UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A hierarchical analysis of historical processes and phylogeographic patterns in Salvelinus (Pisces: Salmonidae) Elz, Anna E. 2003

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

Item Metadata

Download

Media
831-ubc_2003-0596.pdf [ 6.41MB ]
Metadata
JSON: 831-1.0091184.json
JSON-LD: 831-1.0091184-ld.json
RDF/XML (Pretty): 831-1.0091184-rdf.xml
RDF/JSON: 831-1.0091184-rdf.json
Turtle: 831-1.0091184-turtle.txt
N-Triples: 831-1.0091184-rdf-ntriples.txt
Original Record: 831-1.0091184-source.json
Full Text
831-1.0091184-fulltext.txt
Citation
831-1.0091184.ris

Full Text

A hierarchical analysis of historical processes and phylogeographic patterns in Salvelinus (Pisces: Salmonidae) by Anna E . Elz B.Sc, Long Island University, Southampton, 1993  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF T H E REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Zoology)  We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH C O L U M B I A September 2003 © Anna E . Elz, 2003  U B C Rare Books and Special Collections - Thesis Authorisation Form  Page 1 of 1  In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e requirements f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by t h e head o f my department o r by h i s o r her r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d ' t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d without my w r i t t e n p e r m i s s i o n .  Department o f The U n i v e r s i t y o f B r i t i s h Columbia Vancouver, Canada  Date  5tykJkM  2O0J  http://vv^vw.library.ubc.ca/spcoll/thesauth.html  9/29/03  ABSTRACT  Hybrid zones have been studied extensively over the last two decades, but relatively little attention has focused on the historical processes generating hybrid zones. Phylogeography, a subdiscipline of biogeography, evaluates patterns of genetic variation to infer the processes that have shaped the geographic and demographic histories of species and can provide insight into the historical contingencies that facilitated hybridization. Phylogenetic relationships of char (Pisces: Salvelinus) have suggested that hybridization has been a recurring theme throughout their evolutionary history. In this thesis, I used sequence variation in the mitochondrial D N A (mtDNA) control region to perform a hierarchical phylogeographic study that examined intraspecific, interspecific, and trans-species polymorphisms in Salvelinus. I developed a diagnostic assay that identifies bull trout (S. confluentus) mtDNA clades and revealed the presence of both lineages in two coastal watersheds in British Columbia. Headwater stream capture following the last glaciation is suspected to have facilitated this "double invasion". A nested clade analysis (NCA) was performed to evaluate if a widespread hybrid zone between Dolly Varden (S. malma) and bull trout in northwestern North America resulted from two processes: secondary contact of previously isolated lineages and continuous contact resulting from historical introgression in a shared refuge. Despite ambiguity in the statistical parsimony network, N C A provided evidence of secondary contact between Beringian Dolly Varden and bull trout and introgressed (with bull trout mtDNA) Dolly Varden. In addition, I found a degree of phylogeographic substructure amongst bull trout and introgressed Dolly Varden in the Chehalis Refuge, which suggests that introgression has been geographically localized. In a species-level phylogenetic analysis, the "approximately-unbiased test" was used to compare alternate tree topologies and I found that Arctic char lineages were not monophyletic in relation to Beringian Dolly Varden and bull trout. The consensus parsimony tree and the phylogeographic distribution of lineages suggested that both bull trout and Dolly Varden may have experienced historical contact with different Arctic char lineages. I suggest that hybridization has played a significant evolutionary role in the diversification of char. This work presents the most extensive phylogeographic analysis of Salvelinus to date and provides a comparative framework for both small scale and regional studies in northwestern North America.  11  T A B L E OF CONTENTS  Abstract  ii  Table of Contents  iii  List of Tables  v  List of Figures  vi  Acknowledgments  viii  Dedication  ix  Chapter 1: General Introduction  1  Chapter 2: A n analysis of the geographic distribution of divergent mitochondrial D N A lineages in bull trout (Salvelinus confluentus) Introduction  8 9  Material and Methods  11  Results  13  Discussion  14  Chapter 3: Testing phylogeographic hypotheses using nested clade analysis: Origins of a widespread hybrid zone between Dolly Varden (Salvelinus malma) and bull trout (5. confluentus)  29  Introduction  30  Material and Methods  34  Results  41  Discussion  47  Chapter 4: A phylogeographic synthesis of the Salvelinus alpinus-malma species complex Introduction  75 76  )  iii  Materials and Methods  78  Results  80  Discussion......./.  80  Chapter 5: General Discussion  95  References  104  Appendix 3.1: Sample localities and number codes for Figure 3.1  118  Appendix 3.2: Nested clade geographical distance analysis  121  Appendix 3.3: Templeton's (2001) inference key  123  Appendix 3.4: Calculation for time since population expansion  126  iv  LIST O F T A B L E S Table 2.1: Variable sites in mtDNA region between bull tout clades 27 Table 2.2: Sample sites and bull trout clade identification 28 Table 3.1: Mean percent sequence divergence between Salvelinus mtDNA haplotypes 72 Table 3.2: Measures of intrapopulation variability 72 Table 3.3: Summary of geographical associations and inference chain for nested geographical distance analysis 73 Table 3.4: Parameters estimated by the mismatch distribution for bull trout and Dolly Varden lineages 74 Table 3.5: Estimates of the time since expansion for char bull trout and Dolly Varden lineages 74  Table 4.1: Statistical tests of significance for alternative Salvelinus mtDNA phylogenetic tree topologies 94  Table 4.2: Net nucleotide substitutions per site between Salvelinus lineages 94  v  LIST O F FIGURES Figure 1.1: Bull trout (S. confluentus) and Dolly Varden (S. malma) mosaic hybrid zone 7 Figure 2.1: Phylogenetic tree of bull trout mtDNA sequences 23 Figure 2.2: Geographic distribution of bull trout mtDNA lineages 24 Figure 2.3: Map of study area on the southcoast of B.C. 25 Figure 2.4: Distribution of interior and coastal bull trout in three coastal drainages illustrating where stream capture may have occurred 26 Figure 3.1: Sample codes and geographic distribution of bull trout and Dolly Varden mtDNA clades throughout study area 62-63 Figure 3.2: Sequence alignment for the variable site positions defining 28 bull trout and Dolly Varden haplotypes 64 Figure 3.3: Maximum likelihood phylogram of bull trout and Dolly Varden mtDNA sequences 65  ;  Figure 3.4: Nested cladograms (Till) inferred from the 95% parsimony network 66-67  Figure 3.5: Transitions and transversion differentiating haplotypes in the parsimony network illustrating the loops of ambiguity 68 Figure 3.6: Geographic distribution of clade 4-1 in B.C. and adjacent portions of the western U.S. \  6  9  Figure 3.7: Geographic distribution of clade 3.1 in B.C. and adjacent portions of the western U.S. 69  vi  Figure 3.8: Geographic distribution of the DB lineage in B.C. and adjacent portion of the western U.S. 70 Figure 3.9: Geographic distribution of Dolly Varden 3-step clades throughout Russia and North America 70 Figure 3.10: Frequency distribution of the number of pairwise differences for bull trout and Dolly Varden lineages 71 Figure 4.1: Holarctic distribution of phylogeographic lineages in Salvelinus 91 Figure 4.2: Alternative constrained mtDNA tree topologies for Salvelinus 92 Figure 4.3: Maximum parsimony consensus tree for mtDNA control region 93  vii  ACKNOWLEDGEMENTS I knew when I first visited my supervisor, Dr. Eric Taylor, that his lab was the right place to pursue a Master's degree. The enthusiasm and diversity of research made for an ideal environment to learn a new discipline. I am grateful for the experience he afforded me to present my research at the Evolution 2001 conference. Without his encouragement, patience, and financial support, this work could not have been completed. My supervisory committee, Sally Otto, Dolph Schluter, Eric Taylor and Jeanette Whitton provided many thoughtful comments during committee meetings where we finetuned my proposal and research direction. I am grateful that their doors were always open when I had questions. The guidance and suggestions they provided during the writing of this thesis were immensely helpful. Those that have shared my toils most while providing both intellectual and emotional support throughout this process are a wonderful group of labmates and friends: Emily Rubidge, Ladan Mehranvar, Patrick Tamkee, Allan Costello, Ramona DeGraaf, Jason Ladell, Mike Stamford, and Steve Latham. Thanks for keeping me laughing. Being surrounded by such avid "fish-folks" was a blessing when the going got tough. A very special thanks goes to Don McPhail. He often guided me on a virtual field season when I had questions about rivers, terrain, and fish distributions. His firsthand knowledge of char and arctic fish in general was an oft-visited source of information that was much appreciated. I would also like to acknowledge and thank my family and friends that always believed that I could do this, even when I didn't, and encouraged me to pursue a longstanding goal. Finally, I received tissue samples from various sources. Sewall Young at the Washington Dept. of Fish and Wildlife and Bill Mavros at Dept. of Natural Resources, King County, W A each provided many samples that were necessary to make this project work. Vladimir Brykov (Institute of Marine Biology, Far Eastern Division of the Russian Academy of Science) generously sent a variety of Dolly Varden samples from throughout their East Asian range. Jim Reist (DFO, Winnipeg) and Neil Mochnacz from the University of Winnipeg kindly shared bull trout samples from the Northwest Territories.  viii  This work is dedicated in memory of my mom, whose own pursuit of knowledge was a reminder that there are many ways to learn, but the best way is to teach yourself.  ix  Chapter 1:  General Introduction  1  The pursuit to understand how historical, demographic, and behavioural factors have shaped the distribution of biological diversity has been influenced by the advent of phylogeography (Avise 1994; Avise 2000). Using analyses of mitochondrial D N A (mtDNA) and chloroplast D N A (cpDNA) variation, this discipline was initially devised to examine the evolutionary processes responsible for partitioning intraspecific variation (Wilson and Herbert 1998; Alves et al. 2001; Givet and Petit 2002), most notably in studies of human evolution (Caan et al. 1987; Cavalli-Sforza et al. 1994). Comparative phylogeography, an extension of this concept, has been utilized to test biogeographic hypotheses by assessing the geographic patterns of genetic variation in codistributed species (e.g. Avise 1992; Zink 1996). A n accumulation of phylogeographic data detailing broad patterns between two or more species has identified both the geographical and temporal features promoting population differentiation that can potentially lead to speciation (reviewed in Avise 2000). For example, regional assessments of genetic variation in southeastern and northwestern North America have attributed the similar phylogeographic patterns in unrelated taxa to the influence of shared historical forces (Avise 1992; Brunsfeld et al. 2002). A comparison of phylogeographic patterns between regions has also identified broad trends such as lower levels of intraspecific divergence and genetic diversity in fish taxa from glaciated versus unglaciated areas (Bernatchez and Wilson 1998). The genetic structure and ecology both of terrestrial and aquatic north temperate and arctic fauna are intricately linked to the Pleistocene glaciations. The Pleistocene was marked by numerous climatic oscillations over the past 2.4 million years. For approximately the first 1.4 million years, glacial-interglacial cycles occurred every 41,000 years (41- kyr), followed by a 100-kyr frequency (Hewitt 2000) with the most recent of the Wisconsinan glacial advances in northwestern North America beginning about 25-kyra and ending 10-12 kya when the ice retreated (Clague 1991). Species affected by each climatic fluctuation were forced to adjust their ranges by tracking suitable habitat. During cold periods, they were forced into glacial refugia 2  where they went extinct or adapted to refugial environments. As the ice retreated, populations colonized newly available habitat sometimes leaving a genetic "footprint" of population growth or decline (e.g. Costello et al. 2003). Evidence from many north temperate species has accumulated that supports this model of persistence and colonization as the process responsible for secondary contact between intraspecific and interspecific lineages where hybridization and introgression often occurs (Hewitt 2000). The Pleisotcene glaciations severely impacted the distributions of north temperate freshwater fish. Their persistence in refugia and subsequent dispersal opportunities were dictated by meltwater drainage connections that allowed them to expand their ranges considerably. Hence, geological history and the genetic structure of aquatic species are closely intertwined making them ideal for phylogeographic studies. The often isolated and dendritic nature of aquatic habitats also implies that the formation of hybrid zones will be dependent on historical contingencies, such as watershed exchange, that bring previously allopatric taxa into contact (Taylor 2003). This patchwork of connectivity and habitat heterogeneity within a watershed usually results in a mosaic hybrid zone pattern defined by localized areas of sympatry with possibly unique ecological and historical components. Elucidating the temporal and geographical juxtaposition of such hybrid zone formation is a key goal of biogeographic analysis and can contribute to understanding the process of reproductive isolation (Rieseberg 1998). Mitochondrial D N A is usually the marker of choice for phylogeographic studies because, for the most part, it does not recombine and experiences rapid evolution due to a smaller effective population size compared to nuclear loci. The process of lineage sorting is therefore expected to proceed more rapidly in the mitochondrial genome. Notwithstanding recent advances in nuclear gene phylogeography (Hare 2001) and inherent limitations of mtDNA analyses (e.g. Avise 1994; Maddison 1997; Ballard and Whitlock 2003), mtDNA variation within and between closely related species has led to many novel insights into the biogeographic history of species and areas 3  (reviewed in Avise 1994; Avise 2000). For example, Taberlet et al.'s (1997) comparison of the distribution of mtDNA lineages in European taxa identified concordant refugial histories and multiple areas where secondary contact has occurred between intraspecifc lineages. Study system Char, fishes of the salmonid genus Salvelinus, have intrigued fish biologists for over the past century. Within the Salmonidae, they are arguably the most morphologically and behaviourally diverse group of fish (Behnke 1972). Reconstructing the evolutionary history of Salvelinus has thus proven difficult and resulted in much taxonomic confusion, making them an alluring subject for evolutionary studies. Part of the problem stems from localized studies on either side of the Pacific that has prevented comparisons of potential conspecifics throughout their range. Hybridization is also believed to have obscured the evolutionary relationships within the genus (Behnke 1980; Phillips et al. 1999; Westrich et al. 2002) Benhke (1980, 1984) recognized three major subgroups Salvelinus. S. (subgenus Cristovmer) namaycush and S. (subgenus Baione) fontinalis, are endemic to North America. The Arctic char complex makes up the subgenus Salvelinus and, includes S. alpinus (Arctic char), S. malma (Dolly Varden), S. leucomaenis (Japanese white-spotted char) and S. confluentus (bull trout). This group has a vast Holarctic distribution with some forms constrained to either side of the Pacific Ocean. S. leucomaenis is found only in eastern Asia, while S. confluentus, originally thought to be Dolly Varden (e.g. McPhail 1961), are distributed throughout the northwestern North America from the Northwest Territories, Canada, to Nevada, USA. Within the S. alpinus-malma complex, Dolly Varden are recognized as consisting of three subspecies. The northern form, S. m. malma, is found in eastern Russia and north of the Alaskan Peninsula east to the Mackenzie River in North America, whereas the southern form, 5. m. lordi, occurs south of the Alaskan Peninsula to Washington State. A southern Asian form, S. m. krashenikovi, has a limited range south of the Amur River on mainland Russia to Sakhalin 4  Island. Arctic char have a circumpolar, arctic distribution including Europe, arctic Canada, Alaska, and Russia and have many described subspecies that do not necessarily match with presumed phylogeographic lineages (Wilson et al. 1996; Brunner et al. 2001). A n alternative view, suggests that the forms comprising the S. alpinus-malma complex are not discrete taxonomic units, but rather a single super species complex that has diverged very recently (Savvaitova 1980; 1995). Taxonomic assignments based on morphology have been revisited with the implementation of molecular tools. Thus far, numerous genetic studies have attempted to resolve the relationships within Salvelinus using a molecular phylogenetic approach. There has been some success in identifying sister groups, such as S. malma and S. alpinus, but the evolutionary history of S. confluentus remains muddled. Even though hybridization is thought to be responsible for much of the confusion surrounding char, little work has been done exploring where areas of secondary contact may have occurred. Although char phylogeography and systematics has been extensively studied, there have been no comparative phylogeographic studies that have examined the formation of hybrid zones across the full geographic range of component species and few attempts to reconstruct the phylogenetic relationships of all putative taxa across their range. The extensive background information on hybridization between Dolly Varden and bull trout (e.g. Redenbach and Taylor 2002; Redenbach and Taylor 2003) and the existing mtDNA data for many species of Salvelinus (e.g. Grewe et al. 1990; Brunner et al. 2001) provides a strong base to explore these uncertainties.  Study Objectives M y thesis presents a hierarchical phylogeographical analysis of Salvelinus. First, I begin with an intraspecific study of bull trout (S. confluentus) in northwestern North America. A previous study (Taylor et al. 1999) identified two phylogenetic mitochondrial D N A lineages 5  resulting from isolation in separate refugia. I developed a diagnostic assay that distinguishes each clade and extended the geographic coverage of bull trout throughout their range to identify localities where both lineages occur. Second, patches of hybridization have been shown to occur between bull trout and Dolly Varden throughout a zone of parapatry (Fig. 1.1). The phylogeographic structure of the two species and evidence of historical introgression suggest that some parts of the hybrid zone are the result of secondary contact while other areas represent areas of continuous contact (Redenbach and Taylor 2002). I wanted to test this hypothesis by performing a comparative phylogeographic study using nested clade analysis. Third, numerous phylogenetic studies have focused on the systematics within the genus Salvelinus. Often, however, a single geographical sample is used and all members of the complex are usually not included (especially bull trout), which provides only limited information about the evolutionary history of such a diverse group of fish. Conflicting phylogenetic patterns across loci within and between species are often interpreted to be the product of introgressive hybridization (Arnold 1997), but given the limited geographical coverage of these studies, the spatial and temporal juxtaposition of these encounters are unknown. To address such questions, a previous study of the Arctic char phylogeography by Brunner et al. (2001) was used as a template onto which I incorporated data for bull trout and Dolly Varden from throughout much of their range. A geographic assessment of the phylogenetic patterns amongst lineages was used to infer potential areas where contact and hybridization occurred throughout the species' history. Finally, I present potential research directions for studies of char hybrid zones and discuss how hybridization may have been a potential evolutionary force in the diversification of char.  6  Figure 1.1. The parapatric distribution of Dolly Varden and bull trout showing areas of localized sympatry (white circles) and hybridization (hatched circles). Figure adapted from Baxter et al. (1997) and Taylor et al. 2001.  