Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Dynamics of species’ range : adaptation and gene flow in Sitka spruce Mimura, Makiko 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_2006-198971.pdf [ 6.91MB ]
Metadata
JSON: 831-1.0075047.json
JSON-LD: 831-1.0075047-ld.json
RDF/XML (Pretty): 831-1.0075047-rdf.xml
RDF/JSON: 831-1.0075047-rdf.json
Turtle: 831-1.0075047-turtle.txt
N-Triples: 831-1.0075047-rdf-ntriples.txt
Original Record: 831-1.0075047-source.json
Full Text
831-1.0075047-fulltext.txt
Citation
831-1.0075047.ris

Full Text

DYNAMICS OF SPECIES' RANGE: ADAPTATION AND GENE FLOW IN SITKA SPRUCE by MAKIKO MIMURA B.Sc. Osaka Prefecture University, 1998 M.Sc. Washington State University, 2001 THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n The Faculty of Graduate Studies (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA March 2006 © Makiko Mimura 2006 Thesis Abstract In this thesis I investigate the evolutionary dynamics of species' range to determine the effects of gene flow on adaptation and range expansion. Sitka spruce (Picea sitchensis) was employed as the model species in this empirical study. It has a narrow distribution along the Pacific coast covering over 22° of latitude along a relatively linearly heterogeneous environment. Gene flow can increase fitness by increasing genetic diversity or masking recessive deleterious alleles, but also decrease fitness by homogenizing adaptive divergence. I investigated degrees of adaptation and genetic structure between central and peripheral populations as well as between isolated and continuous populations. Common garden experiments suggested strong adaptive divergence along environmental gradients in some growth traits. Pollen genetic structure based on microsatellite loci indicated that recent gene flow has been limited, while other studies based on more conservative markers in Sitka spruce suggested little genetic structure. These imply that during rapid postglacial migration, long-distance dispersal and high gene flow may have initially provided genetic diversity to founder populations. This increased genetic diversity enhances the efficiency of local selection relative to genetic drift, resulting in adaptive divergence along environmental gradients. Sitka spruce populations adapt to local conditions to some extent, but may not adapt optimally because of the homogenizing effect of gene flow. Gene flow and selection can counteract each other, depending on the levels of gene flow and the intensity of selection. At range peripheries, the effects of gene flow may be stronger than selection, limiting further adaptation beyond the range limits. Mating system analysis indicated higher selfing rate and a dramatic decline in effective pollen donor number in peripheral compared to central populations. This difference in population size and resulting density gradients may cause asymmetric gene flow from the centre to the range peripheries. At both the southern and northern range peripheries, despite high inbreeding, the isolated populations showed significantly higher juvenile fitness compared to the continuous populations in the same environments. Isolation has increased the efficiency of selection at soft species borders. The positive and negative effects of gene flow, its circumstances and consequences are discussed in this thesis. Table of Contents Abstract ii Table of Contents Hi List of Tables v List of Figures vi Acknowledgements vii Co-Authorship Statement viii Chapter 1: Research Objectives and Related Literature Review Introduction 1 Literature Review Species' ranges 1 Range shifts and their consequences 5 Sitka spruce and Pacific North America in the Pleistocene 8 Distribution and genetic variation of Sitka spruce 9 Genetic markers for Sitka spruce 11 Statistical approaches 12 Thesis introduction 15 Literature cited 17 Chapter 2: Increased selfing and decreased effective pollen donor size in peripheral relative to central populations in Sitka spruce Introduction 23 Material and Methods 25 Results 26 Discussion 28 Literature Cited 39 Chapter 3: Isolation by distance with postglacial migration and adaptive gradients in growth traits in Picea sitchensis (Pinaceae) Introduction 42 Material and Methods 44 Results 48 Discussion 50 Literature Cited 62 Chapter 4: Gene flow inhibits local adaptation in peripheral populations of Sitka spruce Introduction 65 Material and Methods 67 Results 69 Discussion 71 Literature Cited 83 Chapter 5: Thesis Conclusion Introduction 85 Major Findings Tradeoffs between adaptation and gene flow on the species' borders 86 Response to environmental changes in tree species 87 Future Research 88 List of Table Table 2.1. Classification of population sampled 33 Table 2.2. Allele frequency data for five microsatellite loci studied in this thesis 33 Table 2.3. Summary of mating system of Sitka spruce 35 Table 2.4. Estimated number of effective pollen donors 35 Table 3.1. Summary of sample populations with geographic origin 54 Table 3.2. Pair-wise genetic distance between tested populations 54 Table 3.3. Summary of growth traits among populations transferred along the coast 55 Table 3.4. Pearson correlations of population means in growth traits 55 Table 3.5. Eigenvectors of the first (PCI) and second (PC2) components from PCA 56 Table 3.6. Summary of result from the regression of PCA with geographic variables 56 Table 4.1. Comparison of pairs of continuous and isolated, central and peripheral population 76 Table 4.2. Regression of population means with distance along the coast 76 Table 4.3. Among- and within- population variation in growth and phenology traits 77 Table 4.4. Adaptive performance in isolated and continuous peripheral populations 78 v Lists of Figures Figure 1.1. Sitka spruce natural range map with putative historical migration routes 16 Figure 2.1. Sitka spruce range map showing populations sampled for Chapter 2 36 Figure 2.2. Pollen haplotype diversity (Hh e) 37 Figure 2.3. Genetic parameter estimates for mating system of Sitka spruce 38 Figure 3.1. Sitka spruce natural range showing population sampled for Chapter 3 57 Figure 3.2. Cluster analysis based on genetic distance, with U P G M A and NJ methods 58 Figure 3.3. Geographic distance versus genetic distance among populations 59 Figure 3.4. Plot of the first (PCI) and second (PC2) components of PCA 60 Figure 3.5. Regression analysis of the first two components with climate variables 61 Figure 4.1. Range map of Sitka spruce showing population sampled for Chapter 4 79 Figure 4.2. Population means in selected growth response variables 80 Figure 4.3. Reaction norms of populations across the tested environments 81 Figure 4.4. Juvenile Fitness Index estimates for populations 82 vi Acknowledgement I first acknowledge, with great appreciation, my most supporting and encouraging supervisor, Dr Sally Aitken for sharing her extraordinary knowledge in forestry and conservation genetics and encouragement to read and think various research topics to deepen my knowledge. I would also like to express gratitude to my committee members, Dr Mike Whitlock who shared his ideas, helped statistics and taught me what scientific fairness is, Dr Peter Arcese who shared his ideas and provided useful comments throughout my thesis chapters from his various scientific views that improved my thesis and Dr Greg Henry who shared his knowledge in paleobotany that brought me to the challenge of understanding migration of tree species. I also thank Dr Carol Ritland who patiently shared her knowledge on my thesis proposal and the laboratory work, Dr John Alden who provided his seed collection for this research and shared his wisdom especially in biology of Sitka spruce, Dr John King who kindly shared his seed collection, Christine Chourmouzis who helped my intensive field work and writing with her great patience and shared her ideas and encouragement, Joanne Tuytel who helped me in the field experiments and very intensive laboratory work, and Drs Andress Hamman and Tongli Wang who helped with the statistical analysis and concepts. My acknowledge also goes to the former and current Aitken lab members, my colleagues and especially to my friends and family who are always encouraging without asking me about science. This thesis would not have been completed without them. vii Co-Authorship Statement This thesis was written as a series of related manuscripts on the advice of my supervisory committee. Chapter 1 provides the context, literature review and research objectives. Chapter 2, 3 and 4 are manuscripts to be submitted to journals, and followed a typical scientific journal style including introduction, materials and methods, results and discussion. Chapter 5 summarizes results and discussion of this thesis and addresses further research directions. For all chapters of this thesis, I took the lead in developing the ideas, collecting and analyzing data, and producing draft manuscripts, figures, and tables. However, this research would not have done without the contribution of my supervisor, Sally Aitken. She initially developed the concept for chapters 2 and 4 of this project and obtained funding through her NSERC Discovery Grant proposal, helped me plan my thesis, provided all facilities required for experiments, discussed ideas, provided useful comments, and made editorial corrections in my writing with her great patience, throughout my academic years. To recognize her efforts, I have included her as a co-author for the manuscript chapters. viii CHAPTER 1 Research Objectives and Related Literature Review Introduction Anthropogenic climate change and local adaptation of populations are topics of global concern in the fields of natural resource production and conservation. Changing selection pressures and population dynamics affect the genetic structure of populations and ultimately the ability of populations to respond to environmental changes. A species' habitat is rarely uniform or continuous throughout its range, and is often distributed across heterogeneous environments. Peripheral populations are likely to have different evolutionary potentials from central populations due to physical distance from central populations, differences in selection, population size, and phylogeographic constrictions such as historically reduced genetic diversity at the tips of migration (Lesica and Allendorf 1995). Since peripheral populations are most likely be the first populations that are extirpated or expand in response to environmental changes, it is important to understand the nature of adaptation in peripheral populations, and the evolutionary consequences of peripheral and central population positions for prediction of possible future outcomes. In this thesis, I investigate the factors that limit the adaptation and expansion of peripheral populations, and the process of postglacial migration, by comparing adaptation and population structure between central and peripheral populations and between isolated and continuous populations. I employed Sitka spruce (Picea sitchensis (Bong.) Carr.), a tree species with a wide latitudinal range, from California to Alaska, as a model for this research. Literature review The species' ranges Factors for limiting species range: Theories Species' ranges may be limited by selection, reduced genetic variation, antagonistic relationships among adaptive traits, or gene flow (Barton 2001; Holt 2003; Wiens 2004; Case et al. 1 2005). Al l factors can act differently in core and peripheral populations, causing population divergence, and thus become determinants of species' range. Selection can be a strong factor limiting species' range. Environmental suitability may decrease with distance from the center of the range (Lawton 1993), thus regimes of natural selection in marginal environments are likely to be different than those in central environments. Divergent selection pressures may cause population divergence between central and peripheral populations (Lesica and Allendorf 1995). An extreme environment at a margin can limit population expansion. Interspecific interaction can also impose strong selection pressures and limit species adaptation and expansion (Case and Taper 2000; Case et al. 2005; Holt et al. 2005). In some species, reduced genetic variation, which may be associated with reduced fitness, is more commonly observed in marginal than in central populations. Peripheral populations may have suffered eroded genetic diversity through sequential re-colonization during Quaternary postglaciations, and population genetic structure may be imprinted by founder effects. Founder effects may create a maladapted phase at the margins of range expansion, due to initially low genetic variability. This reduces the genetic opportunity for adaptation to new environments. As migration proceeds, peripheral populations can become central populations. During rapid migration, reduced genetic diversity compared to ancestral populations can be found in all populations through sequential founder effects, for instance, in Pinus coulteri (Ledig 2000). Sequential founder effects during intensive migration may create significant maladaptation in marginal populations and limit opportunities for expansion of species' range. Peripheral populations also tend to be small and relatively isolated (Mayr 1963), and small populations are subject to substantial genetic drift. Smaller populations have a higher chance to fix alleles randomly (Hartl and Clark, 1997). As a result, Inbreeding rates may be higher in peripheral populations than central populations (Mitka 1997). Higher selection pressure at the margins of the range may also reduce genetic variation (Lawton 1993). Similar intensity and direction of selection can drive population divergence if populations have different genetic backgrounds due to genetic drift (Cohan and Hoffmann 1986). These events force peripheral populations to genetically differentiate from central populations. While both selection and reduced genetic diversity may be important factors, negative correlations between adaptive traits can also slow the rate of adaptation beyond the species' range limit (Etterson and Shaw 2001). When the genetic correlation between traits is antagonistic to the direction of selection, the response to the selection requires genetic recombination and beneficial mutation, which could substantially slow down local adaptation. On an evolutionary time scale, however, beneficial mutation should eventually arise and overcome these effects to allow adaptation to peripheral environments. 2 The homogenizing effect of gene flow is an alternative explanation for limiting species range, by disrupting increasing frequency of locally beneficial alleles and introducing maladaptive alleles to peripheral populations. If peripheral populations suffer lower fitness due to reduced genetic variation, increased gene flow could provide new variation and augment populations with new individuals or gametes (Barton 2001). The general effect of gene flow, however, can homogenize populations across environments, and under high gene flow, adaptive divergence between populations cannot be maintained. Theoretical studies have explored the strength of gene flow in counteracting local adaptation. Mayr (1963) proposed that gene flow with alien alleles from the centre of species' ranges can disrupt adaptation at range peripheries. Models of Kirkpatrick and Barton (1997) revealed that half the populations across the entire species ranges could act as demographic sink populations in a linearly heterogeneous environment. In this situation, population density would be highest where the populations are most adapted (e.g., the population mean phenotype closest to the phenotype of maximum fitness for that environment), and lowest in sink populations. This results in asymmetric gene flow of greater magnitude outward from source to sink populations than in the reverse direction. Thus, gene flow is generally higher from the center of the range to marginal populations than vice versa, when central populations have higher density than peripheral populations. Populations of lower density tend to be sink populations. These populations sometimes have higher immigration rates, which increase local population size above local carrying capacity, creating 'pseudo-sink' populations (Dias 1996). Pseudo-sink populations may be formed beyond the species' niche. While pseudo-sink populations may have a large population size or high density, they are characterized by poor reproductive rates and are dependent upon high immigration rates (Watkinson and Sutherland 1995). Climatic conditions, such as a prevailing wind, can also cause asymmetric gene flow. Because sink populations eventually consist of source populations' descendants, selection favours beneficial alleles for source populations even though these alleles are antagonistic in sink populations (Kawecki and Holt 2002). Thus, gene flow can have negative effects on the local adaptation of peripheral populations by introducing maladaptive genes from core populations, i.e. alleles adapted to core, not peripheral environments. From this theory, it is predicted that maladaptive traits will be seen in marginal populations of species with a relatively large and linearly heterogeneous range. Similar models also predict that the physical isolation of peripheral populations from central ones can allow for rapid local adaptation (Garcia-Ramos and Kirkpatrick 1997). The resulting rate of adaptation depends on the genetic potential within an isolated population as long as the population is large enough to incur only weak genetic drift. Their model combining isolation and selection explains rapid evolution in such populations better than the hypothesis of Mayr (1963) which primarily 3 involves genetic drift. The effects of gene flow in counteracting species' range expansion are still controversial, and few empirical studies have been done. Gene flow can also increase fitness by releasing populations from inbreeding depression or reducing the negative effects of genetic drift in small populations, although these effects may be weaker than the effects of introducing maladaptive genes (Wiener and Feldman 1993; Barton 2001). Outcomes would depend on the genetic variation in peripheral populations (Garcia-Ramos and Kirkpatrick 1997), the magnitudes of allelic effects, and the levels of gene flow between sink peripheral and source core populations (Kawecki 2000). Empirical studies for core and peripheral populations Empirical evidence supports the idea that peripheral populations are often genetically or morphologically divergent from others as a result of population history, selection intensity, number and interaction of genes involved, or immigrant alleles. Some conifers show lower genetic diversity in peripheral than in central populations; e.g., Chamaecyparis lawsonii (Millar and Marshall 1991), Pinus contorta (Yeh and Layton 1979), Pinus. ponderosa (Hamrick et al. 1989) and Pinus rigida (Schnabel and Hamrick 1990). Higher inbreeding coefficients and population substructure were also found in peripheral compared to central populations in Picea sitchensis (Gapare et al. 2005; Gapare and Aitken 2005). Peripheral populations may also differ in mating system and reproductive morphology. In the grass species Buchloe dactloides, monoecious forms are more common in geographically peripheral populations (Huff and Wu 1992). In the angiosperm Leavenworthia alabamica, self-compatibility evolved in peripheral populations while self-incompatibility is common in central populations (Busch 2005). Marginal populations of trees can, however, have the same or similar levels of genetic diversity compared to central populations, for instance in Picea abies (Mouna et al. 1990). Life history traits, such as fecundity and dispersal rate, may contribute to the amount of genetic diversity. Hamrick et al. (1992) suggested that tree species have generally higher genetic diversity within populations compared to other life forms, and that range size as well as mating system and dispersal strategies can predict levels of genetic diversity to some extent. For example, pollen can disperse further, on average, via wind than insect vectors, and smaller seed may be more mobile than larger and heavier seeds. Thus, relatively small winged-seeded species may be able to more quickly overcome founder effects from historical migration. 4 Empirical studies for the effects of gene flow on fitness Testing the effects of gene flow on local adaptation at range peripheries is an empirical task (Hoffmann and Blows 1994). Reduced genetic diversity in peripheral populations compared to central populations is sometimes found in plant species; however, associating fitness with reduced genetic diversity is controversial (Holt 2003; Jump and Woodward 2003). When gene flow swamps local adaptation, sink populations may show similar phenotypes to source populations even though adaptive peaks are different between these populations. Directional gene flow caused by environmental factors may maintain sink populations in a permanent maladaptive state (Stanton and Galen, 1997). In birds, there is empirical evidence that gene flow from different environments disrupted optimization of adaptive traits (Garant et al. 2005; Postma and van Noordwijk 2005). The influence of gene flow on adaptation is also evidenced by local adaptation in isolated populations. Isolated populations can also diverge more quickly for quantitative traits than for neutral genetic markers because of selection and genetic drift (Witter and Carr 1988; Merila and Crnokrak 2001). Cody and Overton (1996) observed the evolution of limited dispersal in several plant species in the Asteraceae in a newly founded island population within just a few generations. Range shifts and their consequences Natural selection increases the frequencies of beneficial alleles and advantageous gene combinations, thus increases mean fitness of a population in a particular environment. Geographical heterogeneity can result in strong differential natural selection. Quantitative trait clines associated with environmental gradients in temperature and moisture that are related to latitude, distance to coast and elevation have been reported for many plants (e.g. L i et al. 1998) for trees (reviewed by Morgenstern 1994) and for some animals (Jonas and Geber 1999). The distribution of plant species is strongly limited by environmental gradients and ecological differences (Stephenson 1990; Chuine and Beaubien 2001). Physical barriers such as oceans, mountains, rivers, and man-made structures, settlements, and transport corridors also affect the levels of gene flow between populations. The magnitude of the effect depends on the power of pollen and seed to disperse over these barriers. Climatic variation over time and space results in changes in natural selection and species' distribution. Changes in a species' range may lead to speciation via geographic isolation (e.g. Crawford et al. 1985), hybridization by encountering related species (e.g. Sang et al. 1995) and polyploidization (e.g. Brochmann et al. 1996). Climatic fluctuations over time have created shifting favorable refugia during glacial periods and periods of warming. When populations encounter environmental changes in their habitat, the population will (1) migrate to favourite habitats, (2) adapt to new environmental conditions, or (3) be extirpated. Migration or local extinction of 5 sub-populations is the most likely outcomes for many organisms (Hoffmann and Hercus 2000). Phylogeographic and phylogenetic studies indicate rapid expansion of species along with climate changes (Hewitt 1996; Hewitt 2000). Rehfeldt et al. (2001) observed that populations of Pinus contorta tend to inhabit colder environments than the optimum. This may be explained by rapid migration and adaptational lag. When migration is led by expansion of leading edge populations, isolation by distance from large, core populations precedes genetic differentiation from the core populations via drift or local adaptation of the leading populations. The rapid changes in species' distribution can result in reduced genetic diversity in small or leading-edge populations and even speciation in isolated populations. For instance, the reduced genetic diversity of red spruce (Picea rubens) may reflect speciation from progenitor sub-populations of black spruce (P. mariana) (Perron et al. 2000). Identifying the locations of previous refugia will contribute to understanding phylogeographic evolutionary demographics There are several lines of evidence for the existence of Pleistocene glacial refugia along the northern Pacific coast of North America, including fossil pollen records (Pielou 1991), a combination of pollen spectra and climate record (Adams and Faure 1997), and phylogeographic studies using genetic markers (Avise, 2000). Much palaeoclimatic and palaeoecological evidence suggests the existence of several major refugia in North America (Comes and Kadereit 1998). The Pacific coast north as far as Puget Sound, parts of northern Vancouver Island, as well as parts of the Queen Charlotte Islands and central Alaska may not have been fully glaciated when the west coast of Canada and southwest coast of Alaska were covered by glaciers 15,000 C 1 4 years ago (Pielou 1991), and when sea level was more than 150m lower than now in this area until around 12,400 1 4 C years BP (Barrie and Conway 2002). The fossil record provides physical evidence of past population dynamics. Fossils are generally defined as the prehistoric physical remains (such as seeds, needles and pollen) or trace of life (such as tracks and trails of animals) which are preserved by natural processes. Pollen grains with a nearly indestructible substance called sporopollenin are usually well preserved in sediments at the bottom of lakes, bogs and wetlands. Microscopically visible characteristics of the pollen surface are genus- or species-specific (Hebda 1996). Sometimes only genera can be identified from fossil pollen, for example Pinus ssp. and Picea ssp. However, observed pollen distribution generally matches with known species geographical distributions (Allen et al. 1999), and species can be identified from current geographical niches when species are allopatrically distributed. While fossil records provide the physical evidence of past distribution for phylogenetic and phylogeographic studies, there are some limitations, thus interpretations should be made with caution. For example, the low taxonomic resolution of the fossil record often limits knowledge of the details of species colonization routes. 6 Also plants with specific life history traits such as wind-dispersal or habitats such as wetlands or peat lands are more favourably preserved than others. Species with a low abundance of pollen, due to low density or fragmented distribution, may not be detected in the fossil pollen record, resulting in an underestimation of their past distribution (McLachlan et al. 2005). Therefore, molecular data is also used as an alternative or supplementary tool for phylogeographic studies. Chloroplast D N A (cpDNA) is widely used to provide insight into past distribution and migration for phylogeographic studies in angiosperms, while mitochondrial D N A is a important tool for detecting colonization routes in animal (Comes and Kadereit 1998). Soltis et al. (1997) studied cpDNA in both herbaceous and woody plant species and found a largely consistent phylogeographical pattern in Pacific coastal North America. Chloroplast haplotypes in six of the seven plant species were distinguishable into northern or southern genotypes. The exception to this pattern was Picea sitchensis, in which chloroplast are paternally inherited (Soltis et al. 1997). They concluded that the southern genotypes were likely descendants of big southern refugia during the last glacial period. According to Soltis et al. (1999), the origins of northern genotypes are explained by two hypotheses: "the north-south recolonization hypothesis" and "the leading edge hypothesis." The north-south recolonization hypothesis implies that two widely separated areas of remgia existed during glaciations: southern refugia and possible northern remgia (northern remgia are generally assumed to have had smaller population sizes than southern one). In response to postglacial warming, populations in possible northern and south refugia expanded their ranges and merged with each other in the uncolonized space, resulting in the current, continuous distributions. The leading edge hypothesis supports the survival and subsequent directional migration of limited founders with certain genotypes from a large single southern refugium. Thus, northern populations consist of certain genotypes which are descendants of the southern ancestral populations. Under this hypothesis, recolonization of available deglaciated areas likely resulted in a series of bottlenecks. Thus, northern populations should tend to have lower genetic diversity compared with southern populations under both hypothesized models. In both southwestern North America (e.g., Liston et al. 1992) and eastern North America (Parks et al. 1994; Sewell et al. 1996), lower genetic diversity in the north and higher in the south, has been observed. The highest levels of genetic diversity are often found in populations close to putative refugia, unless current high gene flow has eliminated those signatures or secondary contacts have occurred among populations from different migration routes and refugia (Petit et al., 2003). In Pacific North America, many studies have found that northern populations, assumed to be descended from relatively smaller northern refugia or founders originating from a southern refugium, have lower genetic diversity than southern populations. 