As discussed in Chapter 2, climate change is likely to result in a decrease in available alpine habitat and an increase in the unsuitable, low elevation, habitat matrix surrounding populations. If extensive low elevation habitat results in the limitation of dispersal and decrease in potential for population rescue, it is possible the South population is in serious threat of extirpation. 96 Table 4.1. Comparison of allelic richness (An) and diversity (HE) in two Vancouver Island white-tailed ptarmigan populations. Values presented are means and values in brackets are the ranges. P-values were determined by Wilcoxon paired-samples tests. Population Sampling Period Inventory Intensive P South A n 5.7 6 .0 0 . 9 5 ( 3 . 6 - 8 .0 ) ( 3 . 0 - 8 .0 ) H E 0 . 7 6 0 . 7 7 0 . 6 8 ( 0 . 6 8 - 0 . 8 4 ) ( 0 . 4 6 - 0 . 9 0 ) Central East A„ 5.4 6 .9 0 . 0 2 (3 .5 - 8 .7 ) ( 4 . 7 - 1 2 . 0 ) H E 0 . 6 9 0 . 8 2 0 .01 ( 0 . 3 9 - 0 . 9 1 ) ( 0 . 0 7 2 - 0 . 9 5 ) South and Central East A„ 5.6 6 .5 0 . 0 4 (3 .5 - 8 .7 ) ( 3 . 0 - 1 2 . 0 ) H E 0 . 7 3 0 . 7 9 0 . 0 5 ( 0 . 3 9 - 0 . 9 1 ) ( 0 . 4 6 - 0 . 9 5 ) 9 7 Table 4.2. Analysis of Molecular Variance among 7 Vancouver Island white-tailed ptarmigan populations. Source of variation d.f. Sum of squares Variance % components of variation P Among populations 6 FST 4 6 0 . 1 2 3 . 4 < 0 . 0 0 1 RST 1 9 5 1 4 7 2 8 .6 < 0 . 0 0 1 Within populations 2 5 9 FST 9 1 4 3 . 5 3 9 6 . 6 RST 1 9 9 6 2 7 7 7 1 9 1 . 5 Total 2 6 5 FST RST 9 6 1 2 1 9 1 4 2 3 . 6 5 8 4 3 98 Table 4.3. P a i r w i s e populat ion comparisons o f FST and RST values. FST values are presented on the lower diagonal and R S T values above the diagonal . N o n - s i g n i f i c a n t values are i n b o l d . U n d e r l i n e d values are significant at the 5% level and a l l other values at the 1% l eve l . N e g a t i v e values are due to s a m p l i n g error and indicate gene f low without l i m i t a t i o n . South South West Beaufort Central South Central West Central East North (S) (SW) (B) (CS) (CW) (CE) (N) South (S) - 0.153 0.232 0.192 0.077 0.038 0.161 South West (SW) • 0.029 ~ 0.012 -0.032 0.018 0.059 0.106 Beaufort (B) 0.024 0.015 - 0.072 0.075 0.088 0.193 Central South (CS) . 0.049 0.017 0.012 - 0.038 0.097 0.083 Central West (CW) 0.058 0.042 0.030 0.015 - 0.007 0.152 Central East (CE) 0.042 0.028 0.019 0.024 0.007 - 0.079 North (N) 0.052 0.040 0.032 0.035 0.026 0.015 ~ Table 4.4. Means and 95% confidence intervals of the posterior distributions for migration rates between Vancouver Island white-tailed ptarmigan populations. Values along the diagonal are the proportions of individuals derived from the source population in each generation and are underlined. All standard deviations were < 0.05. Migration rates to other populations > 0.10 are in bold. Migration rates < 0.10 but still with a positive value for the lower 95% confidence interval are in italics. Rate to South South West Beaufort Central South Central West Central East North Rate from (S] (SW) (BJ (CS) (CW) (CE) (N) South (S) 0.937 0.030 0.069 0.079 0.055 0.008 0.012 (0.856,0.992) (0,0.113) (0.006,0.179) (0.017,0.170) (0.002,0.137) (0,0.037) (0,0.063) South West (SW) 0.003 0.697 0.013 0.010 0.008 0.003 0.012 (0,0.018) (0.668,0.769) (0,0.071) (0,0.056) (0,0.043) (0,0.017) (0,0.061) Beaufort (B) 0.022 0.069 0.775 0.156 0.074 0.013 0.055 (0,0.070) (0.011,0.161) (0.691,0.888) (0.053,0.258) (0.010,0.170) (0,0.042) (0.002,0.168) Central South (CS) 0.003 0.012 0.012 0.690 0.008 0.003 0.011 (0,0.016) (0,0.064) (0,0.069) (0.667,0.746) (0,0.041) (0,0.016) , (0,0.066) Central West (CW) 0.004 0.066 0.025 0.021 0.822 0.125 0.014 (0,0.021) (0.009,0.162) (0,0.112) (0,0.078) (0.729,0.913) (0.054,0.195) ' (0,0.075) Central East (CE) 0.029 0.115 0.094 0.036 0.025 0.847 0.200 (0,0.096) (0.034,0.225) (0.014,0.214) (0,0.168) (0,0.113) (0.779,0.923) (0.079,0.295) North (N) 0.003 0.012 . 