7  Chapter 2:  An analysis of the geographic distribution of divergent mitochondrial DNA lineages in bull trout (Salvelinus confluentus)  c  Introduction  The Pleistocene glaciations had a profound effect on the genetic structure and ecology of north temperate species in North America (Pielou 1991; Hewitt 2000). Phylogeographic studies of many diverse taxa in northwestern North America have shown congruent patterns of intraspecific diversification resulting from isolation in separate refugia (Soltis et al. 1997; Stone and Cook 2000; Demboski and Cook 2001). As the glaciers retreated, differentiated forms colonized the ice-free areas and sometimes met in areas of secondary contact. Identifying where r  contact zones exist and how they formed is an integral component to understanding evolutionary processes promoting diversification. The genetic structure of aquatic species expanding into newly available habitat following the last (Wisconsinan) glaciation was especially influenced because of their restriction to freshwater dispersal routes. Meltwater from retreating glaciers severely impacted the hydrological networks in* previously glaciated areas of north temperate North America (Armstrong, 1981). The resulting shifts in connectivity via stream capture, in which a tributary from one watershed joins with a tributary from another, and or proglacial lakes allowed fish to extend their ranges great distances. These processes have created an intimate relationship between geological history and the distribution of genetic diversity in many species of freshwater fish (Bernatchez and Dodson 1991; Wilson and Herbert 1998; McCusker et al. 2000). While concordant phylogeographic patterns across many species can often delineate refugial origins and major dispersal corridors, anomalous patterns of genetic variation in deglaciated areas may also reveal cryptic colonization routes that are currently unavailable. Intraspecific phenotypic and behavioural variation is also a common trend in north temperate fish. Variation in life history traits can be a product of isolation in distinct refugia (Teel 2000), divergence via natural selection (Schluter 1996; Rundle et al. 2000), or possibly 9  both (Travisano et al. 1995; Taylor and McPhail 2000). Areas of secondary contact between divergent lineages are often associated with increased levels of diversity, niche partitioning, and differences in life history traits (such as stream resident vs. anadromy), which may, in part, explain the high occurrence of sympatric ecomorphotypes in fish (reviewed in Taylor 1999). In addition, historical contingency can be responsible for shaping interactions with other species and can indirectly contribute to phenotypic shifts in life histories. For instance, a predationmeditated selective environment resulting from historical hydrological connections was strongly correlated with phenotypic divergence in the Utah chub (Gila atraria) (Johnson 2002). Identifying whether these differences arise due to historical or ecological factors are relevant for understanding the roles of isolation and /or natural selection in promoting divergence. Bull trout (Salvelinus confluentus) are distributed across northwestern North America and exhibit morphological (Haas and McPhail 2001), behavioural (i.e. fluvial, adfluvial, anadromous) (Haas and McPhail 2001), and genetic (Costello et al. 2003) variation amongst populations throughout their range. To assess their evolutionary history, Taylor et al. (1999) surveyed mtDNA polymorphism in bull trout populations throughout much of their range in northwestern North America. They concluded that bull trout contain two distinct evolutionary lineages ("coastal" and "interior") that stem from isolation in separate Wisconsinan glacial refugia, but some gaps in the distribution of these lineages remain and they detected no areas of sympatry. While many of the coastal clade bull trout populations are thought to be sea-run and those of the interior clade largely consist of freshwater-resident populations, there is still not a clear understanding of how refugial history influenced the distributions of these life-history types because the life histories of many of these supposed sea-run populations have not been confirmed. Interestingly, in British Columbia (BC), interior clade bull trout were found in large coastal flowing rivers such as the Klinaklini, Skeena, Stikine, and Nass rivers (Fig. 1.1). Given 10  the proximity of these samples to interdigitating, headwater tributaries, Taylor et al. (1999) proposed that the interior clade bull trout likely entered these systems as a result of faunal exchanges with interior drainages during deglaciation rather than colonization via the sea. Many major coastal systems, however, were not included in their analysis and many of these systems could contain both bull trout lineages given their potential interconnectedness with interior drainages and proximity to rivers bearing only coastal haplotypes. To date, however, only two very large systems, the Fraser and Columbia rivers, contain both lineages. Identifying other areas of "double invasion" and describing the extent that each clade dispersed postglacially (especially along the coast) and the routes they employed is crucial to understanding the roles of historical and ecological factors in governing phenotypic and behavioral diversity in this enigmatic species. The objectives of this study were threefold. First, I wanted to develop a diagnostic restriction fragment polymorphism assay that could easily identify each of the mtDNA clades to avoid sequencing large numbers of individuals. Second, I extended the sample coverage from that in Taylor et al. (1999) to more clearly define the extent and routes of postglacial colonization. Third, I discuss how historical contingency versus ecological determinism may be influencing the occurrence of different life history traits in bull trout.  Materials and Methods  Samples Initially, 17 bull trout D N A samples from Taylor et al. (1999) that comprised two clades based on phylogenetic analysis were used to test for any diagnostic differences in the mithochondrial D N A control region. Additional tissue samples were received opportunistically from Washington State, British Columbia, and the Northwest Territories (Table 1.1) via different sources (private consultants, WDFW, DFO) to describe further the geographic distribution of 11  bull trout clades throughout their range. Sample sizes ranged from 1 to 19 individuals per location. Sampling methods are unknown and likely varied as tissue samples ranged in size from small adipose clips to larger portions of the caudal or pelvic fins. To verify that samples were actually bull trout, those individuals identified as "char" by field consultants were additionally scored for two PCR-based, species diagnostic markers as in Taylor et al. (2001) to confirm that the samples were not Dolly Varden or hybrids. DNA extraction D N A was extracted from approximately 20 mg of fin tissue by overnight digestion at 55°C using the G E N T R A Purgene Cell Lysis solution with 7ul of lmg/mL Proteinase K solution followed by protein removal with Purgene Protein Precipitation solution. D N A was precipitated in isopropanol, washed in 70% ethanol, and the pellet was left to dry overnight. The pellet was resuspended in 50-100ul TE buffer (pH 8.0) depending on the size of the pellet and stored at 20°C. Sequencing, phylogenetic, and RFLP analysis PCR amplification of the entire mtDNA control region (-1000 base pairs, bp) was performed as outlined in Bernatchez et al. (1992). Purification of 150ul (3 X 50ul per sample) PCR products was carried out using Quiagen's Quiaquick PCR Purification kit. Sequences were generated by automated sequencing on an A B I 377 following termination of PCR reactions with Applied Biosystems, BigDye ver. 3 and Centri-Sep purification. A l l sequencing reactions were performed from the 5' end using the Tpro2 primer (Bernatchez et al. 1992). Complimentary sequences were not obtained. Multiple D N A sequences were aligned using C L U S T A L in P C G E N E (Bairoch 1991). Phylogenetic analysis was performed to test for resolution of the two previously described mtDNA clades using the PHYL1P package (Felsenstein 1995). The sequences were analyzed with the distance tree-building program, NEIGFfBOR. The data were subjected to 1000 bootstrap replicates using SEQBOOT and C O N S E N S E constructed the 12  majority rule consensus tree with Salvelinus namaycush chosen as the outgroup. RESRICT in P C G E N E (Higginsl992) was used to identify restriction enzymes that produced diagnostic restriction fragment length polymorphisms (RFLP's) between clades. To test the clade specific RFLP's, restriction digests were performed according to the manufacturer's instructions (GIBCO), except that I used 7ul of PCR product in a total volume of 20ul. The restriction fragments were visualized using 2 % agarose gels stained with ethidium bromide. A l l newly acquired samples were diagnosed with two different restriction enzymes.  Results mtDNA phylogenetic and diagnostic sequence variation A total of -552 bp were available for phylogenetic analysis. Ten positions were variable in the 17 individuals from Taylor et al. (1999) (Table 2.1) that produced seven haplotypes (5 coastal and 2 interior) and resolved the two previously described clades with 73% support for the interior clade and 49% support for the coastal clade (Fig. 2.1). Variation in the mtDNA control region also identified two independent diagnostic sites. Sna&l cut the coastal clade mtDNA at position 188 (G : A transversion), and A v r l l cut the interior clade mtDNA at position 232 (C : T transition), each producing fragment sizes of approximately 800 bp and 200 bp. Clade distribution A total of 23 geographic localities were assayed for bull trout clade identification (Table 2.2). Only the two southcoast B C sites, the Homathko and Southgate rivers, contained both the interior and coastal bull trout clades (Fig. 2.2). Nine lower Fraser River localities spanning approximately 200 km between the mouth of the Fraser River and Birkenhead Lake west of the divide were found to contain only coastal bull trout. Seven localities (numbered 15-21 in Fig. 2.2), all bearing interior clade mtDNA, sampled in Washington State considerably extend Taylor et al.'s (1999) coverage of the mid-upper Columbia region. The remaining three Washington 13  State samples west of the continental divide contained coastal clade bull trout. Both samples from the Northwest Territories (NWT) consisted of interior clade bull trout.  Discussion Geographical distribution of mtDNA clades Comparing Taylor et al's. (1999) findings to my results revealed some new insights into the geographic distribution of bull trout lineages. First, localities 22 and 23 (Fig. 2.2) represent a range extension of the interior bull trout clade into the NWT. Previously, all char in the Mackenzie River basin were thought to be Dolly Varden (Salvelinus malma) or Arctic char (Salvelinus alpinus), but it has recently been shown using morphology that bull trout do indeed exist in the N W T (Reist et al. 2002). The inclusion of these samples in my survey supports the hypothesis that interior bull trout from the Columbia Refuge utilized interior drainages to colonize the Peace/ Liard River system (Lindsey and McPhail 1986; McPhail and Lindsey 1986; Baxter et al. 1997; Taylor et al. 1999). Alternatively, these fish may have colonized northern drainages from the proposed Nahanni Refuge (Haas and McPhail 2001), but sequencing revealed that they shared the most common haplotype with other Columbian clade bull trout (see Chapt. 3). Only one individual was sampled at each of the localities in the NWT, and therefore further sampling is required to more accurately describe levels of genetic variation for the NWT. The distribution of bull trout clades in Washington State (localities 12-21 in Fig. 2.2) and the Fraser River Valley support the hypothesis of expansion from two glacial refugia, the Columbia and Chehalis River valleys (Taylor et al. 1999). First, bull trout east of the Cascade Crest comprise the interior clade while the coastal clade occurs west of this drainage divide. M y additional samples from the unglaciated portion of the Columbia River extend Taylor et al's. (1999) coverage for central and southeast Washington State and are important for examining colonization routes from the source refugial population. M y sampling did not, however, allow a 14  more detailed examination of the area encompassing the lower Columbia River. It appears that much of the Lower Columbia has been colonized by coastal bull trout (Taylor et al. 1999) and given that some are amphidromous (i.e., making short forays to the sea), there exists the possibility that gene flow has occurred or continues to occur between the Chehalis and Columbian refugial groups. Furthermore, evidence of deep lineage splits in the coastal clade suggests the presence of further substructure in the lower Columbia. (Taylor et al. 1999, also see Chapter 3). Consequently, further sampling of watersheds in these areas is required to better understand how historical and/or ecological factors have promoted intraclade divergence. Headwater transfer and double invasion As is often the case in phylogeographic studies, extending sample coverage can bolster previous findings and also uncover new colonization patterns. M y results supported the hypothesis that some localities have been colonized by both bull trout lineages representing "double invasions" of bull trout, specifically, the Homathko and Southgate rivers on the southcoast of B C (Figs. 2.3, 2.4). Both of these rivers could have been colonized by either dispersal through the sea or via headwater capture. The general lack of interior lineages in downstream areas of other coastal rivers (e.g. Fraser, Columbia, Squamish) makes it unlikely that interior clade fish dispersed through the sea. Rather headwater exchanges are a more likely explanation. Both of these rivers flow west to the Pacific Ocean and discharge into Bute Inlet (Fig. 2.4). Their headwaters begin east of the Coast Mountains near the headwaters of the Chilcotin River system, which drains east into the Fraser River (interior plateau). The presence of interior bull trout mtDNA in both of these rivers, therefore, probably reflects an historical drainage connection rather than a current one. Temporary drainage connections such as glacial lakes and headwater capture were common during deglaciation and allowed fish to expand their ranges into previously inaccessible areas (Hocutt and Wiley 1986). Consequently, historical hydrological connections sometimes created the opportunity for isolated groups to meet in an 15  area of sympatry (Bernatchez and Dodson 1990; Hurwood and Hughes 1998; Waters and Wallis 2000). Given the limited geological data for the southcoast region, I can only hypothesize about the possible scenarios that could have resulted in these headwater connections. There is no evidence that a glacial lake persisted in this area, but this should not be discounted. Instead, it appears most likely that through headwater capture, in which a tributary from one watershed joins with a tributary from another, interior clade bull trout migrated into these coastal flowing rivers. Two scenarios could account for this phenomenon: flow reversals and joining by flooding. First, flow reversals can result from recent or ongoing geological uplift, promoting the formation of new drainage divides. For example, Waters et al. (2001) found evidence for drainage evolution based on phylogeographic structure of New Zealand's freshwater galaxid fishes. The Coast Mountains, however, formed during the late Miocene to the Pliocene, and the present east-west flow pattern of its rivers is identical to when the rivers were first formed (Lindsey and McPhail 1986; Montgomery and Greenberg 2000). Second, during deglaciation, ice dams could reverse flow patterns if the main outflow is blocked and the tributaries rise in areas that became ice-free first as was the case with the Skeena River and its two major tributaries (McPhail and Lindsey 1986). If the same was true for the tributaries of the Southgate and Homathko rivers, one would expect to find coastal bull trout mtDNA in interior drainages due to an eastern flow pattern, which I did not find. Because, however, this seems to be a possible example of headwater exchange, further sampling should be concentrated in the headwaters of the adjoining interior drainages to better test this hypothesis. Deglaciation of the Coast Mountains likely occurred by downcasting, whereby the upland areas become deglaciated first followed by the adjacent valleys, which maintained some residual dead ice (Fulton 1967; Clague 1981). The last stage, the dead ice phase, consisted of stagnant ice confined to valleys. It seems most likely that meltwater from decaying glacier ice in the 16  interior flowed west to the Pacific and facilitated headwater capture by flooding many of the interdigitating tributaries east of the coastal divide (Fig. 1.4). This is especially true of the Homathko River since its headwaters begin at the western edge of the plateau (pers. comm. J.J. Clague, Geology, Simon Fraser Univ.). Headwater taxa are potentially highly susceptible to stream capture (Waters 1994) and numerous faunal exchanges by headwater capture has also been suggested to occur in areas north of this region (Skeena, Stikine, Nass rivers) as evidenced by the distribution of other Columbian freshwater fauna in coastal rivers (Nelson 1968; McPhail and Lindsey 1986; Baxter et al. 1997). Again, these data reinforce how phylogeographic patterns can reveal cryptic colonization routes and test geological hypotheses. Two alternative scenarios could have facilitated the spread of interior clade fish in these coastal rivers, namely multiple individual exchanges across the divide, or a single colonization followed by spread to other coastal rivers. The Klinaklini River was possibly penetrated via Tatla Lake tributaries. Similarly, the Homathko River may have been colonized by interior clade fish from Chilko River / Choelquoit Lake via Tatlayoko Lake. The Southgate River likely received fauna from Chilko Lake, but since it also shares its inlet drainage with the Homathko River, fish may have spread from the Homathko River into the Southgate River. Likewise, interior bull trout that reached the Klinaklini River may have dispersed into Homathko River via Tatayoko Lake. It is not possible to discern which of these scenarios is more likely with my data, but these alternatives could be tested using microsatellites to assess population relatedness amongst the different watersheds. I found no evidence of coastal bull trout in interior drainages near the headwaters of the Southgate and Homathko rivers, but my sampling was not extensive in these watersheds. Similarly, Taylor et al. (1999) only looked at five individuals from the Chilko River and found only interior haplotypes. One explanation for the apparent lack of coastal mtDNA clades in interior drainages is that coastal bull trout, being amphidromous, had not accessed the 17  headwaters during flooding. Secondary contact of Columbian and Coastal Refuge rainbow trout (Onchorynchus mykiss), however, was found to occur in the Chilko River (McCusker et al. 2000) indicating that exchange in both directions may have been possible. Potential limitations to dispersal Despite extensive sampling (Fig. 2.2, Table 2.2) no interior clade fish have been identified in the lower-middle Fraser River. This result is interesting because "double invasions" of the lower Fraser River have occurred both in longnose dace (Rhinichthys cataractae) (McPhail 1997) and longnose suckers (Catostomus catostomus) (McPhail and Taylor 1999). Hell's Gate Canyon (Fig. 2.2 B) in the Fraser River is an effective barrier to upstream migration of coastal clade fish, but probably does not prevent downstream dispersal of interior clade fish, which suggests that perhaps competitive exclusion by established populations in the lower reaches limits downstream dispersal of interior clade fish as McPhail and Lindsey (1986) suggested for the lower Columbia River. What ecological or temporal processes, therefore, allow the interior and coastal lineages to co-exist in the Homathko and Southgate rivers? Alternatively, introgression of coastal mtDNA into interior clade fish could explain why no interior clade bull trout have been found in the lower Fraser and Columbia rivers. Bute Inlet drains the Southgate and Homatko rivers both of each which contains both bull trout lineages and may represent a biogeographic boundary for both the southern extent of interior fish experiencing headwater transfer and the northward postglacial dispersal of coastal bull trout. The Squamish and Toba rivers, which only contain coastal haplotypes, do not extend up into the continental divide and, therefore, likely did not experience headwater transfer (Figure 2.4). It is more difficult to explain the apparent absence of coastal bull trout in the Klinaklini River and more northerly localities as well as their curious absence from Vancouver Island. Notwithstanding limited sampling, there are many possible explanations for the limited spread of bull trout along the coast (up to Homathko River). First, fish dispersing from the Chehalis 18  Refuge likely utilized two routes; glacial lake connections through the Puget Sound lowlands to the lower Fraser (Thorson 1980; McPhail and Lindsey 1986) and /or near-shore marine areas (Cavender 1997). Because coastal haplotypes have not been found in any interior drainages it is unlikely that Chehalis bull trout colonized coastal rivers through the interior via headwater exchanges. Therefore, if Chehalis bull trout were indeed amphidromous and colonized coastal rivers via the sea, as did many salmon species dispersing from southern refugia (Taylor et al. 1996; Smith et al. 2001), one possible explanation for their limited northward dispersal is selection based on temperature. Several instances of mtDNA "capture" indicate that unidirectional introgression in species with different ecological requirements or geographical distributions offers a selective advantage to the introgressed population, usually in the form of adaptation to temperature differences. For example, in both cyprinids (Dowling et al. 1997) and char (Glemet et al. 1998), the species with the more southern distribution carrying northern mtDNA may have been better equipped for survival at colder temperatures allowing them to expand their ranges. Similarly, under the selectionist hypothesis (Ballard and Kreitman 1995; Rand 2001), it can be argued that coastal bull trout mtDNA evolved under thermal constraints (Rand 1994), which has restricted their geographic distribution. If this was the case, it is expected that genes with functional constraints would evolve at different rates than neutral genes and lineage-specific differences in the patterns of silent, synonymous, and nonsynonymous substitutions should exist. Given that the environmental conditions between Bute and Knight inlets are probably not that different in terms of temperature, a more plausible explanation is that amphidromous bull trout may not be able to navigate deep channels. No bull trout have been found on Vancouver Island either, making this a reasonable explanation. In addition, the physiographic features of the entrance to Knight Inlet may hinder movement further north. While a long distance dispersal event from the Chehalis Refuge to the Homathko River must have occurred historically, it is not 19  known if bull trout presently undergo extensive migrations at sea. Based on limited evidence that suggests anadromy in bull trout is poorly developed, in the sense that they make only short forays out to sea (amphidromy) (Haas and McPhail 1991), it seems that short migrations between rivers would be more probable, similar to a stepping stone model of migration (Kimura and Weiss 1964). Radio-tracking data or genetic tests of isolation-by-distance would be required to test this hypothesis. In the Puget Sound, W A region, sea-run populations are known to exist, but the extent of migration between drainages is uncertain (pers. comm. Bill Morrow, King County Dept. of Natural Resources). Furthermore, the Fraser and Squamish rivers' bull trout are suspected of anadromy based on their larger size (McPhail and Taylor 1995), but additional work is required to characterize the distribution and dispersal patterns of sea-run bull trout along the coast in B.C. Multiple invasions, secondary contact, and parallel evolution Much evidence has accumulated on the potential impact that multiple invasions have had in creating patterns of pheno/ecotypic variation across taxa (reviewed in Mayr 1963). Svardson (1961) described several instances of phenotypic and behavioral differences between sympatric populations of northwestern European fish possibly resulting from multiple invasions. Phylogenetic evidence (Taylor and McPhail 2000) provides some support for McPhail's (1993) 'double invasion' hypothesis regarding the formation of sympatric 'benthic' and 'limnetic' threespine stickleback species pairs (Gasterosteus aculeatus) in B C lakes. One of the best-known cases involves Darwin's finches, which independently colonized the Galapagos Islands and in the process established sympatric communities (Petren et al. 1999). Similarly, Shaw's (2002) nDNA phylogeny of the cricket genus Laupala substantiates the hypothesis of allopatric divergence followed by multiple invasions resulting in sympatric communities on many Hawaiian Islands.  20  Morphological and behavioral (i.e. amphidromy vs. adfluvial) differences between bull trout lineages exist throughout their range (Cavender 1997; Haas and McPhail 2001), suggesting that vicariance and subsequent isolation in separate refugia may be only partially responsible for the observed variation. A persistent evolutionary dilemma concerning such polymorphisms revolves around whether the two forms represent secondary contact of two distinct lineages throughout their range or perhaps are the result of parallel evolution. For example in different species of whitefish (Coregonus artedi) (Turgeon and Bernatchez 2001) and (C. clupeaformis) (Lu et al. 2001), sympatric ecotypes (dwarf and normal) may have evolved from distinct or the same phylogenetic lineage. The factors that contribute to the ecological and morphological variation in bull trout are not well known (but see Haas and McPhail 2001), and therefore sympatry of the two lineages presents the possibility for examining interesting ecological and genetic interactions. Given that we know little about the extent and origin of amphidromy in bull trout, one might hypothesize that this life history type is genetically fixed for the coastal clade and interior clade fish are restricted to freshwater. For example, Pigeon et al. (1997) found that dwarf and normal whitefish (C. clupeaformis) from Cliff Lake were fixed for Acadian and Atlantic mtDNA respectively. Alternatively, if allopatric bull trout in the lower reaches of the Klinaklini River and other northcoast rivers, which were likely colonized by the interior clade via headwater transfer (discussed above), are indeed amphidromous then they may have evolved this life history trait independently of the coastal clade. Parallel evolution of adaptive ecotypes has been reported extensively in fish exploiting specific trophic niches such as whitefish (Bernatchez and Dodson 1990; Bernatchez et al. 1996), sticklebacks (Gasterosteus aculeatus) (McPhail 1993), and smelt (Osmerus mordax) (Taylor and Bentzen 1993). Likewise, Taylor et al. (1996) generated a phylogeny based on allele frequencies from mtDNA and minisatellites found that sockeye (sea-run) and kokannee (resident) salmon (Onchorynchus nerka) grouped more by 21  geographical proximity than by life history type. In addition, Arctic char, (Salvelinus alpinus), display remarkable trophic polymorphism, resulting in much taxonomic confusion (Behnke 1980; Savvaitova 1980). Elucidating the contributions of historical processes and replicate evolution to the life history polymorphism exhibited by bull trout is pertinent to the study of how natural selection acts in shaping phenotypic variation. In summary, I have developed a diagnostic procedure to identify bull trout clades throughout their range and to explore further their postglacial history. The observation of interior clade bull trout in two southcoast rivers in B C raised hypotheses about hydrological connections that likely impacted the distribution of other aquatic fauna and should be tested further since biological and geological data are both necessary for establishing drainage evolution. Elucidating the role vicariance has played in shaping bull trout's life history variations will require additional field studies that track the movements of fish within and between coastal rivers. Finally, my data represent the first finding of sympatric bull trout lineages and allow the opportunity to study their intraspecific interactions for comparison to allopatric populations in other coastal systems. Because assigning conservation priorities requires knowledge of both phylogenetic and life history variation, these sympatric populations create a unique opportunity to further our understanding of how history and ecological factors have contributed to the diversity present in bull trout.  22  -BEAR 73  WIGWAM BT-I  HOTEL ANDERSON KLINAKLINI SQUAMISH ELWAH CHILLIWACK  BT-C  METOLIUS 49  — PUYALLUP PITT S. namay.  Figure 2.1. Phylogenetic tree of bull trout mitochondrial D N A based on neighbor-joining analysis of Kimura 2-parameter distance inferred from 552 base pairs of control region. Bootstrap support out of 1000 re-samplings is indicated for the two mitochondrial D N A clades (interior and coastal). Sequences are as per Table 2.2. S. namaycush was designated as an outgroup in consensus tree building.  23  A  B  Figure 2.2. A and B (inset) showing the distribution of coastal (black) and interior (stippled) clade bull trout defined by mtDNA. Numbered ellipses refer to study sites in Table 1.1 and other locations are from Taylor et al. 1999. Shaded squares are areas of hypothesized secondary contact.  24  Figure 2.3. Map of the study area on the southcoast of British Columbia, Canada showing the broad scale main coastal rivers in relation to the coastal mountain crest (dashed line) and interior drainages. Inset is shown in Figure 2.4.  25  SO 1  o 1  1  BOWtonrtM  =•  A  Figure 2.4. The distribution of interior (stippled) and coastal (black) bull trout mtDNA lineages observed in three coastal drainages and the interdigitating lake/stream networks where stream capture events may have occurred.  26  cn o H * m O  U*  *  * U.U U U U  <! o o *  *  o  cn  in  * * * * *  CN  cn y  <! *  *  *  *  CN CN  cn CN  * * * * *  ^ * * * * *  oo OJ  *  00  *  # *  *  *  *  # *  uuuu u  * o o o o o  OJ  >  CN 00  <J O * * * * * * # * *  ON  O < <  oo <5 o *  * *  CN  ho •<! *  rn  H  * * *  u o  CS  B  O  * *  # *  *  #  O * * # * * * * * H  M  e c4 es  < < < < o  &  ^  =a *-> ^ .-a  N  r-H  . Pi Pi 9 00  c3 > > .g  §1 CO  u i  H pq  CN  i-H  Table 2.2. Sample sites, number code (see Fig. 1.1), and diagnostic clade identification (interior or coastal) for bull trout mtDNA at each locality.  Sample  Drainage  Interior  Coastal  South Coast, B . C . South Coast, B . C . lower Fraser River lower Fraser River lower Fraser River lower Fraser River lower Fraser River lower Fraser River lower Fraser River lower Fraser River Fraser River  7 7 0 0 0 0 0 0 0 0 0  1 9 3 16 1 1 1 19 1 9  W. Hood Canal Snohomish Lewis Methow Methow Methow Wenatchee Yakima Tucannon Walla Walla  0 0 0 10 10 10 5 5 10 10  7 5 8 0 0 0 0 0 0 0  Mackenzie R. Mackenzie R.  1 1  0 0  coastal B C  Homathko R. (1) Southgate R. (2) Birkenhead R. (3) Owl Cr. (4) Pemberton (5) Joffe Cr. (6) Gowan Cr. (7) Fire Cr. (8) Chehalis Lake (9) Harrison R. (10) Lower Fraser (11)  6  Washington State  Skokomish R. (12) North Fork Skykomish R. (13) Swift Resivoir (14) Early Winters Cr. (15) U . Goat Cr. (16) Lost River (17 ) Phelps Creek (18) Box Cr. Canyon (19) Tucannon River (20) Mill Cr. (21) Northwest Territories  Funeral Cr. (22) Keele R. (23)  28  Chapter 3:  Testing phylogeographic hypotheses using nested clade analysis: Origins of a widespread hybrid zone between Dolly Varden (Salvelinus malma) and bull trout (Salvelinus confluentus)  29  Introduction  Traditional phylogeographic methods rely on making biological inferences based on a visual assessment of the geographic distribution of haplotype relationships (Avise 2000). These inferences are strengthened when adopting a comparative framework that examines many codistributed taxa across a particular region (Avise 1992; Zink 1996; Bernatchez and Wilson 1998). The development of nested clade analysis (NCA) by Templeton et al. (1995) has strengthened this approach by incorporating an analytical framework for assessing geographic and genetic associations at different evolutionary levels on a haplotypic network, thereby allowing the separation of current population structure from population history. The ability of N C A to separate unique historical processes, such as fragmentation and range expansion, from recurrent genetic structure, relies on predictions derived from coalescent theory (Crandell and Templeton 1993; Templeton 1998; Posada and Crandall 2001). Templeton et al's. (1995) definition of an historical event states "such a range alteration must have occurred only once or sporadically during the time interval marked by the coalescence time of the gene region under investigation". Consequently, the timing of coalescence of the gene region under study will dictate the temporal history recovered. N C A ' s analytical properties and nested design make it one of the most promising techniques for testing phylogeographic hypotheses to date (Posada and Crandall 2001; but see Knowles and Maddison 2002). Its primary utility lies in the ability to reconstruct the probable sequence of events that has created the current patterns of genetic variation in a species. Unfortunately, N C A does not have, as one of its properties, the ability to assess the probability of different historical events. This method, while originally designed and widely used for intraspecific analyses (eg. Hammer et al. 1998; Durand et al. 1999; Bernatchez 2001; Pfenninger and Posada 2002), has also been found to work for deeper divergences (Crandell 1994), therefore making its application available to a broad spectrum of evolutionary questions including species 30  concepts (Gomez-Zurita et al. 2000) and the evolution of asexual populations (Law and Crespi 2002). Each new application of N C A to organisms with different population and demographic histories permits an evaluation of the method itself such that its limitations are recognized and further explored (for examples see Paulo et al. 2002; Masta et al. 2003). A knowledge of population histories, specifically the ability to detect allopatric fragmentation and range expansion, should also help determine the origins of hybrid zones between lineages subject to N C A analyses (Cruzan and Templeton 2000; Templeton 2001). Hybrid zones have received much experimental attention because they are seen as one of the keys to understanding the processes important to the evolution of reproductive isolation. Determining the origin(s) of hybrid zones, however, has been a fundamental problem for evolutionary biologists for some time (Harrison 1993) because historical and ecological factors may each play a role (especially in spatially complex "mosaic" hybrid zones). Whereas traditional phylogeographic methods have been successful at providing some insights into the location of glacial refugia and dispersal patterns (Hewitt 2000), which are then used to infer areas of secondary contact, N C A provides the opportunity to explore theses processes more thoroughly. The few studies that have detected hybridization using N C A (Maskas and Cruzan 2000; Pfenninger and Posada 2002), have only dealt with secondary contact and current introgression (but see Matos and Schaal 2000). Because some hybrid zones may have a more complex history, it would be useful to examine how well N C A performs under a greater variety of situations. Hybrid zones between divergent lineages often form in recently deglaciated areas (Hewitt 2000). Glacial contraction and expansion cycles during the Pleistocene greatly influenced the genetic architecture of species in northwestern North America and formed "suture zones" where many differentiated lineages across a broad range of taxa came into contact (Remington 1968). Dolly Varden (Salmonidae: Salvelinus malma) and bull trout (Salvelinus confluentus) are two 31  such species that have largely parapatric distributions in northwestern North America with known patches of hybridization (a mosaic) occurring throughout a broad zone of sympatry that may comprise part of an aquatic suture zone (Redenbach and Taylor 2002). Previous phylogeographic studies have suggested that bull trout and Dolly Varden each consist of two major phylogenetic lineages stemming from isolation in separate Wisconsinan glacial refugia. Bull trout were hypothesized to have been isolated in the Pacific (Columbia River valley) (BT-I) and Chehalis River valley refugia (BT-C) (Taylor et al. 1999), while Dolly Varden are suggested to have survived in the Bering Refuge (DV-N) and the Chehalis River valley ("southern Dolly Varden" DV-S) (Redenbach and Taylor 2002). In addition, DV-S and B T - C mtDNA haplotypes were found to be paraphyletic as a result of introgression of bull trout mtDNA into Dolly Varden. The mitochondrial and nuclear D N A phylogenies were also discordant indicative of historical introgression (Arnold 1997; Redenbach and Taylor 2002). The phylogeographic evidence and a "footprint" of historical introgression, suggest that the current broad zones of hybridization between Dolly Varden and bull trout may be the result of two processes, a range expansion from an area of sympatry (Chehalis River valley) and secondary contact between allopatric lineages (Pacific and Bering refugia). The possibility of two different origins of current hybrid zones is interesting because it means that a portion of the range of both species may have been in continuous contact while other parts of the range have experienced secondary contact. From the standpoint of hybridization studies, these two processes are important because of the potential for differences in the timing and extent of hybridization, which has implications for reproductive isolation and hence speciation. The principal focus of this chapter is to use N C A to investigate the population history of Dolly Varden and bull trout and to assess phylogeographic hypotheses about the origins of a widespread, mosaic hybrid zone between the two species. To meet this objective I employed two complementary sets of techniques: (1) maximum likelihood and maximum parsimony to 32  resolve deeper divergences among mitochondrial D N A haplotypes, and (2) statistical parsimony coupled with nested clade analysis designed for shallow divergences and reconstructing historical processes. Cladogram ambiguities resulted in alternative inferences for the N C A , and because these events are recent, I discuss the limitations of N C A when inadequate genetic variation and population bottlenecks influence the polarity of the cladogram. I also performed analyses of genetic diversity of mtDNA and a mismatch distribution analysis to infer the demographic history of the two species throughout their range. Habitat contraction and expansion cycles affected population demography (Hewitt 1996). Geographic range expansions are often associated with population expansions, and because each species was likely affected differentially by the glaciations throughout the Pleistocene, I expected the timing of population expansion to vary amongst lineages. In addition, differences in each lineage's demographic history may be influenced by life-history variation and modes of colonization. For example, northern Dolly Varden, which are mainly anadromous, are expected to have higher levels of genetic diversity and likely reflect a population at demographic equilibrium due to greater connectivity amongst populations. In contrast, interior bull trout, which utilized freshwater, inland dispersal routes, may show reduced genetic diversity associated with a demographic expansion following a population bottleneck. I also expected that introgressed "southern" Dolly Varden would exhibit a signal of population expansion separate from coastal bull trout resulting from divergence of their mtDNA following an historical introgression event. New results combined with previously published data for Dolly Varden extended the geographic coverage of both species to explore these questions. Finally, now that bull trout lineages can be identified easily (Chapter 1), I examined known hybrids in a watershed that contained introgressed Dolly Varden and both lineages of bull trout to assess if hybridization occurred with BT-C or BT-I. If hybridization is mainly the result of historical isolation and secondary contact, then I predicted that hybridization should be more common with BT-I. Alternatively, if ecological factors are 33  more important in structuring where hybridization occurs then I would expect hybrids to possess both interior and coastal bull trout mtDNA. Materials and Methods:  Sample Collection Bull trout and Dolly Varden samples were obtained opportunistically from throughout their respective ranges (Fig. 3.1, Appendix 1). Many of these included D N A samples from earlier studies as well as newly acquired samples from B C , Washington State, Yukon Territory, and eastern Asia via several different sources. Species were identified based on morphology or collection locality, and sympatric individuals were confirmed by nuclear D N A analysis (see Redenbach and Taylor 2002). Additional Dolly Varden and Salvelinus spp. sequences from Brunner et al. (2001) were obtained from Genbank. DNA isolation and sequencing D N A was extracted from approximately 20 mg of fin tissue by overnight digestion at 55°C using the Purgene D N A extraction kit (Gentra Systems, Inc.) D N A was precipitated in isopropanol, washed in 70% ethanol, and the pellet was left to dry overnight. The pellet was resuspended in 50-100ul T E buffer (pH 8.0) depending on the size of the pellet and stored at 20°C. PCR amplification of the entire mtDNA control region (~1000bp) was performed as outlined in Bernatchez et al. (1992 ). Purification of 150ul (3 X 50ul per sample) PCR products was carried out using Qiagen's Qiaquick PCR Purification kit. Sequences were generated by automated sequencing on an A B I 377 following termination PCR reactions with Applied Biosystems, BigDye ver. 3 and Centri-Sep purification. A l l sequencing reactions were performed only from the 5' end using the Tpro2 primer described in Bernatchez et al. (1992). Multiple D N A sequences were aligned with C L U S T A L (PCGENE) (Higgins et al. 1992).  