7 However, reduced genetic diversity is not always found toward the end of migration routes, especially in tree species. Sometime there is equal diversity in both northern and southern populations (Mouna et al. 1990; Gamache et al. 2003; Muir et al. 2004), or even higher diversity in the northern populations (Hamann et al. 1998). This could be due to multiple migration routes resulting secondary contact (Petit et al. 2003) or the ability of new populations of trees to sustain an initially small population size while receiving sufficient gene flow from neighboring populations to provide genetic variation for adaptation and inbreeding avoidance (Davis and Shaw 2001). The rapid migration rates of trees in the past were driven by long-distance dispersal (Cain et al. 1998; Clark 1998; Nathan et al. 2002). Paleoecologists estimated that the rate of spread exceeded 102 m yr 1 at the end of the Pleistocene (Davis 1976; Macdonald 1993). Long-distance dispersal is the major force of rapid forest migration and suggests that the migration rate exhibited in the past is less than the maximum capacity for the species estimated from models (Clark 1998). Range shifts have served as an alternative to adaptation. Selection is still required to increase the frequency of beneficial alleles in new environments, especially when populations encounter different combinations of environmental factors such as new species of parasites (Davis and Shaw, 2001). Sitka spruce and Pacific North America in the Pleistocene It is well established that there were at least several glaciations during the Pleistocene and that plant and animal populations re-colonized ice-free areas in the interglacial periods (Comes and Kadereit 1998). Fluctuations of temperature and moisture over time and space during the Quaternary deglaciation affected forest formation and may explain some historical aspects of forest population dynamics. The ice sheet reached the northern half of North America (Canada and interior northern US) around 18,000 - 15,000 radiocarbon years ago (rya) (Last Glacial Maximum). Adams and Faure (1997) reconstructed the earth's physical geography based on inter-disciplinary interpretation of evidence from plant fossils, soil and sedimentological analyses. During the Last Glacial Maximum, the west coast of Canada was covered with ice, except for some ice free areas, a possible refugium of some species, e.g., parts of the Queen Charlotte Islands (Mathewes 1989), and perhaps northern Vancouver Island. The ice sheet reached south to Puget Sound, WA. At that time, the northwestern US was probably covered by Alpine tundra and polar desert, where only very scattered remnants of spruce may have existed (Whitlock and Bartlein 1997). Further south, the Pacific Coast was dominated by a conifer mosaic woodland at lower elevations (Thompson et al. 1993). By 15,000 rya, an ice-fee corridor began to form along coastal southwest BC, Canada. Vegetation was similar 13,000 rya although additional ice-free corridors appeared in interior and coastal Canada and Alaska. By 12,000 rya, ice-free corridors widened and boreal forests were dominant in lower coastal BC, 8 Washington and Oregon (Thompson et al. 1993). At this time, sea level was low enough, putatively, to connect some coastal islands to the mainland (e.g. Queen Charlotte Islands and Vancouver Island) (Barrie and Conway 2002). Around 11,000 rya, pollen spectra data indicate cooler and drier climatic conditions predicted (Anderson 1997), which may have slowed species' expansions. By 11,000 rya, conifer forests were distributed up to coastal southeast Alaska from around 40°N; but remaining glaciers fragmented this area. Adams and Faure (1997) suggested the forest at this time might have been mainly or partly dominated by spruce (Picea spp). By 10,000 rya, the cool, arid phase ended and temperature increased. Pacific Coastal North American forests were dominated by cool conifer species (Adams and Faure 1997). Around 8,000 rya, temperatures were slightly higher than present. Present vegetation was established by 5,000 rya. Conifer species are thought to have recolonized Pacific North America quickly as temperature increased and glaciers retreated. Sitka spruce is a pioneer species in North America that extended its range from its southernmost populations, at least 22 km south of San Francisco Bay, California, in the early Pleistocene, reviewed in Daubenmire (1967), to its current northern limit on Kodiak Island (57.8°N), Alaska. According to Mann and Hamilton (1995), Sitka spruce maintained scattered populations in Puget Sound, very close to the boundary of ice sheets 15,000 rya, and reached northern Vancouver Island, at the latest, by 10,000 rya (Figure 1.1). Fossil needles and seeds were found on the Queen Charlotte Islands, BC, from layers estimated at 9,600 rya (Pellatt and Mathewes 1994). It had reached the Alaska panhandle by 7,000 rya, and at the Prince William Sound, the Inlet of central south Alaska, at the latest by 3,000 rya (Mann and Hamilton 1995). Sitka spruce may have arrived on Kodiak Island around 300 to 600 years ago (J Alden, US Forest Service, personal communication). On the island, the Sitka spruce populations are migrating approximately 160 m per year (Griggs 1937). Distribution and genetic variation of Sitka spruce Distribution Sitka spruce is one of the most commercially valuable trees along the Pacific Coast. Its wood has a high strength to weight ratio and excellent sound qualities (Burns and Honkala 1990). It is also an ecologically important tree species. Sitka spruce has a continuous, long and narrow distribution along the Pacific Coast (Little 1953; Peterson et al. 1997), which demonstrates its ability to adapt to photoperiod and temperature across a wide range of latitudes. The species occurs in hypermaritime to maritime cool mesothermal climates on nitrogen-rich soils, and its density decreases with increasing elevation associated with drought and cold (Klinka et al. 1989). Northern California is the southern tip of the species' range, potentially containing descendants of survivors from a putative southern 9 refugium that would have had a larger population size during glacial periods. Kodiak Island and the coastal side of Katmai National Park and Preserve, Alaska are the northern limits (or "migration tip") of the species range. Quantitative trait variation Sitka spruce exhibits strong clinal variation along with latitude and altitude. The populations from the Queen Charlotte Islands and Vancouver Island showed strong clinal variation in bud set timing with latitude in a common garden experiment (Falkenhagen, 1977). Height growth was also correlated with latitude and altitude, but not longitude (Falkenhagen, 1977). Temperature and photoperiod are major factors explaining population differentiation for British Columbia (Xu et al., 2000). Provenance experiments in Alaska (N 58-59 °C, W l 52-154 °C) for the populations from southeast Alaska (Juneau, 58-59 °N latitude) and southwest Alaska (Afognak Island, 58-59 °N latitude) showed that the growth performance was correlated with longitude (Alden 1998). Although the southeast Alaska population showed higher growth rates than the southwest Alaska population, 20% of SW populations produced cones while none of SE population did 16 years after planting (in 1994). Long-term provenance tests revealed clinal patterns of variation for Sitka spruce where early growth associated with latitude shifted to longitudinal clines (coast-inland) at later ages (Ying 1997). Greener foliage was associated with higher growth rate (Copes 1996; Alden 1998). The colour of needles was correlated with latitude and longitudes in the B C populations (Falkenhagen, 1977). Darker foliage in colder environments may be a result of developed cold hardiness in cold environment, or possibly due to hybridization with white spruce, which is more adapted to cold environments than Sitka spruce. Genetic diversity and hybridization with white spruce Long-lived perennial woody plants tend to have higher genetic diversity within populations and less population differentiation than other herbaceous annual and perennial species (Hamrick et al. 1992). This is thought to be due primarily to long distance wind-born pollen flow or longevity, homogenizing among-population variation and increasing within-population variation relative to among-population variation. Sitka spruce is reported to have high genetic diversity for isozymes (Yeh and El-Kassaby 1980), random amplified polymorphic D N A (RAPD) (Vandeven and McNicol 1995), and sequence-tagged-site (STS) makers (Gapare et al. 2005). Yeh and El-Kassaby (1980) reported 92% of total neutral genetic variation was within populations. Conifers generally have a high (>90%) outcrossing rate (Mitton 1992), and the outcrossing rate in a Sitka spruce seed orchard was reported to be 92% (Chaisurisri et al. 1994). Reported measures of population differentiation of P. sitchensis vary 10 with the types of genetic markers used: GST was estimated from 0.079 to 0.082 for 8-10 populations with 10-13 isozyme loci (Yeh and El-Kassaby 1980; Chaisurisri and El-Kassaby 1994) but only 0.03 for 8 populations with 8 Sequence-Tagged-Site (STS) loci (Gapare et al. 2005). Generally, conifers have weak population differentiation (Hamrick et al., 1992). Interspecific hybridization is common in spruce species where the ranges overlap. White spruce (P. glauca) and Engelmann spruce (P. engelmannii) hybridize from central B C to eastern Washington and Yellowstone National Park (Daubenmire 1974). P. sitchensis introgresses with white spruce in northwestern B C and southwestern Alaska where the two species ranges overlap. Hybrids usually exhibit mid-parent performance and are often less adaptive in their parental environments. However, intermediate environments for the parent species may be more suitable for the hybrids. Introgressed genotypes between Sitka and white spruce in B C showed better performance than their parents in the hybrid zone between maritime and continental climates where the margins of parental species overlap. This indicates that bounded hybrid superiority maintains introgressed populations in a habitat intermediate to that of parental species (Bennuah et al. 2004). Genetic markers for Sitka spruce Selectively neutral genetic markers are used to determine genetic diversity, genetic structure and the effects of genetic drift and gene flow. D N A markers represent the genetic diversity in nucleotide sequences among and within individuals in a population. Many genetic markers are now available for tree species. Nuclear co-dominant makers are preferable because they generally behave in a Mendelian manner, allowing for unambiguous genotyping, and are suitable for the accurate determination of paternal contribution and estimation of heterozygosity. Currently available muclear markers for Picea sitchensis include allozymes (Chaisurisri et al. 1994), and PCR-based STS markers (Perry and Bousquet 2001; Bennuah et al. 2004). Nuclear microsatellite (SSR) makers have also been developed in Picea sitchensis (Vandeven and McNicol 1995), P. glauca (Hodgetts et al. 2001; Rajora et al. 2001), and Picea ssp. (Rungis et al. 2004). Microsatellites are species-specific codominant markers. They are often used in population genetics because of their hypervariability. High polymorphic loci are also favored to assess gene flow because it can identify individuals. The problem with microsatellites is that there is a high probability of alleles being identical in state but not by descent due to their high mutation rates. Microsatellites sometimes have null alleles, perhaps due to mutations on flanking regions or failure to amplify repeats that are too long. Rajora et al. (2001) found an average of 14.5 alleles per microsatellite locus in white spruce. Co-dominant markers are more informative than dominant markers for studying population genetic structure, population history and gene flow. In Sitka spruce, relatively little genetic 11 differentiation among populations was detected by either A F L P markers for cpDNA (Soltis et al., 1997) or STS markers (Gapare et al., 2005). cpDNA is, at least predominantly, inherited paternally in most conifers (e.g. Dong and Wagner 1994; Neale and Sederoff 1989). Since pollen flow can be dispersed substantially long distances by wind, cpDNA may not detect population differentiation, especially with their low mutation rate. The STS marker studies of Gapare et al. (2005) were developed based on EST (expression sequence tagged) sites, for the southernmost and northernmost populations of Sitka spruce. These eight STS markers showed low polymorphism (3.3 allele per locus on average), and did not detect strong population differentiation (F ST = 0.03). Eighteen isozyme polymorphic loci estimated F S T as 0.063 (Yeh and El-Kassaby 1980). The weak population structure may result from low levels of neutral variation at STS loci or low mutation rates of the STS markers. Organelle D N A may provide insights into population history and pollen- and seed-dispersal, especially for conifer species whose chloroplast (cp) D N A is paternally inherited and mitochondrial (mt) D N A is maternally inherited. Hypervariable cpSSRs (Vendramin et al. 1996) and mtSSRs (Hodgetts et al. 2001) developed in white spruce and Norway spruce, respectively, were applied to Stika spruce D N A by the author; however, cpSSRs showed very high monomorphism and amplification of mtSSR was unsuccessful (Mimura unpublished data). In this study, I used nuclear microsatellite loci, which generally show higher polymorphism than STS loci, to study population structure and mating system. Statistical approaches: estimation ofpopulation differentiation and gene flow among populations Population differentiation is caused by natural selection and genetic drift. These forces are, however, difficult to separate. The comparison of widely used standardized measures of genetic differentiation for neutral marker genes (such as F S T) and for quantitative traits (QST) can provide some resolution. F S T estimates the degree of differentiation among populations, by quantifying the proportion of total neutral allelic variation that is found among populations. QST is analogues to F S j , but is based on the amounts of additive genetic variance for quantitative traits among and within populations (Spitze, 1993). F S T generally represents neutral variation shaped by genetic drift and gene flow, and Q S j reflects variation in phenotypic traits shaped by selection, genetic drift and gene flow. If FST < QST, the degree of observed differentiation in quantitative traits exceeds that due to genetic drift alone, and natural selection is inferred to favor different genotypes in different populations or environments. If these measures are not different, then the differentiation observed may be generated by genetic drift alone, but it is not possible to distinguish whether the quantitative differentiation is caused by genetic drift, gene flow, or weak selection. If F S T > QST, there may be no selection shaping population structure, or stabilizing selection favors the same genotype in all populations. Merila and 12 Crnokrak (2001) conducted a meta-analysis of F S Tand Q S T including studies in plant, invertebrate and vertebrate species. Their meta-analysis suggests that Q S T usually exceeds F S T , suggesting the effects of selection in the differentiation of natural populations. The difference between F S T and Q S T (FST < QST) is larger for allozymes than microsatellite based estimates of F S T (Merila and Crnokrak 2001). Choice of genetic markers may impact estimates of the strength of selection by comparing F S T and QST, and microsatellite maker may underestimate degrees of selection. Although Q S T usually exceeds F S T , Merila and Crnokrak (2001) found a high correlation between Q S T and F S T regardless of the types of markers, for example allozymes (r = 0.81), microsatellite (r = 0.87) and combined (r = 0.75). The trend may be due to isolation by distance with the assumption that environments changes with geographic distance or genetic markers linking with selective loci. When gene flow is geographically restricted among populations, Q S T can also increase by genetic drift without selection. Further investigation is needed to explain this pattern (Merila and Crnokrak 2001). Migration rate affects the genetic differentiation of populations and variance of population size by homogenizing or facilitating the "genetic differentiation of source and focal populations. Migration rate can be estimated by several approaches, such as the widely accepted Wright's F-statistics, which estimate number of successful migrants per generation, m, from the equation, F S T = l/(4Nm + 1). However, F S T cannot accurately estimated migration rate due to several assumptions which are often violated in natural populations, where effects of geography, features of mating and dispersal, and effects of historical migration on F S T cannot be ignored (Whitlock and McCauley 1999). Better statistical approaches for estimating gene flow are being developed, for instances, two-locus descent measures (Vitalis and Couvet 2001), coalescent methods (Nielsen and Wakeley 2001), isolation by distance methods (Rannala and Mountain 1997), and migration matrix likelihood models (Beerli and Felsenstein 2001). These methods often assume that populations of interest are at equilibrium in a system of genetic drift and migration and require parameters that can hardly be estimate in natural populations. We cannot ignore the effects of migration in natural populations that are not likely at equilibrium and experience fluctuating interactions with other organisms and environments. Wang and Whitlock (2003) developed a new method to estimate effective population size and migration rate. This method is based on moment and likelihood methods that separate directional effects of migration from random effect of drift that change allele frequencies using genotypes from more than two generations in the same populations. It assumes that a population provides immigrants into a focal population in which we can estimate effective population size and immigration rate. This method is free from the assumption that populations of interest are in equilibrium, thus it can estimate migration rates more accurately than other methods. This model assumes all source populations that 13 providing gene flow into a focal population are identified. There are other methods that estimate effective pollen donor, which provide a rough but big picture estimate of the impacts of population size and gene flow. TwoGener analysis estimates pollen flow among families (Smouse et al. 2001). Pollen grains have high mobility and are a significant agent of gene flow, especially for wind-dispersal species. This method is a hybrid of paternity analysis and genetic structure analysis from maternal and progeny genotypes. It estimates the strength of pollen flow among mothers within a population, <PST from A N O V A based on pollen gamete haplotypes contributed to families, including number of effective pollen donor per family for the population (1/20ST = Nep, number of fathers contributing a mother) based on analysis of molecular variation (AMOVA), average distance traveled by a successful male gamete, and effective neighborhood Opollination) area. O s t is also an analog to F S T , but O S X estimates among-family variation in pollen genotypes within a population. Smouse and Sork (2004) recently reviewed and compared TwoGener and parentage analysis in tree species. Parentage analysis can give details of dispersal distance by identify exact fathers' locations. However, gene flow from outside neighbourhoods or populations is usually considerable in tree species (reviewed in Smouse and Sork 2004). They concluded the usefulness of estimation of number of effective pollen donor from pollen gamete variance among families relative to total variance. Empirical applications of this method suggest that the relative contributions of different fathers to a focal mother decreases with inter-tree distance (Sork et al. 2002). In addition, the number of fathers can be affected by isolation and population size; the Nep of 8 for Quercus alba (Smouse et al. 2001) and 10 for Pinus echinata (Dyer and Sork 2001) are higher than 4.57 for Quercus lobata (Sork et al. 2002) whose populations have become highly fragmented during the past 50 years due to logging. A problem with this estimate is that using a small number of marker loci will overestimate Nep in a relatively inbred populations (e.g. populations that consist of a limited number of genotypes such as founder populations) regardless of lower effective population size. Current TwoGener analysis does not deal with selling (Smouse and Sork 2004). Ritland proposed the estimation of outcrossing rate under a mixed mating model (Ritland, 1989; Ritland, 1990). This model assumes no selection and no assortative mating effects on pollen dispersal and fertilization. To meet these assumptions, the method can be used only on wind-pollinated, self-compatible species. The program M L T R calculates multi-locus outcrossing rate using maximum likelihood by fitting an observed proportion of genotypes of progenies from families to the expected proportion under a mixed mating model (Ritland 2002). Correlation of paternity (rp) for populations is another measure of genetic differences among families, and can be used to estimate the number of effective pollen donor as 1/ rp. It may be reasonable to use Ritland's MLTR for species 14 with mixed mating systems. Thesis introduction Gene dispersal is essential when populations are shifting range, e.g. into recently deglaciated areas. High gene flow can depress local adaptation by homogenizing populations over different environments. Although there is substantial theoretical evidence that gene flow inhibits local adaptation, empirical studies have not investigated the effects of gene flow at species' range peripheries as a factor limiting the range. In populations where the effects of gene flow may be strong compared to those of selection, such as at soft boundaries, we might see tradeoffs between gene flow and selection: gene flow swamping the increases of beneficial allele frequency due to selection. The dynamics of peripheral populations at species boundaries may provide insight into the nature of adaptation during range shifts and range expansion. In the following chapters, I investigate some of the evolutionary dynamics of species' range, especially the effects of gene flow on local adaptation in isolated and continuous populations, using Sitka spruce as a model species, by 1) measuring degrees of local adaptation at range peripheries and 2) characterizing the mating system and population genetic structure of peripheral compared to central populations. In CHAPTER 2,1 focus on effective population size and genetic diversity, which can affect asymmetric gene flow among populations. In CHAPTER 3,1 infer dynamics of postglacial migration and adaptation from pollen genetic structure and adaptive divergence of the species. In CHAPTER 4,1 simulated central and peripheral environments to investigate the relative fitness of isolated and continuous peripheral populations to local environments. My major findings are summarized and future research prospects are discussed in CHAPTER 5. 15 Figure 1.1. Sitka spruce natural range map with putative historical migration routes. Estimates of time of first post-glacial arrival are based on fossil pollen, fossil needles, and tree age observation. (J. Alden. US Forest Service, pers. comm.; Daubenmire 1967; Griggs, 1937; Mann and Hamilton, 1995; Pellatt and Mathewes, 1994) AK 60C 50' 40° 30C * Kodiak Island 300-600 BC (tree ages) * Prince William Sound 3,000 rya (fossil pollen) •Alaska Punhandle * * • 1 7,000 rya (fossil pollen) •Queen Charlotte Islands 9,600 rya (fossil needles) NWT BC * Northern Vancouver Island m m - m 10,000 rya (fossil pollen) * Puget Sound * mm • WA Glacier Limit at GLM OR •22 km south Of San Francisco The least southern limit in early Pleistocene* _ CA • » * NV 130': 120- 110C 16 Literature cited Adams, J.M. and H. Faure (1997). Preliminary vegetation maps of the world since the Last Glacial Maximum: an aid to archaeological understanding. Journal of Archaeological Science 24: 623-647. Alden, J.N. (1998). Growth and adaptability of Afognak and Juneau Sitka spruce provenances on Afognak Island, Alaska. In R. Wheeler. Alaska Extension Review: 1997 Trial season. Alaska Cooperative Extension, University of Alaska Fairbanks, Fairbanks. Allen, G.B., K.J. Brown and R J . Hebda (1999). Surface pollen spectra from southern Vancouver Island, British Columbia, Canada. Canadian Journal of Botany 11: 786-799. Anderson, D.E. (1997). Younger Dryas research and its implications for understanding abrupt climatic change. Progress in Physical Geography 21: 230-249. Avise, J. C. (2000). Phylogeography: the history and formation of species. Harvard Univ. Press, Cambridge. Barrie, J.V. and K.W. Conway (2002). Rapid sea-level change and coastal evolution on the Pacific margin of Canada. Sedimentary Geology 150: 171-183. Barton, N.H. (2001). Adaptation at the edge of a species" range In J. Silvertown and J. Antonovics. Integrating Ecology and Evolution in a Spatial Context. Blackwell Science. Beerli, P. and J. Felsenstein (2001). Maximum likelihood estimation of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach. Proceedings of the National Academy of Sciences of the United States of America 98: 4563-4568. Bennuah, S.Y., T.L. Wang and S.N. Aitken (2004). Genetic analysis of the Picea sitchensis x glauca introgression zone in British Columbia. Forest Ecology and Management 197: 65-77. Brochmann, C , T. Nilsson and T . M . Gabrielsen (1996). A classic example of postglacial allopolyploid speciation re-examined using RAPD markers and nucleotide sequences: Saxifraga osloensis (Saxifragaceae). Symbolae Botanicae Uppsalensis 31: 75-89. Busch, J.W. (2005). The evolution of self-compatibility in geographically peripheral populations of Leavenworthia alabamica (Brassicaceae). American Journal of Botany 92: 1503-1512. Cain, M . L . , H . Damman and A. Muir (1998). Seed dispersal and the Holocene migration of woodland herbs. Ecological Monographs 68: 325-347. Case, T.J., R.D. Holt, M.A. McPeek and T.H. Keitt (2005). The community context of species' borders: ecological and evolutionary perspectives. Oikos 108: 28-46. Case, T.J. and M . L . Taper (2000). Interspecific competition, gene flow, environmental gradients, and the coevolution of species borders. American Naturalist 155: 583-605. Chaisurisri, K. and Y.A . El-Kassaby (1994). Genetic diversity in a seed production population vs natural populations of Sitka spruce. Biodiversity and Conservation 3: 512-523. Chaisurisri, K., J.B. Mitton and Y . A . El-Kassaby (1994). Variation in the mating system of Sitka spruce (Picea sitchensis) - evidence for partial assortative mating. American Journal of Botany 81: 1410-1415. Chuine, I. and E.G. Beaubien (2001). Phenology is a major determinant of tree species range. Ecology Letters 4: 500-510. Clark, J.S. (1998). Why trees migrate so fast: Confronting theory with dispersal biology and the paleorecord. American Naturalist 152: 204-224. Cody, M . L . and J.M. Overton (1996). Short-term evolution of reduced dispersal in island plant populations. Journal of Ecology 84: 53-61. Cohan, F .M. and A.A . Hoffmann (1986). Genetic divergence under uniform selection. 