0.012 0.009 0.008 0.003 0.696 . (0,0.015) (0,0.060) (0,0.068) (0,0.046) (0,0.049) (0,0.016) . (0.668,0.772) o o Table 4.5. Direct movement and dispersal distances for Vancouver Island white-tailed ptarmigan determined from aerial and ground radio-telemetry data. The diagonal represents movements within a population. The first line presents the number of movements within the population/the total number of movements observed. The number of individual birds contributing to the movement analysis are then presented in brackets. The second line on the diagonal presents the mean distance moved (km) and the range of distances in brackets. The lower diagonal presents the geographic distance between the populations (km). Vancouver Island white-tailed ptarmigan were never observed moving beyond their population boundaries. Rate to Rate from South (S) South West (SW) Beaufort (B) Central South (CS) Central West (CW) Central East (CE) North (N) South South West Beaufort Central South Central West Central East North (S) (SW) (B) (CS) (CW) (CE) ' (N) 93/93 (31) 1.1 (0-6.8) 50.5 39.8 81.2 93.6 79.2 137.6 32/32 (11) 0.8(0-3.6) 37.2 8/8 (4) 2.8 (0-10.1) 41.0 46.9 6/6 (3) 1.4(0.3-5.2) 60.9 55.0 22.0 72/72 (18) 0.9(0-10.0) 55.0 39.6 26.6 17.9 130/130 (45) 1.1(0-10.5) 167.7' 128.0 96.9 77.0 88.6 7/7 (6) 0.6 (0.2-2.0) Figure 4.1. Neighbour-joining consensus dendrogram showing population clustering for Vancouver Island white-tailed ptarmigan based on the genetic distance of Nei (1978). The one partition (90%) groups both sampling periods (inventory and intensive) in the South together and those in the Central East together. The number on the branch is percentage of 1000 bootstrap replicates. Inventory South Intensive South Intensive Central East 102 Figure 4.2. Isolation by distance pattern for Vancouver Island white-tailed ptarmigan. Regression of genetic differentiation [estimated by FST/(1-FST)] against the logarithm of geographical distances (km) for all pairs of populations. Circled points refer to data causing deviations from a pattern of migration-drift equilibrium and demonstrate restricted gene flow relative to geographic distance. 103 F i g u r e 4.3. Principal components analysis on multilocus genotypes of Vancouver Island white-tailed ptarmigan. The first/three principal components (PC) explained 72% of the total variance. ™ o ~ 0 South West (SW) • North (N) Central West (CW) South (S) • Beaufort (B) Central Sou* Central East (CE) (CS) I South West (SW)# • Beaufort (B) *Central South (CS) Central West (CW) • • South(S) North (N)* * • Central East (CE) I 0 PCI (42%) 1 0 4 F i g u r e 4.4. Schematic depicting the results of pairwise FST comparisons and the between population migration rates (m) inferred by ByesAss for Vancouver Island white-tailed ptarmigan. R E F E R E N C E S Armstrong, D. P. 2005. Integrating the metapopulation and habitat paradigms for understanding broad-scale declines of species. Conservation Biology 19: 1402-1410. Baguette, M . 2004. The classical metapopulation theory and the real, natural world: a critical appraisal. Basic and Applied Ecology 5: 213-224. Baker, A. ML, P. B. Mather and J. M . Hughes. 2001. Evidence for long-distance dispersal in a sedentary passerine, Gymnorhina tibicen (Artamidae). Biological Journal of the Linnean Society 72: 333-342. Beerli, P. 2004. Effect of unsampled populations on the estimation of population sizes and migration rates between sampled populations. Molecular Ecology 13: 827-836. Berthier, K., N. Charbonnel, M . Galan, Y. Chaval and J.-F. Cosson. 2006. Migration and recovery of the genetic diversity during the increasing density phase in cyclic vole populations. Molecular Ecology 15: 2665-2676. Blais, J. M . , D. W. Schindler, D. C. G. Muir, L. E. Kimpe, D. B. Donald and B. Rosenberg. 1998. Accumulation of persistent organochlorine compounds in mountains of western Canada. Nature 395: 585-588. Boone, M . D. and O. E. Rhodes. 1996. Genetic structure among subpopulations of the eastern wild turkey (Meleagris gallopavo silvestris). American Midland Naturalist 135: 168-171. Bradford, D. F., A. C. Neale, M . S. Nash, D. W. Sada and J. R. Jaeger. 2003. Habitat patch occupancy by toads (Bufo punctatus) in a naturally fragmented desert landscape. Ecology 84: 1012-1023. Caizergues, A., A. Bernard-Laurent, J.-F. Brenot, L. Ellison and J.-Y. Rasplus. 2003. Population genetic structure of rock ptarmigan Lagopus mutus in Northern and Western Europe. Molecular Ecology 12: 2267-2274. Ciofi, C , G. A. Wilson, L. B. Beheregaray, C. Marquez, J. P. Gibbs, W. Tapia, H. L. Snell, A. Caccone and J. R. Powell. 