34  Sequence variation and phylogenetic relationships Estimates of variability were computed with A R L E Q U I N version 2.0 (Schneider et al. 2001). Haplotype diversity (H), nucleotide diversity (TT.), and mean number of pairwise differences among haplotypes assuming a Tamura and Nei (1983) model of sequence evolution with an empirically determined ti:tv ratio of 3:1, which is the average ratio across all lineages. I reconstructed phylogenetic relationships among mitochondrial D-loop haplotypes using neighbor-joining (NJ), maximum-likelihood (ML) and maximum parsimony (MP) methods. Analyses were performed using P A U P * version 4.0b 10 (Swofford 2000). A hierarchical likelihood ratio test implemented in the program Modeltest 3.06 (Posada and Crandall 1998) was used to select the model of substitution that best fitted the data for the M L analyses. I also applied the parameters estimated by this model to the distance criteria settings in P A U P * to compute the pairwise distance matrix amongst haplotypes. NJ, M P and M L cladograms were constructed using a heuristic search and confidence in those topologies was assessed by 1000, 1000, and 200 bootstrap replicates respectively (Felsenstein 1985). Statistical parsimony and nested clade analysis To test hypotheses concerning the spatial distribution of mtDNA haplotype variation, I performed Templeton et al's. (1995) nested clade analysis (NCA). This method uses coalescent theory and patterns of dispersal to discriminate between historical processes and contemporary gene flow. In particular, if dispersal is limited, traces of fragmentation and range expansion should be detected. First, a haplotype network was generated using the program TCS (Clement et al. 2000), which implements the algorithm of Templeton et al. (1992). The algorithm first estimates the maximum number of differences among haplotypes resulting from single substitutions with 95% statistical confidence, or the parsimony limit. Next, all haplotypes that differ by one change are connected, then those differing by two, three and so on until all the haplotypes are connected or 35"  the parsimony limit is reached (Posada and Crandall 2001). Two or more networks can result when the haplotypes differ by more steps than the parsimony limit. The haplotype network is then nested by hand following the algorithm of Templeton et al. (1987) and Templeton and Sing (1993). In general, sets of n-step clades joined to a common ancestor by one mutational step make up each n+l-step clade, where n designates the total number of mutational steps necessary to define the clade. The procedure begins by nesting all haplotypes (0-step clades) separated by one mutational step into 1-step clades beginning from the outside of the network and moving inward. The 1-step clades are combined in 2-step clades and so on with the culmination of the total cladogram enclosing the highest nesting levels. Since parsimony assumes that all site changes are equally likely, multiple haplotypes differentiated by changes at the same site will form a "loop of ambiguity". A closed loop can be nested in the same one-step clade (Templeton and Sing 1993). When multiple loops occur, they need to be resolved to construct the nested cladogram. Alternative resolutions to the loops are based on predictions derived from coalescent theory using three criteria: frequency, topology, and geography (Crandell 1994; Crandell and Templeton 1993; Pfenninger and Posada 2002). Haplotypes are more likely to be connected to a high frequency haplotype than to a singleton because common haplotypes are expected to be older and therefore have more time to give rise to newer haplotypes. Similarly, haplotypes are more likely to be connected to interior haplotypes than tip ones because the older haplotypes are found in the interior of the cladogram. Finally, haplotypes are more likely to be connected to other haplotypes from the same region because newly arisen haplotypes are not expected to move far, unless there is a high level of gene flow. The nested design can be subjected to two different tests of geographical association. The simplest method uses an exact permutation contingency analysis, outlined in Templeton and Sing (1993) and Templeton et al. (1995), that tests the null hypothesis of no association of clades with geographic location. Each geographical location is treated as a categorical variable that is 36  tested against each clade within a nested category. The observed X value is compared to a 2  distribution of X values generated from permutations of the original data. This test, however, 2  does not take into account information about geographical distances amongst clades. A more objective geographical distance analysis of each clade comprises four distance measures Dc(X), Dn(X), I-Tc(Y), and I-Tn(Y), where X and Y refer to specific n and n+1- step clades respectively. Dc(X) is the clade distance which measures how geographically widespread haplotypes are within clade X . Dn(X) is the nested clade distance and measures how geographically distant haplotypes in clade X are from all other haplotypes in the next higher level (Y) (i.e. those that they are nested within). Finally, the I-Tc (Y) and I-Tn (Y) statistics represent the average Dc and Dn values for all interior clades within the nesting clade Y, minus the average Dc and Dn values for all tip clades within clade Y . The distributions of these distance measurements under the null hypothesis of no geographical associations within the nesting clade are determined by 10,000 permutations of the nesting clades vs. the sampling location. This procedure independently tests for statistically significant large and small distances (at the 5% level) for each clade within a nested group. Both nested contingency analysis and nested geographical distance analysis were performed using the program Geodis version 2.0 (Posada et al. 2000). An inference key (Templeton 1998, updated February 2001, see Appendix 3) was used to interpret statistically significant patterns for potential causes of geographical association. A temporal polarity in the network of haplotypes is required to perform the distance analysis and interpret any significant patterns. According to coalescent theory, the polarity is provided by the topology of the network and the haplotype frequencies (Castelloe and Templeton 1994). As discussed above, haplotypes at higher frequencies are expected to be older as are interior clades with two or more connections to the younger tip clades. The TCS program also calculates outgroup probabilities, which can also be used to assign interior status to a particular clade. The second type of polarity is the nested design itself, such that the higher the clade level, 37  the older it is in relation to the clades within the nested series (Templeton et al. 1995). Both the contingency and geographical distance analyses apply only to clades with geographic and genetic variation. The three reasons why the null hypothesis of no associations between haplotypes and geographic location fails to be rejected are: 1) the populations are panmictic and have not experienced expansion or fragmentation events, 2) sampling is inadequate either in terms of both the numbers per location and the number and position of geographic locations, or 3) there is insufficient genetic variation in the sampled populations. Predictions: I outline below a number of explicit zoogeographic hypotheses that can be assessed using N C A . Recall that D V - N represents Beringian Dolly Varden, BT-I represents bull trout from the Columbia Refuge and DB consists of coastal bull trout (BT-C) and introgressed DollyVarden (DV-S) co-occurring in the Chehalis refuge. First, if D V - N survived the Wisconsinan glaciations only in Beringia as proposed by Redenbach and Taylor (2002), a signal of allopatric fragmentation should exist between northern D V and the remaining haplotypes, evidenced by a larger than average number of steps separating them at the highest clade level (possibly different networks). Similarly, if interior bull trout (BT-I) indeed survived in a refuge distinct from DB, their haplotype distribution should exhibit a signal of past fragmentation at a lower clade level (more recent) from coastal bull trout and introgressed Dolly Varden. Following isolation, interior bull trout may have exhibited long distance colonization, in which haplotypes from the expanding population moved far away from those in the ancestral range. Secondly, if coastal bull trout and introgressed Dolly Varden underwent a range expansion north from a common southern refuge, as implied by the results of Redenbach and Taylor (2002), then a clade encompassing both BT-C and DV-S should consist of northern haplotypes that are younger (tip) than southern ones (interior) and there should be fewer widespread haplotypes in northern areas and more variation in the southern, pre-expansion regions (Templeton et al. 1995; Templeton 1998). If geographic sampling is adequate, the distance values can differentiate the 38  processes of contiguous range expansion via individual short-distance dispersal (concordant Dc and Dn values) and colonization achieved through individual long-distance dispersal (i.e. small Dc and large Dn). Finally, I expected that restricted gene flow could explain some of the variation in lower level clades. Under this scenario, haplotypes are associated by age such that older haplotypes are more widespread. If introgressed Dolly Varden and coastal bull trout have diverged under a model of restricted gene flow, then the younger D V haplotypes should be less widespread than the older bull trout ones. In addition, because the ancestral bull trout haplotypes are expected to be most frequent near their geographical origin, the derived Dolly Varden haplotypes should also occur within the same range. Molecular clock and divergence Pairwise sequence divergence estimates (p) between phylogenetic groups were corrected for within-group variation using the formula, PAB(net)  = PAB  - 0.5(PA  + -PB),  where P andP are A  B  the mean sequence divergence among haplotypes in phylogenetic groups A and B respectively and P  A B  is the mean sequence divergence between haplotypes of these groups (Wilson et al.  1985). A common rate of nucleotide substitution in salmonids of 1-2% per million years (Bermingham and Avise 1986; Grewe et al. 1990; Smith 1992) was used to convert  P B(net) A  into  estimated times since divergence. Demographic History A signature of population growth or decline can be detected from estimates of the frequency distribution of nucleotide site differences between every pair of sequences (Rogers and Harpending 1992; Harpending 1994). A mismatch distribution analysis was performed on each lineage to test for historic episodes of demographic expansion. This approach is possible because population structure has a limited effect on the mismatch distribution (Rogers 1995). Concordance of the observed frequency distribution of pairwise differences with the expected distribution underlying a model of sudden expansion (Rogers 1995) was assessed for each data 39  set in A R L E Q U I N using a least-squares approach (Schneider et al. 2001). The validity of the model was assessed by comparing the sum of squares deviation between the observed distribution (PSSD ) and the estimated distribution obtained for each simulated data set obs  (PSSD). In addition, the "raggedness index" was calculated to indicate the smoothness of the mismatch distribution used to differentiate between an expanded population and one at equilibrium (Harpending 1994). For distributions that did not significantly differ (P> 0.05) from the sudden expansion model, I used T (an estimate of the mode of the mismatch distribution), which is an index of time since expansion expressed in units of mutational time to compare the timing of demographic expansions amongst lineages. The age of population growth was calculated using the relationship x =2ut where w is the mutation rate per generation for the sequence, and t is the time since expansion expressed in generations (for sample calculation see Appendix 4). Estimates of x indicate the largest expansion because smaller, subsequent expansions are to some extent overshadowed by these events. Confidence intervals were estimated from 5000 bootstrap replicates. The pattern of pairwise differences between haplotypes is predicted to form a unimodal wave in an expanding population, whereas a population at demographic equilibrium is expected to yield a multimodal pattern (Slatkin and Hudson 1991; Rogers and Harpending 1992). Because the geographic ranges of bull trout lineages were differentially affected by Pleistocene glaciations, I would expect either an older expansion event or rejection of the model of sudden expansion for coastal bull trout compared to interior bull trout, which may show signs of a recent, sudden expansion following a bottleneck. Northern Dolly Varden are expected to have maintained a relatively constant population size given their greater dispersal ability and population connectivity. Southern Dolly Varden may show signs of a recent population expansion because the complete introgression of bull trout mtDNA and subsequent divergence caused a bottleneck effect in the mtDNA genome. Finally, the combined lineage consisting of 40  both coastal bull trout and introgressed Dolly Varden was tested to assess if it underwent an expansion episode from one ancestral population, post-introgression, indicated by a unimodal distribution, or whether it resembles a pattern of formerly isolated populations now in secondary contact, which is predicted to be bimodal.  Results Control Region Variation and phylogenetic relationships Ninety-nine individuals were sequenced for 550 bp of the mitochondrial control region. Excluding the three outgroup taxa, whose sequences were obtained from Brunner et al. (2001), thirty-one positions were polymorphic, including two indels, defining 28 haplotypes (Fig. 3.2). Fifty-four bull trout sampled at forty-seven localities yielded ten haplotypes, whereas 45 Dolly Varden from 39 different locations produced 12 haplotypes (if Brunner et al's (2001) Dolly Varden are included, 20 haplotypes from 58 individuals at 47 localities) (see Appendix 1). A l l individuals from the same population had the same haplotype, and bull trout and Dolly Varden (defined as such by nuclear loci) shared two haplotypes. The model chosen from the 56 models of evolution (described in Swofford et al. 1996) evaluated by likelihood ratio tests was the TrN+I+G model (Tamura and Nei 1993), a restriction of the general time-reversible model (Rodriguez et al. 1990). For the combined data set, the proportion of sites assumed to be invariant was 0.7533; identical sites were removed proportionally to base frequencies estimated from all sites, and rates (for variable sites) were assumed to follow a gamma distribution (Yang 1996) with a shape parameter (a) of 0.7856. The NJ, M P and M L analyses produced similar topologies, but terminal branching patterns differed where bootstrap support was low. A 50% majority-rule consensus M L tree shows that Clade N Dolly Varden was monophyletic and had 76 % bootstrap support (Fig 3.3). Bull trout haplotypes are paraphyletic with the interior bull trout (BT-I) forming a monophyletic 41  clade with 66% bootstrap support, and the coastal bull trout having a paraphyletic relationship with the Clade S Dolly Varden. Support for this group, which I refer to as clade D B , was low (34%); no derived mutation distinguished DV-S from BT-C, and they actually shared two haplotypes (6-G and 10-J) D V - N was approximately equally divergent from bull trout and introgressed Dolly Varden (2.7-3.0%) (Table 3.1). The corrected mean percent sequence divergence within DV-S (0.43) is slightly less than the divergence between bull trout lineages (0.61). The genetic diversity amongst lineages and within species was highly variable (Table 3.2). As expected, D V - N had the greatest haplotype and nucleotide diversity (0.92 +/- 0.3 and 0.0071 +/- 0.0041) followed by DB (0.89 +/- 0.3 and 0.0057 +/- 0.0033). Within D B , DV-S had similar haplotype diversity to BT-C, but lower nucleotide diversity and less sequence divergence among haplotypes. BT-I had extremely low values of both haplotype and nucleotide diversity (0.06 +/0.05 and 0.0001 +/- 0.0003) compared to the other lineages. Parsimony network and resolving cladogram ambiguities The statistical parsimony (SP) analysis found that haplotypes separated by up to ten mutational steps had a 95% probability of being connected parsimoniously. Consequently, two separate networks were produced (Figure 3.4) because northern Dolly Varden were separated by more than 10 steps. Of the 31 variable sites, only 23 sites were parsimony informative. Both networks contained ambiguous connections. In the D V - N cladogram, the loop connecting B9 to haplotype C and the missing intermediate (B6) could be resolved according to either the frequency and topology criteria or based on geography because B9 and B6 are found in the same general region compared to haplotype C. The alternative connections did not affect the overall nesting strategy or change the results. The loop connecting B l to interior haplotype A and BER2 was resolved based solely on the geographical criterion because both B E R 1 and 2 are found in  42  the Kamchatka River. This resolution subsequently maintains the connection with B4 and B5, also found in the same region. The BT-I / DB statistical parsimony network (clade 4-1) contained three interconnected "loops of ambiguity" (Fig. 3.5) that could be resolved in different places with equal probability according to the three criteria outlined above. I therefore created a set of plausible cladograms by breaking various combinations of mutational connections and identified which cladograms led to different nesting designs compared to keeping the loops intact. The resulting cladograms led to twelve different nesting designs and consequently, potentially different inferences about population history. Next, I explored how inferences drawn from each of the different nesting designs compared to those with the loops intact and identified where inferences of population histories were not affected by alternative nesting strategies.  For the sake of simplicity, I will  focus on the results that were identical across all nesting designs and use three different nesting designs to illustrate which haplotypes affected geographical associations at the lower-level clades. I will also highlight alternative inferences in the portion of the cladograms where uncertainty about temporal polarity exist because it is difficult to distinguish tip and interior status of clades at the higher nesting levels. Nested clade analysis and inferred population history The results of the nested contingency and nested geographical distance analyses are given in Table 3.3 and Appendix 2. The nested contingency test did support an association of genetic variation and geographical location for the two main clades. At the total cladogram level, the oldest event inferred was a past fragmentation between Clade 4-1 (BT-I/DB) and 4-2 (DV-N). The contingency analysis did not support the geographical structuring of Dolly Varden haplotypes at any clade level beside clade 1-14. Conversely, the nested geographical distance analysis did find significant geographical patterns to infer population history for each Dolly Varden clade with the exception of clade 1-14. The contingency test rejected the null hypothesis 43  of no geographical association for all bull trout and introgresssed Dolly Varden clades clades except their highest nesting level, clade 4-1. Each 4-step clade contained two 3-step clades. Within clade 3-1, some coastal bull trout haplotypes (2-2) were more closely related to interior bull trout (2-1) than they were to other coastal or introgressed Dolly Varden haplotypes (3-2). The inference key suggests restricted gene flow with isolation-by-distance (RGF) as the process responsible for the two 3-step clades rather than past fragmentation because their ranges overlap (Fig. 3.6). Regardless of how clades 2-1 and 2-2 are configured (see below), clade 3-1 underwent a continuous range expansion (CRE) to the north if it is assumed that clade 2-2, consisting of BT-C, is the ancestral clade (Fig. 3.7). Alternatively, if interior bull trout (2-1) are actually ancestral, as suggested by the M L tree, then a signal of past fragmentation is detected (Table 3.3). No geographical association was detected for clade 3-2 unless haplotype 10-J is included in which case the pattern is characterized by restricted gene flow with isolation-by-distance (Fig. 3-4,1 and II). Similarly, in clade 2-3 (with 10-J) no distinction can be made between restricted gene flow with isolation-by-distance and past fragmentation (PF) due to inadequate sampling between the lower clade levels. (Fig. 3.4,1 & II). The population history of clade 2-2 and the clades nested within it were also affected by the different nesting designs. When haplotype 10-J is nested in clade 2-2 (Figure 3.4, III), the inferred geographical structure is compatible with restricted gene flow with isolation-bydistance, otherwise there are no significant clade distances. Inadequate sampling prevented the distinction between continuous range expansion and long distance dispersal (LDD) as the process responsible for the distribution of bull trout haplotypes 5 and 9 in clade 1-2 for all the nesting designs (Fig. 3.8). For each of these configurations, clade 2-2 (BT-I) did have a significant geographical association but no significant clade distances (Table 3.3). Some strategies were deemed unlikely based on the rules for asymmetry (Templeton and Sing 1993), which allow clades 1-1 and 1-2 to form separate 2-step clades. 