2. different responses to selection for knockdown resistance to ethanol among Drosophila melanogaster populations and their replicate lines. Genetics 114: 145-164. Comes, H.P. and J.W. Kadereit (1998). The effect of quaternary climatic changes on plant distribution and evolution. Trends in Plant Science 3: 432-438. 17 Copes, D.L. (1996). Relation of crown and foliage traits to height growth of Sitka spruce. Western Journal of Applied Forestry 11: 77-80. Crawford, D.J., R. Ornduff and M . C . Vasey (1985). Allozyme variation within and between Lasthenia minor and its derivative species, Lasthenia maritima (Asteraceae). American Journal of Botany 12: 1177-1184. Daubenmire, R. (1967). Some geographic variations in Picea sitchensis and their ecologic interpretation. Canadian Journal of Botany 46: 787-798. Dauberimire, R. (1974). Taxonomic and ecologic relationships between Picea glauca and Picea engelmannii. Canadian Journal of Botany 52: 1545-1560. Davis, M.B. (1976). Pleistocene biogeography of template deciduous forests. Geoscience and Man 13: 13-26. Davis, M.B. and R.G. Shaw (2001). Range shifts and adaptive responses to Quaternary climate change. Science 292: 673-679. Dias, P.C. (1996). Source and sinks in population biology. Trends in Ecology & Evolution 11. Dong, J. and D.B. Wagner (1994). Paternally inherited chloroplast polymorphism in Pinus: estimation of diversity and population subdivision, and tests of disequilibrium with a maternally inherited mitochondrial polymorphism. Genetics 136: 1187-1194. Dyer, R J . and V . L . Sork (2001). Pollen pool heterogeneity in shortleaf pine, Pinus echinata Mill. Molecular Ecology 10: 859-866. Etterson, J.R. and R.G. Shaw (2001). Constraint to adaptive evolution in response to global warming. Science 294: 151-154. Falkenhagen E.R. (1977). Genetic variation in 38 Provenances of Sitka spruce. Silvae Genetica 26: 2-3. Gamache, I., J.P. Jaramillo-Correa, S. Payette and J. Bousquet (2003). Diverging patterns of mitochondrial and nuclear D N A diversity in subarctic black spruce: imprint of a founder effect associated with postglacial colonization. Molecular Ecology 12: 891 -901. Gapare W.J. and Aitken S.N. (2005). Strong spatial genetic structure in peripheral but not core populations of Sitka spruce [Picea sitchensis (Bong.) Carr.]. Molecular Ecology 14: 2659-2667. Gapare, W J . , S.N. Aitken and C E . Ritland (2005). Genetic diversity of core and peripheral Sitka spruce (Picea sitchensis (Bong.) Carr) populations: implications for conservation of widespread species. Biological Conservation 123: 113-123. Garant, D., L.E.B. Kruuk, T.A. Wilkin, R.H. McCleery and B.C. Sheldon (2005). Evolution driven by differential dispersal within a wild bird population. Nature 433: 60-65. Garcia-Ramos, G. and M . Kirkpatrick (1997). Genetic models of adaptation and gene flow in peripheral populations. Evolution 51: 21-28. Griggs, R.F. (1937). Timberlines as indicators of climate trends. Science 85: 251-255. Hamann, A., Y . A . El-Kassaby, M.P. Koshy and G. Namkoong (1998). Multivariate analysis of allozymic and quantitative trait variation in Alnus rubra: geographic patterns and evolutionary implications. Canadian Journal of Forest Research 28: 1557-1565. Hamrick, J.L., H .M. Blanton and K J . Hamrick (1989). Genetic structure of geographically marginal populations of ponderosa pine. American Journal of Botany 76: 1559-1568. Hamrick, J.L., M.J.W. Godt and S.L. Sherman-Brotles (1992). Factors influencing levels of genetic diversity in woody plant species. New Forests 6: 95-124. Haiti, D.L. and A . G . Clark (1997) Second edition. Principles of Population Genetics. Sinauer Associates, Sunderland. Hebda, R J . (1996). Climates and landscapes from plant fossils of the Quaternary In. Life in Stone: Fossils of British Columbia. U B C Press, Vancouver. Hewitt, G. (2000). The genetic legacy of the Quaternary ice ages. Nature 405: 907-913. Hewitt, G . M . (1996). Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58: 247-276. 18 Hodgetts, R.B., M . A . Aleksiuk, A . Brown, C. Clarke, E. Macdonald, S. Nadeem and D. Khasa (2001). Development of microsatellite markers for white spruce (Picea glauca) and related species. Theoretical and Applied Genetics 102: 1252-1258. Hoffmann, A .A . and M.W. Blows (1994). Species borders: ecological and evolutionary perspectives. Evolution 9: 223-227. Hoffmann, A .A . and M.J. Hercus (2000). Environmental stress as an evolutionary force. Bioscience 50:217-226. Holt, R.D. (2003). On the evolutionary ecology of species' ranges. Evolutionary Ecology Research 5: 159-178. Holt, R.D., T .H. Keitt, M.A. Lewis, B.A. Maurer and M . L . Taper (2005). Theoretical models of species' borders: single species approaches. Oikos 108: 18-27. Huff, D.R. and L. Wu (1992). Distribution and inheritance of inconstant sex forms in natural populations of dioecious buffalograss (Buchloe dactyloides). American Journal of Botany 79: 207-215. Jonas, C S . and M . A . Geber (1999). Variation among populations of Clarkia unguiculata (Onagraceae) along altitudinal and latitudinal gradients. American Journal of Botany 86: 333-343. Jump, A.S. and F.I. Woodward (2003). Seed production and population density decline approaching the range-edge of Cirsium species. New Phytologist 160: 349-358. Kawecki, T.J. (2000). Adaptation to marginal habitats: contrasting influence of the dispersal rate on the fate of alleles with small and large effects. Proceedings of the Royal Society of London Series B-Biological Sciences 267: 1315-1320. Kawecki, T.J. and R.D. Holt (2002). Evolutionary consequences of asymmetric dispersal rates. American Naturalist 160: 333-347. Klinka, K., V.J. Krajina, A . Ceska and A . M . Scagel (1989). Indicator Plants of Coastal British Columbia. U B C Press, Vancouver. Kirkpatrick, M . and N.H. Barton (1997). Evolution of a species' range. American Naturalist 150: 1-23. Lawton, J.H. (1993). Range, population abundance and conservation. Trends in Ecology & Evolution 8: 409-413. Ledig, F.T. (2000). Founder effects and the genetic structure of Coulter pine. Journal of Heredity 91: 307-315. Lesica, P. and F.W. Allendorf (1995). When are peripheral populations valuable for conservation? Conservation Biology 9: 753-760. Li , B., J. Suzuki and T. Hara (1998). Latitudinal variation in plant size and relative growth rate in Arabidopsis thaliana. Oecologia 115: 293 - 301 Liston, A. , L . H . Rieseberg and M . A . Hanson (1992). Geographic partitioning of chloroplast D N A variation in the genus Datisca (Datiscaceae). Plant Systematics and Evolution 181: 121-132. Little, E.J.J. (1953). A natural hybrid spruce in Alaska. Journal of Forestry 51: 745-747. Macdonald, G . M . (1993). Fossil pollen analysis and the reconstruction of plant invasions. Advances in Ecological Research 24: 67-110. Mann, D.H. and T.D. Hamilton (1995). Late Pleistocene and Holocene paleoenvironmets of the North Pacific coast. Quaternary Science Reviews 14: 449-471. Mathewes, R.W. (1989). The Queen Charlotte Islands refugium: a paleoecological perspective In R.J. Fulton. Quaternary Geology of Canada and Greenland. Geological Survey of Canada, Geology of Canada, Ottawa. Mayr, E . (1963). Animal Species and Evolution. Cambridge, Harvard University Press. McLachlan, J.S., J.S. Clark and P.S. Manos (2005). Molecular indicators of tree migration capacity under rapid climate change. Ecology 86: 2088-2098. Merila, J. and P. Crnokrak (2001). Comparison of genetic differentiation at marker loci and quantitative traits. Journal of Evolutionary Biology 14: 892-903. 19 Millar, C.I. and K . A . Marshall (1991). Allozyme variation of Port Orford cedar (Chamaecyparis lawsoniana) - Implications for genetic conservation. Forest Science 37: 1060-1077. Mitka, J. (1997). Small, isolated plant populations at geographic range borders: some ecological and genetic processes. Wiadomosci Botaniczne 41: 13-24. Mitton, J.B. (1992). The dynamic mating systems of conifers. New Forests 6: 197-216. Morgenstem, E.K. (1996). Geographic Variation in Forest Trees. U B C Press, Vancouver . Mouna, O., L. Paule, A . E . Szmidt and K. Karkkainen (1990). Mating system analysis in a central and northern European population of Picea abies. Send. J. For. Res. 5: 97-102. Muir, G., A.J. Lowe, C C . Eleming and C. Vogl (2004). High nuclear genetic diversity, high levels of outcrossing and low differentiation among remnant populations of Quercus petraea at the margin of its range in Ireland. Annals of Botany 93: 691-697. Nathan, R., G.G. Katul, H.S. Horn, S.M. Thomas, R. Oren, R. Avissar, S.W. Pacala and S.A. Levin (2002). Mechanisms of long-distance dispersal of seeds by wind. Nature 418: 409-413. Neale, D.B. and R.R. Sederoff (1989). Paternal inheritance of chloroplast D N A and maternal inheritance of mitochondrial D N A in loblolly pine. Theoretical and Applied Genetics 11: 212 — 216. Nielsen, R. and J. Wakeley (2001). Distinguishing migration from isolation: A Markov chain Monte Carlo approach. Genetics 158: 885-896. Parks, C.R., J.F. Wendel, M . M . Sewell and Y . L . Qiu (1994). The significance of allozyme variation and introgression in the Liriodendron tulipifera complex (Magnoliaceae). American Journal of Botany 81: 878-889. Pellatt, M . G . and R.W. Mathewes (1994). Paleoecology of postglacial tree line fluctuations on the Queen Charlotte Islands, Canada. Ecoscience 1:71-81. Perron, M . , D.J. Perry, C. Andalo and J. Bousquet (2000). Evidence from sequence-tagged-site markers of a recent progenitor-derivative species pair in conifers. Proceedings of the National Academy of Sciences of the United States of America 91: 11331-11336. Perry, D.J. and J. Bousquet (2001). Genetic diversity and mating system of post-fire and post-harvest black spruce: an investigation using codominant sequence-tagged-site (STS) markers. Canadian Journal of Forest Research 31: 32-40. Peterson, E.B., N . M . Peterson, G.F. Weetman and P.J. Martin (1997). Ecology and management of Sitka spruce, emphasizing its natural range in British Columbia. U B C Press, Vancouver . Petit, R.J., I. Aguinagalde, J.L. de Beaulieu, C. Bittkau, S. Brewer, R. Cheddadi, R. Ennos, S. Fineschi, D. Grivet, M . Lascoux, A. Mohanty, G . M . Muller-Starck, B. Demesure-Musch, A. Palme, J.P. Martin, S. Rendell and G.G. Vendramin (2003). Glacial refugia: Hotspots but not melting pots of genetic diversity. Science 300: 1563-1565. Pielou, E . C (1991). After the Ice Age. Return of life to glaciated North America. University of Chicago Press, Chicago. Postma, E . and A.J. van Noordwijk (2005). Gene flow maintains a large genetic difference in clutch size at a small spatial scale. Nature 433: 65-68. Rajora, O.P., M . H . Rahman, S. Dayanandan and A. Mosseler (2001). Isolation, characterization, inheritance and linkage of microsatellite D N A markers in white spruce {Picea glauca) and their usefulness in other spruce species. Molecular and General Genetics 264: 871-882. Rannala, B. and J.L. Mountain (1997). Detecting immigration by using multilocus genotypes. Proceedings of the National Academy of Sciences of the United States ofAmerica 94: 9197-9201. Rehfeldt G.E., Wykoff W.R., and Ying C C (2001). Physiologic plasticity, evolution, and impacts of a changing climate on Pinus contorta. Climate Change 50:355-376. Ritland K. (1990). A series of FORTRAN computer programs for estimating plant mating systems. Journal of Heredity U: 235-237. Ritland, K. (2002). Extensions of models for the estimation of mating systems using n independent loci. Heredity 88: 221-228. 2 0 Rungis, D., Y. Berube, J. Zhang, S. Ralph, C E . Ritland, B.E. Ellis, C. Douglas, J. Bohlmann and K. Ritland (2004). Robust simple sequence repeat markers for spruce (Picea spp.) from expressed sequence tags. Theoretical and Applied Genetics 109: 1283-1294. Sang, T., D.J. Crawford and T.F. Stuessy (1995). Documentation of reticulate evolution in peonies (Peonia) using internal transcribed spacer sequences of nuclear ribosomal D N A - Implications for biogeography and concerted evolution. Proceedings of the National Academy of Sciences of the United States of America 92: 6813-6817. Schnabel, A. and J.L. Hamrick (1990). Organization of genetic diversity within and among populations of Gleditsia tricanthos (Leguminosae). American Journal of Botany 77: 1060-1069. Sewell, M . M . , C R . Parks and M.W. Chase (1996). Intraspecific chloroplast D N A variation and biogeography of North American Liriodendron L (Magnoliaceae). Evolution 50: 1147-1154. Smouse, P.E., R.J. Dyer, R.D. Westfall and V . L . Sork (2001). Two-generation analysis of pollen flow across a landscape. I. Male gamete heterogeneity among females. Evolution 55: 260-271. Smouse, P.E. and V . L . Sork (2004). Measuring pollen flow in forest trees: an exposition of alternative approaches. Forest Ecology and Management 197: 21-38. Soltis, D.E., M . A . Gitzendanner, D.D. Strenge and P.S. Soltis (1997). Chloroplast D N A intraspecific phylogeography of plants from the Pacific Northwest of North America. Plant Systematics and Evolution 206: 353-373. Sork, V . L . , F.W. Davis, P.E. Smouse, V.J . Apsit, R.J. Dyer, J.F. Fernandez and B. Kuhn (2002). Pollen movement in declining populations of California Valley oak, Quercus lobata: where have all the fathers gone? Molecular Ecology 11: 1657-1668. Spitze, K. (1993).Population Structure in Daphnia obtusa: Quantitative Genetic and Allozymic Variation. Genetics 135: 367-374 Stanton M . L . and C Galen (1997) Life on the edge: adaptation versus environmentally mediated gene flow in the snow buteercuppanunculus adoneus. American Naturalist 150:143-178. Stephenson, N.L. (1990). Climatic control of vegetation distribution - the role of the water balance. American Naturalist 135: 649-670. Thompson, R.S., C. Whitlock, P.J. Bartlein, S.P. Harrison and W.G. Spaulding (1993). Climatic changes in the western United States since 18,000 yr. B.P. In H.E.J. Wright, J.E. Kutzbach, T. Webb Illet al. Global Climates Since the Last Glacial Maximum. University of Minnesota Press, Minneapolis. Vandeven, W.T.G. and R.J. McNicol (1995). The use of RAPD markers for the identification of Sitka spruce (Picea sitchensis) clones. Heredity 75: 126-132. Vendramin, G.G., L . Lelli, P. Rossi and M . Morgante (1996). A set of primers for the amplification of 20 chloroplast microsatellites in Pinaceae. Molecular Ecology 5: 111-114. Vitalis, R. and D. Couvet (2001). Two-locus identity probabilities and identity disequilibrium in a partially selfing subdivided population. Genetical Research 11: 67-81. Wang, J.L. and M . C . Whitlock (2003). Estimating effective population size and migration rates from genetic samples over space and time. Genetics 163: 429-446. Warner, B.G. and R.W. Mathewes (1982). Ice-free conditions on the Queen Charlotte Islands, British Columbia, at the height of late Wisconsin Glaciation. Science 218: 675-677. Watkinson, A.R. and W.J. Sutherland (1995). Sources, sinks and pseudo-sinks. Journal of Animal Ecology 64. Whitlock, C. and P.J. Bartlein (1997). Vegetation and climate change in northwest America during the past 125 kyr. Nature 388: 57-61. Whitlock, M.C. and D.E. McCauley (1999). Indirect measures of gene flow and migration: F S T not equal to l/(4Nm+l). Heredity 82: 117-125. Wiener, P. and M.W. Feldman (1993). The effects of the mating system on the evolution of migration in a spatially heterogeneous population. Evolutionary Ecology 1: 251-269. Wiens, J.J. (2004). Speciation and ecology revisited: Phylogenetic niche conservatism and the origin of species. Evolution 58: 193-197. 21 Witter, M.S. and G.D. Carr (1988). Adaptive radiation and genetic differentiation in the Hawaiian silversword alliance (Compositae, Madiinae). Evolution 42: 1278-1287. Xu, P., C C . Ying, and Y .A . El-Kassaby (2000). Multivariate Analyses of Causal Correlation between Growth and Climate in Sitka Spruce. Silvae Genetica 49:257-263 Yeh, F . C and Y.A . El-Kassaby (1980). Enzyme variation in natural populations of Sitka spruce (Picea sitchensis) .1. Genetic variation patterns among trees from 10 IUFRO provenances. Canadian Journal of Forest Research 10: 415-422. Yeh, F.C. and C. Layton (1979). Organization of genetic variability in central and marginal populations of Lodgepole pine Pinus contorta spp latifolia. Canadian Journal of Genetics and Cytology 21: 487-503. Ying, C C (1997). Effects of site, provenance, and provenance and site interaction in Sitka spruce in Coastal British Columbia. Forest Genetics 4: 99-112. 22 CHAPTER 2 Increased selling and decreased effective pollen donor size in peripheral relative to central populations in Sitka spruce By Makiko Mimura and Sally N. Aitken Introduction ' Species distributions sometimes end in soft boundaries, at which species density gradually decreases at the edges of the range. It is generally accepted that a species at equilibrium is most abundant in core environments in the central regions of the species' range. Brown (1984) suggested that species are generally most abundant at centre of their range, and that local abundance is a reflection of site suitability for a species. When environmental suitability decreases at the range margins (Lawton 1993), peripheral populations should have smaller population sizes than central populations. Small population size may result in increased inbreeding and genetic drift. Inbreeding may further reduce effective population size by revealing recessive deleterious alleles to selection in a homozygous state (Charlesworfh and Charlesworth 1987). Genetic drift may lead to an erosion of genetic variation within populations, and divergence among populations (Falconer and Mackay 1996). As a consequence, the likelihood of local extinction is thought to be higher for small and isolated populations (Hoffmann and Blows 1994, Lesica and Allendorf 1995), and peripheral populations are likely to have lower genetic diversity than central populations. Another important outcome of smaller population size in peripheral populations than in central populations is asymmetric gene flow due to source/sink population structure. Large central populations can become source populations for smaller peripheral populations. Gene flow homogenizes genetic variation among populations, and source populations eventually have a large genetic influence on sink populations (Kawecki and Holt 2002). Source populations provide more gene copies into sink populations than the reverse, and as a result, sink populations are less locally adapted, and can even limit species' range (Kirkpatrick and Barton 1997). "Central and peripheral hypothesis" predicts these genetic and structural differences between central and peripheral populations. However, results from empirical tests of the central and peripheral hypothesis still remain ambiguous although the decline of plant density and genetic variation toward species' peripheries is theoretically reasonable. While some authors have concluded that genetic diversity 23 declines toward range peripheries (Guries and Ledig 1982, Rajora et al. 2002), there is approximately equal empirical evidence against this hypothesis from studies where no differences in genetic diversity have been found among central and peripheral populations (Mouna et al. 1990, Gamache et al. 2003, Muir et al. 2004). A main reason for this is perhaps because this hypothesis assumes the 'abundant centre' distribution where population size is greatest in central populations (note that central populations in this case are at the centre of the species' niche). However, population size does not always decrease with distance from the centre of the range, especially when a species is distributed over heterogeneous environments with multiple optimum niches, and when niches end due to hard selection such as at water-land boundaries. A test of this hypothesis based on 22 empirical studies ended with uncertainty about the 'abundant centre' distribution (Sagarin and Gaines 2002). Schwartz et al. (2003) claimed that a critical problem in the central/peripheral argument is a lack of operational definitions of peripheral populations for a species. In some species, high gene flow rate and reduced genetic diversity will only be observed at the very edges of the range. Genetic diversity can also be influenced by environmental instability. For instance, the higher genetic diversity in small island populations of Norway spruce (Picea abies) compared to larger island populations was attributed to a lower frequency of fire disturbance on the smaller islands (Wang et al. 2003). Sitka spruce (Picea sitchensis) is a wind-pollinated conifer endemic to the Pacific coast of North America, from northern California to southwest Alaska. The distribution is restricted to a narrow strip along the Pacific coast stretching over 22 degrees of latitude, with soft boundaries at the northern and southern range limits (Figure 2.1). Core populations seem to be widely distributed between southwest Washington and southeast Alaska, and the width of the range from the coast inland tapers towards the northern and southern margins. Contemporary forest tree species have experienced repeated and severe climatic change (Hewitt 1996), and Sitka spruce is thought to have shifted its range northward and southward over the last few glacial cycles. The contemporary populations of the species likely originated from southern glacial refugia. These historical events are also likely to have affected genetic diversity. We examined whether population position within the range plays a role in current levels of genetic diversity, mating system and population size in Sitka spruce. According to concepts based on central and peripheral population structure, we would expect that peripheral populations should have (1) smaller population sizes, (2) higher inbreeding rates, (3) lower genetic diversity, and (4) higher probabilities of being sink populations. Central populations were defined at the centre of the distribution, and peripheral populations at the southern and northern peripheries of the range. The central and peripheral hypothesis was tested by comparing genetic parameters between central and peripheral populations as well as between continuous and disjunct populations. Our objectives are to evaluate the influence of population size and population isolation on genetic diversity and mating system, and to understand how population history and position within the species range shapes population genetic variation. 24 Materials and Methods Six populations were selected for inclusion based on population position within the species' range (Figure 2.1). One continuous/disjunct pair of populations was sampled from each of three different regions: northern peripheral, central, and southern peripheral (Table 2.1). The northern disjunct population was from Kodiak Island, Alaska (AK), isolated over 100km from the nearest continuous populations by ocean, and a southern disjunct population from Fort Bragg, California (CA), currently separated from continuous populations by over 100km, apparently due to local climate. The central disjunct population was from the Queen Charlotte Islands, British Columbia (BC), one of the most productive regions for Sitka spruce, approximately 60km from central mainland populations. Seeds were obtained from maternal trees located at low elevation (<70m) on the coast to eliminate additional environmental effects such as elevation on mating success and to avoid contamination though introgression with white spruce (P. glauca, Bennuah et al. 2004). Seed samples were provided by the US Forest Service, Alaska (samples from Alaska) and the British Columbia Ministry of Forestry (samples from Canada). There were no adequate family-sampled seed collections available from the southern peripheral populations; therefore we collected seeds in the early fall of 2003 near Redwood and Fort Bragg in California. The open-pollinated progeny of 10 to 20 maternal trees (average 16.3 trees) from each of six populations were analyzed for mating system and genetic diversity. Genotyping In conifers, the megagametophyte that surrounds and provides nutrition to the embryo has the same haplotype as the egg resulting in the embryo. Genotyping both embryo and the corresponding megagametophyte with codominant genetic markers for each seed allows the unambiguous determination of the pollen gamete haplotype, the most significant vector of gene flow. In total, 186 to 216 pollen gamete haplotypes were determined for each population. After soaking in water for 48 hours, seeds were stratified at 4°C for three weeks on filter paper in Petri dishes, and then germinated at room temperature in the laboratory in 2003. One to two-week-old seedlings and megagametophytes were stored at -80°C until DNA extraction. DNA extraction followed protocols developed by Hodgetts et al. (2001) for both megagametophytes and embryos. Seven polymorphic nuclear microsatellite markers were selected for genotyping: UAPgAG150 and UAPgAG105 developed for P. glauca (Hodgetts et al. 2001), SPAGC1 and SPAGG3 developed for P. abies (Pfeiffer et al. 1997), EAC7H07 developed for P. abies (Scotti et al. 2002), and WS0073.H08 and WS0061 .K02 developed for P. glauca and P. sitchensis (Rungis et al. 2004). Polymerase chain reactions (PCR) were conditioned at 95°C for 5min. of initial denaturing, followed by 30 cycles of 94°C for 45sec, 53-57°C for 45sec. and 72°C for 45sec, and 72°C for lOmin for final extension. PCR cocktails followed the LiCor 4200 manual with slight adjustments for individual primer sensitivities. PCR products were visualized on a 2 5 LiCor 4200 automated sequencer. Images were scored using Saga Generation 2 (LiCor Inc.) with manual adjustments. One marker that had very low polymorphism (UAPgAG105) and one that had a high frequency of null alleles in the central disjunct and both northern populations (EAC7H07) were not included in further analyses. Data analysis Pollen gamete haplotypes for seedlings were inferred from the genotypes of embryos and megagametophytes, and used for the estimation of a basic diversity index for each population. Nei's (1987) gene diversity from gamete allele frequencies (pollen gene diversity, H q e ; 1 w h e r e pt is a frequency of allele i) was estimated for each population's pollen gamete pool. Pollen haplotype diversity (Hhe) was calculated from haplotype frequencies (1-JA, where ht is a frequency of haplotype i based on the five microsatellite loci), and standard errors estimated from 1000 permutations using Arlequin 2.001. The number of alleles was averaged over loci and weighted by mean sample size over sampled populations. Mating system parameters were estimated from offspring genotypes with maternal gamete information using the MLTR program (Ritland 2002). This program estimates outcrossing rate under a mixed mating model using the maximum likelihood method (Ritland 1990, Ritland 2002). Multi4ocus outcrossing rate (tm), single-locus outcrossing rate (ts), multi-locus and single-locus correlation of paternity among families (r p ( m ) and r p ( s ) , respectively), and correlation of selfing (rs) among families and loci were estimated based on the allele frequencies of all populations using maximum likelihood with the numeric Newton-Raphson method (Ritland 2002). Standard deviations were estimated from 1000 bootstrap resampling of individuals within families. Selfing rate (S), biparental inbreeding rate and number of effective pollen donors were estimated as 1 - tm, tm - ts, and l/r p ( m), respectively. The number of effective pollen donors (Nep) is an indirect measurement of gene flow among families. The significance of differences between populations was tested using a pairwise t-test. Results A total of 61 different alleles was detected in pollen gametes from all populations using the five pairs of nuclear microsatellite primers, while 43 alleles were detected in maternal genotypes. Alleles frequencies for each locus and allelic richness parameters are shown in Table 2.2. Pollen gamete gene diversity (H g e) ranged from 0.53 to 0.70 among populations (Table 2.2). In the continuous populations, the northern population exhibited lower Nei's gene diversity than the central and southern populations. In the disjunct populations, in which the effects of genetic drift are likely more pronounced than in continuous populations, Nei's gene diversity was higher in the central population than in either the northern or southern peripheries. Pollen haplotype diversity (Hh e) showed a similar pattern to the pollen gamete gene 26 diversity (Figure 2.2). The continuous southern population had higher haplotype diversity than other continuous populations. On average, continuous populations had higher Nei's gene diversity than the corresponding disjunct populations, but this trend was more pronounced at the range peripheries than in central populations. On the other hand, the observed number of alleles declined both in continuous and disjunct populations at the range peripheries (Table 2.2) and there were fewer alleles in disjunct than continuous populations. Private alleles are those that are observed only in one of the populations (Table 2.2). Very few private alleles were found in the two peripheral disjunct populations. The central disjunct population had the highest number of private alleles among the tested populations. The number of alleles shared with central populations was higher in continuous peripheral than disjunct peripheral populations, but the difference was slight, especially between the two northern peripheral populations. Mating system parameters varied with population position over the range. All estimates and their standard deviations estimated from bootstrapping are summarized in Table 2.3. Estimates of multi-locus selfing rate were lower at the range centre and higher at the northern and southern peripheries (Figure 2.3a). The highest selfing rate was found in the northern disjunct peripheral population (Kodiak Island, S = 0.352±0.030), while the central continuous population had the lowest selfing rate (Ocean Falls, S = 0.024±0.017). The selfing rates in the disjunct populations (mean S = 0.21) were higher than in nearby continuous populations (mean S = 0.07) in all regions. There was a strong correspondence between population isolation and selfing, and selfing rates of the continuous populations were significantly lower than those of the nearby disjunct populations in all three regions (Central: difference =0.097 s.d.