2006. Phylogeographic history and gene flow among giant Galapagos tortoises on southern Isabela Island. Genetics 172: 1727-1744. Cornuet, J. M . , S. Piry, G. Luikart, A. Estoup and M . Solignac. 1999. New methods employing multilocus genotypes to select or exclude populations as origins of individuals. Genetics 153: 1989-2000. Crochet, P.-A. 1996. Can measures of gene flow help to evaluate bird dispersal? Acta Oecologica 17: 459-474. Doak, D. F. and L. S. Mills. 1994. A useful role for theory in conservation. Ecology 75: 615-626. 106 Ebert, D., C. Haag, M . Kirkpatrick, M . Riek, J. W. Hottinger and V. I. Pajunen. 2002. A selective advantage to immigrant genes in a Daphnia metapopulation. Science 295: 485-488. Elmhagen, B. and A. Angerbjorn. 2001. The applicability of metapopulation theory to large mammals. Oikos 94: 89-100.. 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. Excoffier, L., G. Laval and S. Schneider. 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1: 47-50. Franken, R. J. and D. S. Hik. 2004. Influence of habitat quality, patch size and connectivity on colonization and extinction dynamics of collared pikas Ochotona collaris. Journal of Animal Ecology 73: 889-896. Galbusera, P., L. Lens, T. Schenck, E. Waiyaki and E. Matthysen. 2000. Genetic variability and gene flow in the globally, critically-endangered Taita thrush. Conservation Genetics 1: 45-55. Gerlach, G. and H. N. Hoeck. 2001. Islands on the plains: metapopulation dynamics and female biased dispersal in hyraxes (Hyracoidea) in the Serengeti National Park. Molecular Ecology 10:2307-2317. Giles, B. E. and J. Goudet. 1995. A case study of genetic structure in a plant metapopulation. In: I. A. Hanksi and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 429-454. Goossens, B., L. Chikhi, P. Taberlet, L. P. Waits and D. Allaine. 2001. Microsatellite analysis of genetic variation among and within alpine marmot populations in the French Alps. Molecular Ecology 10: 41-52. Goudet, J. 1995. FSTAT (Version 1.2): A computer program to calculate F-statistics. Journal of Heredity 86: 485-486. Goudet, J., M . Raymond, T. deMeeus and F. Rousset. 1996. Testing differentiation in diploid populations. Genetics 144: 1933-1940. Hanski, I. and M . Gilpin. 1991. Metapopulation dynamics; brief history and conceptual domain. Biological Journal of Linnean Society 42: 3-16. Harper, G. L. , N. Maclean and D. Goulson. 2003. Microsatellite markers to assess the influence of population size, isolation and demographic change on the genetic structure of the U K butterfly Polyommatus bellargus. Molecular Ecology 12: 3349-3357. 107 Harrison, S. 1991. Local extinction in a metapopulation context - an empirical-evaluation. Biological Journal of the Linnean Society 42: 73-88. Harrison, S. and D. A. Taylor. 1997. Empirical evidence for metapopulation dynamics. In: I. A. Hanksi and M . E. Gilpin. Metapopulation Biology: ecology, genetics, and evolution, Academic Press: 27-42. Hartl, D. L. and A. G. Clark. 1997. Principles of population genetics. Sinauer Associates, Inc., Sunderland. Hastings, A. and S. Harrison. 1994. Metapopulation dynamics and genetics. Annual Review of Ecology and Systematics 25: 167-188. Hutchison, D. W. and A. R. Templeton. 1999. Correlation of pairwise genetic and geographic distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53: 1898-1914. Kimura, M . and G. H. Weiss. 1964. The stepping stone model of population structure and the decrease of genetic correlation with distance. Genetics 49: 561-576. Krajick, K. 2004. All downhill from here? Science 303: 1600-1602. Kvist, L. , M . Ruokonen, A. Thessing, J. Lumme and M . Orell. 1998. Mitochondrial control region polymorphism reveal high amount of gene flow in Fennoscandian willow tits (Parus montanus borealis). Hereditas 128: 133-143. Martin, K., P. B. Stacey and C. E. Braun. 2000. Recruitment, dispersal, and demographic rescue in spatially-structured white-tailed ptarmigan populations. Condor 102: 503-516. Martin, K., G. A. Brown and J. R. Young. 2004. The historic and current distribution of the Vancouver Island white-tailed ptarmigan (Lagopus leucurus saxatilis). Journal of Field Ornithology 75: 239-256. Martin, K. M . 2001. Wildlife in alpine and sub-alpine habitats. In: D. H. Johnson and T. A. O'Neil. Wildlife-Habitat Relationships in Oregon and Washington., Oregon State University Press: 285-310. Michalakis, Y . and L. Excoffier. 