44  According to the N C A inference key, clade 4-2 (DV-N) underwent a continuous range expansion throughout the Pacific basin (Table 3.3, Fig. 3.4). The genetic structure of Dolly Varden in clade 3-3 was characterized by restricted gene flow with isolation-by-distance, but sampling was inadequate to discriminate between isolation-by-distance and long distance dispersal for individuals in clade 3-4 (Fig. 3.9). Isolation-by-distance was also responsible for the patterns observed between clades 1-10 and 1-11 in clade 2-5 and between clade 1-13 and 114 within clade 2-7. Most recently, the individuals in clade 1-10 experienced a range expansion. In summary, the N C A analyses indicated that the distribution of Dolly Varden and bull trout haplotypes has been influenced by range fragmentation between D V - N and BT-I, DB. In addition, some effects of restricted gene flow with isolation-by-distance were detected in the distribution of the clade containing all B T and introgressed Dolly Varden (4-1). The inference for interior bull trout and some coastal bull trout grouped in clade 3-1 could be interpreted as either range expansion or past fragmentation depending on which two-step clade is assigned ancestral status (Table 3.3). Finally, inadequate sampling prevented the ability to determine whether coastal bull trout in clade 1-2 underwent continuous range expansion or long distance colonization. Limited sampling and ambiguities in the nesting design also influenced the interpretation of processes responsible for introgressed Dolly Varden's distribution. Mismatch distribution The mismatch distribution for the full data set had four modes (Figure 3.10) representing the differences within and between groups. The mean number of pairwise differences was lowest in BT-I (0.059) (Table 3.2) consistent with the reduced diversity typical of a recent and severe population bottleneck. The demographic parameters estimated by the model of sudden expansion for this group should not be considered reliable since a single haplotype dominated its entire range. For the groups that were genetically variable, only the Dolly Varden lineages had unimodal mismatch distributions typical of a model of population expansion. BT-C and " D B " 45  both had bimodal distributions. A model of population expansion was, however, statistically supported by the SSD test (PSSD  obs  > 0.05) for each of the groupings (Table 3.4). Next, the  observed number of polymorphic sites was compared to the 95% confidence interval generated by 5000 resamplings of the data. BT-C, DV-S, and D V - N had observed values consistent with simulated values suggesting that the parameters estimated by the model accurately described the observed polymorphism for these lineages (Table 3.4). The age of expansion parameter (x ) was substantially different among these three groups. The observed value was greatest for BT-C indicating an older expansion event than D V - N , which in turn expanded earlier than DV-S. M y estimates of the expansion times, based on a molecular clock of 1-2% per million years, were approximately 350,000-700,000 years ago for BT-C, 120,000-230,000 years ago for D V - N , and a more recent expansion of DV-S between 68,000-127,000 years ago (Table 3.5). The magnitude of the change in effective population size could not be accurately estimated due the large confidence intervals on 9]. Clade identification of known hybrids Previously identified hybrids (Hatfield Consultants Ltd. 2001; Taylor et al. 2001) between Dolly Varden and bull trout from the Southgate River on the southcoast of B C were assayed for bull trout clade identification (see Chapt. 1 for method). This was done to assess if hybridization was more common with the interior clade bull trout. Of the seven hybrids (out of a total of 37 fish) assayed, two possessed interior clade mtDNA and five had coastal clade mtDNA.  46  Discussion Phylogeography and population history The lineages resolved by the M L analysis were concordant with, but had greater bootstrap support than, previous analyses of char mtDNA (Redenbach and Taylor 2002) (Fig. 3.2). The network approach was used to portray the relationships among sequences and to aid in inferences regarding phylogeographic and historical processes. The lack of bootstrap support amongst the terminal branches was manifested as extensive loops, particularly between BT-I and DB haplotypes in the statistical parsimony networks (Fig. 3.4). Network ambiguities are not uncommon in N C A and often either do not affect the nesting design or may be resolved easily by applying the three criteria outlined earlier: geography, topology, and frequency (Pfenninger and Posada 2002). At the other extreme, homoplasy may be so pervasive throughout the data that the nesting procedure fails (e.g. Vila et al. 1999). The data for bull trout and Dolly Varden fell somewhere between these two extremes and thus I chose to focus on the inferences that were consistent across the majority of the different nesting strategies. In the following, I discuss why some of these results may be incorrect based on the stochastic nature of the coalescent process and or limited sampling. Finally, I describe why the nested design did not support the hypothesis that the hybrid zone is the result of both secondary contact and continuous range expansion, but rather, appears to have identified substructure amongst coastal bull trout and introgressed Dolly Varden and an unexpectedly close relationship between coastal and interior bull trout haplotypes. The population history of bull trout and Dolly Varden as inferred by N C A appears complex, involving both historic events and recurrent gene flow. Despite the uncertainty in the cladograms, inferences for clades 3-1,4-1, and the total cladogram were repeatable across the different nesting strategies (Fig. 3.4). The earliest and most reliable event detected by N C A supported the first prediction of an allopatric fragmentation event between northern Dolly Varden in Beringia (4-2) and bull trout (4-1) in a southern refuge and probably dates back to the 47  late Pliocene or early Pleistocene based on the sequence divergence between D V - N and bull trout (2.7-3.0%) (Table 3.1). Any associations within each of these major clades would have been subsequently influenced by Pleistocene or post-Pleistocene climatic events. Attempts to date more recent events should be interpreted with caution due to the inherent variability of the phylogenetic branching pattern and variation introduced by demographic factors. Bull trout and introgressed Dolly Varden The population history between bull trout and introgressed Dolly Varden could not be unequivocally resolved at both the 4-step and 3-step levels because of uncertainty about the polarity of the cladogram. The grouping of coastal bull trout and southern Dolly Varden into a single clade (DB) as proposed by the M L analysis was not evident in the nested cladograms, but the close relationship between some coastal bull trout and southern Dolly Varden was reflected in the nested cladograms. I had predicted that interior bull trout would form a separate clade from both coastal bull trout and introgressed Dolly Varden resulting from isolation in separate refugia. The most striking aspect of the cladograms is the two 3-step clades in which some coastal bull trout are more closely related to interior bull trout (clade 3-1) while others are quite divergent and are closer or share haplotypes with DV-S (clade 3-2) (Fig. 3.4). Restricted gene flow with isolation-by-distance characterized the current population structure of all bull trout and introgressed Dolly Varden (4-1) because the geographical range of haplotypes from the tip clade (3-2) are nested within the range of clade 3-1 (Fig. 3.6). Under this model, the ancestral haplotypes are expected to be more widespread, but because it takes time for haplotypes to spread, the clade distance (Dc) should also increase with clade level (Templeton et al. 1995; Templeton 1998) and this is not the case because the Dc values for lower level clades are actually greater or the same as those from higher clade levels (Appendix 2). When this occurs, the null hypothesis of no geographical association should be accepted (Templeton et al. 1995), implying that the rate of gene flow is high enough to distribute older haplotypes throughout the 48  range. Recall that the contingency analysis also rejected the hypothesis of population structuring within clade 4-1. This pattern, however, may also be an artifact of the widespread distribution of clade 2-1 (BT-I) nested within clade 3-1 than a consequence of gene flow. A n alternative possibility is that the method has failed because of a lack of genetic variation in clade 2-1 caused by haplotype extinctions in the Columbia refuguim or subsequent founder events has hindered the temporal scaling of historical events relative to the mutation rate. The small number of mutational differences between interior and coastal bull trout suggests that if these groups were actually isolated in two separate refugia, it occurred recently. Population bottlenecks caused by expansion into newly available habitat can, however, lead to an erroneous conclusion of recent ancestry (Nichols 2001). Clearly, clades 2-1 (BT-I) and 2-2 (largely BT-C) have restricted ranges (Figure 3.7) with only slight overlap along the south coast of B C suggestive of a past fragmentation event. A range expansion from the coast to the interior was inferred instead. This outcome relies on the presumption that BT-C mtDNA (clade 2-2) is ancestral based on its interior placement in the nested cladogram. Coastal bull trout are probably the descendants of lineages that resemble ancestral bull trout mtDNA given their much greater haplotype diversity (0.85) and sequence divergence amongst haplotypes (0.79) compared to interior bull trout (0.06 and 0.19, respectively) (Tables 3.1 and 3.2; also see Taylor et al. (1999). Interior bull trout are more geographically widespread and genetically depauperate having only two haplotypes, in which a single haplotype dominated the entire distribution from Nevada to the Northwest Territories. Costello et al. (2003) presented microsatellite D N A data for interior populations of bull trout that showed clear signals of population bottlenecks during postglacial recolonization of much of interior British Columbia. The mismatch distribution (Figure 3.10) also provides evidence that interior bull trout have experienced a recent and severe population bottleneck and hence appear derived compared to coastal bull trout. It should be noted, however, that when BT-I are treated as the ancestral state (as indicated in the M L tree), a signal of past 49  fragmentation between the two clades was inferred because the interior-tip distance is significantly large rather than significantly small indicative of restricted ranges. In addition, haplotype 2 from the Snake River had the greatest outgroup weight designating BT-I as the ancestral state, but this evidence is not strong because one-step loops seriously affect the outgroup probabilities (Castelloe and Templeton 1994). N C A also failed to detect range expansion or long distance dispersal by interior bull trout. The main reason is apparent in the assumptions of the inference key regarding the distribution of tip and interior haplotypes. Under the null hypothesis of no geographical association, ancestral haplotypes are more widespread than derived ones. Consequently, range expansions should be detected by the inversion of this pattern, as in Caan et al.'s (1987) human population data, such that tips have an extended range. When a bottleneck (and associated haplotype extinction) occurs, as is the case for BT-I, the founder events that are associated with range expansion are more likely to fix the more frequent haplotype, which is ancestral. Under these circumstances N C A cannot detect range expansion and hence long distance colonization (Templeton 1998; Paulo et al. 2002). In contrast, the extinction of haplotypes in intermediate areas owing to climatic change can lead to a false inference of long distance colonization (Masta et al. 2003). Both of these results exemplify how inferences about population history at different temporal scales can be strongly influenced by stochastic demographic factors. Evolutionary history of bull trout Examination of existing zoogeographical, geological, and paleoentological information provides the basis for interpreting how demographic factors may have influenced the temporal polarity of the cladogram resulting in alternative historical inferences (clade 3-1) and a potentially false inference about current population structure (clade 4-1). To address bull trout's demographic history and the resulting uncertainty about the polarity of clade 3-1,1 will outline two possible scenarios to reconcile the genealogical with the geographic history of bull trout. 50  If bull trout originated in the Columbia River basin as proposed by Benhke (1980), and subsequently colonized other coastal watersheds (such as the Chehalis River) then the extinctions and bottlenecks that were associated with extensive Pleistocene glaciation and postglacial colonization in the upper Columbia River likely resulted in the extinction of older and intermediate haplotypes (Lindsey and McPhail 1986; McPhail and Lindsey 1986). Persistence of bull trout in a coastal refuge that did not experience the same conditions could have contributed to the greater antiquity of bull trout in this region (Taylor et al. 1999). Alternatively, bull trout could have originated in a coastal watershed and then spread throughout the Columbia River basin. Under this scenario, historical bottlenecks could be caused by either founding events during dispersal into the middle and upper Columbia River or extinctions as outlined above. The earliest known fossil evidence of Salvelinus comes from northern Nevada, is at least ten million years old (Cavender 1980), and may represent an early ancestor to bull trout in North America, implying BT-I is the ancestral state. In contrast, based on phylogenetic analysis of ribosomal nuclear D N A (Pleyte et al. 1992) and growth hormone loci (Westrich et al. 2002), the bull trout's closest sister taxon is presumed to be S. leucomaenis, the white-spotted char, found only in eastern Asia. Behnke (1980) also suggested that bull trout could share a common ancestor with the "stone charr" of the Kamchatka River based on morphological and developmental features in the head, mandibula, and premaxillary. Not enough is known about the stone charr to support this hypothesis, but I suggest if bull trout actually separated from a common ancestor with either of the Asian char, it seems plausible that they could have a coastal origin and this should be further explored to understand better the deep phylogenetic history of bull trout in relation to other Salvelinus species. It is interesting to note that the estimated time frame for coastal bull trout's population expansion based on the mismatch distribution analysis (350,000700,000 yrs. ago, Table 3.5) coincides with the approximate divergence time between bull trout lineages and could have occurred under either of the scenarios presented above. 51  The evidence for restricted gene flow with isolation-by-distance for clade 4-1 is inconsistent with the idea that bull trout were isolated in the Columbia refugium as outlined above, with expansion predicted to occur on a recent time scale relative to the mutation rate. The hydrological history of the region and comparative phylogeographic studies on other taxa also belay the inaccuracy of this inference. North temperate fish were capable of rapidly expanding into newly available habitat following glacial retreat (e.g. Bernatchez and Dodson 1994) yet were restricted to available dispersal routes. Zoogeographical evidence from other fish that likely survived in the Columbia (Pacific) Refuge (Lindsey and McPhail 1986; McPhail and Lindsey 1986), and the distribution of haplotype 1, as well as microsatellite data (Costello et al. 2003), support the hypothesis that bull trout utilized interior drainages to disperse north. These authors suggested that bull trout from the Columbia Refuge colonized the Fraser River through the Arrow Lakes late in deglaciation after the connection between the Fraser and Columbia rivers via the Okanagan River was severed because no bull trout are currently found in the Okanagan valley. From the upper Fraser River, bull trout probably utilized connections into the Peace River, which then experienced headwater transfer with coastal rivers farther north (Redenbach and Taylor 2002; Costello et al. 2003). This evidence further suggests that the inference of restricted gene flow with isolation by distance may not accurately reflect bull trout's population history. Introgression, persistence, and colonization Hybridization and the unidirectional transfer of the bull trout mtDNA genome into Dolly Varden have been well documented (Baxter et al. 1997; McPhail and Taylor 1995; Taylor et al. 2001; Redenbach and Taylor 2002). The fact that bull trout and Dolly Varden shared some haplotypes does not necessarily mean that hybridization is common between the two species. Actually, in the Southgate River where hybridization is known to occur (Redenbach 2000; Taylor et al. 2001), coastal bull trout and introgressed Dolly Varden have quite common yet 52  divergent haplotypes 9 and E, respectively. Haplotypes shared by both species (6-G and 10-J) are each connected by one mutational step to both a bull trout and a Dolly Varden haplotype suggesting that this is an old rather than recent hybridization event. Each of these haplotypes is, however, found within the same watershed, and therefore does not exclude the possibility that hybridization is ongoing. Redenbach and Taylor (2003, in press) recently found two probable F l hybrids in the Southgate River to support the latter scenario. Ancestral Dolly Varden from Beringia and coastal bull trout likely first came into contact during a warm interstadial period within the same range that southern Dolly Varden occupy today. Some of these Dolly Varden could have persisted with coastal bull trout throughout the subsequent glaciations in the Chehalis Refuge. M y estimated mean percent sequence divergence amongst DV-S haplotypes (0.43%) fell within the range calculated for other portions of the mtDNA genome (Redenbach and Taylor 2002), which coincides with an introgression event approximately 200,000-400,000 years ago. Assuming that the hybrid zone is similar to when it first formed, I would argue that the introgression event was geographically localized rather than widespread because derived DV-S haplotypes are not found throughout the cladogram (i.e. in the lower Columbia River). If Dolly Varden became introgressed in a localized area that was isolated from other refugial BT-C populations (Long Cr., Metolius R. and Swift R.) then that could explain the bimodal mismatch distribution of the DB clade. Populations that have been isolated from each other for a period of time are predicted to yield at least two peaks in the mismatch distribution (Rogers and Harpending 1992). This pattern was clearly evident for " D B " (Fig. 3.10), such that the first peak represents variation within the introgressed population and the second peak represents comparisons of haplotypes from the two different populations. Given Dolly Varden's present distribution, they most likely persisted with some bull trout in the unglaciated portions of western Washington State as this is where the most derived haplotypes occur. 53  NCA and char range expansions N C A did not detect a range expansion event for coastal bull trout (clade 2-2) or introgressed Dolly Varden clade 3-2 in any of the nesting strategies. Even when BT-I was removed from the analysis, a significant chi-square value (P=0.005) for their combined clades suggests that there is a geographical and genetic association, but no distance values were significant (Appendix 2). Clearly, both species have expanded into a previously uninhabitable (glaciated) range, so why was N C A unable to detect it? Theoretically, the expectations for range expansion rely on widespread tip clades (large Dc and Dn) with some older haplotypes restricted to the ancestral range (small Dc and large I-T). We would also expect the more northern haplotypes to be younger and expect more haplotype variation in the southern, pre-expansion range. The distribution of the introgressed Dolly Varden clade (tip) falls within the range of the coastal bull trout clade (ancestral) (Fig. 3.8) making their Dc and Dn values essentially the same (Appendix 2). I suggest that the reason N C A was unable to detect range expansion is because derived (tip) Dolly Varden diverged in situ rather than upon expansion and have become established in their ancestral range due to a shift in life history traits. For instance, most Dolly Varden are anadromous, but in sympatry with bull trout they tend to take on a stream resident life history and bull trout assume a migratory life history, either adfluvial or amphidromous (e.g. Hagen and Taylor 2001). Restricted gene flow and drift would also allow fixation of different haplotypes resulting from isolation of Dolly Varden populations in different streams (see below). A resident life history could also explain their limited spread north with the exception of the anomalous haplotype E, which is inferred to be ancestral and was found only along the south coast of B C and further north. Demographic parameters estimated by the mismatch distribution analysis indicated that a population for DV-S occurred recently (approx. 68,000-127,000 yrs ago, Table 3.3) and probably coincides with a period of warming previous to the last glaciation. 