=0.027; South: difference =0.072 s.d =0.032; North: difference =0.256 s.d.=0.037; all standard deviations estimated from bootstrapping). The biparental inbreeding rate (tm -ts), the estimated frequency of mating among relatives that was included in the multi-locus outcrossing rate estimate, was highest on average in northern populations (Table 2.3; Figure 2.3b). There was a significant increase in biparental inbreeding from south to north in the continuous populations, while it was lowest in the centre and higher in both the north and south in the disjunct populations. Nineteen to 24% of matings were among relatives in the northern populations. However, it should be noted that biparental inbreeding rate (tm -ts) is a function of the amount of polymorphism and the estimate is sensitive to the number of loci and types of markers (Ritland 2002). For example, loci with lower levels of polymorphism estimate higher biparental inbreeding. The correlation of paternity (rp(m)) is the probability that two randomly chosen progenies from the same mother tree are full-sibs. There is a significant pattern of strong correlation of paternity in peripheral populations (rp(m) from 0.163 to 0.549) but not in central populations (rp(m) = 0.048 and 0.059) (Table 2.3, Figure 2.3c). All pair-wise differences were significant except between the central, continuous and the central, disjunct populations. The correlation of paternity was lower in continuous (mean rp ( m) = 0.127) than in discontinuous populations (mean r p ( m ) = 0.374). Thus, the number of effective pollen donors per mother tree within populations (1/ rp(m)) showed a dramatic decline in peripheral populations, 27 especially for disjunct populations at both the northern and southern peripheries (Table 2.4). The central populations had a high number of effective pollen donors (Nep= 14.1-40.0 in the continuous, 11.4-33.3 in the disjunct), yet a very low N e p was estimated for the peripheral populations (4.8-8.2 for continuous peripheral, 1.7-2.2 for disjunct peripheral). There was no strong paternal substructure within populations except in the southern peripheral population, indicated by the difference between multi-locus and single-locus correlation of paternity (rp(s) -rP(m)> Table 2.3). The central and southern disjunct populations had a slightly lower correlation of selfing among families compared to their nearest continuous populations, yet the correlation of selfing was not statistically different between any populations tested. The correlation of selfing among loci indicates 'apparent' selfing due to biparental inbreeding. The central disjunct population had a moderately high selfing rate (S=0.121±0.022) and approximately half of this was caused by pure selfing rather than biparental inbreeding (rs among loci = 0.501±0.105, Table 2.3). The peripheral continuous populations had similar selfing rates (approximately 10%), but pure selling rate was higher in the south than in the north although the difference was not significant (difference in r s among loci = 0.253±0.135). The peripheral disjunct populations had relatively high selfing rates, yet only small portions of these were caused by uniparental selfing (Table 2.3). Discussion Central and peripheral hypothesis in genetic diversity Genetic and structural differences between central and peripheral populations (central and peripheral hypothesis) lead to an expectation of reduced genetic diversity toward range peripheries. This is due to a combination of lower population size, higher inbreeding rate, higher instability of environment and lower environmental suitability at range periphrasis than in central populations. In this study, the pollen gene diversity (Nei's gene diversity), as measured by Hge and H h e , roughly decreased toward the north in continuous populations, and did not support the central and peripheral hypothesis. Genetic diversity is also influenced by population history, including migration, founder and bottleneck events. Genetic diversity is expected to be reduced away from refugia (Hewitt 1996), but it may also be higher in populations where descendants from two different migration routes are mixed (Petit et al. 2003). The directional reduction of average gene diversity from south to north in continuous populations may reflect the last Quaternary migration path and species' range shift. At the Last Glacial Maximum, the ice sheet had reached south to Puget Sound, WA (latitude 43°). It is reasonable to assume that southern populations are older than central and northern populations. Directional migration following the retreat of the ice sheet may have reduced gene diversity toward the migration tips. High genetic diversity and low biparental inbreeding in the southern continuous population suggest that this population is composed of many unrelated individuals, which perhaps were descendants from a large refugial population during the 28 Quaternary glaciations. This decrease in genetic diversity toward the north is also found in Pinus rigida (Guries and Ledig 1982). Reduced Nei's gene diversity in the disjunct population compared to a continuous population at the southern periphery (Table 2.2, Figure 2.2) supports a recent population reduction. Allelic diversity is more sensitive to a demographic bottlenecks than Nei's gene diversity (Nei et al. 1974, Spencer et al. 2000, Widmer and Lexer 2001) and here showed the predicted pattern for the central and disjunct hypothesis. The substantial reduction in allelic diversity toward the range peripheries was stronger in disjunct than in continuous populations, and corresponds to the current species' distribution and effective population size. An allele with lower frequency has a higher probability of being lost due to genetic drift, but the loss of low frequency alleles has little effect on other genetic diversity estimates such as expected heterozygosity and gene diversity. Nei's gene diversity in the southern populations may ultimately be reduced due to a lower number of effective pollen donors, and genetic diversity may then correspond to prediction of central and peripheral hypothesis in the future. There is no strong trend in number of private allele between continuous central and peripheral populations. The central disjunct populations had the highest number of private alleles while peripheral disjunct populations had fewer or no private alleles. The numbers of private alleles were higher in the continuous than disjunct populations at range peripheries, and were all rare alleles (p < 0.03). Gapare et al. (2005) detected 'rare but localized' alleles more in disjunct than in continuous populations at the Sitka spruce range peripheries, using STS markers. Their definition of localized alleles was alleles that occurred in less than 50% of the tested populations, and they concluded that disjunct population, especially at the range peripheries, were affected by recent population bottlenecks. On the other hand, our result in the number of private alleles suggested higher genetic diversity in larger populations (e.g. continuous populations) resulting in a higher probability that two randomly chosen alleles from a population are different. The higher number of private alleles in the central disjunct population suggests a large effective population size and high genetic diversity. The maintenance of high genetic diversity in the isolated condition supports the idea of a putative refugium on Queen Charlotte Islands, where lowered sea level create ice-free land approximately between 12,500-14,600 years ago (Barrie and Conway 2002) that were dominated by terrestrial and aquatic plants about 16,000 years ago (Warner and Mathewes 1987). Fossil records suggest that Sitka spruce arrived about 11,200 years ago on the island (Mann and Hamilton 1995, Warner and Mathewes 1987). The high genetic diversity on the Queen Charlotte Islands (the central disjunct population) may be a result of historical population establishment during glaciation or of the large population size with high density. Further investigation with increased numbers of populations sampled will be needed to answer this question. 29 Central and peripheral hypothesis in mating system and pollen donor size The mean selfing rate estimated for Sitka spruce in this study was rather high for a conifer (mean S = 0.14, range from 0.024 to 0.352) compared to the selfing rate estimated in a seed orchard for this species (S = 0.08, Chaisurisri et al. 1994). Conifers generally have low (<10%) selfing rates (Mitton 1992). Since half of our sample populations were disjunct, where effective population size is likely smaller than the majority of continuous populations, the mean selfing rate was overestimated in this study. Increased selfing and increased correlation of paternity in peripheral and disjunct populations suggests that these populations have either lower population size or higher within-population structure than central populations. Selfing rate is frequency- and density- dependent, and depends on the frequency of genotypes surrounding individuals (Holsinger 1991). It increases when effective population size is small or plant density is low (Loveless and Hamrick 1984). The generally high correlation of paternity observed in peripheral and disjunct populations suggests a considerably higher probability of full-sib occurrence in an open-pollinated family in these populations compared to central populations. This may also result from lower population size, lower species density or fecundity, or a combination of these factors. Negative correlations between selfing rate and population size have also been found in other tree species (Farris and Mitton 1984, Raijmann et al. 1994, Rajora et al. 2002). However, populations with a relatively wide range of local plant densities have shown no differences in outcrossing rate (Neale and Adams 1985, Morgante et al. 1991). Only large differences in size or density may affect mating system (Robledo-Amuncio et al. 2004). Although N e p is usually much smaller than N e (Smouse and Sork 2004), the observed pattern in effective pollen donor size indicates a smaller effective population size at the range periphery than at the centre of the range for Sitka spruce. Peripheral populations, especially disjunct populations, seem to have relatively small effective population sizes and considerable isolation from the central continuous populations. Southern and northern peripheral populations of Sitka spruce mature stands are more spatially structured compared to central population (Gapare and Aitken 2005). Despite differences in climatic selection pressures and population ages between the northern and southern peripheries, they share similar levels of inbreeding and effective pollen donor size. This supports a strong effect of population position in conjunction with population size on mating system. Decreased distribution area and the reduction of N e p at the edges of the species' range may cause asymmetric gene flow due to density dependence. The number of shared alleles with the central populations were lower in peripheral disjunct than peripheral continuous populations both in the south and north (Table 2.2). These results imply that continuous populations may be receiving more alleles from central populations than disjuct populations at range peripheries. The concept of the central and peripheral hypothesis of genetic diversity has been argued, and there are several reasons for deviations from the hypothetical abundant centre distribution. Hard selection, such as at water/land boundaries, determines species' boundary without a change in population density at 3 0 the edge. When a species has discrete or multiple landscape optima across geographic ranges, plant density may not be a function of population position. For example, increased plant density in Aquilegia canadensis in northern peripheral populations compared to the central populations may be caused by geographic restrictions of habitat in central regions (Herlihy and Eckert 2005). The genetic model for species distribution predicting a decrease in density toward range peripheries assumes a linearly heterogeneous environment for a species (Kirkpatrick and Barton 1997). We found striking reductions in effective pollen donor size and allelic richness, perhaps due to reduction in population sizes near soft boundaries of Sitka spruce. Sitka spruce has reduced range width at its peripheries and^ has an almost linearly heterogeneous environment along the coast at lower elevations (Figure 2.1). These factors make Sitka spruce appear to fit the model well. However, the majority of the current Sitka spruce range resulted from recent postglacial expansion (for the past 15,000 years), and we may not see strong central and peripheral structure when a species has not yet reached equilibrium. This can also be seen in our results as gene diversity reflects geographic patterns due to postglacial recolonization more than central and peripheral population structure (Table 2.2, Figure 2.2). The strong central and peripheral population structure in allelic richness, mating system and pollen donor size may eventually shape gene diversity to better fit the central and peripheral model. Selfing in isolated populations It is somewhat surprising that the disjunct populations at range peripheries exhibited very high selfing rates (35%), in contrast to the typical mating system of wind-pollinated temperate conifers (less than 10%). Although a large proportion of the 'apparent' selfing was likely caused by biparental inbreeding (Table 2.3), these peripheral disjunct populations may experience severe inbreeding depression. However, even the northern isolated population (with 35% estimated selfing) is extending the species limit and colonizing non-forested area on the Kodiak Island (Griggs 1937, J Alden (Univ. of Alaska) personal communication). The northern population also appears to have higher juvenile fitness compared to the nearby continuous populations (CHAPTER 4). In wind-pollinated conifer species, it is generally accepted that inbreeding depression is severe at the individual level for fitness-related traits such as growth and seed production; however, this effect may not be important at the stand level (Williams and Savolainen 1996, Wang et al. 2004). Isolated conifer stands of reproductive age in natural populations sometimes do not show a loss of fitness due to inbreeding. For example, isolated old-growth populations of red spruce (P. rubens) have significant selfing (Rajora et al. 2000) but exhibit growth performance comparable to large, continuous populations (Mosseler et al. 2000). In tree species, inbred deleterious genotypes may be effectively purged through natural selection due to low competitive ability (Plessas and Strauss 1986), during their relatively long juvenile period. This study suggests that population position may be able to predict effective pollen donor size, mating system and allelic richness, but not Nei's gene diversity, in Sitka spruce. Smaller effective 31 p o p u l a t i o n s i z e m a y h a v e a s t r o n g i m p a c t o n m a t i n g s y s t e m . E v i d e n c e o f i n c r e a s e d s e l f i n g at r a n g e p e r i p h e r i e s i m p l i e s that t h i s w i l l e v e n t u a l l y s h a p e g e n e d i v e r s i t y to b e h i g h e r at the r a n g e c e n t r e a n d l o w e r at the p e r i p h e r i e s . T h i s s t u d y s u p p o r t s the c e n t r a l a n d p e r i p h e r a l h y p o t h e s i s at e q u i l i b r i u m f o r sof t s p e c i e s b o u n d a r i e s , a n d s u g g e s t s f a c t o r s l e a d i n g to c e n t r a l a n d p e r i p h e r a l p o p u l a t i o n s t r u c t u r e m a y h a v e s i g n i f i c a n t e v o l u t i o n a r y c o n s e q u e n c e s f o r p e r i p h e r a l p o p u l a t i o n s . 32 Table 2.1. Classification of population sampled. Disjunct populations are geographically separated from continuous populations by at least 60 km. QCI: Queen Charlotte Islands. Population State Sample size Range classification Family Seed Region Position Type Rocky Bay AK, USA 19 228 north peripheral continuous Kodiak Island AK, USA 20 216 north peripheral disjunct Ocean Falls BC, Canada 18 189 central central continuous QCI BC, Canada 10 186 central central disjunct Redwood CA, USA 17 203 south peripheral continuous Fort Bragg CA, USA 14 201 south peripheral disjunct Table 2.2. Allele frequency data for five microsatellite loci studied in this thesis. Pollen gamete haplotypes were pooled within population, and averaged over five microsatellite loci. Private alleles is weighted number of alleles that only occurred in the sample population. North Central South Conti. Disjunct Conti. Disjunct Conti. Disjunct n = 227 n = 208 n= 192 n= 175 n= 192 n= 196 Locus Allele WS0073.H08 1 - - - - 0.021 -2 - - - - 0.005 -3 0.167 0.053 0.031 0.063 0.083 0.174 4 0.018 0.091 0.135 0.171 0.219 0.260 5 0.009 0.298 0.281 0.269 0.167 0.148 6 0.612 0.226 0.307 0.314 0.375 0.311 7 0.097 0.019 0.078 0.086 0.068 0.005 8 0.018 0.188 0.026 0.051 0.021 -9 0.009 - - - - -10 - - 0.010 - - -11 - - - - 0.005 -12 0.004 - - - - -WS0061.K02 1 0.110 0.077 0.370 0.349 0.240 0.179 2 0.802 0.909 0.500 0.526 0.698 0.786 3 0.009 - 0.016 0.023 0.016 -4 - 0.005 0.005 - - -SPAGC1 1 - - 0.052 0.057 0.089 0.041 2 - - 0.005 - - 0.005 3 - - - - - 0.031 4 - - - 0.006 - -5 - - - 0.029 0.094 0.122 6 - - 0.005 - - -7 - - - 0.006 - -8 0.004 - 0.005 0.029 - -3 3 9 - - - 0.006 - -10 0.031 0.005 - 0.006 0.010 -11 0.009 0.039 0.016 0.017 0.021 0.010 12 0.458 0.236 0.193 0.171 0.417 0.454 13 0.189 0.630 0.328 0.366 0.135 0.102 14 0.018 - 0.073 0.063 0.010 -15 0.004 0.019 0.026 0.063 0.083 0.138 16 - 0.010 0.026 0.034 0.068 0.010 17 - - 0.099 0.063 0.021 0.026 18 0.044 - 0.021 0.023 0.005 -19 0.013 - - - - -20 - - 0.005 - - -21 0.167 0.029 0.026 - - -22 - - - - 0.010 -SPAGG3 1 0.339 0.067 0.260 0.269 0.130 0.010 2 - - 0.010 0.040 0.125 0.112 3 0.004 - 0.005 - - -4 0.075 0.014 - 0.040 0.115 0.122 5 0.075 0.034 0.089 0.109 0.073 0.015 6 0.040 0.221 0.104 0.063 0.068 0.041 7 0.062 0.039 0.109 0.149 0.099 0.036 8 0.018 - 0.073 0.063 0.037 0.393 9 0.110 0.010 0.052 0.017 0.078 -10 0.004 0.005 0.031 0.006 0.037 -11 0.013 0.524 0.021 0.046 0.068 -12 0.137 0.029 0.104 0.074 0.047 -13 - 0.005 0.010 0.017 0.016 0.071 14 0.004 0.005 0.010 0.034 0.005 0.122 15 - - 0.005 0.006 0.005 -16 0.009 - 0.021 - 0.037 -17 0.009 - 0.010 0.006 - -18 - - - 0.006 - -UAPgAG150B 1 0.013 - - - - -2 0.269 0.197 0.464 0.480 0.344 0.143 3 0.256 0.159 0.302 0.331 0.438 0.719 4 0.326 0.524 0.146 0.109 0.182 0.092 5 - - - 0.011 - -Total number of alleles*1 61 34.0 28.6 45.4 49.8 43.3 29.3 Average number of alleles per locus*1 6.8 5.7 9.1 10.0 8.7 5.9 Private alleles*1 3.5 0.0 3.1 4.5 4.1 1.0 Shared allele with centrals*1 30.5 28.6 - - 39.2 28.3 Gene diversity (H K e )* 2 0.6 0.5 0.7 0.7 0.7 0.6 A l l observed number of alleles were weighted by average sample size (n=198). Gene diversity was estimated as l-£p,-2 where pt is a frequency of allele , at a locus, and then average over loci. 34 Table 2.3. Summary of mating system of Sitka spruce. tm: multi-locus outcrossing rate; ts: single-locus outcrossing rate; tm-ts: biparental inbreeding; r p(m ): multi-locus correlation of paternity; and rs: correlation of selfing. Standard deviation in parentheses was estimated from bootstrapping 1000 times, resampling individuals within families. Northern peripheral Central Southern peripheral Continuous Rocky Bay Disjunct Kodiak Island Continuous Ocean Falls Disjunct QCI Continuous Redwood Disjunct Fort Bragg tm t s t ,„-ts rp(m) r P( S) rP(s) - rP(m) rs among family r„ among loci 0.904 (0.022) 0.711 (0.026) 0.193 (0.021) 0.169 (0.040) 0.196 (0.047) 0.028(0.031) 0.229 (0.115) 0.105 (0.073) 0.648 (0.030) 0.409 (0.024) 0.239 (0.019) 0.514(0.064) 0.445 (0.099) -0.069 (0.078) 0.300 (0.064) 0.165 (0.052) 0.976 (0.017) 0.828 (0.023) 0.149 (0.024) 0.048 (0.023) 0.059 (0.034) 0.011 (0.029) 0.368 (0.299) 0.037 (0.062) 0.879 (0.022) 0.772 (0.028) 0.107 (0.019) 0.059 (0.029) 0.065 (0.031) 0.006 (0.018) 0.128 (0.076) 0.501 (0.105) 0.915 (0.020) 0.844 (0.025) 0.071 (0.019) 0.163 (0.041) 0.168 (0.040) 0.005 (0.019) 0.322 (0.123) 0.357 (0.115) 0.848 (0.026) 0.661 (0.026) 0.186 (0.020) 0.549 (0.055) 0.664 (0.080) 0.114(0.046) 0.163 (0.059) 0.086(0074) Table 2.4. Estimated number of effective pollen donors per family (Nep). The range was calculated from the standard deviations for correlation of paternity. Standard deviations were estimated from bootstrapping 1000 times, resampling individuals within families. N e D Location Continuous Disjunct Northern peripheral 4.8-7.6 1.7-2.2 Central 14.1-40.0 11.4-33.3 South peripheral 4.9-8.2 1.7-2.0 35 Figure 2.1. Sitka spruce range map showing populations sampled. Kodiak Island, QCI and Fort Bragg were classified as disjunct populations; and Rocky Bay, Ocean Falls and Redwood as continuous populations. Figure 2.2. Pollen haplotype diversity (Hh e ) , estimated as Nei's gene diversity from pollen haploid frequencies based on 5 microsatellite loci. The error bars are standard deviation from 1000 permutations. 1.00 0 > <D Q. >> O C L 03 JZ a 0) O CL North Central South 37 Figure 2.3. Genetic parameters for mating system: (a) multi-locus selfing rate (l-tm), (b) biparental inbreeding rate (tm-ts), and (c) correlation of paternity (rp). Standard deviation was estimated from bootstrapping 1000 times, resampling individuals within families. (a) 0.5 0.4 a> 0.3 Ul § 02 <D CO 0.1 [~1 Continuous • Disjunct North Central South (b) (c) North Central South North Central South Literature Cited Barrie, J.V. and K.W. Conway (2002). Rapid sea-level change and coastal evolution on the Pacific margin of Canada. Sedimentary Geology 150: 171-183. Bennuah, S.Y., T.L. Wang and S.N. Aitken (2004). Genetic analysis of the Picea sitchensis x glauca introgression zone in British Columbia. Forest Ecology and Management 197: 65-77. Brown J.H. (1984). On the relationship between abundance and distribution of species. American Naturalist 124: 255-279. Chaisurisri, K., J.B. Mitton and Y.A. El-Kassaby (1994). Variation in the mating system of Sitka spruce (Picea sitchensis) - evidence for partial assortative mating. American Journal of Botany 81: 1410-1415. Charlesworth, D. and B. Charlesworth (1987). Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18: 237-268. Falconer, D.S. and T.F.C. Mackay (1996). Introduction to Quantitative Genetics. Pearson Education, Harlow. Farris, M.A. and J.B. Mitton (1984). Population density, outcrossing rate, and heterozygote superiority in ponderosa pine. Evolution 38: 1151-1154. Gamache, I., J.P. Jaramillo-Correa, S. Payette and J. Bousquet (2003). Diverging patterns of mitochondrial and nuclear DNA diversity in subarctic black spruce: imprint of a founder effect associated with postglacial colonization. Molecular Ecology 12: 891-901. Gapare, W.J., S.N. Aitken and C E . Ritland (2005). Genetic diversity of core and peripheral Sitka spruce (Picea sitchensis (Bong.) Carr) populations: implications for conservation of widespread species. Biological Conservation 123: 113-123. Gapare W.J. and Aitken S.N. (2005). Strong spatial genetic structure in peripheral but not core populations of Sitka spruce [Picea sitchensis (Bong.) Carr.]. Molecular Ecology 14: 2659-2667. Griggs, R.F. (1937). Timberlines as indicators of climate trends. Science 85: 251-255. Guries, R.P. and F.T. Ledig (1982). Genetic diversity and population-structure in pitch pine (Pinus rigida Mill). Evolution 36: 387-402. Herlihy, C R . and C . G Eckert (2005). Evolution of self-fertilization at geographical range margins? A comparison of demographic, floral, and mating system variables in central vs. peripheral populations of Aquilegia canadensis (Ranunculaceae). American Journal of Botany 92: 744-751. Hewitt, G M . (1996). Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58: 247-276. Hodgetts, R.B., M.A. Aleksiuk, A. Brown, C. Clarke, E. Macdonald, S. Nadeem and D. Khasa (2001). Development of microsatellite markers for white spruce (Picea glauca) and related species. Theoretical and Applied Genetics 102: 1252-1258. Hoffmann, A.A. and M.W. Blows (1994). Species borders - ecological and evolutionary perspectives. Trends in Ecology & Evolution 9: ITi-lll. Holsinger, K.E. (1991). Mass-action models of plant mating systems - the evolutionary stability of mixed mating systems. American Naturalist 138: 606-622. Kawecki, T.J. and R.D. Holt (2002). Evolutionary consequences of asymmetric dispersal rates. American Naturalist 160: 333-347. Kirkpatrick, M. and N.H. Barton (1997). Evolution of a species' range. American Naturalist 150: 1-23. Lawton, J.H. (1993). Range, population abundance and conservation. Trends in Ecology & Evolution 8: 409-413. Lesica, P. and F.W. Allendorf (1995). When are peripheral populations valuable for conservation? Conservation Biology 9: 753-760. Loveless, M.D. and J.L. Hamrick (1984). Ecological determinants of genetic structure in plant populations. Annual Review of Ecology and Systematics 15: 65-95. Mann, D.H. and T.D. Hamilton (1995). Late Pleistocene and Holocene paleoenvironmets of the North Pacific coast. Quaternary Science Reviews 14: 449-471. Mitton, J.B. (1992). The dynamic mating systems of conifers. New Forests 6: 197-216. Morgante, M. , G.G. Vendramin and P. Rossi (1991). Effects of stand density on outcrossing rate in 2 3 9 Norway spruce (Picea abies) populations. Canadian Journal of Botany 69: 2704-2708. Mosseler, A., J.E. Major, J.D. Simpson, B. Daigle, K. Lange, Y.S. Park, K.H. Johnsen and O.P. Rajora (2000). Indicators of population viability in red spruce, Picea rubens. I. Reproductive traits and fecundity. Canadian Journal of Botany 78: 928-940. Mouna, O., L. Paule, A.E. Szmidt and K. Karkkainen (1990). Mating system analysis in a central and northern European populations of Picea abies. Scandinavian Journal of Forest Research 5: 97-102. Muir, G , A.J. Lowe, C C . Fleming and C. Vogl (2004). High nuclear genetic diversity, high levels of outcrossing and low differentiation among remnant populations of Quercus petraea at the margin of its range in Ireland. Annals of Botany 93: 691-697. Neale, D.B. and W.T. Adams (1985). The mating system in natural and shelterwood stands of Douglas-fir. Theoretical and Applied Genetics 71: 201-207. Nei, M . (1987). Molecular Evolutionary Genetics. New York, Columbia University Press. Nei, M. , T. Maruyama and R. Chakrabory (1974). The bottleneck effect and genetic variability in populations. Evolution 29: 1-10. Petit, R.J., I. Aguinagalde, J.L. de Beaulieu, C. Bittkau, S. Brewer, R. Cheddadi, R. Ennos, S. Fineschi, D. Grivet, M . Lascoux, A. Mohanty, G.M. Muller-Starck, B. Demesure-Musch, A. Palme, J.P. Martin, S. Rendell and GG. Vendramin (2003). Glacial refugia: Hotspots but not melting pots of genetic diversity. Science 300: 1563-1565. Pfeiffer, A., A . M . Olivieri and M. Morgante (1997). Identification and characterization of microsatellites in Norway spruce (Picea abies K). Genome 40: 411-419. Plessas, M.E. and S.H. Strauss (1986). Allozyme differentiation among populations, stands, and cohorts in Monterey pine. Canadian Journal of Forest Research 16: 1155-1164. Raijmann, L.L. , N . C Vanleeuwen, R. Kersten, J.GB. Oostermeijer, H.C.M. Dennijs and S.B.J. Menken (1994). Genetic variation and outcrossing rate in relation to population size in Gentiana pneumonanthe L. Conservation Biology 1014-1026. Rajora, O.P, A. Mosseler and J.E. Major (2000). Indicators of population viability in red spruce, Picea rubens. II. Genetic diversity, population structure, and mating behavior. Canadian Journal of Botany 78: 941-956. Rajora, O.P., A. Mosseler and J.E. Major (2002). Mating system and reproductive fitness traits of eastern white pine (Pinus strobus) in large, central versus small, isolated, marginal populations. Canadian Journal of Botany 80: 1173-1184. Ritland K. (1990). A series of FORTRAN computer programs for estimating plant mating systems. Journal of Heredity 81: 235-237. Ritland, K. (2002). Extensions of models for the estimation of mating systems using n independent loci. Heredity 88: 221-228. Robledo-Arnuncio, J.J., R. Alia and L. Gil (2004). Increased selfing and correlated paternity in a small population of a predominantly outcrossing conifer, Pinus sylvestris. Molecular Ecology 13: 2567-2577. Rungis, D., Y. Berube, J. Zhang, S. Ralph, C E . Ritland, B.E. Ellis, C. Douglas, J. Bohlmann and K. Ritland (2004). Robust simple sequence repeat markers for spruce (Picea spp.) from expressed sequence tags. Theoretical and Applied Genetics 109: 1283-1294. Sagarin R.D. and S.D. Gaines (2002). The 'abundant centre' distribution: to what extent is it a biogeographical rule? Ecology Letters 5: 137-147. Schwartz, M.K., L.S. Mills, Y. Ortega, L.F. Ruggiero and F.W. Allendorf (2003). Landscape location affects genetic variation of Canada lynx (Lynx canadensis). Molecular Ecology 12: 1807-1816. Scotti, I., F. Magni, GP. Paglia and M . Morgante (2002). Trinucleotide microsatellites in Norway spruce (Picea abies): Their features and the development of molecular markers. Theoretical and Applied Genetics 106: 40-50. Smouse, P.E. and V.L. Sork (2004). Measuring pollen flow in forest trees: An exposition of alternative approaches. Forest Ecology and Management 197: 21-38. Spencer, C C , J.E. Neigel and RL. Leberg (2000). Experimental evaluation of the usefulness of microsatellite DNA for detecting demographic bottlenecks. Molecular Ecology 9: 1517-1528. Wang, T , S.N. Aitken, J.H. Woods, K. Polsson and S. Magnussen (2004). Effects of inbreeding on coastal 40 Douglas fir growth and yield in operational plantations: a model-based approach. Theoretical and Applied Genetics 108: 1162-1171. Wang, X.R., V.E. Chhatre, M.C. Nilsson, W. Song, O. Zackrisson and A.E . Szmidt (2003). Island population structure of Norway spruce (Picea abies) in northern Sweden. International Journal of Plant Sciences 164: 711 -717. Warner, B.G. and R.W. Mathewes (1982). Ice-free conditions on the Queen Charlotte Islands, British Columbia, at the height of late Wisconsin Glaciation. Science 218: 675-677. Widmer, A. and C. Lexer (2001). Glacial refugia: sanctuaries for allelic richness, but not for gene diversity. Trends in Ecology & Evolution 16: 267-269. Williams, C . G and O. Savolainen (1996). Inbreeding depression in conifers: Implications for breeding strategy. Forest Science 42: 102-117. 41 CHAPTER 3 Isolation by distance with postglacial migration and adaptive gradients in growth traits in Picea sitchensis (Pinaceae) By Makiko Mimura and Sally N. Aiken Introduction In response to climate change during the Quaternary, species' ranges repeatedly shifted northward and southward (Hewitt 1996; Hewitt 2000). Climate change over time and space shifted the intensity and direction of natural selection toward new local optima. Although a whole species range could shift to occupy the same climate envelope as defined by the same range of temperature and moisture as before the change, the species would still have to adapt to new set of environmental conditions including biotic agents, species interactions or a new photoperiod. Species' range shift driven by climate change is not just a movement of an entire range, but involves migration and adaptation to new environments (Davis and Shaw 2001). There are many reports of geographic divergence among populations of forest trees. Genetic divergence among populations is shaped by selection, genetic drift, or a combination of these factors (Coyne 1992). Genetic drift randomly fixes alleles in small populations. If there are relatively few founders during migration, strong genetic divergence from ancestral populations may result (Gavrilets and Hastings 1996; Gavrilets and Boake 1998). While drift has random effects on divergence among populations, natural selection tends to favour similar phenotypes in similar environments. The clinal changes in adaptive phenotypic values along environmental gradients that many species show suggest selection is often stronger than genetic drift. Fossil pollen records suggest rapid migration of tree species in response to current postglacial warming (Davis 1981; Macdonald 1993; King and Herstrom 1997). Migration rates estimated from several tree species exceed lOOm/year (Clark 1998). Long-distance dispersal may be a key feature of rapid migration of tree species (Cain et al. 2000; Cain et al. 2003), and populations may have spread as described by the fit-tail dispersal model (Clark 1998; Clark et al. 2003). High gene flow and high 42 genetic variation within populations, typical in tree species (Hamrick et al. 1992), may have been enhanced by long-distance dispersal. However, long-distance dispersal and high gene flow initially homogenizes genetic variation among populations. Despite past rapid migration during the Quaternary, tree species exhibit clinal phenotypic values along environmental gradients (Morgenstern 1996). It is a paradox that tree species migrate so fast yet adapt locally to new environments. Davis and Shaw (2001) proposed that gene flow from neighboring populations plays a crucial role in species' range shift in response to linear environmental shifts. During such environmental changes, gene flow from neighboring populations can provide somewhat pre-adapted alleles to founder populations, and with genetic recombination and selection, may enhance local adaptation. A migrating population may adapt to a new environment through a combination of gene flow and selection. Such species should show isolation-by-distance genetic structure following the stepping stone migration model in relatively few generations, while exhibiting adaptive population divergence across the range. The concept of isolation by distance (IBD) was originally addressed by Wright (1943). In this model, the level of gene flow (or number of migrants) is expected to decrease with increasing distance between populations under equilibrium conditions. In theory, IBD should be more pronounced at equilibrium between genetic drift and gene flow (Slatkin 1993), thus it is expected that IBD will be seen in long-established populations but may not be detectable in recently established populations (e.g. Genton et al. 2005; Sharbel et al. 2000). In addition, Peterson and Denno (1998) conducted mata-analysis in host insect species and found that high and low dispersal species showed less IBD compared to moderate dispersal species. We expect to see IBD when gene flow from neighboring populations is stronger than long-distance gene flow in population genetic structure for long-established populations. Migration in response to postglacial warming has left signatures on population structure in many species. A study of chloroplast D N A (cpDNA) of herbaceous and woody plants distinguished northern and southern genotypes for 6 of 7 species along the Pacific coast of North America (Soltis et al. 1997). This north-south population structure suggests either leading edge migration or glacial refugia in both northern and southern areas. Similar patterns have been shown for many species in both southwestern and eastern North America (Liston et al. 1992; Parks et al. 1994; Sewell et al. 1996). Sitka spruce (Picea sitchensis (Bong.) Carr.) did not, however, show the same phylogeographic pattern and had little geographic structure in cpDNA (Soltis et al. 1997). Sitka spruce is a wind-pollinated, outcrossing conifer species, endemic to North America. The current range of this species spans more than 22 degrees of latitude from northern California to Kodiak Island 43 in southwest Alaska. At the Last Glacial Maximum, the southern range limit of Sitka spruce reached at least 22 km south of San Francisco Bay, California (Daubenmire 1967). Sitka spruce extended its northern range limit from south of Puget Sound, Washington, in the early Pleistocene, to the current northern limit on Kodiak Island in Alaska. Growth performance of populations from central regions (southern British Columbia to southeast Alaska) exhibits a strong correlation with environmental gradients such as latitude (Falkenhagen 1977). The absence of genetic differences among populations in cpDNA in Sitka spruce, which is paternally inherited in the Pinaceae and maternally inherited in the other species studied by Soltis et al. (1997), implies that high pollen flow across the wide range has homogenized population genetic structure. Similar results were found using 7 Sequence-Tagged-Site (STS) markers in nuclear D N A (F ST = 0.030; Gapare et al. 2005). Some other widely distributed tree species have also shown little population genetic structure in nuclear DNA, for examples, Picea abies (Mouna et al. 1990) and Picea mariana (Gamache et al. 2003). Although sequential migration decreases genetic diversity in founding populations, self-sustaining populations may establish over a few overlapping generations if the colonizers are sufficiently long-lived to allow for capturing gene flow primarily via pollen from neighboring populations. However, failure in detecting population genetic structure in Sitka spruce may be also due to insufficient polymorphism in the 8 STS markers (averaging 3.3 allele per locus and a total of 26 alleles over the loci, Gapare et al. 2005). Eighteen isozyme markers produced a higher F S T in Sitka spruce populations than these STS markers (F ST = 0.063, recalculated with polymorphic loci, which produced 46 alleles in total, from Yeh and El-Kassaby 1980). We observed adaptive traits and genetic structure among populations across the entire range of Sitka spruce to understand the mechanisms involved in tree migration in response to past environmental change. Sitka spruce is an ideal species for studying these dynamics because its distribution is across a generally linear environmental gradient from south to north along the coast within the range (Figure 3.1) simplifying the historical patterns of migration and selection. Our objectives were to: 1) determine the degree of local adaptation to a wide range of environments using a common garden experiment, 2) evaluate the influence of recent migration on population structure, and 3) investigate the implications of gene flow for migration and adaptation. Materials and Methods Study Area Sitka spruce (Picea sitchensis) is widely distributed from north to south along the Pacific Coast of North America where climates are relatively warm and moist (Figure 3.1). Away from the 44 coast and at higher elevations in the northern portion of the range, forests are dominated by introgression between P. sitchensis and P. glauca (Bennuah et al. 2004). In this study, samples were collected from low elevations (< 70m) close to the Pacific Ocean to avoid hybrid contamination. Six populations sampled for genetic analysis overlap with those used for quantitative trait analysis, while the quantitative traits were analyzed for seventeen populations to test clinal trends and geographic patterns. Seeds were obtained from the British Columbia Ministry of Forests, the University of Alaska, Fairbanks, USDA Forest Service, and California Department of Forestry and Fire Protection, and represented populations from across the full range of the species. Pollen Genetic Structure Two populations from each of the southern range periphery, the centre of the distribution, and the northern range periphery surplus to the need for the common garden experiment were selected to assess genetic diversity and population genetic structure (Figure 3.1). Seed was re-collected from two southern peripheral populations in September, 2003, while stored seed collections were previously available for the central and northern populations. Each of six population samples consisted of 7 to 20 offspring from 10 to 20 families (Table 3.1). D N A was extracted from both one-week-old germinants and the corresponding haploid megagametophyte from the seed using the method of Hodgetts et al. (2001). The pollen gamete contributing to each fertilization producing a seed was determined by genotyping both the embryo (offspring genotype) and corresponding megagametophyte (maternal haplotype). A total of 175 to 227 pollen gamete haplotypes were determined per population and pooled for each population. Al l samples were genotyped for five polymorphic nuclear microsatellite markers: UAPgAG150 developed for Picea glauca (Hodgetts et al. 2001), SPAGC1 and SPAGG3 developed for P. abies (Pfeiffer et al. 1997), and WS0073.H08 and WS0061.K02 developed fori? glauca andP sitchensis (Rungis et al. 2004). Two additional loci, EAC7H07 (Scotti et al. 2002) and UAPgAG105 (Hodgetts et al. 2001), were also genotyped, but were excluded from the analysis due to low polymorphism (UAPgAG105) and a high frequency of null alleles (EAC7H07). Common Garden Experiments Seedlings in a common garden experiment for nine of the seventeen populations consisted of open-pollinated progeny of 10 to 13 seed parents for each population. Montague Island (MI, AK) , Icy Bay (IB, AK), and Columbia River (CR, OR) were bulked to average 30.6 individuals from 12.3 families per populations. The remaining five populations were represented by bulk wild-stand open-pollinated seedlots for reforestation for which seeds were not identified by mother tree (Figure 45 3.1). Bulk seedlots comprise seed from at least 10 seed parents, usually many more. Seeds were germinated in 2002 on fdter paper in Petri dishes at room temperature after soaking in water for 48 hours and stratifying for three weeks at 4°C. Germinants were transplanted into Super Leach Tubes® and grown in a greenhouse at U B C in Vancouver, BC, for one growing season. In February 2003, the seedlings were transplanted into outdoor raised nursery beds at UBC. The beds were 0.75m tall and filled with sandy loam topsoil. Since the southern and northern populations exhibited considerable differences in height in the first year, populations were grouped into southern, central, and northern regions to avoid shading effects. A split-plot block design was used with eight blocks, where the main plots were regions and the subplots were populations within region. Population subplots were noncontiguous with individuals from populations randomized within main plots. Bud phenology (timing of bud break and bud set) and height were recorded repeatedly throughout the growing period in 2003. Final height was also obtained at the end of the growing season in 2004. Bud break was defined as the Julian date when needle primordia first emerged though the bud scales. Bud set was defined as the Julian date when brown bud scales were first visible to the naked eye at the primary shoot apex on the leader. To eliminate first year effects, height in the second and third year was recorded as the length between first year or second year terminal bud scale scars and the shoot apex. Cold hardiness test Cold hardiness was determined by artificial freeze testing on a subset of trees after bud set in the fall of 2003 using the methods described in Hannerz et al. (1999). One or more branches were harvested, and four needle segments 5mm in length cut at both ends were placed in each tube with 0.2ml of distilled water and a trace of silver iodide for ice nucleation. A total of 446 individuals from 17 populations were sampled on November 3rd, and 420 individuals from 17 populations sampled on November 10th. For the samples harvested on Nov 3rd, the test temperatures were -16°C and -26°C, and for the samples harvested on Nov 10th, they were -20°C and -30°C due to anticipated acclimation between the two sample dates. At the planting site in Vancouver, the first natural frost fell on the week of November 3 r d in 2003. Increased cold hardiness generally develops after the first exposure to frost (Sakai and Larcher 1987). Thus, we lowered the test temperature for the second week of testing (Nov. 10th). For both test dates, chamber temperature was lowered at a rate of 4°C/h from an initial temperature of 4°C, and held at the first test temperature for one hour. One complete set of samples was then removed, and the temperature was further decreased by 4°C/h to the second test temperature, and again held for one hour before the second complete set was removed. Control samples for all genotypes were kept at 4°C for the duration of freezing test. After freezing, samples were thawed 46 overnight at 4°C, diluted with 3ml of distilled water, shaken for one hour and measured for electrolytic conductivity. Samples were then heat killed in 95°C water bath, shaken for one hour and remeasured for maximum conductivity. Flint's Index of Injury (cold injury index) was estimated as a percent of maximum injury possible from the electrolytic leakage of the corresponding control, frozen and heat killed samples (Hannerz et al. 1999). Data Analyses Genetic data Two statistical measures of genetic distances, F S T (Wright 1951; Nei 1987) and R S T (Slatkin 1995), were estimated from components of genetic variation within and among population, and for pollen gamete haplotypes pooled within populations, using Arlequin var. 2.001 (Schneider et al. 2000). F S T is an estimate of the genetic distance derived from variation in allele frequencies among populations based on the infinite allele model. R S T is an analog of F S j , developed specifically for microsatellite markers that follow the stepwise mutation model, and takes into account allele size. R S T is generally recognized as a better estimator when mutation rate is high relative to migration rate, while F S T is a conservative estimate under most conditions if sample size is small (Gaggiotti et al. 1999). In this study, both estimators were used to analyze effects of genetic drift and mutation on population structure. The correlation between genetic and geographic distance matrices was tested using a Mantel test with 1000 permutations. For cluster analysis of pollen gamete populations, R S T and F S T were estimated using MICROSAT (Minch 1995) with 1000 bootstrap tests. The U P G M A technique was used to generate dendrograms with the program PHYLIP (Felsenstein 1993). As nodes in the resulting dendrogram were not strongly supported, we used the Neighbor Joining (NJ) method to confirm the reliability of the U P G M A dendrograms. Quantitative growth trait data Height measurements were log-transformed to generate a more normal distribution. To understand the pattern of variation in growth traits across the geographic range, analyses of variance were calculated for 2 n d and 3 r d year height, bud break, bud set, growth period (days between date of bud break and date of bud set), mean daily growth rate (height increment divided by growth period), and cold injury index, using the model; yiji = u + bj + pj + f(p)jk + bi*pj where b is the effect of block i, pj is the effect of population j and f(p)jk is the effect of family k within a population j . The regional plots in the experimental design to avoid shading effects were not included in this model to allow for quantification of clinal pattern rangewide and their association 47 with environmental gradients. Preliminary analyses indicated continuous clinal patterns in population means across the regional groupings. Cold injury index for the freezing tests on the two different dates were highly correlated phenotypically (r = 0.96), thus cold hardiness scores from each date were standardized, combined and analyzed as a single trait. Least squared population means were estimated with this model using PROC G L M in SAS Ver. 9.0 (SAS Institute Inc. 2000). The procedure estimated simple linear regression coefficients of population means with distance along the coast as an independent variable. Correlations among traits were estimated for population means using PROC CORP.. Principal component analysis was applied to quantitative data for population means for all variables using PROC PRTNCOMP. Principal component analysis extracts orthogonal variables that are linear combinations of the original variables best explaining the total variance in the data set. Using PROC REG, multiple regressions were then performed on the first two principal component scores as dependent variables with geographical and environmental factors (distance along the coast, mean annual temperature, mean warmest month temperature, mean coldest month temperature, growing degree days (above 5°C), and precipitation) as independent variables. Principal components were included in regression models when their eigenvalues were greater than 1.0. We compared standardized measures of genetic differentiation for neutral genetic markers (F S T or R S T) with quantitative traits (QST) to evaluate the relative strengths of natural selection and genetic drift on population differentiation (Media and Crnokrak 2001). Q S T was estimated for the same 6 populations used for estimating population genetic structure, following Spitze (1993) for family-structured populations from additive genetic variances estimated using the PROC VARCOMP procedure of SAS with the restricted maximum likelihood option (METHOD = REML). The additive genetic variances for each trait were also used to estimate within population individual heritabilities. Results The genetic distance measures FST and RST showed similar degrees of population differentiation (Table 3.2, F S T = 0.11, R S T = 0.09). The U P G M A dendrogram based on F S T clustered central populations and continuous populations together while the disjunct peripheral populations were outliers (Figure 3.2a); however, the dendrogram based on NJ method was not consistent with the U P G M A dendrogram (Figure 3.2c). RST clustered pairs of disjunct and continuous populations regionally for both the central and southern regions, but did not cluster the northern disjunct and continuous populations with U P G M A method (Figure 3.2b). Similarly, NJ method clustered populations from same regions (Figure 3.2d). For both measures of genetic distance, the two central 48 populations, regardless of degree of isolation, are core in terms of genetic structure. Correlations of geographic and genetic distances between populations were statistically significant (Figure 3.3). Isolation by distance (IBD) was stronger for RST (Mantel test: r=0.73, p<0.01) than for F S T (Mantel test: r=0.52, p<0.01). Individual heritabilities were moderate and fairly uniform, ranging from 0.23 to 0.32 (Table 3.3). The lowest heritability was for daily growth rate, while bud break date showed the highest heritability. Heights at ages 2 and 3 showed similar heritabilities (0.25 and 0.26, respectively). Regressions of quantitative traits with distance along the coast revealed strong, significant clines for all traits except bud break timing (Table 3.3). Southern populations had much longer periods of primary growth periods than northern populations. Height, days to bud set and fall cold injury significantly increased toward the south, while average daily growth rate increased significantly to the north. Both height increment ('height') at age 2 (2003) and at age 3 (2004) increased at a rate of 1.05 mm per 100 km south (Table 3.3). Julian days to bud set increased by 3.2 days, and standardized cold injury index increased by 0.07 points for every 100 km south. Mean daily growth rate ('growth rate') decreased 0.008 mm per day for every 100 km south (Table 3.3). Bud beak timing and daily growth rate appeared to be genetically independent traits (Table 3.4). In contrast, significant negative phenotypic correlations were found between population mean growth rate and bud set timing (r=-0.56, Table 3.4), and between growth rate and growth period (days from bud break to bud set; r=-0.58, Table 3.4). This means individuals that terminate primary growth earlier (set buds earlier) tend to grow faster within the limited growing periods. Overall, populations from the north set bud earlier, grow less overall in height but more per day, and are hardier to cold events in fall than southern populations. Principal component analysis of quantitative growth traits from all population means revealed strong geographic relationships (Figure 3.4). Principal component 1 (PCI) represented 72.2% of the overall variation in growth (height at age 2 and age 3, bud break timing at age 2, bud set timing at age 2, and average fall cold injury index week 1 and week 2) (Table 3.5). Principal component 2 (PC2), primarily represented growth rate and bud break timing, and accounted for 17.9% of the total variation (Table 3.5). PCI seemed to explain population differentiation due to geographic distribution. PC2 was hard to interpret as both southern peripheral and northern peripheral populations showed similar high scores, and central populations had lower scores (Figure 3.3). Regression analysis was performed on the first two principal components (PCI and PC2) whose eigenvalues were more than 1.0. PCI had a significant linear relationship with coastal distance, mean annual temperature (MAT; Figure 3.5a), mean warmest month temperature (MWMT), mean coldest month temperature (MCMT), and growing degree days (GDD) (Table 3.6). PC2 had a significant relationship only with M W M T 49 (p=0.04; Table 3.6, Figure 3.5b). In contrast with other variables, M W M T does not show a linear relationship with latitude. It is higher in the south-central Vancouver area (lat. 49.16°) and lower at the northern and southern margins, inverse to the U-shaped north-south distribution of PC2. Mean population differentiation for the observed growth traits (QST) was 0.56 (Table 3.3). Bud set timing and growth period exhibited extremely strong differentiation among populations (QST = 0.87 and 0.86, respectively) and these traits were highly correlated (Table 3.4). Daily growth rate showed lower but still substantial differentiation (Q S T = 0.29; Table 3.3). Height at age 3 had higher differentiation among populations than height at age 2. Al l Q S T estimates were based on only the family-structured populations as they allowed estimation of additive genetic variance within populations. Al l QST estimates were substantially higher than estimates of population differentiation for the neutral genetic markers, F S T and R S T (Table 3.2 and Table 3.3). Discussion Adaptive divergence across the range Gradual clines in growth traits of Sitka spruce with latitude have long been recognized (Harris 1978; Xu et al. 2000). Strong geographic clines in all quantitative traits except bud break timing were also found in this study. By testing a wide range of populations, strong differences in population adaptation to local environments have been characterized. For example, individuals from southern populations grow taller than those from northern populations, while individuals from the north grow faster per day than those from southern populations. These growth strategies correspond to available frost-free growing periods in the original environments of these populations. The date with a 50 percent probability of first frost (0°C) occurs on or before September 30th in Kodiak Island and Kitoi Island, Alaska (Western Regional Climate Centre, USA), but not until December 16th in Fort Bragg, California (Koss et al. 1988). The growth response in the northern populations appears to be an adaptive strategy well-suited to a short growing season and low temperature environment: reach maximum growth as quickly as possible within the limited favourable period, while minimizing risk of cold injury. Summer