1996. A generic estimation of population subdivision using distances between alleles with special reference for microsatellite loci. Genetics 142: 1061-1064. Moilanen, A., A. T. Smith and I. Hanski. 1998. Long-term dynamics in a metapopulation of the American pika. American Naturalist 152: 530-542. Newman, R. A. and T. Squire. 2001. Microsatellite variation and fine-scale population structure in the wood frog (Rana sylvatica). Molecular Ecology 10: 1087-1100. 108 Oyler-McCance, S. J., N. W. Kahn, K. P. Burnham, C. E. Braun and T, W. Quinn. 1999. A population genetic comparison of large- and small-bodied sage grouse in Colorado using microsatellite and mitochondrial DNA markers. Molecular Ecology 8: 1457-1465. Pannell, J. R. and B. Charlesworth. 2000. Effects of metapopulation processes on measures of genetic diversity. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 355: 1851-1864. Piertney, S. B., A. D. C. MacColl, P. J. Bacon and J. F. Dallas. 1998. Local genetic structure in red grouse (Lagopus lagopus scoticus): evidence from microsatellite DNA markers. Molecular Ecology 7: 1645-1654. Raymond, M . and F. Rousset. 1995. Genepop (Version-1.2) - population-genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248-249. Ricketts, T. H. 2001. The matrix matters: effective isolation in fragmented landscapes. The American Naturalist 158: 87-99. Rousset, F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 1219-1228. Segelbacher, G. and I. Storch. 2002. Capercaillie in the Alps: genetic evidence of metapopulation structure and population decline. Molecular Ecology 11: 1669-1677. Segelbacher, G., J. Hoglund and I. Storch. 2003. From connectivity to isolation: genetic consequences of population fragmentation in capercaillie across Europe. Molecular Ecology 12: 1773-1780. Slatkin, M . 1987. Gene flow and geographic structure of natural populations. Science 236: 787-792. Slatkin, M . 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139: 1463-1463. Smith, M . A. and D. M . Green. 2005. Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28:110-128. Watson, A. and R. Moss. 2004. Impacts of ski-development on ptarmigan (Lagopus mutus) at Cairn Gorm, Scotland. Biological Conservation 116: 267-275. Weir, B. S. and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370. Wilson, G. A. and B. Rannala. 2003. Bayesian inference of recent migration rates using multilocus genotypes. Genetics 163: 1177-1191. 109 C H A P T E R 5 Conclusions Through this research I found that white-tailed ptarmigan on Vancouver Island (Lagopus leucura saxatilis) have unique habitat selection behaviour compared to other populations and subspecies of white-tailed ptarmigan. My comparison of competing models of habitat selection showed that Vancouver Island white-tailed ptarmigan pursue a generalist approach to habitat selection, choosing sites that provide for multiple requirements for survival and reproduction. These findings represented the presumed tradeoffs between foraging, predator avoidance (Hik 1995, Sih and Christensen 2001, Verdolin 2006), and perhaps thermal regulation. All previous studies of white-tailed ptarmigan habitat use revealed a strong preference for willow (Salix sp.; Choate 1963, Weeden 1967, May and Braun 1972, Herzog 1977, Herzog 1980, Giesen and Braun 1992, Frederick and Gutierrez 1992, Allen and Clarke 2005). Given the absence of abundant willow on Vancouver Island, white-tailed ptarmigan choose sites based on other habitat characteristics during the brood rearing period. My habitat selection analyses revealed that Vancouver Island white-tailed ptarmigan select sites with three key habitat variables: 1) ericaceous shrubs, 2) boulder cover, and 3) proximity of water. The flowers of the dominant ericaceous shrubs, white mountain heather (Cassiope mertensiana) and pink mountain heather (Phyllodoce empetriformis), were the preferred food during the brood rearing period and used sites contained a greater than random proportion of these species, highlighting the importance of these food species. Predator avoidance is an important component of white-tailed ptarmigan life history since they tend to follow a survivor or bet-hedging life history strategy (Sandercock et al. 2005b). Therefore, Vancouver