54  The only range expansion event detected in all of the nesting strategies occurred recently within clade 1-2 (BT-C), but inadequate sampling between Oregon and Puget Sound, W A made it impossible to distinguish whether movement between Oregon and the south coast of B C was due to long distance colonization or contiguous range expansion (Fig. 3.10, haplotypes 5 & 9). r  More extensive sampling in the ancestral range may also help to clarify the relationships amongst bull trout in clade 2-2 and identify colonization routes. If more ancestral variation was uncovered in the southern portion of the refugium, then some of the bull trout haplotypes identified here may actually be derived and therefore provide evidence of range expansion. Restricted gene flow rather than range expansion characterized the introgressed DB clade 3-2, for one of the nesting strategies (Fig. 3.4,1) despite it's spread into north-central B.C. This inference can be explained by the restricted distribution of tip haplotypes in the ancestral range around Puget Sound, W A while the older Dolly Varden haplotype, E , dispersed extensively along the south coast of B C to and including the headwaters of the Zolap and Omineca rivers in northern B C . Since this inference was only supported in two of the nesting designs, it should be treated as just a hypothesis and would be better supported if haplotype E was also found in the Olympic Peninsula. Other inconsistencies amongst the nesting strategies involve haplotypes that are potentially connected to haplotype E. Because it has a restricted distribution, any connection with other geographical locations affects the distance values, changing their significance. Resolution of these relationships is required to differentiate between alternate hypotheses and requires more genetic and geographic variation. Sufficient genetic variation is necessary to define the relationships amongst haplotypes, which could be attained through a combination of markers for the same individuals rather than sampling more individuals per population. With that said, extensive sampling throughout the refugial areas is also required to detect geographical associations (Cruzan and Templeton 2000).  55  Even though evidence of range expansion was not explicit, the prospect of localized introgression raises the question of what constitutes continuous contact in a shared refuge? If introgressed Dolly Varden were isolated from bull trout in the lower Columbia River during the most recent glaciation and each subsequently colonized the southcoast of B C , then that region may be an area of secondary contact between both bull trout lineages as well as southern Dolly Varden. I had predicted that if coastal bull trout and Dolly Varden were in continuous contact, then hybridization would be less common because more time had passed to develop reproductive isolating barriers. Preliminary data, however, suggest that hybrids in the Southgate River possess mitochondrial D N A from both bull trout lineages implying that ecological factors may be more important than historical contact in promoting hybridization. Range Expansion in Northern Dolly Varden Extending the analysis to include Dolly Varden from B C and the Kuril Islands confirm and extend previous findings from Brunner et al. (2001). Overall, the recovered pattern for northern Dolly Varden (clade 4-2) was one of continuous range expansion, when the origin of the expansion is from clade 3-3, presumably from a once continuous distribution across the Bering Land Bridge during the Illinoian glaciation. An inferred demographic expansion between 118,000-236,000 years ago may coincide with population growth during the Sangamon as sea levels rose and provided extensive dispersal opportunities, which is also evident in their much greater mtDNA genetic diversity. Following the expansion, each clade of Dolly Varden became isolated by distance, but it is not possible to determine if some long distance colonization occurred between the Kuril Islands and British Columbia since the intermediate areas were not sampled thoroughly (Fig. 3.9). More extensive sampling throughout Alaska is needed to assess if Dolly Varden, like Arctic char (Brunner et al. 2001) experienced a trans-Pacific colonization. The inference of a recent expansion in clade 1-10, which implies that Dolly Varden spread from  56  the north coast (QCI) into Alaska and Russia could, however, be a false inference due to limited sampling in eastern Russia and Alaska where we would expect more ancestral variation. Implications and utility of NCA: Scale, timing, and polarity Clearly, my nested clade analysis of the bull trout and Dolly Varden hybrid zone produced alternative inferences at many different clade levels, which essentially created more hypotheses instead of testing the original one. Claims that N C A is capable of testing evolutionary hypotheses (Posada and Crandell 2001) need to be reevaluated (Knowles and Maddison 2002). The statistical properties behind N C A lie solely within the creation of the parsimony network and the two permutation tests used for assessing geographical associations. Any inferences drawn from the key are based on biological principles of dispersal but do not have any statistical merit especially when considering alternate population histories (Knowles and Maddison 2002). For example, in a study that simulated allopatric divergence, Knowles and Maddison (2002) found that N C A usually correctly identified geographic associations but often inferred the wrong historical process because it does not take into account the stochastic nature of the coalescent process. Any processes inferred using N C A , therefore, provide only spatial and temporal evolutionary hypotheses because there is no means to differentiate alternate hypotheses with any statistical support. The inference key does, however, incorporate measures to recognize when the data is inadequate to distinguish between two or more alternative processes as seen in clades 1-2 and 3-4 (Table 3.3). Inferences should also be scrutinized based on comparative phylogeographic data, geological data, and species ecological requirements. A method that quantifies the likelihood of any given inference is required to actually test phylogeographic hypotheses. The nested contingency and geographical distance re-sampling tests did not always yield the same results when identifying clades with a geographical association. Because the contingency test is weaker, it is common to reject the null hypothesis of a random distribution of 57  haplotypes, as was evident in clades 4-1 and 4-2, when the higher-level nestings are composed of such divergent lineages. At the lower-level nestings, however, inadequate sampling can produce the same result of a non-random distribution of haplotypes by a lack of statistical power of the test. This seemed to be the case for almost all of the northern Dolly Varden clades (Table 3.3). A significant, but weak, spatial correlation was found for bull trout and introgressed Dolly Varden two and three step clades because the sampling density per geographic area was greater. Because my objective was to test the population history of the two species across their entire range, it might be expected that broad scale sampling would hinder the resolution of geographic and genetic associations at smaller spatial scales. Rooting an intraspecific network is inherently difficult because of the inability to place a well-differentiated haplotype in a network of closely related haplotypes with any statistical justification (Templeton and Sing 1993). It is therefore important for the network to have some temporal polarity so that clades can be assigned interior and tip status. This criterion is based on the frequency of the haplotypes and the network topology such that tip clades are connected to the cladogram by only one mutational connection, and interior clades are those connected by two or more mutational connections (Crandell and Templeton 1993). Within this framework, neutral coalescent theory predicts that haplotypes with a higher frequency are older than derived singletons, and tip clades are almost always younger than interior clades. Another form of polarity is in the nesting design itself, such that lower level clades are younger than the higherlevel clades that they are nested within. At the higher clade levels, it is inherently difficult to recognize what is a tip clade versus what is an interior one because often the clade that connects the ingroup to the outgroup is not reliably rooted (A.R. Templeton, Washington University, pers com.). Outgroup weights can also be used as a proxy to help identify the root of the tree (Castelloe and Templeton 1994), but the loops of ambiguity hampered this method. In my data set, interior bull trout (clade 2-1) were presumed to represent a tip clade based on a single 58  connection to BT-C. It is possible, however, to imagine that the high frequency haplotype 1 may have mutational offshoots that were not sampled in the mid-Columbia region. If more haplotype variation was found for interior bull trout, my prediction that BT-I would form a separate clade from D B would likely stand true. Given that N C A does not take into account the stochastic nature of the coalescent process, the demographic history of the taxa under study can greatly affect the polarity of the cladogram, particularly at the higher levels when it is already difficult to assign interior and tip status. These problems should be taken into consideration when performing nested clade analysis. The criteria for inferring range expansion (widespread tip clades and some older clades restricted to the ancestral range) are the least theoretically justified, but N C A has correctly inferred range expansions for multiple species that occur in previously glaciated areas (reviewed in Templeton 1998). I suggest that these criteria be re-examined for species that experienced severe population bottlenecks and associated haplotype extinctions causing only a few haplotypes to dominate the current range. A selective sweep could also cause this pattern. Alternatively, chance alone could cause the ancestral haplotypes to spread first. In such instances, the inference key does not take into account the stochasticity of the coalescent process. The ability of N C A to infer recent events, particularly range expansions, requires the detection of new mutations in the expanding population. For N C A to have the greatest power to detect Pleistocene events it is desirable to assay a rapidly evolving sequence. I expected that the mtDNA control region would provide the necessary variation based on its elevated rate of evolution found in other organisms (Howell et al. 1996; Hundertmark et al. 2002), including fish (Turner et al. 1996; Brunner et al. 2001). Contrary to these findings, I found relatively little Dloop variation (Figure 3.2) as has been found in intraspecific studies of chum salmon, Oncorhynchus keta (Park et al. 1993), lake whitefish, Coregonus clupeaformis (Lu et al. 2001), brook char (Bernatchez and Danzmann 1993), and rainbow trout, Oncorhynchus mykiss 59  (McCusker et al. 2000) with few informative sites and the presence of homoplasy which hampers phylogenetic reconstruction. It is therefore necessary to consider if the use of parsimony is justified. Conclusions Nested clade analysis supported the prediction that northern Dolly Varden and bull trout were isolated in separate refugia during the Pleistocene glaciations. Based on the results of the geographical distance analysis (Table 3.3, Appendix 2), it was inferred that Northern Dolly Varden experienced range expansions at different temporal scales. There was also some evidence of a recent range expansion in coastal bull trout. M y data also suggested that population size expansions occurred in both Dolly Varden lineages and coastal bull trout. The population history of bull trout and introgressed Dolly Varden, however, could not be resolved at two different clade levels where I had expected to find support for fragmentation and range expansion. One of nested clade analysis' supposed strengths is that it does not require a priori assumptions about the past to make inferences about population history (Templeton 1998). For areas that were recently glaciated, however, some expectations of fragmentation and expansion are warranted given an organism's distribution in a previously uninhabitable range. Because the Pleistocene was dynamic, the ability of N C A to examine multiple processes makes this method attractive for deciphering complex histories. M y attempt to examine the origins of a mosaic hybrid zone, however, demonstrates some of the inherent problems of using N C A for identifying processes when a severe population bottleneck has occurred. In such situations, it is necessary to incorporate as much information as possible such as paleoentological and life history data, and for fish, hydrographic evidence of dispersal routes. The inability of N C A to support the hypothesis of range expansion from a common refuge for some lineages may be due in part to the recency of events, such that new mutations have not arisen and, in part, limited sampling in 60  the ancestral range. Other studies have had difficulty in detecting range expansion (Alexandrino et al. 2002; Seddon et al. 2001) and fragmentation (Masta et al. 2003; Turgeon and Bernatchez 2001) during the most recent Pleistocene glaciation. The uncertainty in the cladograms provides, at best, multiple hypotheses about the geographical basis of historical introgression between bull trout and Dolly Varden. Unfortunately, N C A is not able to assess the certainty of the inferred events nor does it provide the means to test alternative inferences with any statistical rigor (Knowles and Maddison 2002). Finally, the history presented here is representative of a single genealogical history and other markers should be considered. Nuclear gene phylogeography is becoming increasingly popular for testing complex phylogeographic histories by providing concordance across loci (Hare 2001). Before any further phylogeographic research continues, the demographic history of bull trout, in particular, should be explored more thoroughly. More rigorous methods that take into account the genealogical history of the data are available for evaluating population histories thereby providing more accurate estimates of historical population size, growth , and migration (Emerson et al. 2001). M y study did not allow this approach because it requires independent population-level data.  61  70°  Russia  I  Alaska  inr  n n  n n 60°  Bering Sea  "4  no nn  \  v  Of detail  150°  180°  •150°  50°  'i 1  •120° 0  400  t  N  A Yukon R-  A  Liard R.  A 60°  Peace R Athabasca R.  no Fraser R.  3,33,50 "  North Pacific  14, 34, 51A XEk  Ocean  .Upper (Columbia, Rhfer  '50°  •ft Columbia R  Snake R.  © -140°  -130°  -120°_ _ km_ " 200  62  40° | 110° 4(H) IM  B  50°  <  Fraser R. 38,57,58  Vancouver Island  49°  1 ®  *' ® w  ©  48°  Olympic Peninsula  Puget Sound  Chehalis River Valley -125°  -124°  47° -121°  -123°" 0  50  100  c  Figure 3.1. Geographic distribution of Dolly Varden (white) and bull trout (black) mtDNA clades throughout study area. Area of detail in map A is shown is map B and map B's area of detail in southwestern B C and western W A is shown in map C. D V - N are squares, BT-I are triangles, and DB are ovals. Sample codes are identified in Appendix 1.  63  11122222233344445555 123345888833444506703 690003 23546574402 680514501531099012 7 Haplotype  BT-1 BT-2 BT-5 BT-6,DV-G BT-7 BT-8 BT-9 BT-10,DV-J BT-11 BT-12 DV-E DV-F DV-H DV-I DV-A,BER3 DV-B DV-C DV-D DV-Y DV-Z BERl BER2 BER4 BER5 BER6 BER7 BER8 BER9  CA-AACAATAGGGCCCATGTAGACGATGCTA -. . .G .G ..A.T. G.A.TA .A.T. .A.T. ..A.T. .T. G.A.TA .T. G.A.TA ..A.T. .T. G.A.TA .T. . - .G. G.A.TAT . .A.T .G T. A.T. T. -C.G. C .AAA.T C.C.CGT T. .C .G. C .AAA.T C.C.CGT T. -C.G. CGAAA.T C.C.CGT T. -C.G. C .AAA.T C...CGT T. -C.G. CGAAA.T CAC.CGT T. -C.G. GC .AAA.T C.C.CGT C.C.CGTA T. -C.G. C .AAA C.C.CGTA T. -C.G. CGAAA C.C.CGTAT T. -CCGT C .AAA C.C.CGTAT T.C -CCGT AAA C.C.CGT.T T. -C.G. AAA C.C.CGT GTTC -C.G. AAA -C.G. AAA C.C.CGT..GCTC -C.G. CGAAA C.C.CGT.T..T.  Figure 3.2. Mitochondrial sequence alignment for the variable positions from 550 bp of the control region defining 28 haplotypes. Dots indicate identity with haplotype 1 in bull trout. Dolly Varden B E R sequences are from Brunner et al. 2001.  64  HAPLO 6 G — HAPLO F HAPLO 11 — HAPLO 10 J — HAPLO I HAPLO E  DB Chehalis Refuge  | HAPLO 5 ~L  HAPLO 9  HAPLO 8 HAPLO H 93,89,76  — I  HAPLO 7 HAPLO 12  BT-I Columbia Refuge  HAPLO 1 L-  78,70,66  HAPLO 2  94,89,90  BER4 BE BER5  BER6 51,48, 57  BER9 | BER2  "~l— BER1  DV-N Beringia  HAPLO C HAPLO Y  95,87,92 92,90,76  BER7 BER8 HAPLO A BER3 — HAPLO Z 77,53, 76  HAPLO B HAPLO D  S. m. kraschen 79,71,58  S. namaycush S. leucomaenis S. fontinalis  0.01  Figure 3.3. Maximum likelihood phylogram estimated under the TRN-I-G model of evolution showing relationship among 28 control region haplotypes among Dolly Varden (HAPLO-letters) and bull trout (HAPLO-numbers) rooted with three outgroup species. Numbers at each node indicate bootstrap support greater than 50% for NJ, M P , and M L analyses. "BER" haplotypes are from Brunner et al. 2001. Bars adjacent to the haplotypes designate proposed lineages resulting from isolation in separate refugia. Clade DB reflects a paraphyletic relationship between bull trout and southern Dolly Varden due to historical introgression.  65  Nesting Design I  l-n (B6>  BT-I and DB  1-13  4-1 RGF 1-3  2-5 RGF  <8>—©  1-6  CD  2-3 1-2 CRE or LW^  3-2 * RGF  2-4  , 1-15 1-12  2-2 *  3 1-1,  i  ?-1  BT-I  O  DV  O  BT  O  Both  j 2-6  3-1 *  DV-N  CRE or AF  RGF  3-3  1-16  2-8  RGF or LDD  3-4  4-2 CRE  Figure 3.4. Nesting designs I-III inferred from the 95% parsimony network of the 28 haplotypes detected for bull trout and Dolly Varden. Each line represents one mutational change, open circles are inferred intermediate haplotypes, and thin dashed lines indicate alternative parsimonious connections. The numbers identify bull trout haplotypes, letters are Dolly Varden, and B# haplotypes are Dolly Varden from Brunner et al. (2001) as per Appendix 1. Clade levels (e.g. 3.1) are given in each box along with inferences for clades that had significant clade distances. Those that were significant at the 5% level by a chi-square test of geographical association are indicated with an asterisk. CRE= contiguous range expansion, RGF=restricted gene flow with isolation by distance, PF= past fragmentation, and LDD=long distance dispersal. Clades 4-1 and 4-2 could not be connected because they exceeded the 95% limit of 10 mutational steps, and the heavy dashed line represents allopatric fragmentation. Nesting design II and III only illustrate the different connections in clade 4-1.  66  Nesting Design III  Nesting Design II  Figure 3.4 continued  67  Figure 3.5. Transitions and transversions differentiating haplotypes in the parsimony network illustrating the loops of ambiguity, for example between haplotypes 8,10-J, E, and H .  68  I 65°  Yukon R.  Liard R.  Peace R.  55°  Athabasca R.  4-1 RGF  Figure 3.6. Geographic distribution of clade 4-1 in B.C. and adjacent portions of the western U.S. Clade 3-1 (I) is the ancestral clade and 3-2 (T) is the tip clade from nesting design I.  3-1 (I)  •  <S» 3-2 fT) i 45°  Columbia R,  Snake R  -140°  -130"  -120°  km 200  -110" 400  65°  Yl^cnR.  Mackenzie R  I  Liard R.  Athabasca R.  Fraser  3-1 CRE •  2-1  (T)  0  2-2  (I)  ^hrf  -3:O Columbia R.  i•  -+\  J  o  45°  o o  -140°  55°  -130°  -120°  69  Snake R  km 200  -110° 400  Figure 3.7. Geographic distribution of clade 3-1 from nesting design I. Clade 2-1 (T) designates BT-I as the tip clade and Clade 2-2 (I)  E  •  0 9 •  l  50  9  r  • 10-J  §  6-G. 1 • ft  (P S  ,:  6-G 48°  6-G  Figure 3.8. Geographic distribution of the D B lineage in B.C. and adjacent portions of western U.S. for nesting design I (Fig. 3.4).  D B ° 8  ^ 6-G  y  ,o  C l a d e 2-2  O  46'  50  C l a d e 2-3  44  (;  C l a d e 2-4 12  42°  -12-1°  -126"  -128  o -120  e  0  50  100  70°  Russia  Alaska  A-Ber3 Serf I Bert  B ®- . ~  D 60°  Ber7,8  W Berl,2 Bei4,5  150°  50°  -150°  180°  #  Clade 3-4  ^  Clade 3-3  A  Clade 3-2  -120° km  0  70  400  Figure 3.9. Geographic distribution of Dolly Varden 3step clades throughout Russia and North America from nesting design I in Fig. 3.4.  0  1  2  3  4  5  0  1  2  3  4  5  6  7  8  Number of pairwise differences  Figure 3.10. Mismatch distribution based on 550 bp of the mtDNA control region plotted as the frequency of pairwise substitution difference amongst haplotypes in all lineages combined (top left panel), and separately for each lineage within species and the paraphyletic lineage, DB, representing DV-S and BT-C. Diamonds represents the observed data, the model fitted to the data is the bold line, and the dashed lines represent the 2.5 and 97.5 percentile values of 5000 simulated samples.  71  9  Table 3.1. Mean percent sequence divergence between mitochondrial D-loop haplotypes, corrected for within-group variation. Where the identical headings intersect, the values shown represent the variation within a group. Bull trout are subdivided into interior bull trout (BT-I) and coastal bull trout (BT-C), and Dolly Varden are subdivided into northern Dolly Varden (DVN) and introgressed Dolly Varden (DV-S). D B are the combined haplotypes from BT-C and DV-S. Other Salvelinus sp. sequences are from Brunner et al. (2001).  Corrected Percent Seq uence diver gence  BT-I BT-I  0.19  BT-C DV-S DB  0.61 1.06 0.72  DB  DV-S  0.79 0.1  0.43  —  —  0.76 2.3 3.5  0.78 3.3  8.9 8.2  6.6 .5.3 8.4  DV^ S. m. krasch.  3 3.3  2.7 3.2  2.7 4.1  S. nam. S. leuco. S. font.  8.1 7.5  8.8 8 12.1  9.2  11.6  8.9 12.6  12.3  S. m. krasch. S. nam.  DV^J  BT-C  .  S. leuco.  