temperature is an indirect indicator of growth rate and timing of bud break, where lower summer temperatures of population source environments are associated with later bud break but faster growth rate (Table 3.6). All traits we observed are assumed to be under differential selection for local adaptation as all Q S T estimates exceeded values for the neutral genetic distance measures F S T and RST (Merilaand Crnokrak 2001). Bud phenology is key for the adaptation of conifers to local climate (Howe et al. 2003) as fall and spring frosts can severely damage current year shoot and needles (Redfern and Cannell 1985; 50 Peterson et al. 1997). Before hardening, or when de-hardening occurs during warm periods in winter, tissues become more susceptible to frost (Cannell et al. 1985), thus cold hardiness is an adaptive trait for optimal growth (Howe et al., 2003). For conifer species, bud set timing is generally determined by photoperiod (critical night length), and bud break by heat sum accumulation in spring following adequate chilling in winter (Aitken and Hannerz 2001). In this study, there was no significant variation in bud break timing among the populations. Therefore, although both bud set and bud break timing are major determinants of seasonal growth cycle in Sitka spruce, bud set timing (which shows different critical photoperiods among populations for initiation) is more important when populations are migrating or when environments are changing. Cold hardiness is highest in northern populations. The southernmost populations showed little, if any, development of cold hardiness. Hardiness would probably be less beneficial in the California environment where temperatures rarely drop down below 0°C throughout year (average minimum temperature of the coldest month 4.4°C at Fort Bragg, Normal Climate Data 1971-2000, National Climate Data Center of US). The steep cline observed for bud set timing is evidence for differential adaptation to the risk of frost. Isolation by distance in recently migrated populations The microsatellite markers used detected relatively strong genetic differentiation among populations of Sitka spruce (FS T=0.11, Rsi^O.09). Conifers generally have low genetic differentiation for nuclear loci, for instance GST averages 0.055 in Picea, 0.065 in Pinus, and 0.063 in Abies (Hamrick et al. 1992; Le Corre and Kremer 1998). The weak genetic differentiation of conifer populations is usually attributed to high gene flow via wind dispersed pollen, as well as large population size, preventing differentiation due to genetic drift. In this study, the genetic differentiation detected by microsatellites for 6 pollen gamete population pools was much higher than that found for 8 populations from the same geographic range (Fort Bragg, C A to Kodiak Island, AK) with a large sample of mature trees genotyped by the 8 STS markers (F ST = 0.03, Gapare et al. 2005). This does not agree with a general expectation: a high stepwise mutation rate on microsatellite loci could increase homoplasy and decrease F S T (Balloux and Lugon-Moulin 2002). The stronger population structure detected with microsatellites than with STS markers may be due to the high mutation rate at microsatellite loci reflecting genetic changes on shorter time scales than STS markers, which have lower mutation rates. This may indicate that gene flow in this species has been more limited recently than historically. However, our estimate is relatively similar to the isozymed-based estimate for 10 populations from a slightly smaller geographic range (F S T = 0.063, recalculated with 18 polymorphic loci from the allele frequency data in Yeh and El-Kassaby 1980). The STS markers were developed based on EST sites and reflect mostly intro-length polymorphism (Perry and 51 Bousquet 1998a, b). The lack of population differentiation reflected by the STS markers is likely due to their low total polymorphism and low mutation rate. Our estimates of F S T and RST among all populations were very similar, on average, but pair-wise comparisons between populations illustrated different features of these estimates. F S T, assuming mutation follows the infinite alleles model, can provide a measurement of population differentiation caused by genetic drift (Reynolds et al. 1983), yet a primary problem of F S T is its sensitivity to mutation rate when gene flow is low (Ballouz and Lugon-Moulin 2002). It is assumed that the microsatellite mutation rate is generally high, and that mutations mostly follow the stepwise mutation model (SMM). R S T takes into account differences in the number of motif repeats in microsatellite alleles resulting from such stepwise mutations (Slatkin 1995). Under the strict S M M , RST is independent of the mutation rate. Ballouz and Lugon-Moulin (2002) note that the factor influencing measurements of F S T and R S T is not mutation rate itself but the magnitude of mutation relative to migration, and mutation has little influence on the measurement of F S T and R S T when migration is high. Therefore, the values of these two measurements should become similar as relative migration rate increases. In this study, R S T better reflected the effect of gene flow on population structure than F S T in this species. Isolation by distance should be more pronounced at equilibrium, and it may take populations considerable time to achieve equilibrium (Slatkin 1993). High dispersal insect species also tend to show weak isolation by distance (Peterson and Denno 1998). Strong isolation by distance may not develop in Sitka spruce due to 1) long-distance dispersal via pollen may inhibit population divergence, or 2) the great northward range expansion that occurred since the last glacial maximum (< 15,000 years ago). The genetic diversity in these populations may not have always decreased toward the north (the migration tips), but rather may have resulted from current isolation associated with bottlenecks and other historical reductions in population size (CHAPTER 2). Despite the species' rapid expansion and loss of alleles in the isolated populations (CHAPTER 2), strong isolation by distance in this species suggests that the populations can to some extent reach drift-migration equilibrium fairly quickly, at least at microsatellite loci, and that the majority of contemporary gene flow is restricted to within geographic regions, showing stepping stone migration along the coastal distribution of the range. Isolation by distance may imply the accumulation of genetic divergence with time since divergence from a single ancestral population; however, we have too few populations to identify putative refugia in this study. We conclude that gene flow within geographic regions has been an important factor influencing current population structure, despite possible long-distance dispersal during past migration. 52 Considering that the rapid past migrations of tree species can not be explained without invoking long-distance dispersal, yet gene flow from neighouring rather than distant populations is crucial for adaptation, we hypothesize adaptation in conjunction with postglacial migration. In the first phase of postglacial migration, long-distance dispersal of seed may have initially established populations, followed by pollen flow from neighboring populations that increased genetic diversity within the new populations. In this phase, the effective population size of the founder population would increase, while hard selection (e.g. frost events) would act on traits essential for survival (e.g. bud set timing). Selection is more efficient in larger and more variable populations. Following population establishment, soft selection would increase the frequency of locally favoured alleles associated with more competitive growth, which perhaps would originate from neighboring populations rather than distant populations. Ultimately, this results in geographic clines. This occurs when environmental gradients are steep enough for selection to counteract gene flow. Conifer species such as Sitka spruce, with wind pollination, high fecundity, high longevity and high gene flow, may have effective population sizes that approach census size. This gives Sitka spruce the ability to colonize new areas and sustain founder populations until population size and within-population variation reach sustainable levels that can respond to selection efficiently. Through these processes, wind-pollinated conifers have had the capacity to adapt to rapid environment changes in the past. However, it should be noted that CHAPTER 2 suggested contemporary mating system was strongly influenced by geographic position, potentially associated with population size. It appears that effective population size and outcrossing rate declines toward the northern and southern range peripheries, especially in the disjunct peripheral populations (CHAPTER 2). This is also suggestive of the sensitively of the species to current population size and gene flow. Sitka spruce exhibits a wide range of adaptive trait variation despite rapid Quaternary migration, and has maintained or recovered relatively strong population genetic structure in the past 15,000 years. Gene flow could be partially responsible for the rapid migration of Sitka spruce by providing diversity for more effective adaptation to new environments encountered during migration, although current gene flow appears relatively limited, possibly contributing to the maintenance of adaptive divergence. The degree of adaptation may depend on rates of gene flow and the steepness of environmental gradients. This tree species has demonstrated a high capacity to respond to selection and adapt to past climate change along the Pacific Coast of North America. 53 Table 3.1. Summary of sample populations with US states or a Canadian province of population origins. Isolation (Isol.): Cont. indicates population is located in the continuous portion of species' range; Disj. indicates it is disjunct, separated from other populations by 60-100 km. Seedlings (Seed) were nested in families (Fam.) unless indicated as 'Bulk'. Distance (Dist.): actual distance in km from southernmost location. Polllen indicates the number of seedlings for which the paternal contribution (haplotype) was genotyped. MAT: mean annual temperature; M W M T : mean warmest month temperature; M C M T : mean coldest month temperature; DD: degree days; and Prec: annual precipitation. Common Marker Garden analysis Geographic and climatic variables Population (code) State Isol. Fam. Seed Fam. Pollen Dist. MAT MWMT MCMT DD Prec Valdez (VL) AK Cont. 10 114 - - 3129 3.5 12.9 -5.6 815 1712 Icy Bay (IB) AK Cont. Bulk 30 - - 2925 4.2 12.0 -3.4 721 4074 Montague (MI) A K Cont. Bulk 30 - - 3272 3.9 12.5 -4.1 786 2445 Rocky Bay (RB) AK Cont. 10 114 19 227 3540 4.1 12.7 -2.2 735 1706 Kodiak Island (KI) AK Disj. 10 124 20 208 3693 4.7 12.8 -1.3 769 1914 Lot 71 (71) A K Cont. Bulk 30 - - 2335 7.2 14.0 1.61 1141 219 Lot 95 (95) A K Cont. Bulk 30 - - 2235 5.8 14.5 -3.0 1141 163 Lot 167 (167) AK Cont. Bulk 30 - - 2124 6.6 14.2 -1.2 1184 202 Lot 111 (111) AK Cont. Bulk 30 - - 2080 6.6 14.0 -0.1 1127 288 Prince Rupert (PR) BC Cont. 12 105 - - 1652 7.1 13.5 1.3 1226 2594 Queen Charlotte Islands (QC) BC Disj. 10 87 10 175 1472 8.3 15.0 3.2 1462 1398 Ocean Falls (OF) BC Cont. 10 116 18 192 1429 8.0 16.8 -1.0 1652 1702 Vancouver Island (VI) BC Cont. 11 116 - - 1240 9.7 17.6 3.0 1960 1179 Vancouver (VA) BC Cont. 13 121 - - 1099 10.0 17.1 3.6 2027 1277 Columbia River (CR) OR Cont. Bulk 30 - - 741 10.6 16.0 5.8 2019 1705 Redwood (RW) CA Cont. 10 112 17 192 292 11.6 14.8 8.8 2395 968 Fort Bragg (FB) CA Disj. Bulk 39 14 196 0 11.8 14.7 9.1 2479 1041 Table 3.2. Pair-wise genetic distance between populations sampled. F S T estimates are above the diagonal and R S T estimates are below in Italics. South Central North Continuous Disjunct Continuous Disjunct Continuous Disjunct Region Isolation Population RW, CA FB, CA OF, BC QC, BC RB, A K KI, A K South Continuous RW, CA 0.06 0.03 0.03 0.05 0.14 Disjunct FB, CA 0.00 0.13 0.12 0.14 0.24 Central Continuous OF, BC 0.02 0.06 0.00 0.10 0.16 Disjunct QC, BC 0.01 0.03 0.00 0.10 0.15 North Continuous RB, A K 0.17 0.23 0.09 0.12 0.17 Disjunct KI, A K 0.19 0.21 0.17 0.17 0.19 54 Table 3.3. Summary of clinal variation in growth traits among populations. The change in all traits per 100 km along the Pacific Coast of North America from north to south is also estimated. Significant regressions (p < 0.001) are indicated in bold. North-South cline Growth traits Heritability Qst R2 P Change per 100km 0.25 0.48 0.24 <.0001 1.05 0.26 0.72 0.48 <.0001 1.05 0.32 0.16 0.05 0.3864 0.06 0.27 0.87 0.94 <.0001 3.24 0.25 0.86 0.71 <.0001 3.00 0.23 0.29 0.39 0.0073 -0.008 n/a n/a 0.58 <.0001 0.07 Height age2 (mm) Height age3 (mm) Bud break (days) *3 Bud set (days) 3 Growth period (days) Growth rate (mm) *4 Cold injury index *' The change in trait per 100km to the south. *2 Analyzed as log height but back transformed for changes per 100km *3 Julian date (days from Jun 1st) *4 Mean daily growth = (height at age 2)/(bud set Julian date - bud break Julian date) *5 Analyzed as standardized cold injury index of the freezing tests on the two dates Table 3.4. Pearson correlations among population means for traits measured in 17 Sitka spruce populations in a common garden. Height age2 Height age3 Bud set Bud break Growth Period Growth Rate Cold injury index Height age2 * 1 0.85 0.90 0.30 0.90 -0.18 0.94 Height age3 * 1 0.95 0.3 0.94 -0.56 0.92 Bud set 1 0.25 0.99 -0.56 0.94 Bud break 1 0.18 0.16 0.44 Growth Period 1 -0.58 0.92 Growth Rate 1 -0.35 Cold injury index 1 * log transformed 55 Table 3.5. Eigenvectors of the first (PCI) and second (PC2) components from Principal component analysis. Eigenvectors Traits PCI PC2 Height (age 2)*' Height (age 3)*1 0.41 0.19 0.43 -0.05 Bud set 0.44 -0.07 Bud break 0.15 0.71 Growth period 0.44 -0.13 Growth rate -0.23 0.64 Cold injury index *2 0.43 0.17 Eigenvalue 5.05 1.25 72.2% 17.9% log transformed standardized and combined from the two tests on the different dates Table 3.6. Summary of results from the regression of the first two Principal components (PCI and PC2, refer in Table 3.5) extracted from quantitative growth traits on geographic and climatic variables. Distance refers to the distance along the Pacific coast from north to south. Variables in bold have p values <0.05. Dependent Variable Predictor Variable R Slope P PCI Distance (km) 0.94 -0.2 <.001 Mean Annual Temperature (°C) 0.95 0.80 <.001 Mean Wannest Month Temperature (°C) 0.44 0.89 0.004 Mean Coldest Month Temperature (°C) 0.90 0.50 <.001 Degree Days (°C) 0.96 0.004 <.001 Precipitation (mm) 0.06 -0.0005 0.36 PC2 Distance (km) 0.02 0.014 0.59 Mean Annual Temperature (°C) 0.02 -0.05 0.62 Mean Wannest Month Temperature (°C) 0.26 -0.34 0.04 Mean Coldest Month Temperature (°C) 0.002 0.011 0.87 Degree Days (°C) 0.01 -0.000 0.79 Precipitation (mm) 0.22 0.0005 0.06 56 Figure 3.1. Picea sitchensis natural range and populations sampled. Symbols indicate population sampled for quantitative growth traits. Circles indicate progeny-sampled populations and squares bulk populations for quantitative trait analysis. RB and KI from Alaska, PR and OF from British Columbia, and FB and RW from California were also sampled for genetic structure analysis. The star indicates Vancouver, the location of the common garden experiment. 130° 120° 110° 57 Figure 3.2. Cluster analysis based on genetic distance, with UPGMA method a) F S T , and b) R S T , and with NJ method c) F S T , and d) R S T , as pair-wise genetic distance among populations. The scale length indicates 0.01 of coefficient of distance measures. Numbers on the dendrogram indicate confidence on nodes from 1000 bootstrap resamplings. KI, AK (northern isolaseo) Ki, AK (northern isolated) RW. CA (southern continuous} 58 Figure 3.3. Geographic distance versus genetic distance among populations for FST and RST-Correlations and probabilities were estimated from a Mantel test with 1000 repeats of bootstrap resampling. a) 0.25 0.2 H 0.15 0.1 .a ? 0.05 CO 13 -0.05 r = 0.52 • p < 0.01 • • • •• • • • • • * • 0 1000 2000 3000 4000 Geographic distance (Km) b) te 025 0.2 0.15 S »-1 TJ y 1 0.05 -J <» -0.05 r = 0.73 • p < 0.01 • •• • • • • • • • • • • 1000 2000 3000 Geographic distance (km) 4000 59 Figure 3.4. Plot of the first (PCI) and second (PC2) components estimated from Principal component analysis for quantitative traits (Table 3.5). Population codes are defined in Table 3.1. RW, CA_ CM O 0 CL -2 IB, AK Mt, AK Kl, AK • RB, AK CR, CA VL, AK m 95, AK 71, AK 167, AK^ 111.AK PR, BC ORBC QC. BC # V A , B C VI, BC FB, CA P C 1 60 Figure 3.5. Regression analysis of the first two Principal components of quantitative traits on the climate variables tested in Table 3.6. (a) o CL 6 5 4 3 2 1 0 -1 -2 -3 -4 R =0.93 p < 0.001 5 10 Mean Annual Temperature 15 10 12 14 16 18 Mean Warmest Month Temperature 6 1 Literature Cited Aitken S. N . and M . Hannerz (2001). Genecology and gene resource management strategies for conifer cold hardiness. In F. J. Bigras, and S. J. Colombo [eds], Conifer Cold Hardiness, 23-53. Kluwer Academic Publishers, Dordrecht. Balloux, F. and N. Lugon-Moulin. 2002. The estimation of population differentiation with microsatellite markers. Molecular Ecology 11: 155-165. Bennuah, S.Y., T L . Wang and S.N. Aitken (2004). Genetic analysis of the Picea sitchensis x glauca introgression zone in British Columbia. Forest Ecology and Management 197: 65-77. Cain, M . L . , B . G Milligan and A . E . Strand (2000). Long-distance seed dispersal in plant populations. American Journal of Botany 87: 1217-1227. Cain, M.L . , R. Nathan and S.A. Levin (2003). Long-distance dispersal. Ecology 84: 1943-1944. Cannell, M.G.R., L.J. Sheppard, R. Smith and M.B. Murray (1985). Autumn frost damage on young Picea sitchensis. 2. Shoot frost hardening, and the probability of frost damage in Scotland. Forestry 58: 145-166. Chaisurisri, K., J.B. Mitton and Y . A . El-Kassaby (1994). Variation in the mating system of Sitka spruce (Picea sitchensis) - evidence for partial assortative mating. American Journal of Botany 81: 1410-1415. Clark, J.S. (1998). Why trees migrate so fast: Confronting theory with dispersal biology and the paleorecord. American Naturalist 152: 204-224. Clark, J.S., M . Lewis, J.S. McLachlan and J. HilleRisLambers (2003). Estimating population spread: What can we forecast and how well? Ecology 84: 1979-1988. Coyne, J.A. (1992). Genetics and speciation. Nature 355: 511-515. Crispo E. and A . P Hendry (2005). Does time since colonization influence isolation by distance? A meta-analysis. Conservation Genetics 6: 665-682. Daubenmire, R. (1967). Some geographic variations in Picea sitchensis and their ecologic interpretation. Canadian Journal of Botany 46: 787-798. Davis, M.B. (1981). Quaternary history and the stability of forest communities. In D .C. West and H. H. Shugart and D.B. Botkin [eds], Forest Succession: Concepts and Application. Springer. New York. Davis, M.B. and R.G. Shaw (2001). Range shifts and adaptive responses to Quaternary climate change. Science 292: 673-679. Felsenstein, J. (1993). PHYLIP (Phylogeny Inference Package) version 3.5c. Distributed by the author. Department of Genetics, University of Washington, Seattle. Gaggiotti, O.E., O. Lange, K. Rassmann and C. Gliddon (1999). A comparison of two indirect methods for estimating average levels of gene flow using microsatellite data. Molecular Ecology*: 1513-1520. Gamache, I., J.P. Jaramillo-Correa, S. Payette and J. Bousquet (2003). Diverging patterns of mitochondrial and nuclear D N A diversity in subarctic black spruce: imprint of a founder effect associated with postglacial colonization. Molecular Ecology 12: 891-901. Gapare, W.J., S.N. Aitken and C E . Ritland (2005). Genetic diversity of core and peripheral Sitka spruce (Picea sitchensis (Bong.) Carr) populations: implications for conservation of widespread species. Biological Conservation 123: 113-123. Gavrilets, S. and C.R.B. Boake (1998). On the evolution of premating isolation after a founder event. American Naturalist 152: 706-716. Gavrilets, S. and A. Hastings (1996). Founder effect speciation: A theoretical reassessment. American Naturalist 147: 466-491. Genton B.J., J.A. Shykoff and T. Giraud (2005) Hight genetic diversity in French invasive populations of common ragweed, Ambrosia artemisiifolia, as a result of multiple sources of introduction. Molecular Ecology 14: 4275-4285. Hamrick, J.L., M.J.W. Godt and S.L. Sherman-Broyles (1992). Factors influencing levels of genetic 62 diversity in woody plant species. New Forest 6 : 95-124. Hannerz, M . , S.N. Aitken, J.N. King and S. Budge (1999). Effects of genetic selection for growth on frost hardiness in western hemlock. Canadian Journal of Forest Research 29: 509-516. Harris, A.S. (1978). Distribution, genetics, and silvical characteristics of Sitka spruce. Proceedings of the IUFRO joint meeting of working parties, British Columbia Ministry of Forestry. Vancouver, Canada. Hewitt, G.M. (1996). Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58: 247-276. Hewitt, G..M. (2000). Trie genetic legacy of the Quaternary ice ages. Nature 405: 907-913. Hodgetts, R.B., M . A . Aleksiuk, A. Brown, C. Clarke, E . Macdonald, S. Nadeem and D. Khasa (2001). Development of microsatellite markers for white spruce (Picea glauca) and related species. Theoretical and Applied Genetics 102: 1252-1258. Howe, G.T., S.N. Aitken, D.B. Neale, K.D. Jermstad, N.C. Wheelern and T.H.H. Chen (2003). From genotype to phenotype: unraveling the complexities of cold adaptation in forest trees. Canadian Journal of Botany 81: 1247-1266. King, R.A. and A.A . Herstrom (1997). Holocene migration rates objectively determined from fossil pollen data. In B. Huntley, W. Cramer and A.V. Marsa. Past and future environmental changes: the spatial and evolutionary responses of terrestrial biota. Springer. New York. Koss, J.R., J.R. Owenby, P.M. Steurer and D.S. Ezell (1988). Freeze/Frost Data. Climatography of the U.S. No.20, Supplement No. L Le Corre, V. and A. Kremer (1998). Cumulative effects of founding events during colonisation on genetic diversity and differentiation in an island and stepping-stone model. Journal of Evolutionary Biology 11: 495-512. Lines, R. (1987). Seed origin variation in Sitka spruce. Proceedings of the Royal Society of Edinburgh Section B-Biological Sciences 93: 25-39. Liston, A. , L . H . Rieseberg and M.A. Hanson (1992). Geographic partitioning of chloroplast D N A variation in the genus Datisca (Datiscaceae). Plant Systematics and Evolution 181: 121-132. Macdonald, G M . (1993). Fossil pollen analysis and the reconstruction of plant invasions. Advances in Ecological Research 24: 67-110. Merila, J. and P. Crnokrak (2001). Comparison of genetic differentiation at marker loci and quantitative traits. Journal of Evolutionary Biology 14: 892-903. Minch E. (1996). MICROSAT. Version 1.4. Stanford University Medical Centre, Stanford, USA. Morgenstern, E. K. (1996). Geographic Variation in Forest Trees. U B C Press, Vancouver. Mouna, O., L. Paule, A . E . Szmidt and K. Karkkainen (1990). Mating system analysis in a central and northern European population of Picea abies. Scandinavian Journal of Forest Research 5: 97-102. Nei, M . (1987). Molecular Evolutionary Gentics. Columbia University Press, New York. Parks, C.R., J.F. Wendel, M . M . Sewell and Y.L. Qiu (1994). The significance of allozyme variation and introgression in the Liriodendron tulipifera Complex (Magnoliaceae). American Journal of Botany 8 1 : 878-889. Peterson, E.B., N . M . Peterson, G.F. Weetman and P.J. Martin (1997). Ecology and Management of Sitka spruce, Emphasizing its natural range in British Columbia. U B C Press, Vancouver. Peterson, M.A. and R.F. Denno (1998). The influence of dispersal and diet breadth on patterns of genetic isolation by distance in phytophagous insects. American Naturalist 152: 428-446. Perry, D.J. and J. Bousquet (1998a). Sequence-tagged-site (STS) markers of arbitary genes: development, characterization and analysis of linkage in black spruce. Genetics 149: 1089-1098. Perry, D.J. and J. Bousquet (1998b). Sequence-tagged-site (STS) markers of arbitrary genes: the utility of black spruce-devived STS primers in other conifers. Theoretical and Applied Genetics 97: 735-743. Pfeiffer, A., A . M . Olivieri and M . Morgante (1997). Identification and characterization of 63 microsatellites in Norway spruce (Picea abies K). Genome 40: 411-419. Raynolds J., B.S. Weir and C C . Cockerham (1983). Estimation of the coancestry coefficient: basis for a short-term genetic distance. Genetics 105: 767-779. Redfern, D.B. and M.G..R. Cannell (1982). Needle Damage in Sitka spruce caused by early Autumn Frost. Forestry 55: 39-45. Rungis, D., Y. Berube, J. Zhang, S. Ralph, C E . Ritland, B.E. Ellis, C. Douglas, J. Bohlmann and K. Ritland (2004). Robust simple sequence repeat markers for spruce (Picea spp.) from expressed sequence tags. Theoretical and Applied Genetics 109: 1283-1294. Sakai, A. and W. Larcher (1987). Frost Survival of Plants. Springer, Heidelberg, Berlin. SAS Institute Inc. (2000). SAS OnlineDocR, Version 8, Cary, NC. Schneider, S., D. Roessli and L. Excoffier (2000). Arlequin: A software for population genetics data analysis. Ver 2.001. Genetics and Biometry Lab, Dept. of Anthropology, University of Geneva. Scotti, I., F. Magni, G.P. Paglia and M . Morgante (2002). Trinucleotide microsatellites in Norway spruce (Picea abies): Their features and the development of molecular markers. Theoretical and Applied Genetics 106: 40-50. Sewell, M . M . , C.R. Parks, and M.W. Chase (1996). Intraspecific chloroplast D N A variation and biogeography of North American Liriodendron L (Magnoliaceae). Evolution 50: 1147-1154. Sharbel, T.F., B. Haibold and T. Mitchell-Olds (2000). Genetic isolation by distance in Arabidopsis thaliana: biogeography and postglacial colonization of Europe. Molecular Ecology 9: 2109-2118. Slatkin, M . (1993). Isolation by distance in equilibrium and non-equilibrium populations. Evolution 47: 264-279. Slatkin, M . (1995). A measure of population subdivision based on microsatellite allele frequencies. Genetics 139: 457-462. Soltis, D.E., M.A. Gitzendanner, D.D. Strenge and PS. Soltis (1997). Chloroplast D N A intraspecific phylogeography of plants from the Pacific Northwest of North America. Plant Systematics and Evolution 206: 353-373. Spitze, K. (1993). Population structure in Daphnia obtuse: Quantitative genetic and allozymic variation. Genetics 135: 367-374. Wright, S. (1951). The genetic structure of populations. Annu. Eugenics 15: 355-420. Xu, P., C C Ying and Y.A. El-Kassaby (2000). Multivariate analyses of causal correlation between growth and climate in Sitka spruce. Silvae Genetica 49: 257-263. Yeh, E C and Y. A. El-Kassaby (1980). Enzyme variation in natural populations of Sitka spruce (Picea sitchensis .) 1. Genetic variation patterns among trees from 10 IUFRO provenances. Canadian Journal of Forest Research 10: 415-422. 