Island white-tailed ptarmigan likely selected boulder cover because it provides cover from predators. Boulder cover may also provide shade to aid in thermoregulation. The selection for areas closer to water suggests Vancouver Island white-tailed ptarmigan prefer moist areas and is consistent with other white-tailed ptarmigan habitat use studies (Choate 1963, Herzog 1977, Frederick and Gutierrez 1992). Moist areas likely provide greater insect availability, which can influence the survival of grouse chicks (Hudson 1986). Moist areas may also be closer to snow fields and thus assist with thermoregulation. My analyses of population performance revealed significant regional variation, with the central region populations generally performing better than southern region populations. The central region showed higher female reproductive success than the southern region. Adult 110 white-tailed ptarmigan have higher survival and reproductive output than yearling birds (Wiebe and Martin 1998, Sandercock et al. 2005a) and the central region had a higher proportion of adult to yearling birds than the southern region. Furthermore, white-tailed ptarmigan are a monogamous species with typically male biased sex ratios (Braun et al. 1993). The southern region was much more male-biased in its sex ratio than the central region, suggesting lower female survival and reproduction in the southern region. These results combine to clearly show higher population performance in the central region than in the south. Studies of metapopulation dynamics have shown that the probability of local extinction (a reflection of population performance) can often be predicted by the coarse-scale measures of patch area and isolation (Thomas and Jones 1993, Hanski 1998, Moilanen and Hanski 1998). However, other studies indicate that environmental variables, such as habitat quality, have a greater influence on the metapopulation dynamics of a species (Fleishman et al. 2002, Matter and Roland 2002, Walker et al. 2003). I identified Vancouver Island white-tailed ptarmigan preferences for several key habitat variables by examining their resource selection behaviour. Animals likely select habitats that maximize individual fitness (Manly et al. 2002) and therefore correlations between preferred habitat variables and population performance may reveal habitat features important to the reproductive success and survival of a species. My analyses of regional variation in key habitat variables revealed that only the proximity to water correlated in the predicted direction with regional variation in population performance. This result provides some suggestion that available moisture can influence population performance, perhaps through its influence on insect availability or plant productivity in late summer. Unfortunately, experimental manipulations of water availability in late summer were unfeasible given the remoteness of my study sites. Future studies could measure insect availability during the brood rearing period to further clarify the relationship between moisture availability and population performance. Vancouver Island white-tailed ptarmigan have a unique population genetic structure. All seven populations demonstrated high levels of diversity (mean H E = 0.78) combined with high F]s values and significant heterozygote deficiencies. Examples of terrestrial vertebrate populations with high molecular diversity and generalized heterozygote deficiencies are very rare (see Chapter 3). I provided strong evidence that heterozygote deficiencies were not caused by either null alleles or short allele dominance. I also concluded that observed heterozygote deficits were not likely the product of undetected subpopulation structure (the Wahlund effect). The apparent paradox of high diversity combined with high Fis and generalized heterozygote 111 deficiencies are best explained by two scenarios. First, sampling may have captured a snapshot of a group of populations on their way to severe isolation suggesting that significant geographic isolation among populations may have existed long enough for the increase of inbreeding but not long enough for drift to result in strong population differentiation and a corresponding decrease in diversity. The second scenario evokes a pattern of infrequent