n/a 5.1 4.7 8.1  n/a 3.6 9.3  n/a 8.6  Table 3.2. Measures of intrapopulation variability for the numbers of individuals sequenced (N) in each phylogenetic grouping.  Lineage  N  No.  Haplotype  haplotypes diversity (H) BT-I  34  2  BT-C  20  DB DV-S  34 14  8 12  DV-N  31  6 14  Nucleotide  Mean pairwise  diversity (%)  differences  0.06 (0.05) 0.0001 (0.0003) 0.85 (0.06) 0.0052 (0.0032)  0.059 (0.13)  0.89 (0.03) 0.0057 (0.0033) 0.77 (0.09) 0.0027 (0.0019) 0.92 (0.03) 0.0071(0.0041)  3.11 (1.65) 1.49 (0.96) 3.92 (2.02)  72  2.84 (1.56)  o c  o  CJ oo  <D  P  "So P  P JO  JS  CS Cu  cS  PH P H  £  U CJ CJ  00 CU CJ  CS  cS  T3  T3  OO  J=  <-i  .£  *  ^  -a  ^n  ^  C  00  -t-»  Cu X CJ  c3  CJ  CJ  _cS  JS  * 9  03  CH  PH P H  0 a 01 oi >  0  Pi  01  u  1  i  O fc  73 73  c  o  CS CJ  ifH C cs ti CN ok ^ 60 00 T—I | T-H '« ± t ^ J2 H CO H o o I I I U (N CN CN fc fc  £  ,*H  03  >  C  © 2 ° f c  CJ  J=,  a oi u  -4—»  -o -a  OJ c ,*> c  JS  CJ C H  -4—»  0 01  P  CO  O  o fc I  i Tf T f CN ' I  t  CO  0\  O fc  o fc vo 1 m  *-H ^H  o  O fc I ^ CN I  T-H  I 1 I I I cn cn cn cn cn ^ cn i i i i i i i cn cn i CN C N C N C N C N C N C NCN CN CN  OV VO vo —. t— r~~ CN CN cn O r- i n m o o cn 2  O © d u  ti ©  oi-  5  cn m d d  CN 00 0\ CN O o cn >n TJ- >n oo vo o O o cn ^ H o O p d  d d d d d d d  cc  cs '-3 g o o g p p p p d d i n CN © d S ov CN cn  ci ^ -a a  O cn . o d d ° d  o o o o o o o o o _^ o r - cn t— CN i n CN ov ^H ^  3  ^  s  ,  SV  OH 00 O  4J  T3  O Tf  CS  03  i  * J  i  i  i  i  i  i  SJ  C J C J C J C J C J C J C J C J  OO  c S c S c S c S c S c S c S c S  z u u u u u u u u  i  I  CN m I  1  Tf  r-H I  CN  "O "O  CJ T3 cS  I  i  cn cn cn cn T f T f CJ CJ OJ CJ  CJ  T3 cS  o3 03 -a o 03u u o u u  CS  T3 cS  u  cn  o  T3  o c  CD  .  o  « ca  "  U O CL O d  cu '3  2  cu '% oo 5 •S & 3  t  U  3g  ^3  *  CO C „3 s OH  co cu  2  a  &  ta  .s S  % JO  .2 -4—>  X  co S3 CD o  7 3  co  '-2  S «  3!  Q  T3  CO  O CL  E O  £3  JS  e  fl\ t I  o  I  U  cu CCaUL 2 w Cl  T3  CU  9c2 * .2  «a « to cu o  JS  —  U  O .y  Q 43  s S O Is! ca O E Jcu S co o  fO  &  ±  3 O  o  q  CU  '•^ s  2 H cu 6 o ca CL T 3  -  8  ON"  CN  1  00  CO CN CN CO i i CN1  O r^ —^ ON  /  s  co 1  T^  CN  _  3  3 CL  O  CL  CU ca cu 3  3  ^  U  W pq  CQ  Q  Q  w  CL CL ca  >  JO  CU T 3  O  0  >  >n -a '-acu ca op co JJ J O ca o jo G cu  !?!  CO  ca *  2 «> CEU .5 o co  T> ^ oC  ^s  I  CU ^  U  .s u 3  M .5 co Co 4g 3  3  00  o in o cu o " *s  0)  o X o d 3  o E  § CL X W  3  ca ?u  '-O ~  .2  IO ON  3  CU  E ^  c#  co  e 2 3  U  3  trj o  .O  3 3  CO  ca CU  U  ^ JO  S3 «  * 3 t5B O  ca CL X  ca  • u ca co  03  S "£  CU  3  cu  3  00o u  3 O  JjS ^3  3  o« c  3  .2  t:"°  3 3  •S to CU  .  3 O  -D O  ^  o  VH  1  03  u cu co co  .3;  ON  O  00•o~  ca  t Tt T CN T i t Tt Tt ql d CN CN q O OTt d CN CO  3  CU  J=) O  Tt  coi co co o  § CL X  ca  ^  JS  4^ JS  ON  cu  *Z2 co E "3 •—i ^— »  o  <u ca JS o  •a 5 ~vc>u  T—H  ON  CN  Q  ON ON  ON  X  o  <  CN  ON  ON  ON  o  3 O  JS  . co  cu w-i vCL ^ c •o  3  ^  o d °  JS  3  p  ^  oo  P  CO ON O  o  T3  ^ w  E  o cu  co  9 6 So o  w  J3 CU  -  cu ca  ^  •4—1  Tt  CO  •c  ^,  jo  ON  CO  IO ON  ca c •— 3 cu u ^—»  ^>  ^ Tt ~  o  CU  ^  H  CN  E  bo c &c ca ^o_ u  - ca u •2 T3 J S  oo  CL  3 J2>  JS  cu  >o  ° cu "  ^  .5 '5b cu 3  c#  3  U CU  ON  O  23 cu o ca ta > ca rv D  ^  o d d d  N CN  x  ^ 3  oo CN  T3  CO  >  O  00o  3  o  ^  oo  °  J^  <+H cu  00S  o ^  U  CN o oo  w  ° T3o -S JCcSUo •t-' CU Cd •4—» ca  •S Q  s s ^  ^  ••S  .2 ° 3 CD  p.  &  co &, °  O  ca ca JS  J2 S  (= T3 cu SS '9 ^ •fi 't o 3 ,3 CL  W  g CU  JS  ri.  t  CN oo  Tt o iT nt <d o T t rCN o oo mTt nt o" oo" iT Tt i n ON  T3 3  a cq  CO  T3  £  O  E ^  S3 c3  m  tn r~ oCN o o d d O d d d  £? i3 £  . CD  CO  c  3  1  E  oo  »5  X CU CU C  v*  CL  > 2 00' P O  ca  s « c c a 3 CD cu c cu b  oo  E  p  E S ca o 2 S3 . 2  bo  ->  .s <£ •S x 'co CD 2 JO jP CD O .N  3 T3 CU  •2 o ^  oo  cu  J S CU CO *-* J S U  3  oo  CN  H  CO  3  1  :D  N 0\  o d o  (U  X)  cu S CO  n.  ca  ON  _o in  'cO 3  ca  CL X W  00  o, ca  o ON  1/3  CN CN  ON  co ro VO  o CN co r- CN ^ H  00Tt CN 00 H >> pq Q Q  U  1  Chapter 4: A phylogeographic synthesis of the Salvelinus alpinus-malma species complex  75  Introduction  Char inhabit harsh northern latitudes and exhibit striking morphological and ecological polymorphisms related to resource use and life history traits (Behnke 1972). Such diversity has created considerable taxonomic confusion at both the species and subspecies levels within the genus Salvelinus (Behnke 1980; Savvaitova 1980). Attempts to organize this diversity and define evolutionary relationships amongst taxonomic entities remain problematic. A myriad of morphological (McPhail 1961; Cavender 1978), allozyme (Crane et al. 1994), karyological (Cavender and Kimura 1989; Frolov 2001), and D N A studies (Grewe et al. 1990; Pleyte et al. 1992; Radchenko 2002; Salmenkova et al. 2000) have tried to disentangle the evolutionary relationships of char, but often reveal conflicting relationships, particularly amongst members of the Salvelinus alpinus-malma complex. For example, data from allozymes and nuclear ribosomal D N A support a sister relationship between S. confluentus (bull trout) and S. leucomaenis (white-spotted char) (Pleyte et al. 1992; Westrich et al. 2002), but mtDNA revealed a close relationship between S. confluentus and S. alpinus erythrinus (Arctic char) (Grewe et al. 1990). Likewise, S. malma (Dolly Varden) were found to have a paraphyletic relationship with both S. alpinus (Brunner et al. 2001) and S. confluentus (Redenbach and Taylor 2002) using mtDNA. Studies of the ITS (rDNA) and Growth Hormone (GH) regions, however, showed S. malma to be reciprocally monophyletic in relation to S. confluentus (Redenbach and Taylor 2002) and paraphyletic with S. alpinus (Phillips et al. 1999). Inconsistencies between morphological and molecular data can be attributed to convergent evolution and phenotypic plasticity, whereas discordance between nuclear and mtDNA markers often indicate historical introgression because mtDNA has been shown to cross species boundaries more readily than nuclear loci (Avise 1994). Alternatively, differential lineage sorting amongst loci can cause the same patterns in populations that have been separated for a short time relative to their effective  76  population size (Maddison 1997). For closely related species, distinguishing the difference between these two processes can be difficult. Indeed, past hybridization events between the different groups has been put forth as an explanation for the discordant patterns (Behnke 1980; Phillips et al. 1999; Westrich et al. 2002). Arctic char are known to hybridize even with more divergent species such as lake trout (Salvelinus namaycush, Wilson and Herbert 1993) and brook trout (S.fontinalis, Bernatchez et al. 1995). Because most of these studies focused on small geographic areas and on the interactions of only two species, a broad geographical and temporal context of diversification and subsequent introgression is lacking. A partial exception is a recent study by Brunner et al. (2001) that identified five S. alpinus mtDNA phylogeographic lineages throughout their entire range (Fig. 4.1). S. malma were included and clustered in a distinct lineage (BER) whose status was unresolved in relation to the other S. alpinus lineages. The distribution of S. malma from throughout much of their range in British Columbia and Washington was, however, not included, nor was the southern Asian Dolly Varden subspecies, S. m. krascheninnikovi. The relationship between S. confluentus and S. a. erythrinus (stated above) also makes it necessary to consider a historical interaction between these two species in any attempt to understand the evolutionary divergence in the S. alpinus-malma complex. This chapter presents a comparative phylogeographic analysis that incorporates a subset of my mitochondrial D N A sequence data from chapter 3 with Brunner et al.'s 2001 S. alpinus sequences to provide a more complete evolutionary picture of mtDNA diversity in the S. alpinus-malma complex. Given some uncertainty in the mtDNA phylogeny at deeper nodes, I evaluated the relationships amongst ;well-supported clades with the "approximately unbiased test" (AU, Shimoodaira 2002) to test the relative positioning of S. confluentus and S. malma in relation to S. alpinus in more detail. Specifically I wanted to test the hypothesis that S. confluentus and the arctic lineage (ARC) of S. alpinus would form a monophyletic clade due to historical contact and hybridization suggested 77  by some earlier phylogenetic research (Grewe et al. 1990). These results provide a geographical context for comparisons to previous observations and help to identify geographical regions that deserve further attention. Finally, I discuss the importance of broad geographic surveys (including sympatric populations) using a combination of analytical tools required to elucidate complex histories.  Materials and Methods  Sample locations Sample locations for the BTDB and BER haplotypes are given in Chapter 3, Appendix 1, and the remaining lineages are described in the Appendix from Brunner et al. (2001). Briefly, they sampled Arctic char from eastern Asia and Siberia, northern Europe, eastern North America, Alaska, and along the Canadian Arctic coast (Fig. 4.1). I retrieved Brunner et al.'s S. alpinus and S. malma sequences from GenBank. Molecular Methods Laboratory methods for D N A extraction, PCR, and sequencing analysis for 552 base pairs of the mitochondrial D N A control region are not described here as they are thoroughly described in Chapter 3 (pp.34). Data analyses Divergence between lineages was calculated as the net number of nucleotide substitutions per site (Da) (Nei 1987) with DnaSP ver.3.5 (Rosas and Rozas 2000). Neighbor-joining, parsimony, and maximum-likelihood analyses were conducted with P A U P version 4.10b (Swofford 2000). Salvelinus fontinalis, S. leucomaenis, and S. namaycush were included as the outgroup species since they are considered the most basal taxa (Phillips et al. 1995). The model of D N A substitution that best fit the data was chosen using Modeltest 3.0 (Posada and Crandall 1998). The model selected was the TRN+I+G model and these parameters were used in 78  constructing both the distance and likelihood trees. The data were resampled 1000 times using a heuristic search for both the neighbor-joining and parsimony trees and a fast-step heuristic search for the likelihood trees to test for bootstrap support (Felsenstein 1985). Because the deeper nodes exhibited low bootstrap support, I compared four alternative topologies including the consensus tree generated by all three clustering methods, using the approximately unbiased (AU) test (Shimodaira 2002) implemented in the program C O N S E L (Shimodaira and Hasegawa 2001). This test is an "unbiased" modification of the SH-test (Shimodaira and Hasegawa 1999). Like the SH-test, it too controls for type-1 error owing to comparisons of multiple trees, including the a-posteriori generated tree (Goldman et al. 2000). But the AU-test also reduces the selection bias that makes the SH-test overly conservative (Shimodaira 2002). The AU-test utilizes a multi-scale bootstrap technique based on the theory of Efron et al. (1996) to compare the loglikelihood values for each tree and calculates approximately unbiased P-values that assess significance for competing hypotheses. Using this approach I examined the phylogenetic position of S. confluentus and S. malma by taking the topology reflected by the maximum parsimony consensus tree in which S. confluentus and the S. alpinus Arctic lineage form a monophyletic group (BTARC mono) (shown in Figure. 4.2) and evaluated it against three constraint trees which shifted S. confluentus and S. malma to different phylogenetic positions. First, I enforced monophyly for all S. alpinus (ARC mono), which would indicate that bull trout and Dolly Varden had remained isolated from Arctic char. Next, I enforced monophyly for S. confluentus only (BT mono) which assumes that Dolly Varden and Arctic char have a shared history, but bull trout has remained distinct. Finally, I allowed for polyphyly amongst all three species such that bull trout is intermediate between Dolly Varden and the A R C lineage suggesting hybridization with either or both lineages in the past.  79  Results  M y additional northern Dolly Varden (DV-N) samples clustered with the Bering lineage (BER) (Fig. 4.3) reported by Brunner et al. (2001). D V - N haplotype A from B C and Alaska was identical to BER3 from the Seutakan River on the Chukotka Peninsula. Bull trout and introgressed southern Dolly Varden (BTDB) formed a monophyletic clade with 91% bootstrap support. A l l three phylogenetic methods produced topologies that supported the major phylogenetic lineages resolved by Brunner et al. (2001), but bootstrap support was very low at the nodes defining the ordering of the lineages. M y analysis with the additional samples resulted in an unconstrained 50% majority-rule consensus maximum parsimony tree (Fig. 4.3) that placed the Siberian and Atlantic lineages of Arctic char in a monophyletic group compared to Brunner et al. (2001) who found that the Atlantic and Acadian lineages grouped together. The positions of S. m. krascheninnikovi and Salvethymus svetovidovi (SVET) varied across the three methods, so to maintain consistency, I constrained them to a basal position in relation to all the remaining lineages. The AU-test performed on the constrained topologies ranked B T A R C as the best tree; however, the polyphyletic and B T monophyletic trees could not be rejected. The tree that constrained all S. alpinus into a monophyletic clade (ARC mono) was rejected (p=0.035) (Table 4.1). While this is not the most exhaustive search for different topologies, test is not overly "liberal" in comparison to the S H test which accepted all of the topologies as equally likely. Other potentially competing arrangements pertaining to the constrained basal taxa, were not considered here.  Discussion  This study presents the most extensive mitochondrial D N A phylogeny for Pacific basin char to date by building on the earlier work of Brunner et al. (2001). M y extended coverage of S. malma (Dolly Varden) throughout much of its North American range and the incorporation of S.  80  confluentus (bull trout) provides a more complete phylogeographic framework for assessing previous observations regarding evolutionary relationships from both a zoogeographic and a taxonomic perspective. As described in the Introduction, the relationships between Arctic char and Dolly Varden and bull trout remain unresolved because of the lack of genealogical concordance amongst loci within and between species. Hybridization is the most common reason given for these patterns. In this study, I explore the geographical basis for potential hybridization and provide several alternative explanations for discordant gene genealogies with regard to the formation of species. In addition, Malyarchuk (2002) analyzed Brunner et al's (2001) data for mutational hotspots and also retrieved the major lineages but found reticulation caused by parallel and reverse mutation was the cause of alternative topologies amongst lineages. Therefore, it will be necessary to examine if concordant patterns occur throughout the mitochondrial genome. Out of four topologies compared using the AU-test, the only tree that was rejected was the one constraining all Arctic char lineages to monophyly (Table 4.2). Each of the other three trees showed that Beringian Dolly Varden are paraphyletic with Arctic char, consistent with Brunner et al.'s (2001) mtDNA phylogeny. The most parsimonious tree (Fig. 4.2, Table 4.1) suggested a close relationship between bull trout and the Arctic lineage (ARC) of Arctic char similar to Grewe et al.'s (1990) mtDNA data. The S. a. stagnalis from Baffin Island, NWT studied by Grewe et al. (1990) is accepted as S. a. erythrinus found throughout the Canadian Arctic (Phillips et al. 1995) and therefore is probably included in the Arctic lineage. The tree constraining bull trout (BTDB) to a monophyletic lineage could not, however, be rejected with any statistical support. Paraphyletic and polyphyletic relationships between mtDNA lineages can result from two processes: introgressive hybridization and incomplete lineage sorting.- First, it is possible that bull trout may have experienced contact and hybridization with the Canadian Arctic char lineage via an interior drainage connection, while Dolly Varden, being mostly anadromous, 81  may have shared a coastal connection with the remaining Arctic char lineages. These inteipretations imply that both Dolly Varden and bull trout have experienced hybridization at both ends of their ranges, which each other in the south (McPhail and Taylor 1995; Baxter et al., 1997; Redenbach and Taylor, 2002) and with Arctic char in the north. Alternatively, these sequences may reflect ancestral polymorphisms that predate the separation of each species. Phylogeography and introgression in char As discussed in Chapter 2, the geographic range of bull trout was recently shown to extend north in the south-central Mackenzie River valley in the Northwest Territories (Reist et al. 2002). Following the last glaciation, bull trout dispersing from the Columbia River basin probably colonized the headwaters of the Mackenzie system (i.e. Peace and Liard rivers) through a series of well-documented connections between these and river systems to the south (Lindsey and McPhail 1986; McPhail and Lindsey 1986; Haas and McPhail 2001; Costello et al. 2003). Ten thousand years ago proglacial Lake McConnel existed on the fringe of the retreating Laurentide ice sheet east of the Mackenzie River connecting these drainages (Lindsey and McPhail 1986) and could have facilitated contact between bull trout and Arctic char dispersing south from an Arctic refugium. The presence of bull trout in these drainages suggests that their range may have been similar during previous interglacial periods. I found no reports of Arctic char in the main stem of the Mackenzie River (see below), but their range may have been further south early in deglaciation and subsequently contracted due to warming climatic conditions. For example, Arctic char mtDNA was found in brook trout (S. fontinalis) populations in eastern Canada where Arctic char do not exist anymore (Bernatchez et al. 1995). The mtDNA phylogeny grouping bull trout and the A R C lineage is discordant with the nuclear ribosomal D N A internal transcribed spacer region (rDNA ITS-1) phylogeny for bull trout, Dolly Varden subspecies, and Arctic char (Nauyak Lake, NWT) that showed that bull trout were monophyletic (Redenbach 2000) and supported a sister relationship with S. leucomaenis 82  (Phillips et al. 1994). As historical introgression is often the best explanation for discordances between nuclear and mitochondrial D N A results (Arnold 1997), I propose that the close relationship between bull trout and the Arctic lineage of S. alpinus could be the result of historical hybridization. The tree that constrained bull trout to a monophyletic clade, however, could not be rejected with strong statistical support; therefore, this phylogeny should be treated as a hypothesis. It is generally accepted that Dolly Varden is more closely related to Arctic char than to any other Salvelinus spp., but the phylogeographic history of the S. alpinus-malma complex is not well understood. Ordering the evolutionary divergences within and between these groups is problematic due to morphological and behavioral plasticity as well as conflicting signals at molecular loci. The phylogenetic evidence presented here suggests that the Bering lineage (BER), identified by Brunner et al. (2001) as Dolly Varden, is intermediate between A R C and the remaining Arctic char mtDNA lineages, but that the southern Asian Dolly Varden, S. m. krascheninnikovi, is distinct and equally as distant from both S. malma (BER) and bull trout (BTDB) as it is from Arctic char (Table 4.2). Interestingly, the newly described genus Salvethymus svetovidovi (Chereshnev and Skopetz 1990, cited in Brunner et al. 2001) from Lake Elgygytgyn on the Chukotka Peninsula was found to have similar divergence estimates from S. malma and Arctic char and was closely related to S. m. krascheninnikovi raising the question of its taxonomic status. Based on a molecular clock of 5-10% divergence per million years between lineages for the mitochondrial control region, Brunner et al. (2001) estimated that the A R C group split from an ancestral group containing the SEB, A T L and A C D lineages approximately 300,000-700,000 years ago. The Acadian lineage split next from the Eurasian group, which subsequently diverged into the Atlantic and Siberian groups. They did not include the Bering lineage in these estimates, so the relationship between Dolly Varden and Arctic char remains to be reconciled. The BER 83  lineage was separated from both the A R C and SIB lineages by the same number of minimum mutations (also see Malyarchuk 2002), and an alternative configuration linking the Atlantic and Bering groups could not be rejected (Brunner et al. 2001) indicating that Dolly Varden split off from Arctic char sometime during the middle to late-Pleistocene epoch. If we accept a more conservative molecular clock that estimates the divergence time for mtDNA lineages in salmonids (Smith 1992) and bony fish in general (Bermingham and Avise 1986; Grewe et al. 1990) as 1-2% per million years, respectively, then the initial split between the A R C and the remaining lineages occurred approximately 0.75-3.5 million years