64 CHAPTER 4 Gene flow inhibits local adaptation in peripheral population of Sitka spruce By Makiko Mimura and Sally N . Aitken Introduction Peripheral populations at the soft boundaries of a species' range, where the species gradually disappears, may have different evolutionary potentials than central populations due to the physical distance between core and peripheral populations and to environmental heterogeneity across the range (Lesica and Allendorf 1995; Lammi et al. 1999). Current changes in climate will pressure species to shift their range to track favorable environments. As a result, peripheral populations may be key in facing and responding to environmental changes (Lesica and Allendorf 1999; Hampe and Petit 2005). Peripheral populations, however, tend to be small and relatively isolated (Mayr 1963), and they can be subjected to substantial founder effects and genetic drift. As a consequence, inbreeding may be higher in peripheral than in central populations, and drift may create genetic divergence among populations. Environmental suitability for a species normally decreases with distance from the center of its distribution (Lawton 1993), thus regimes of natural selection in marginal environments are likely different from central ones. Even the same direction and intensity of selection can result in genetic divergence among populations when populations have different genetic backgrounds as results of drift and inbreeding (Cohan and Hoffmann 1986; Hoffmann and Cohan 1987). Al l these factors lead to genetic divergence between central and peripheral populations; thus evolutionary potential can be different for central and peripheral populations, and among peripheral populations with differing degrees of isolation. Reduced population size will often result in reduced total reproduction for peripheral compared to central populations. Differences in reproduction among populations may result in asymmetric gene flow, with more alleles moving from higher to lower density populations than the reverse, resulting in a source-sink population structure (note that a sink population in this study is defined as a population in which the number of immigrants exceeds the number of emigrants). Gene 65 flow homogenizes gene frequencies among populations (Slatkin 1987). Source populations have a larger influence over sink populations than the reverse due to asymmetric gene flow, and counteract local adaptation in sink populations (Kawecki and Holt 2002). Thus, theoretical studies suggested that asymmetric gene flow due to central/peripheral population structure may counteract local selection in peripheral populations, resulting in less efficient selection and weaker adaptation to local conditions at the range periphery than in central populations (Haldane 1956; Holt and Gomulkiewicz 1997; Kirkpatrick and Barton 1997; Lenormand 2002). This effect of migration on fitness has been called "migration load." As a result, the amount of gene flow from central populations can limit the expansion of species' ranges (Kirkpatrick and Barton 1997; Wiens 2004). An allele with antagonistic effects in different environments can be maintained at migration-selection equilibrium. Selection must be strong enough to maintain adaptive divergence among populations or species (yet not too strong to threaten population persistence), or migration must be low enough not to counteract selection (Ronce and Kirkpatrick, 2001). Population genetic models have proposed that peripheral populations will develop phenotypes close to the optimum when isolated from other populations if populations are distributed across a linearly heterogeneous environment (Garcia-Ramos and Kirkpatrick 1997). If asymmetric gene flow counteracts local selection in a sink population, reduced gene flow may contribute to the efficiency of local selection. This will only happen when isolated populations can persist with limited gene flow and when there is enough variation within populations to respond to local selection. If isolated populations suffer genetic drift and inbreeding, then immigrant genes may be favored to increase population size and genetic diversity (Alleaume-Benharira et al.; Barton 2001). However, there is little empirical evidence for gene flow limiting a species' range. Sitka spruce, Picea sitchensis, is a conifer with a long, narrow distribution from south to northwest along the Pacific coast of North America. The distribution is restricted to coastal areas where maritime climates are relatively warm and moist compared to areas farther inland. The species' range extends from 39° to 61° N latitudes, and tapers in width towards the southern and northern edges, which we consider to be soft boundaries (Figure 4.1). This species has high genetic variation for both neutral genetic markers and quantitative traits within populations, typical of many wind-pollinated woody species, and also has relatively high differentiation among populations for a conifer (CHAPTER 2; CHAPTER 3; Hamrick et al. 1992). Our previous studies of Sitka spruce found a strong pattern of decreasing numbers of effective pollen donors, allelic richness, and outbreeding rates from central to peripheral populations, with particularly low estimates for these parameters for isolated peripheral compared to central populations (CHAPTER 2). The high genetic variation and strong central/peripheral population structure of Sitka spruce make it an appropriate 66 model species to explore the effects of gene flow at a species' borders. We observed the degree of local adaptation of populations to peripheral climates in Sitka spruce, using a growth chamber experiment simulating two peripheral environments (southern extreme and northern extreme) and one central environment. We compared the performance of isolated peripheral populations to the nearest continuous peripheral populations, to test the effects of gene flow on local adaptation of peripheral populations. Compared to continuous peripheral populations, isolated peripheral populations may have: 1) lower fitness to local climate due to lower population size resulting in higher inbreeding rates; or 2) higher fitness due to isolation from populations favouring alleles maladapted to the local peripheral climate. Materials and Methods We focused on differences between isolated and continuous populations at the northern and southern range peripheries in Sitka spruce. To do this, we selected one pair of isolated and continuous populations from each of three different geographic locations and determined the climatic conditions for these locations in order to simulate these environments in three growth chambers (Figure 4.1). Climate data were obtained from Daily Climate Normals 1971-2000 for Fort Bragg, California (CA; southern extreme environment), Sandspit, Queen Charlotte Islands, British Columbia (BC; central environment) and Kodiak Island, Alaska (AK; northern extreme environment). Temperature and photoperiod were averaged for every two week period from April 7 t h to December 10Ih, and those averages were used to set growth chamber conditions for two-week intervals. Daily temperature was ramped from daily normal minimum at midnight to daily normal maximum at noon. Seedlings were kept well watered and were fertilized regularly in all chambers. Seed collections from 17 populations spread across the entire range of Sitka spruce, from California to Alaska, were employed in this study (Figure 4.1). Al l seeds were collected from low elevations near the ocean to eliminate elevational effects on adaptive traits, and to avoid sampling populations with introgression from P. glauca (Bennuah et al. 2004). Seeds were soaked in water for 48 hours, stratified at 4°C for three weeks to break dormancy, and germinated on filter paper in Petri dishes. Germinants were transplanted into Super Leach Tubes® and grown in a greenhouse at the University of British Columbia, Vancouver, Canada. First-year seedlings were taken outside to experience chilling and bread bud dormancy in the 2002/2003 winter. Plants were then moved into three Conviron E15 growth chambers (Controlled Environments Inc.) on April 7, 2003. A complete randomized design was used. Each chamber contained a total of 490-491 seedlings from 17 populations with an average of 27.6 seedlings from at least 10 and up to 20 open-pollinated families per population. 67 Quantitative growth traits measured included 2" year height increment, date of bud break, date of bud set, growth period (days between bud break and bud set), mean daily growth rate (height increment divided by growing period), and total biomass at the end of the growing season. Climate variables at population origins were obtained from Monthly Climate Normals from 1971-2000 (National Climate Data Center of US and Environment Canada; Table 4.1). Data analysis Growth performance was analyzed separated for each chamber. Least squared population means were estimated using PROC G L M (SAS Institute Inc. 2000) with population as an independent variable. Geographic clines were tested using PROC R E G with population means. Previous research identify bud set timing before the date of first frost as an important trait for adaptation to seasonal cycles (reviewed in Howe et al. 2003) and a major determinant of summer growth periods for Sitka spruce (CHAPTER 3). In addition, Sitka spruce is less shade tolerant compared to one of its common competitors, western hemlock (Tsuga heterophylla), and prefers open canopy gaps for regeneration (Harris 1990). Taller seedlings will more effectively compete for light than shorter seedlings. Therefore, we defined 'better adapted individuals' as those individuals that set bud before the estimated first frost date in the simulated environment, and also grow taller relative to others. The first frost date was estimated from Climate Normals as the first date in the fall with a 50% probability of below 0°C temperatures. These dates were September 30 th in the northern peripheral environment (Kodiak Island and Kitoi Bay, A K , Western Regional Climate Centre of US), December 11th in the southern peripheral environment (Fort Bragg, C A , Western Regional Climate Centre of US), and November 3r d in the central environment (Sandspit, B C , B C Ministry of Forests and Range). All seedlings that did not start forming a terminal bud before the estimated first frost date were considered susceptible to cold injury and thus maladapted; the remaining individuals were considered seasonally adapted. We assumed that the surviving seasonally-adapted individuals would suffer density-dependent selection due to competition for light. Among those remaining individuals, the tallest in each environment were considered 'better adapted individuals.'At age 100 years (southern Oregon environment site productivity index = 48m, Alaska environment site productivity index = 33m, and estimated B C (53°N) site productivity index = 41m, Fair and Harris 1979), we estimated 100 year-old survivability to be 25% in the southern sites, 35% in the central sites and 40% in the northern sites based on the Sitka Spruce Natural Stand Density Management Diagram (Farnden 2001). Stand density management diagrams predict self-thinning relative to competition within a relatively monospecific stand based on site productivity and initial plant density. The initial seedling density at 68 establishment was assumed to be 2000 seedlings/ha (the natural regeneration density after clearcutting in Peterson et al. 1997). Taller individuals are more likely to survive competition for light and contribute to reproduction. Thus, the top 25%, 35% and 40%> tallest individuals among seasonally fit individuals were considered 'better adapted individuals' for the southern peripheral (CA), central (BC), and northern peripheral (AK) environments, respectively. We defined Juvenile Fitness Index as the proportion of 'better adapted individuals' in each population, indicating how many members of the population do well in each environment. To estimate the standard error of Juvenile Fitness Index, individuals were resampled by bootstrapping within populations within treatments 1000 times. For testing the effect of isolation at the range peripheries on the fitness index, two peripheral populations from each original peripheral chamber environment were analyzed as a combined data set. They were the southern isolated and continuous populations, Fort Bragg and Redwood, from the C A chamber, and the northern isolated and continuous population, Kodiak Island and Rocky Bay, from the A K chamber. Since replications of isolated and continuous populations were from different environments, the ratio of isolated to continuous populations' fitness index in each peripheral environment was tested to determinate if it was significantly greater than one using a t-test, with growth chamber as replication. The significance of the difference in fitness index between isolated and continuous populations was also tested binomially with individuals as replications. Results Variation and growth strategies in different environment Most quantitative traits showed a general clinal pattern of variation with distance along the coast (Table 4.2). Bud set timing exhibits strong clinal variation along the coast in all chamber environments in regression analysis (Table 4.2, Figure 4.2a). Generally, bud set timing was earlier in northern populations and later in southern populations, as well as earlier in the C A (southern) chamber environment and later in the B C (central) and the A K (northern) chamber environments. Variation among populations was responsible for 75.9 to 79.0% of all variation within chambers (Table 4.3). Bud break timing varied less among populations than bud set, and was highly variable within populations (Table 4.3). Bud break timing of all populations was generally earlier in the C A environment, intermediate in the B C environment and later in the A K environment, but there was no significant cline along the coast in C A and B C environments (Table 4.2, Figure 4.2b). Growth period was calculated as the number of days between bud break and bud set, and the geographic trends were similar to those for bud set timing. Growth period showed relatively strong among-population variation (Table 4.3). It generally increased from the southern to the northern environment in all populations (Table 4.2, Figure 4.3a). 69 There was a significant cline in 2-year height increment ('height') along the coast in the C A environment. This trend was significant but weaker in the A K environment and not significant in the B C environment (Table 4.2, Figure 4.2c). The majority of variation in height was within-population, except in the southern chamber environment (Table 4.3). Reaction norms in height showed strong population-by-environment interaction. The southern populations were shorter in the cooler environment (BC and AK) than in the C A environment, while the northern populations were taller in the cooler environments (BC and AK) than in the warmer (CA) environment (Figure 4.3b). The central populations were slightly shorter in their original (BC) environment than in the other environments. Biomass had relatively high among-population variation (25% to 37.2%; Table 4.3) with a significant cline along the coast in all chamber environments (Table 4.2, Figure 4.2d). The southern populations had slightly less biomass in the A K than in the C A chamber environment, while the northern populations seemed not to differ among chambers for biomass (Figure 4.3c). The southern and central populations showed reduced biomass in the B C chamber. The majority of variation for mean daily growth rate ('growth rate') was also within populations (Table 4.3). Growth rate showed a strong clinal response along the coast in the B C chamber and a weak cline (p = 0.07) in the A K chamber, but no trend in the C A chamber (Table 4.2, Figure 4.2e). Growth rates in southern populations were dramatically reduced in the B C and A K environments compared to the C A environment, while the northern populations showed less reduction in growth rate from the C A to A K chamber environments (Figure 4.3e). The central populations grew faster in the B C and A K chambers compared to other populations; however, again their growth rate was somewhat depressed in the B C chamber. Population fitness in continuous and isolated populations Juvenile Fitness Index showed clinal variation in the degree of adaptation to marginal environments (Figure 4.4). Populations originating closest to the simulated environment had higher fitness, and mean population fitness decreased with distance populations originated from that environment in the peripheral environments (CA: R2=0.77, p<0.001; AK: R2=0.85, p<0.001), but not in the central (BC) environment. For every 100km closer to the simulated environment a population originated, the number of better adapted individuals in a population increased 1.9% in the C A environment and 1.7% in the A K environment. In the southern peripheral environment (CA), all individuals finished their active growth within the estimated available growing season before the date of 50% probability of first frost. The isolated peripheral population (Fort Bragg) had the highest fitness (Juvenile Fitness Index) of all populations in the C A environment, and the fitness of the southern isolated peripheral population 70 (0.864) was higher than that of the nearest continuous peripheral populations (0.593) (Table 4.4, Figure 4.4a). In the central environments (BC), 9% of seedlings (all from the southernmost populations, FB and RW populations) did not meet the seasonal criteria and were actively growing on the estimated first frost date (Nov. 3). Fitness was not significantly different between central isolated and continuous populations that originated in this environment (Figure 4.4b). In the northern peripheral environment (AK), more than half of the individuals were still actively growing on the estimated first frost date (Sep.30), thus were considered likely to be killed or damaged by frost in that environment. Fitness was higher in the northern peripheral isolated population (0.500) than in the nearest continuous peripheral population (0.367) in the A K environment (Table 4.4, Figure 4.4c). For population mean height and bud set date for all individuals, the southern isolated population grew taller and later in the fall than the continuous populations in the southern (CA) environment (Table 4.4). The northern isolated population also grew taller, but it set bud slightly earlier in the fall than the continuous population in the A K environment. In a joint analysis of the northern and southern isolated and continuous populations, the ratio of the fitness index of isolated populations to continuous populations in the original peripheral environments was significantly greater than one (mean fitness ratio = 1.41, p=0.036, Table 4.4). A significant effect of isolation on mean height of all and 'better adapted individuals' was also detected (Table 4.4). A binomial test compared how many individuals met the fitness criteria (1) or not (0) in isolated and continuous populations with the combined data of the northern and southern peripheral populations. The proportion of "better adapted individuals" was 0.65 in isolated populations (sample size = 52) and 0.47 in continuous populations (sample size = 57). It also showed significantly higher fitness of isolated populations (p=0.026). Discussion Gene flow and adaptation in peripheral populations A previous study of mating system in the three pairs of isolated and continuous Sitka spruce populations studied here suggested a decreasing number of effective pollen donors toward the range peripheries (CHAPTER 2). The geographic species' range map also suggests a range narrowing toward the northern and southern margins (Figure 4.1), and anecdotal evidence indicates species density appears towards the decrease to peripheries as well (Gapare and Aitken 2005). This change in population size may cause asymmetric gene flow from central to peripheral populations due to density dependence (Kirkpatrick and Barton 1997). Gene flow can provide variation within populations, but it also can inhibit further adaptation as rate of gene flow increases. Garcia-Ramos and Kirkpatrick (1997) demonstrated that phenotypic value in isolated populations at range 71 peripheries should approach a local optimum when gene flow is limiting species ranges. We have shown higher juvenile fitness in isolated than in continuous populations in both northern and southern peripheral environments (Figure 4.4a, c). The fitness of isolated populations was significantly greater than continuous populations (Table 4.4). Besides frost kill, a higher survival rate is expected in the long term for established seedlings in northern than in southern environments (survival at 100 year predicted as 40% in the north and 25% in the south). This is due to slower tree growth resulting in weaker density-dependent competition effects in the north. Applying a harsher selection intensity in the north (25% survival by competition) showed similar results (data not shown). Considering that the continuous peripheral populations in this study have similar environments to nearby isolated populations (Table 4.1), isolated populations appear to have better performance than the nearby continuous populations in similar peripheral environments. Overall mean height and bud set date in peripheral populations reflect local selection (Table 4.4). The southern isolated population grew taller and later than continuous populations, which are considered favoured phenotypes in the southern environment. The northern isolated population had slightly taller individuals that set bud slightly earlier than those in the continuous population. This implies that local selection in the north favours individuals that are taller, but set bud earlier to avoid frost risk. When the immigration rate is low in an isolated population, population size tends to be small. Small populations are subject to both genetic drift, resulting in fixation of deleterious alleles (Lande 1994), and to inbreeding, resulting in inbreeding depression (Charlesworth and Charlesworth 1987). Our previous study (CHAPTER 2) found a lower number of effective pollen donors (N e p) and a higher selfing rate (S) in the isolated northern (Kodiak Island, A K ) (N e p= 1.7-2.2, S = 35%) and southern (Fort Bragg, CA) populations (N e p= 1.7-2.0, S = 15%) compared to continuous northern (Rocky Bay, AK) (N e p= 4.8-7.6, S = 10%) and southern (Redwood, CA) populations (N e p= 4.9-8.2, S = 8%, CHAPTER 2). These selfing rates in the isolated populations are considerably higher than previously estimated for Sitka spruce (S =8%, Chaisurisri et al. 1994). Higher or equal fitness in the isolated compared to continuous populations suggests more effective local selection in the isolated populations, and that either purging of recessive deleterious alleles or improved local adaptation outweighs inbreeding depression. Inbreeding depression is a common feature of outcrossing tree species and inbreeding in the isolated peripheral populations may have facilitated purging of recessive deleterious alleles. The northern isolated population is thought to have arrived on Kodiak Island only 300 to 600 years ago -approximately 3 to 5 generations. This estimation is based on tree ages and the observation that there were many living 300 year-old or older trees on the Island but no decayed fallen trees (Griggs 1937; J Alden, pers. comm. 2006). Empirical study has showed that purging can happen within a few 72 generations in a selfing population (Barrett and Charlesworth 1991). However, this leading edge population of the postglacial migration may have required initial gene flow to become established (CHAPTER 3), and purging by itself may not be able to increase fitness in this isolated population within so few generations with an intermediate selfing rate (35%; CHAPTER 2). Mixed mating systems (20 to 80% of outcrossing) often have evolved regardless of severe inbreeding depression (Goodwillie et al. 2005). Intermediate mixed mating system may ensure reproductive success and also could have both advantages of outcrossing that increases phenotypic variation and selfing that could purge deleterious alleles, which both enhance local selection as long as populations persist. Enhanced local selection by isolation and intermediate mixed mating may have increased the frequency of better adapted individuals within relatively a few generations. And thus it may mask inbreeding depression, if any occurs. As a result, inbred individuals with depressed fitness in this study may be at a low frequency so that they had little effect on the Fitness Index that only counts better performing individuals. Interestingly, Griggs (1937) briefly cited the findings of the US Forest Service, which states that the growth of the northern isolated populations on Kodiak Island was 'not less but actually somewhat greater than' the populations in southeastern Alaska. This, with the results in this study, supports positive effects of isolation on local adaptation that can be seen within a few generations in this species. There is similar empirical evidence in populations of great tits (Parus major) where asymmetric gene flow can drive the phenotypic value in sink populations close to that in source populations (Garant et al. 2005), and inhibit the phenotypic value from moving to the local optimum when there is divergent selection between populations (Postma and van Noordwijk 2005). In blue tits in Mediterranean, the low density populations were under the influence of the high density populations (Dias and Blondel 1996). Increased or similar fitness in isolated populations with relatively high inbreeding in this study suggests that the efficiency of selection increases with isolation. This results in an increased frequency of beneficial alleles which increase population fitness and counteract the negative effects of inbreeding. Migration can also increase fitness by masking deleterious alleles (Whitlock et al. 2000). Empirical research also suggest the positive effects of gene flow. For instance, the migration of parasites between different hosts increases host-parasite coevolution when migration is not high enough to homogenize parasite populations (Morgan et al. 2005). CHAPTER 3 also proposed the positive effects of gene flow on persistence of founder populations. It is initially important to increase genetic variation in founder populations. The consequences of isolation depend on selective gradients and migration rate between populations, population size and genetic variation within populations. 73 Adaptive divergence in growth traits Significant clines in growth traits along the Pacific coast of North America strongly suggest that populations have diverged as a result of adaptation to local environments. Clinal variation along geographic and climatic gradients is common in tree species (Morgenstern 1996). For example, bud set timing and growing period showed strong clinal relationships along the coast (Table 4.2, Figure 4.2a). Growth period may be maximized to fit the frost free period in a populations' own environments, especially in northern populations where frost hardiness is critical for growth and survival. Longer growing seasons are advantageous for achieving more height growth in warmer environments; however, this advantage is reduced or eliminated in colder environments due to frost damage resulting from delayed growth cessation. Differentiation among populations is thus driven by differential selection for growth rhythm to fit the available frost free period in their environment of origin and by competition among individuals for light at young ages. Adaptive divergence may not be expressed as clinal variation in phenotypic values along ecological gradients in a common garden experiment when there is a tradeoff in adaptive performances between different environments. When populations adapt to their local environment, they may lose their competitive advantage in other environments. For example, the reaction norms for height in tested environments were different among populations (Figure 4.3b). Southern populations had reduced height in the colder environment, while northern populations had greater height in the northern compared to the southern environment. As a result, the difference in height between southern and northern populations was smaller in the A K environment (Figure 4.2c). This suggests divergent selection for height performance, perhaps with a trade-off between height and adaptation of growth rhythm to seasonal cycles (Rehfeldt 1989). Growth rate had different cline along the coast among the chambers (Table 4.2). This suggests that growth rate responds to temperature, but with different threshold or optimal temperatures among populations (Chuine et al. 2001). Thus, although growth rate is generally lower in all populations from cooler environment, the reduction of growth from a warmer to a cooler environments was smaller in northern than in southern populations. Generally, southern populations had higher fitness than northern populations in southern environments, and vice versa (Figure 4.4). Such genotype-by-environment interaction strongly suggests local adaptation among populations. Reduced growth was observed in the B C chamber mostly for southern and central populations. In theory, the central populations should perform best of all populations in the central environment, and this result is somewhat difficult to explain. In a subsequent experiment with other species, the growth chamber used for the central environment has had lower average growth than expected for unknown reasons (unpublished data). Some malfunction in this chamber may explain the 74 reduced overall growth in this chamber, but does not explain why the central populations didn't do the best in the chamber. One possible reason is that the intermediate temperature and photoperiod in the central environment might encourage all populations to grow well in the controlled chamber and may have induced extra competition for light within the chamber. CHAPTER 3 showed moderate genetic differentiation for five microsatellite loci among populations of Sitka spruce (Fsf=0.11 and Rsf=0.09), higher than estimates for most outcrossing conifers (Hamrick et al. 1992). Population differentiation in quantitative traits was very high (mean Q S T for 6 traits = 0.56, CHAPTER 3). Population differentiation in quantitative traits and genotype-by-environment interaction indicates recent limited effective gene flow, likely caused by local selection against immigrant alleles. Our results support the theory that gene flow can restrict species' range by preventing local adaptation in peripheral populations to some extent. However, this theory remains somewhat controversial not only because isolated population often have small effective population sizes that have depressed fitness due to inbreeding, but also because there are other possible factors limiting further adaptation in peripheral populations, such as phylogeographic constraints, metapopulation dynamics (Holt et al. 2005) and interspecific interactions (Case and Taper 2000; Case et al. 2005). Further study on these factors may clarify the evolutionary trajectories and constraints at species borders. 