dispersal among populations sufficient to maintain high levels of diversity, combined with low densities and limited mate choice resulting in the relatively quick accumulation homozygosity levels within populations. Dispersal affects the demographic and adaptive trajectories of animal populations and is fundamental to understanding population dynamics. My research investigated the patterns of dispersal and identified barriers to gene flow for this threatened subspecies of white-tailed ptarmigan. I examined dispersal patterns and population connectivity among seven populations using 10 microsatellite loci. I also addressed patterns of dispersal using direct measures by following the inter-seasonal movements of radio-collared birds. The genetic results showed low, but significant, genetic differentiation between most populations, and both direct and genetic estimates of dispersal suggested limited gene flow among populations. The stepping-stone model of dispersal was the most likely explanation for dispersal patterns in Vancouver Island white-tailed ptarmigan. Analysis of molecular data also showed a generally consistent pattern of isolation by distance. However, large areas of unsuitable low elevation habitat might act as barriers to gene flow. My research applied a broad range of techniques in order to understand the processes that influence the patterns of distribution and abundance in animals (Krebs 1994). Coarse-scale parameters such as the size of the habitat matrix surrounding suitable habitat patches influenced the probability of dispersal among Vancouver Island white-tailed ptarmigan populations. The population genetics implied a pattern of infrequent dispersal sufficient to maintain high levels of diversity and suggested that populations can persist despite heterozygote deficiencies. In this research I also highlighted the importance of the different time scales on which the processes of drift and the accumulation of inbreeding operate and discussed how this variation could influence population genetic structure. Fine-scale habitat selection revealed a generalist approach and a capacity to 'compensate' for the lack of willow in Vancouver Island alpine habitats. Identification of key habitat variables suggested that moisture availability might have an important influence on regional variation in population performance. 112 Future research directions In Chapter 2,1 identified preferred components of white-tailed ptarmigan habitat on Vancouver Island. However, the relative influence of fine-scale distribution and availability of resources and the coarse-scale attributes of patch area and isolation on the regional variation in population performance remains unclear. To approach this issue, I intend to use Landsat imagery to classify white-tailed ptarmigan habitat based on the ground vegetation plots (habitat use and availability plots) described in Chapter 2. Vegetation plots will be classified based on their relevance to Vancouver Island white-tailed ptarmigan habitat selection and key habitat variables. Example habitat classes are: 1) primarily food species coverage, 2) primarily boulder cover, and 3) bare ground; up to approximately 8 classes. This classification scheme will be extrapolated to other parts of Vancouver Island containing potentially suitable habitat via supervised classification of satellite imagery. I will do the GIS work in collaboration with Dr. Yuri Zharikov of Simon Fraser University. By combining these data with existing information on population densities, we can assess whether the resulting vegetation maps accurately predict bird abundance. I will also compare the distribution of preferred habitat variables and patch area and isolation to address which components likely have the highest impact on population performance. For example, if the distribution of preferred habitat variables is similar between the southern and central regions of the Island, then it is likely coarse-scale patch area and isolation and their influences on dispersal etc. that are the most important proximate causes of lowered population performance in the southern region of the Island. Further studies into the population genetics of white-tailed ptarmigan and other high elevation grouse species holds potential to influence both theoretical and applied aspects of population genetics. Yukon populations of white-tailed ptarmigan are less fragmented than Vancouver Island populations. I am interested in collecting and genotyping more samples from white-tailed ptarmigan populations in the Yukon. I predict that Yukon population genetics will show higher connectivity and exchange of individuals compared to Vancouver Island populations. I also expect fewer instances of heterozygote deficits within the Yukon populations as a result of greater connectivity and migration among populations. However, if Yukon populations show similar patterns of high diversity and heterozygote deficiencies to the Vancouver Island white-tailed ptarmigan populations, I would suggest that Yukon white-tailed ptarmigan follow similar patterns of infrequent dispersal and limited mate choice in congruence with the explanations I detailed in Chapter 3. 113 The Pacific Northwest, including Vancouver Island, is experiencing significant increases in temperature as a result of global climate change (Service 2004). In many mountain areas an increase in mean annual temperature has resulted in lower snow packs and an upward shift of low elevation plant and animal communities to higher elevations (Beniston 2003). In some cases, treelines are advancing to higher altitudes (Wardle and Coleman 1992, Meshinev et al. 2000, Kullman 2001); alpine plants are shifting upwards in elevation from 1-4 m per decade (Grabherr et al. 1994, Parmesan 1996), and lowland birds are extending their distribution from lower mountain slopes to higher areas (Pounds et al. 1999). These changes will result in a decrease in the amount of alpine habitat. The up shifting of low elevation communities and treelines could result in smaller habitat patches for white-tailed ptarmigan and greater distances between populations. Predictions for southern Vancouver Island based on a mean annual temperature increase of ~3°C show an upward elevational shift in biogeoclimatic zones resulting in a significant decrease of available alpine habitat (Hebda 1998). This shift could result in greater separation of the south and central regions of the island with the central region progressing, structurally and functionally to the current patterns observed in the southern region. Many findings of my research suggest detrimental impacts to Vancouver Island white-tailed ptarmigan as a result of climate warming trends. White-tailed ptarmigan on Vancouver Island prefer areas that are closer to water. Increased winter temperature will likely lead to a decrease in spring and summer snow pack, leaving alpine areas, particularly in the southern region, desiccated in late summer during the brood rearing period. Drier habitat will likely produce fewer insects and may affect juvenile survival. The increase in unsuitable habitats in the matrix surrounding habitat patches will also lead to greater separation of alpine habitats and limit the already infrequent dispersal among populations. Research reported in Chapters 3 and 4 suggested little movement between the most isolated South population. Greater unsuitable habitat may lead to similar isolation in other populations. If movement remains limited (or is hindered more with increasing unsuitable habitat) to the South population I expect a decrease in levels of diversity and increase in homozygosity. If this leads to inbreeding depression, then I expect further decreases in the population performance of the South population of Vancouver Island white-tailed ptarmigan. 114 R E F E R E N C E S Allen, T. and J. A. Clarke. 2005. Social learning of food preferences by white-tailed ptarmigan chicks. Animal Behaviour 70: 305-310. Beniston, M . 2003. Climate change in mountain regions: a review of possible impacts. Climate Change 59: 5-31. Braun, C. E. , K. Martin and L. A. Robb. 1993. White-tailed ptarmigan (Lagopus leucurus). In: F. Gill and A. Poole. The Birds of North America, 68. Academy of Natural Sciences, Philadelphia, PA, and American Ornithologists' Union, Washington, D.C.: 22. Choate, T. S. 1963. Habitat and population dynamics of white-tailed ptarmigan in Montana. Journal of Wildlife Management 27: 684-699. Fleishman, E. , C. Ray, P. Sjogren-Gluve, C. L. Boggs and D. D. Murphy. 