75 Table 4.1. A comparison of pairs of continuous and isolated, central and peripheral populations. Populations were classified by location within the range and by isolation from the continuous distribution of the species. Climate variables for the population are MAT: mean annual temperature and MWMT: mean warmest month temperature. Population condition Original climate Key population name State Location Isolation M A T M W M T Rocky Bay (RB) A K Northern peripheral Continuous 4.11 12.72 Kodiak Island (KI) A K Northern peripheral Isolated 4.72 12.78 Queen Charlotte Islands (QC) B C Central Continuous 8.00 16.80 Ocean Falls (OF) BC Central Isolated 8.30 15.00 Redwood (RW) C A Southern peripheral Continuous 11.61 14.83 Fort Bragg (FB) C A Southern peripheral Isolated 11.83 14.67 Table 4.2. Regressions of population means for the observed quantitative traits on distance along the coast (in 100 km from north to south) C A (south) chamber B C (centre) chamber A K (north) chamber Traits slope P slope P slope R Bud set (days) 3.32 <0.01 2.11 <0.01 2.64 <0.0T Bud break (days) 0.12 0.15 0.09 0.32 0.25 0.05 Growth period (days) 3.20 <0.0T 2.02 <0.01 2.30 O.01 Height (mm) 2.90 <0.01 0.12 0.60 1.13 <0.01 Biomass (mg) 0.11 <0.0T 0.05 <0.01 0.07 <0.01 Daily growth rate (mm) -0.0003 0.84 -0.008 <0.01 -0.003 0.07 76 Table 4.3. Among- and within- population variation in growth and phenology traits in the three chamber environments. CA (south) chamber BC (centre) chamber AK (north) chamber Among Within Among Within Among Within populations populations populations populations populations populations Traits Variance % Variance % Variance % Variance % Variance % Variance % Bud set (days) 1540.7 79.0 410.4 21.0 832.8 78.5 228.5 21.5 1052.7 75.9 334.1 24.1 Bud break (days) 11 13.1 73.2 86.9 15.9 11.5 122.6 88.5 34.7 14.3 207.4 85.7 Growth period (days) 1404.7 75.9 445.2 24.1 678 67.7 323.4 32.3 781.6 57.4 580.1 42.6 Height (mm) 1198 44.0 1523.5 56.0 51.8 4.9 1015.3 95.1 202.3 11.0 1631.5 89.0 Biomass (mg) 1.67 37.2 2.82 62.8 0.44 25.0 1.32 75.0 0.79 36.6 1.37 63.4 Daily growth rate (mm) 0.0046 3.0 0.1463 97.0 0.0094 11.9 0.0693 88.1 0.0023 3.0 0.0746 97.0 Table 4.4. Adaptive performance in isolated and continuous peripheral populations (I: isolated, C: continuous population). Results are based on Juvenile Fitness Index (Figure 4.4). Better adapted individuals are individuals that meet the criteria for local adaptation to each environment. Population means in height and bud set date were calculated for all samples and better adapted individuals were identified. Bud set date was counted from when all plants were placed in the chambers (April 7, 2003). N: total number of samples, Na: number of individuals that meet the fitness criteria. Standard deviation for Juvenile Fitness Index was estimated from bootstrap resampling. Student t-test was performed on the ratios of the fitness index of an isolated population to a continuous population (I/C) from two peripheral environments, to test whether the ratios were greater than one. Peripheral Al l individuals Better adapted individuals Juvenile Fitness Index t-test for I/C chamber Population Isolation N Height Bud set (se) (se) Na Height Bud set (se) (se) Index SD (Na/N) (bootstrap) Ratio (I/C) H A : I/C > 1 (df=l) South (CA) Fort Bragg (FB) I 22 163.4 200.8 19 174.0 198.1 0.864 0.086 South (CA) Redwood (RW) C 27 (7.9) 128.8 (4.1) 184.4 16 (6.5) 151.6 (4.8) 183.7 0.593 0.126 1.457 North (AK) Kodiak Island (KI) I 30 (7.1) 82.8 (3.7) 148.1 15 (7.1) 108.3 (5.2) 152.3 0.500 0.141 North (AK) Rocky Bay (RB) C 30 (6.7) 73.1 (6.7) (3.5) 151.0 (3.5) 11 (6.5) 103.5 (8.6) (5.4) 152.4 (6.3) 0.367 0.150 1.364 t= 5.757p=0.036 OO Figure 4.1. Natural range map of Sitka spruce, Picea sitchensis showing locations of populations sampled. 130° 120° 110° Figure 4.2. Population means in selected growth response variables by location along the coast. The city of Prince Rupert, B C is considered the geographic centre (0) of Sitka spruce's range. The chambers indicate the temperature and photoperiod conditions tested in the three growth chambers, a) Date of bud set and b) date of bud break were counted from April 7, 2003, when the growth chamber experiment was started, c) Height increment in the 2 n d year. First year height was subtracted from the final height at the end of the 2 n d year growth, d) Total biomass after the 2 n d year of growth, e) Daily growth rate: 2 n d year height increment divided by growth period. CA (south) chamber BC (centre) chamber AK (north) chamber .5> X (0 E o S TO E E C D 250 200 150 100 50 50 40 30 20 10 160 140 120 100 80 60 40 20 24 23 22 21 20 0 9 0 8 0.7 0.6 0.5 ' X i £ - I t r r . V* % ^ 1 I I - I I • 1 I - I I • , I I • l r hz* • 1 T - l t I r I 1 £ 1 I '. fi / I II1 v r 1 ; ! i i n . i - 2 - 1 0 1 2 - 2 - 1 0 1 2 - 2 - 1 0 1 2 Distance (1000 km) south north 80 Figure 4.3. Reaction norms of populations across growth chamber environments for central and peripheral continuous and isolated populations (see Table 4.1). a) Growth period 250 200 CB T J "g 150 100 so CA (south) • Southern Isolated (FB) o Southern Continuous (RW) • Central Isolated (QC) O Central Continuous (OF) A Northern Isolated (KI) A Northern Continuous (RW) BC (centre) Chamber AK (north) b) Height CA (south) BC (centre) AK (north) Chamber I Growth rate CA (south) BC (centre) Chamber AK (north) CA (south) BC (centre) AK (north) Chamber 81 Figure 4.4. Juvenile Fitness Index for populations. The index was estimated as the proportion of T j e t t e r adapted' individuals in a population that met criteria for timing of bud set and height. Error bars are standard deviation from the bootstrap datasets. Populations are arranged from south to north, from left to right. Shaded bars indicates populations local to the environment simulated in the chamber. Dark grey bars are isolated local populations, and light grey bars are continuous local populations. a) Southern periperal (CA) chamber (25% tallest trees that set bud before the estimated first frost date of Dec. 11) 1 0 9 08 x % 0.7 3 OS a> £ 0 5 E * 0.4 | 0.3 ~* 02 0.1 0 | Isolated local I Continuous local n XL b) FB, RW, CR, VA. VI, OF, QC, P R 111, 167, 71, 95, IB, VL, Ml, RB, KI, CA CA OR BC BC BC BC BC AK AK AK AK AK AX AK AK AK Population Central (BC) chamber (35% tallest trees that set bud before the estimated first frost date of Nov. 3} FB, RW, CR, VA, VI, OF, QC, PR, 111, 167, 71, 95, IB, M_, Ml. RB, KI, CA C A OR BC BC BC BC BC AK AK AK AK AK AK AK AK AK population c) Northern peripheral (AK) chamber (40% tallest trees that set bud before the estimated first frost date of Sep, 30) 0 9 0 8 0.7 0.6 0.5 0 4 0.3 02 0.1 • isolated local • Continuous local n ra m FB, RW, CR, VA, VI, OF, OC, PR, 111, 167. 71, 95, IB, VL, Ml. RB. KI, C A C A O R B C B C B C B C B C A K A K A K A K A K A K A K A K A K Population 82 Literature Cited Alleaume-Benharira, M . , LR. Pen and O. Ronce (in press). Geographical patterns of adaptation within a species' range: interactions between drift and gene flow. Journal of Evolutionary Biology. Barrett, S.C.H. and D. Charlesworth (1991). Effects of change in the level of inbreeding on the genetic load. Nature 352: 522-524. Barton, N.H. (2001). Adaptation at the edge of a species' range In J. Antonovics. Integrating Ecology and Evolution in a Spatial Context. Blackwell, Oxford. Bennuah, S.Y., T.L. Wang and S.N. Aitken (2004). Genetic analysis of the Picea sitchensis x glauca introgression zone in British Columbia. Forest Ecology and Management 197: 65-77. Case, T.J., R.D. Holt, M . A . McPeek and T.H. Keitt (2005). The community context of species' borders: ecological and evolutionary perspectives. Oikos 108: 28-46. Case, T.J. and M . L . Taper (2000). Interspecific competition, gene flow, environmental gradients, and the coevolution of species borders. American Naturalist 155: 583-605. Chaisurisri, K., J.B. Mitton and Y.A. El-Kassaby (1994). Variation in the mating system of Sitka spruce (Picea sitchensis) - evidence for partial assortative mating. American Journal of Botany 81: 1410-1415. Charlesworth, D. and B. Charlesworth (1987). Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18: 237-268. Chuine, I., S.N. Aitken and C C . Ying (2001). Temperature thresholds of shoot elongation in provenances ofPinus contorta. Canadian Journal of Forest Research 31: 1444-1455. Cohan, E M . and A.A. Hoffmann (1986). Genetic divergence under uniform selection. 2. different responses to selection for knockdown resistance to ethanol among Drosophila melanogaster populations and their replicate lines. Genetics 114: 145-164. Dias, P.C and J. Blondel (1996). Local specialization and maladaptation in the Mediterranean blue tit (Parus caeruleus). Oecologia 107: 79-86. Farnden, C. (2001). Timber Production: developing commercial thinning regimes. Stand Density Management Diagrams. Forest Practices Branch, B.C. Ministry of Forestry, Victoria, B.C. Fair, W.A. and A.S. Harris (1979). Site Index of Sitka spruce along the Pacific coast related to latitude and temperatures. Forest Science 25. Gapare, W.J. and S.N. Aitken (2005). Strong spatial genetic structure in peripheral but not core populations of Sitka spruce (Picea sitchensis (Bong.) Carr.). Molecular Ecology 14: 2659-2667. Garant, D., L .E .B . Kruuk, T.A. Wilkin, R.H. McCleery and B.C. Sheldon (2005). Evolution driven by differential dispersal within a wild bird population. Nature 433: 60-65. Garcia-Ramos, G. and M . Kirkpatrick (1997). Genetic models of adaptation and gene flow in peripheral populations. Evolution 51: 21-28. Goodwillie, C , S. Kalisz and C G . Eckert (2005). The evolutionary enigma of mixed mating systems in plants: occurrence, theoretical explanations, and empirical evidence. Ann. Rev. Ecol. Syst. 36: 47-79. Griggs, R.F. (1937). Timberlines as indicators of climate trends. Science 85: 251-255. Haldane, J.B. (1956). The relation between density regulation and natural selection. Proceedings of the Royal Society of London Series B-Biological Sciences 145: 306-308. Hampe A. and R.J. Petit (2005) Conserving biodiversity under climate change: the rear edge matters. Ecology Letter 5: 461-467. Hamrick, J.L., M.J.W. Godt and S.L. Sherman-Broyles (1992). Factors influencing levels of genetic diversity in woody plant species In. New Forests. Kluwer Academic Publishers. Harris, A.S. (1990). Picea sitchensis (Bong.) Carr. Sitka spruce In B.H. Honkala. Silvices of North America Volume 1. Conifers. Washington, DC. Hoffmann, A .A . and F.M. Cohan (1987). Genetic divergence under uniform selection. 3. Selection for knockdown resistance to ethanol in Drosophila pseudoobscura populations and their replicate lines. Heredity 58: 425-433. Holt, R.D. and R. Gomulkiewicz (1997). How does immigration influence local adaptation? A 83 reexamination of a familiar paradigm. American Naturalist 149: 563-572. Holt, R.D., T.H. Keitt, M.A. Lewis, B.A. Maurer and M . L . Taper (2005). Theoretical models of species' borders: single species approaches. Oikos 108: 18-27. Howe, G.T., S.N. Aitken, D.B. Neale, K.D. Jermstad, N.C. Wheeler and T.H.H. Chen (2003). From genotype to phenotype: unraveling the complexities of cold adaptation in forest trees. Canadian Journal of Botany-Revue Canadienne De Botanique 81: 1247-1266. SAS Institute Inc. 2000 SAS OnlineDocR, Version 8, Cary, NC: SAS Institute Inc. Slatkin, M . (1987). Gene flow and the geographic structure of natural populations. Science 236: 787-792. Kawecki, T.J. and R.D. Holt (2002). Evolutionary consequences of asymmetric dispersal rates. American Naturalist 160: 333-347. Kirkpatrick, M . and N.H. Barton (1997). Evolution of a species' range. American Naturalist 150: 1-23. Lamrni, A., P. Siikamaki and K. Mustajarvi (1999). Genetic diversity, population size, and fitness in central and peripheral populations of a rare plant Lychnis viscaria. Conservation Biology 13: 1069-1078. Lande, R. (1994). Risk of population extinction from fixation of deleterious mutations. Evolution 48: 1460-1469. Lawton, J.H. (1993). Range, population abundance and conservation. Trends in Ecology & Evolution 8: 409-413. Lenormand, T. (2002). Gene flow and the limits to natural selection. Trends in Ecology & Evolution 17: 183-189. Lesica, P. and F.W. Allendorf (1995). When are peripheral populations valuable for conservation? Conservation Biology 9: 753-760. Mayr, E. (1963). Animal Species and Evolution. Cambridge, Harvard University Press. Morgan, A.D., S. Gandon and A. Buckling (2005). The effect of migration on local adaptation in a coevolving host-parasite system. Nature 437: 253-256. Morgenstern, E.K. (1996). Geographic Variation in Forest Trees. U B C Press, Vancouver. Peterson, E.B., N . M . Peterson, G.F. Weetman and P.J. Martin (1997). Ecology and Management of Sitka spruce, Emphasizing its natural range in British Columbia. U B C Press, Vancouver . Postma, E . and A.J. van Noordwijk (2005). Gene flow maintains a large genetic difference in clutch size at a small spatial scale. Nature 433: 65-68. Rehfeldt, G.E. (1989). Ecological adaptations in douglas-fir (Pseudotsuga menziesii var glauca): a synthesis. Forest Ecology and Management 28: 203-215. Ronce, O. and M . Kirkpatrick (2001). When sources become sinks: Migrational meltdown in heterogeneous habitats. Evolution 55: 1520-1531. Whitlock, M . C , P K . Ingvarsson and T. Hatfield (2000). Local drift load and the heterosis of interconnected populations. Heredity 84: 452-457. Wiens, J.J. (2004). Speciation and ecology revisited: Phylogenetic niche conservatism and the origin of species. Evolution 58: 193-197. 84 CHAPTER 5 Thesis Conclusions for Dynamics of species' range Introduction The primary objective of this thesis was to investigate some of the factors involving determination of species' range in a tree species, which generally has high genetic variation and strong local adaptation despite a long life span. There are several major factors that may limit species expansion beyond species' borders: 1) selection, 2) reduced genetic diversity, 3) negative correlations between critical traits for adaptation, and 4) dispersal ability. However, as evolutionary consequences, beneficial mutations should eventually rise in marginal populations to overcome these effects. An alternative explanation is the homogenizing effect of gene flow. Local selection will be counteracted by gene flow when selection pressures vary among populations. Plant densities are generally higher at the centre of the species range and lower at the margins. This may generate asymmetric gene flow from central to marginal populations. When gene flow is asymmetric, selection at source populations is favored, and peripheral populations are under the influence of source populations. Thus, selection is less efficient at the range peripheries due to immigrant alleles from the centre of the range where the strength and direction of selection is different from in marginal environments. There are a few reports that provide evidence of the effects of gene flow in preventing local adaptation. However, the effect of gene flow for limiting species' range is controversial, mainly because species' range is not always limited by gene flow, but often by hard selection such as at oceanic borders. Gene flow can likely only counteract selection on soft borders where plant density gradually decreases in a relatively linearly heterogeneous environment. In this thesis, I focus on soft boundaries of Sitka spruce (Picea sitchensis) to investigate the effects of population size and gene flow on local adaptation in two extreme environments (northern and southern borders). I employed nuclear microsatellite markers to observe pollen genetic diversity and mating system in six populations. The degree of local adaptation of seventeen populations from across the range was measured using a field common garden experiment and a controlled growth chamber experiment. The levels of gene flow and local adaptation were observed and comparisons made between central and peripheral populations as well as between continuous and disjunct 8 5 populations. Major findings Tradeoffs between adaptation and gene flow on species' borders I found strong central-peripheral population structure in the number of effective pollen donors, an indication of larger effective population size in central populations and smaller population size in peripheral populations. A corresponding pattern was seen in inbreeding, with higher selfing rates at the range peripheries than in central populations. The disjunct, isolated populations showed high selfing rates (15% in the south and 35% in the north). While the central continuous and disjunct populations had similar numbers of effective pollen donors, the disjunct peripheral populations exhibited significantly lower population sizes than the continuous peripheral populations, suggesting strong effects of isolation at the range peripheries. This strong difference in population size between central and peripheral populations can cause asymmetric gene flow from central to peripheral populations. Despite decreasing population size toward the range peripheries, phenotypic patterns in growth traits were generally clinal along the north-south coast, which was strongly correlated with a gradient in mean annual temperatures. The timing of bud set had high among-population variation (high Q S T ) and significantly increased toward the southern populations. Bud set timing was the major factor (96 %) determining the plant primary growth period in this experiment and is considered a very important adaptive trait for young Sitka spruce to fit the local seasonal cycle and avoid early fall frost injury. Southern and northern populations have different adaptive strategies that reflect their own native environments. Trees from southern populations were generally much taller than trees from northern populations, while trees from northern populations generally grew faster per day but formed terminal buds earlier than southern populations. For young Sitka spruce, competition for light is critical to establish stands. Northern populations grow faster to achieve height quickly during the available favourable growth period but terminate active growth early to increase cold hardiness to early fall frosts. Therefore, I defined "better adapted seedlings" as individuals that grow taller (better competitor) yet terminate active growth (set bud) before the first frost date in the simulated environments in the growth chambers. Three growth chambers simulated the temperatures and photoperiod (based on the 1971-2000 Climate Normals) of the southern peripheral (CA), central (BC), and northern peripheral (AK) environments. In both the C A and A K environments, populations originating closer to the simulated environment generally had higher juvenile fitness and fitness decreased with increasing distance of population origin to the simulated environment. The peripheral isolated populations had 86 higher or the same fitness as peripheral continuous populations, despite high selfing rates in isolated populations. This supports the hypothesis that selection may take place more effectively in isolated populations at the range peripheries that receive limited gene flow from other environments outside of the population than in continuous populations. When a species is distributed across a relatively linearly heterogeneous environment and the density gradually decreases toward the margins (soft boundaries), gene flow may counteract the efficiency of selection and become a crucial factor to limit a species' range. In this case, isolation from other populations adapted to different environments will lead to better local adaptation. The isolated populations must be self-sustaining to achieve higher fitness. Smaller threshold population sizes that severely depress population fitness may lead isolated populations to different fates such as extirpation rather than adaptation. Determination of the relative effects of gene flow and drift is crucial to predict the fates of small peripheral populations. Response to environmental changes in tree species Wind-pollinated tree species generally show very high genetic variation within populations, and are considered to have substantial long-distance dispersal. Considering the longevity of Sitka spruce, high gene flow among populations may explain the past rapid migration in the Quaternary. I observed relatively high genetic distances among populations using neutral markers ( R S T = 0.09 and F S T = 0.11). Genetic distances estimated from quantitative traits were much higher (average Q S T = 0.56), which indicates that differences among populations in these quantitative growth traits are clinal and result from differential selection. Maintenance of adaptive divergence depends on relative degrees of selection and gene flow among populations. A positive correlation between genetic distance ( R S T ) and geographic distance (r = 0.73) suggests that Sitka spruce populations are currently at migration-drift equilibrium and following a stepping stone migration model on microsatellite loci. This suggests that gene flow from neighbouring populations may counteract the effects of long-distance dispersal required for rapid postglacial migration. A strong positive correlation between genetic and geographic distance, and a strong pattern of adaptive divergence among populations along the climate gradients, combined with pollen records of rapid migration, suggest an ability of Sitka spruce to migrate and adapt to climate changes that require a shift of the entire species' range. In addition to the mobility of wind-dispersal pollen and winged seeds, fast population growth have made rapid migration into ice-free areas possible. The migration likely consisted of several steps. First, long-distance dispersal creates a few founder populations in open areas. Then these founder populations may have received high gene flow from neighboring populations, sufficient to achieve relatively high genetic variation among populations. Habitat across 87 the current range of Sitka spruce is relatively heterogeneous in terms of climate. Thus populations may have alleles that are somewhat pre-adapted to their neighbouring population's environments. A combination of pre-adapted alleles, recombination of genes, and natural selection in such populations may have shaped the population mean phenotype close to the local optimum faster than in populations with low gene flow and low genetic variation. Tree species have a high capacity to respond to selection and to sustain founder populations in harsh environments due to their longevity and fecundity, which potentially increase effective population size and adaptive genetic variation. After selection associated with tolerance of biotic stress and competitive ability, the populations may develop a structure that fits the stepping stone migration model. This may be seen only when environmental gradients are steep enough. Prediction of species' range shift requires careful observation of species' genetic variation, dispersal ability, and environmental gradients affecting selection intensity. Future research In this thesis, I have demonstrated the effects of gene flow on local adaptation and migration. Dispersal is essential for range expansion; however, further expansion will be inhibited when the margins are too far away from the centre of species' range due to habitat quality and density dependence. Evolutionary and ecological dynamics of peripheral populations in response to selection, drift and gene flow can lead us to understand processes of adaptation, speciation and coevolution. Comparative studies between core and peripheral populations provide gradients of selection and gene flow across the range to conduct comprehensive experiments. Asymmetric gene flow may allow us to identify which populations are eventually favored by selection regardless of current fitness. Directions and degrees of asymmetric gene flow can be predicted by density. Decreasing density toward species margins can be a relevant indicator for identifying where the effects of gene flow might counteract selection. Effective population sizes in Sitka spruce decreased toward the range peripheries, but not all species follow this so-called centre abundant distribution. Species in heterogeneous environments may have multiple peaks in density across the range. In such species, the effects of gene flow may be weaker or stronger than selection at the margins, depending on selective gradients, distance between margins and closest area of abundance. It would be interesting to determine how gene flow and selection counteract each other in the various selective gradients. In metapopulation structure, the likelihood of extinction and recolonization increases, and the 88 effects of asymmetric gene flow and drift should become stronger than selection as the rate of extinction/recolonization increases at range peripheries. Comparative studies among sufficient numbers of peripheral subpopulations in metapopulations could provide insights into the effects of drift and gene flow on adaptation. Existence of other species may increase at range peripheries when species compete for similar resources. Competition depresses a species' density. Interspecific interactions can be a strong factor limiting species' range, not only as an agent causing biotic selection but also as an agent to create asymmetric gene flow. Studies of interaction among species are very difficult to conduct mainly because of lack of knowledge of factors involved in the interactions. Analyzing numbers of species sharing range peripheries and species' distribution patterns of sympatric and allopatric species may provide the general patterns in the effects of interspecific interaction on species' range. Because peripheral populations are at the margins of species' niches, and may be vulnerable to environmental changes, small populations at peripheries may be good targets for conservation. Classic views of conservation genetics suggest that connections among populations are important to increase genetic variation increases the sustainability of small populations. However, this study uindicates isolation actually can increase the efficiency of selection in natural tree populations. The outcomes depend on the amount of genetic variation within populations and the steepness of environmental gradients. Fundamental observations, not only assessing genetic diversity but also fitness in isolated conditions, will be necessary for both short-term and long-term persistence of populations. Hampe and Petit (2005) hypothesized that species' leading edge and tailing edge might have different evolutionary consequences in response to climate changes. Tailing edge, southernmost populations (in the Northern Hemisphere), are more likely to be fragmented and extirpated than the northernmost populations. Comparative studies between leading and tailing edge populations in the species' range may show the nature of adaptation in response to changing environments. Evolutionary and ecological dynamics at range peripheries can provide various insights of evolution and ecology of species and prediction of the species and ecosystem in a changing world. Such knowledge will contribute to effort- and cost- effective conservation strategies. Literature Cited Hampe A. and R.J. Petit (2005) Conserving biodiversity under climate change: the rear edge matters. Ecology Letter 5: 461-467. 89 

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.831.1-0075047/manifest

Comment

Related Items