2002. Assessing the roles of patch quality, area, and isolation in predicting metapopulation dynamics. Conservation Biology 16: 706-716. Frederick, G. P. and R. J. Gutierrez. 1992. Habitat use and population characteristics of the white-tailed ptarmigan in the Sierra Nevada, California. The Condor 94: 889-902. Giesen, K. M . and C. E. Braun. 1992. Winter home range and habitat characteristics of white-tailed ptarmigan in Colorado. Wilson Bulletin 104: 263-272. Grabherr, G. M . , M . Gottfried and H. Pauli. 1994. Climate effects on mountain plants. Nature 369: 448. Hanski, I. 1998. Metapopulation dynamics. Nature 396: 41-49. Hebda, R. 1998. Atmoshperic change, forests and biodiversity. Environmental Monitoring and Assessment 49: 195-212. Herzog, P. W. 1977. Summer habitat use by white-tailed ptarmigan in southwestern Alberta. Canadian Field-Naturalist 91: 367-371. Herzog, P. W. 1980. Winter habitat use by white-tailed ptarmigan (Lagopus leucurus) in southwestern Alberta, Canada. Canadian Field-Naturalist 94: 159-162. Hik, D. S. 1995. Does risk of predation influence population-dynamics - evidence from the cyclic decline of snowshoe hares. Wildlife Research 22: 115-129. Hudson, P. 1986. Red Grouse: the biology and management of a wild gamebird. The Game Conservancy Trust, Fordingbridge. Krebs, C. J. 1994. Ecology: the experimental analysis of distribution and abundance. HarperCollins College Publishers, New York. 115 Kullman, L. 2001. 20th century climate warming and tree-limit rise in the southern Scandes of Sweden. Ambio 30: 72-80. Manly, B. F. J., L. L. McDonald, D. L. Thomas, T. L. McDonald and W. P. Erickson. 2002. Resource Selection by Animals: statistical design and analysis for field studies. Kluwer Academic Publishers, Dordecht, The Netherlands. Matter, S. F. and J. Roland. 2002. An experimental examination of the effects of habitat quality on the dispersal and local abundance of the butterfly Parnassius smintheus. Ecological Entomology 27: 308-316. May, T. A. and C. E. Braun. 1972. Seasonal foods of adult white-tailed ptarmigan in Colorado. Journal of Wildlife Management 36: 1180-1186. Meshinev, T., I. Apostolova and E. Koleva. 2000. Influence of warming on timberline rising: a case study on Pinuspeuce Griseb. in Bulgaria. Phytocoenologia 30: 431-438. Moilanen, A. and I. Hanski. 1998. Metapopulation dynamics: effects of habitat quality and landscape structure. Ecology 79: 2503-2515. Parmesan, C. 1996. Climate and species' range. Nature 382: 765-766. Pounds, J. A., M . P. L. Fogden and J. H. Campbell. 1999. Biological response to climate change on a tropical mountain. Nature 398: 611-615. Sandercock, B. K., K. Martin and S. J. Hannon. 2005a. Demographic consequences of age-structure in extreme environments: population models for arctic and alpine ptarmigan. Oecologia 146: 13-24. Sandercock, B. K., K. Martin and S. J. Hannon. 2005b. Life history strategies in extreme environments: comparative demography of arctic and alpine ptarmigan. Ecology 86: 2176-2186. Service, R. F. 2004. Water resources: as the west goes dry. Science 303: 1124-1127. Sih, A. and B. Christensen. 2001. Optimal diet theory: when does it work, and when and why does it fail? Animal Behaviour 61: 379-390. Thomas, C. D. and T. M . Jones. 1993. Partial recovery of a skipper butterfly (Hesperia-Comma) from population refuges - lessons for conservation in a fragmented landscape. Journal of Animal Ecology 62: 472-481. Verdolin, J. L. 2006. Meta-analysis of foraging and predation risk trade-offs in terrestrial systems. Behavioral Ecology and Sociobiology 60: 457-464. Walker, R. S., A. J. Novaro and L. C. Branch. 2003. Effects of patch attributes, barriers, and distance between patches on the distribution of a rock-dwelling rodent (Lagidium viscacid). Landscape Ecology 18: 187-194. 116 Wardle, P. and M . C. Coleman. 1992. Evidence for rising upper limits of 4 native New-Zealand forest trees. New Zealand Journal of Botany 30: 303-314. Weeden, R. B. 1967. Seasonal and geographic variation in the foods of adult white-tailed ptarmigan. Condor 69: 303-309. Wiebe, K. L. and K. Martin. 1998. Age-specific patterns of reproduction in white-tailed and willow ptarmigan Lagopus leucurus and L. lagopus. Ibis 140: 14-24. 117 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0074963/manifest

Comment

Related Items