ago. These estimates are similar to those calculated from allozyme genetic distances, which indicated the split between Arctic char and Dolly Varden occurred as early as 0.825 to 3.1 million years ago (Salmenkova et al. 2000). S. m. malma and S. m. krascheninninovi were also found to be monophyletic compared to Arctic char using rDNA ITS-1 sequences (Phillips et al. 1999; Redenbach and Taylor 2002) also suggesting an earlier divergence from Arctic char. I speculate that the BER lineage, consisting also of my D V - N haplotypes, existed before the A R C lineage split from the S I B - A T L - A C D supergroup, resulting from an earlier divergence with ancestral Arctic char in northeastern Asia. Subsequent introgression between Beringian Dolly Varden and Arctic char from the Siberian lineage along the Arctic coast could have resulted in the paraphyly observed between Dolly Varden and Arctic char. Sparse sampling in the Siberian region prevented the identification of possible contact zones between the two lineages. Zoogeography of Arctic char and Dolly Varden in Alaska and Arctic Canada In the western Arctic of Canada, the Mackenzie River appears to be a biogeographic boundary for the B E R lineage, but not for the A R C lineage. Thus the North Slope of Alaska may represent an area of recent secondary contact between the A R C and B E R lineages (Fig. 4.1). East of the Mackenzie River the anadromous form of char is Arctic char, presumably S. a. erythrinus, while west of the Mackenzie River, the anadromous char is Dolly Varden and the 84  lacustrine form is a relictual Arctic char (Reist et al. 1991; Reist et al. 1997). The relictual form, found in several lakes on the North Slope of Alaska (McCart and Craig 1980) and throughout western and southern Alaska (Morrow 1980) is more similar genetically and morphologically to the Arctic char east of the Mackenzie River than to Dolly Varden (Morrow 1980; Reist et al. 1997). Brunner et al.'s (2001) data concur with these generalizations such that the North Slope Alaskan samples fell within the A R C lineage. One of those lakes, surprisingly, contained samples identified as S. malma while the other two contained S. a. erythrinus. M y data, unfortunately, did not include any supplemental Arctic char samples from southwestern Alaska. This form called S. a. taranetzi or Bristol Bay-Gulf of Alaska form of S. alpinus (McPhail 1961) is similar to the Asiatic Arctic char on the Chukotka Peninsula and Kamchatka (Behnke 1984). I would expect, therefore, they would also fall within the Arctic lineage and be closely related to the distinct ARC3 haplotype (also S. a. taranetzi). Interestingly, this form was found with S. m. malma from the B E R lineage in the Seukatan River on the Chukotka Peninsula, suggesting a long period of reproductive isolation. A lacustrine form of Arctic char, called S. a. taranetzi was included in Phillips et al.'s (1995) study where mtDNA and nuclear data both suggested that northern Dolly Varden (S. m. malma) from the North Slope of Alaska was more closely related to 5. a. taranetzi from the Kenai Peninsula than to southern Alaskan Dolly Varden (5. m. lordi). These data, coupled with mine and Brunner et al.'s (2001) observation of paraphyly between the 5. malma (BER) and the S. alpinus SIB mtDNA lineages, suggest that Dolly Varden and Arctic char have experienced hybridization on more than one occasion: historically with Eurasian groups comprising the SIBA T L - A C D lineages and more recently with members of the A R C lineage. If Dolly Varden on the North Slope recently hybridized with S. a. tarenetizi contained within the Arctic lineage, then this could explain why constraining all Arctic char to a monophyletic group was rejected (Table 4-1). 85  One oft-mentioned geographic demarcation between North American Dolly Varden subspecies is the Alaskan Peninsula: S. m. malma occur north and east to the Mackenzie River and S. m. lordi are found south in central and southeastern Alaska to Washington State (Behnke 1980, 1984; Morrow 1980). Reist (1997) proposed that the separation occurred on either side of the Seward Peninsula based on historical drainage patterns of Beringia. He also suggested that there are actually three forms of Dolly Varden in North America; those north of the Seward Peninsula along the North American Arctic coast to the Mackenzie River, which resemble Siberian Arctic coast populations, a southern form found south of the Yukon River in North America, which is similar to those in southern drainages on the Chukotka Peninsula, and an Anadyrian form situated between the south side of Norton Sound, A K and the southern Seward Peninsula. Geological evidence also indicates that the Alaskan North Slope was partially isolated during the Wisconsinan glaciation and more so during the Illinoian, which has given rise to distinctive populations of Thymallus arcticus, Salvelinus namaycush, and S. alpinus (Lindsey and McPhail 1986). Reist's (1997) observation that the northern form is distinct could also be the result of historical or ongoing hybridization between Dolly Varden and Arctic char on the Alaskan North Slope. Stochastic lineage sorting vs. introgression When vicariant events separate ancestral populations and geographic or other types of barriers, impede gene flow over time, isolated groups are expected to reach reciprocal monophyly (at unlinked loci) through the process of lineage sorting. The process is considered incomplete during the intermediate stages of polyphyly followed by paraphyly. Assuming neutrality, genetic drift will dictate the rate of lineage sorting based on the effective population (Ne) size of the gene region under study. Mitochondrial D N A is expected to have a fourfold lower Ne because it is haploid and maternally transmitted, and, therefore, should attain monophyly sooner than nuclear loci. These expectations are, however also influenced by 86  demographic factors such as mating success bias, the degree of population structure, and selection (Ballard and Whitlock 2003). I have suggested that secondary gene flow following vicariance is the most likely explanation for the observed para- and polyphyly in the mtDNA genealogy. The alternative explanation for the rejection of monophyly between Arctic char lineages and Dolly Varden (BER) is that these sequences within the mtDNA control region represent a shared ancestral polymorphism between the two species. Nuclear D N A sequences are also expected to be polyphyletic or paraphyletic, if this were the case. Phillips et al.'s (1999) data from the ITS region found that S. malma were paraphyletic in relation to S. alpinus, but another study using the same gene region indicated that Dolly Varden subspecies are actually monophyletic in relation to S. a. erythrinus (Redenbach 2000). Both trees were constructed under the same model of evolution but did not contain the same representative taxa. For instance, the latter tree included bull trout sequences and provided elevated bootstrap support. One reason for the discrepancy could be differential mutation rates across lineages or species and this should be tested. If we accept that the two species are indeed monophyletic at the ITS locus, historical hybridization is the most likely explanation for the observed paraphyly between Dolly Varden and Arctic char, but alternative interpretations, such as homoplasy, cannot be rejected without further evidence for all subspecies from other loci. A paraphyletic relationship between bull trout and Arctic char mtDNA is not as readily explained by incomplete lineage sorting. The percent sequence divergence between S. alpinus and S. confluentus based on the rDNA ITS region is almost six times greater than the divergence between S. alpinus and S. malma (Pleyte et al. 1992). Because the tree constraining bull trout to a monophyletic clade could not be rejected, it is necessary to evaluate if the paraphyletic relationship is the result of homoplasy in the control region.  87  Dolly Varden and Arctic char: separate species ? The muddled evolutionary history of Arctic char and Dolly Varden can lead one to wonder if they are in fact separate species. The true test of biological species is the degree of reproductive isolation where they occur in sympatry (Mayr 1963). Both the northern and southern forms of Dolly Varden are sympatric with Arctic char in several lakes in Alaska (McPhail 1961; Behnke 1980). As noted above, Dolly Varden mainly exist with the Bristol BayGulf of Alaska form of Arctic char such as on Kodiak Island and the Aleutian chain (McPhail and Lindsey 1970), but sympatry has also been reported in lakes west of the Mackenzie River with the eastern arctic form, S. a. erythrinus (Behnke 1980). Sympatry is also widespread throughout Asia on the Chukotka Peninsula and Kamchatka (Behnke 1984; Salmenkova et al. 2000). Sympatric populations in Kamchatka are similar to those in Alaska, where Dolly Varden are anadromous and Arctic char occur as lacustrine populations. Those found together in lakes and rivers draining the Chukotka Peninsula, however, were often both found to be anadromous (Chereshnev 1979 in Behnke 1984). No strictly sympatric populations were included in this study, but records indicate that hybridization is limited or absent (McPhail 1961; Behnke 1980; Behnke 1984, Salmenkova et al. 2000), and significant differences in morphological characters and life history traits exist between species (McPhail 1961; Behnke 1984; Reist et al. 1997). Finally, allopatric populations of S. a. taranetzi and S. a. erythrinus were more closely related than sympatric populations of Dolly Varden and S. a. taranetzi (Salmenkova et al. 2000). Current evidence of reproductively isolated sympatric populations indicates that S. malma and S. alpinus are indeed separate species. The degree to which historical isolation, secondary contact, and potential hybridization have played a role in creating these types of divergences deserves further attention.  88  Conclusions and future directions  The intention of this work was to present a more thorough phylogeographical analysis of the diversity present in the genus Salvelinus for comparisons to previous observations and to identify geographical regions that deserve further attention. As with most studies of char, this study also poses more questions. The addition of bull trout in this context produced an interesting phylogeographic hypothesis concerning historical contact with Arctic char. Similarly, I propose that Dolly Varden experienced an early Pleistocene connection with the Eurasian lineages of Arctic char along the Arctic coast and a more recent interaction with the Arctic lineage on the North Slope of Alaska. The distribution and relationships between Dolly Varden and Arctic char on the North Slope of Alaska compared to those in the south require further work particularly by examining sympatric populations in both regions. Coregonus clupeaformis also occurs here as sympatric populations and may serve as a basis for a comparative phylogeographic study of the region. The use of broad geographic surveys with a combination of analytical tools is necessary to elucidate complex histories and test phylogeographic hypotheses (Schneider-Broussard et al. 1998; Sota et al. 2001). This is particularly true for species complexes such as the S. alpinusmalma complex, which may have formed as recently as the Pleistocene. Nuclear phylogeography has become increasingly useful as a comparative tool to the traditional mtDNA phylogeography (Hare 2001). When both approaches are combined, the ability to differentiate between incomplete lineage sorting of ancestral polymorphism, homoplasy, and introgressive hybridization becomes possible (Avise 2000). Any further mtDNA phylogeographic analyses should expand sequence collection beyond the mitochondrial control region. It is possible then, that hybridization is just one explanation for the "best" tree identified by my analysis, and other portions of the mitochondrial genome as well as nuclear sequences should be combined and used to test these hypotheses further. 89  On a finer scale, the Bering-Chukotka region has had a dynamic geological history, which has produced much faunal similarity and also generated isolated pockets of diversity. Because the coastal regions were less severely affected, especially during the Illinoian glaciation, minor refugia such as Bristol Bay and the North Slope of the Brooks Range in Alaska have created distinct forms (Lindsey and McPhail 1986; Stamford and Taylor 2003). Further exploration of these areas (including sympatric populations) within the broader phylogeographical context will be necessary to understand the complex evolutionary history of the S. alpinus-malma complex.  90  Figure 4.1. Holarctic distribution of phylogeographic lineages found in Arctic char (Arctic, Siberia, Atlantic, Acadia), Dolly Varden (Bering, S.m. krascheninnikovi), and bull trout /introgressed Dolly Varden (BTDB). The arrow indicates the proposed phylogeographic hypothesis that bull trout and the Arctic lineage experienced introgression in the past. Symbols in white indicate Brunner et al.'s (2001) samples and dark symbols are additional samples included in the present study. Figure adapted from Brunner et. al (2001).  91  ATL  BTARC mono  ATL  ARC mono  SIB —  :  SIB  ACD  ACD ARC  BER/DV-N  BER/DV-N —:  BTDB  —  ARC  BTDB _SVET  ..SVET . S.rn. krssch.  _ S.m. kiasch. Outgroup Outgroup ATL SIB  BTDB  BT mono  Polyp hyly  ATL  ACD —  SIB  BER/DV-N  BTDB ACD ARC  BER/DV-N I—•• ARC  _SVET _ S.m..kiasch.  _SVET  Outgroup  _ S.m. kiasch. Outgroup  Fig. 4.2. Alternative constrained topologies for the mitochondrial D N A control region used to test phylogeographic hypotheses. B T A R C mono grouping bull trout and the A R C lineage is based on the 50% majority-rule consensus tree, A R C mono forces all of the arctic char lineages into a monophyletic group, B T mono constrains only bull trout to monophyly, and Polyphyly constrains all species to a single lineage. S. m. krasheninninovi and the SVET (Salvethymus svetovidovi) haplotypes are constrained to a basal position in relation to all the other lineages.  92  ATL8 — ATL16 ATL18 ATL12 ATL10  63,53,21  — ATL2  ATL7  — ATL3  1  SIB4  t  1  SIB3 — SIB5 SIB2 — SIB1  SIB7 SIB6 SIB8  — ACD5 ACD1 ACD7  56, 24,36  , 87,77  ACD2 BER4  93, 90,86  1  ACD4 ACD3 BER5  — BER6 — BER9 DVC — BER2 — DV Y  BER7 BER8  "* 91,89,90  L / v  51,25,11  BER1 I— D V Z DV B DV D  i — ARC5  ARC12 ARC10 — ARC7  91,81,73 94, 88,62 62, 49,46  ARC9 ARC4 ARC18 ARC17 ARC6 ARC3 BT 11 —I DV F •— DVH I RT 5R BT _| BT 11 BT 2 1  91,89,81  69, 66,37  SVET S.m.krasch. S namaycush S leucomaenis  S fontinalis  Figure 4.3. Maximum parsimony tree depicting the phylogeographic relationships between arctic char (ATL, A C D , SIB, ARC), Dolly Varden (BER/DV), and bull trout/introgressed Dolly Varden (BTDB), and three outgroups. Numbers are bootstrap values generated for distance, parsimony, and maximum likelihood clustering methods.  93  Table 4.1. Statistical tests of significance for competing phylogenetic hypotheses based on parsimony. Tests were performed using the approximately unbiased test (AU) and the Shimodaira-Hasewaga test (SH)(Shimodaira and Hasegawa 2001). A p-value < 0.05 indicates that the constrained tree is significantly worse than the best tree.  Rank  -LnL  Tree  SH  AU P-value  P-value  1 (best)  B T A R C mono  1930.27  0.726  0.905  2  Polyphyly  1930.94  0.484  0.702  3  BT mono  1932.00  0.351  0.632  4  A R C mono  1940.33  0.035  0.114  Table 4.2. Net nucleotide substitutions per site (Da) between lineages of Arctic char: Arctic (ARC), Siberian (SIB), Atlantic (ATL), Acadian (ACD), and bull trout and introgressed Dolly Varden (BTDB), and Beringian Dolly Varden (BER), and S. m. krascheninninovi (S.m. kr.), and Salvethymus svetovidovi (SVET).  ARC BTDB BER SIB ACD ATL S. m. kr. SVET  ARC  BTDB  BER  ~ 0.017 0.015 0.022 0.028 0.026 0.465 0.472  0.018 0.018 0.020 0.021 0.468 0.472  —  0.012 0.013 0.013 0.469 0.469  SIB  ACD  ATL  S. m. kr.  —  0.012 0.005 0.468 0.470  94  —  0.011 0.465 0.472  —  0.469 0.467  —  0.027  Chapter 5:  General Discussion  95  Phylogeography is the study of the geographic distribution of genealogical lineages either within or between closely related species. Although it comprises basic aspects of classical zoogeography, the advent of molecular analysis has stimulated tremendous intellectual and empirical activity in phylogeography (Avise 2000). Given their Hoi arctic distribution and potential for interaction among lineages within species and between species, Salvelinus (char) species have provided rich material for phylogeographic investigations. M y thesis has examined phylogeography of Salvelinus at various levels of biological organization: within species across watersheds, between species within and between watersheds, and systematic relationships and interactions within most species of Salvelinus.  Double invasions and sympatric mtDNA lineages of bull trout  In chapter 2,1 investigated the distribution of two mtDNA lineages of bull trout in northwestern North America and found evidence of "double invasions" of some watersheds in contrast with their largely allopatric distribution described previously (Taylor et al 1999). Multiple invasions have been widely implicated in explaining elevated phenotypic diversity in localized geographic areas (see Mayr 1963, p. 504). Similarly, double invasion and parallel evolution may have been important in generating the life history variation (e.g. amphidromy vs. fluvial) exhibited by bull trout within watersheds. In short, are the different life history forms of bull trout a product of isolation and adaptation in separate refugia followed by secondary contact, or can they arise independently of their evolutionary history by natural selection postglacially? Addressing these alternatives is, however, complicated by the presence of Dolly Varden in coastal watersheds with both allopatric and sympatric lineages of bull trout because the presence of Dolly Varden char appears to be a factor in the expression of amphidromy in bull trout (Haas and McPhail 1991), and the two species do occur in sympatry in 96  many coastal river systems (Taylor et al. 2001). In areas of sympatry, Dolly Varden, which are typically anadromous when allopatric, take on a stream-resident life history while bull trout will maintain the amphidromous or adfluvial type (Hagen and Taylor 2001). The shift to different resource environments during the adult stages is consistent with the idea that natural selection via character displacement is acting to reduce interspecific competition in sympatry. Phenotypic plasticity cannot, however, be ruled out. In addition, differences in body size associated with resource use are believed to contribute to both pre- and postmating isolation between Dolly Varden and bull trout (Hagen and Taylor 2001). Therefore, the life history variation exhibited by bull trout and Dolly Varden in sympatry suggests similarities to resource polymorphism and ecological speciation in many temperate freshwater fish (Schluter 1996; Robinson and Schluter 2000; Redenbach and Taylor 2003). Consequently, my finding of sympatric mtDNA lineages of bull trout in some river systems, coupled with other genetic and life history differences that are associated with these lineages (discussed in Taylor et al. 1999) suggest several potentially interesting research questions that may be addressed when both bull trout lineages coexist with Dolly Varden. For example, have bull trout lineages remained distinct in sympatry with Dolly Varden or do they form a single amphidromous population? If interior clade bull trout are unable to shift to amphidromy, they may retain a fluvial (mainstem river resident) life history, while Dolly Varden remain tributary stream residents. In addition, is it possible that size assortative mating and spatial or temporal differences in spawning activity between forms could contribute to reproductive isolation between lineages of bull trout?  A mosaic hybrid zone between Dolly Varden and bull trout Chapter 3 of my thesis was concerned with comparative phylogeographic inferences and historical demography in the formation of a Dolly Varden and bull trout hybrid zone. This  97  chapter provided important new information for both species largely because my analysis is the most extensive yet in terms of geographical coverage. A nested clade analysis of both species did infer that Dolly Varden were sundered in two separate refugia during the last glaciation and suggested that Beringian Dolly Varden have expanded their range on both sides of the Pacific basin. The possibility of a trans-Pacific colonization was also evident, similar to the phylogeographic patterns seen in sticklebacks (Deagle 1996). Dolly Varden south of the ice sheet shared the Chehalis refuge with coastal bull trout (DB clade), where a historical introgression event appears to have been localized in northwestern Washington. The genetic and geographic substructuring of the DB clade indicates that bull trout in the lower Columbia may have remained distinct from other coastal bull trout and probably colonized the southcoast of B.C. separate from populations on the Olympic Peninsula. Interior clade bull trout have clearly experienced a recent and severe population bottleneck. While more extensive than previous studies, even better geographic coverage of the three refugial areas will be necessary to understand some of the finer scale phylogeographic patterns such as genetic substructure in and dispersal from the Bering and Chehalis refugia. Combining the control region sequences with other portions of the mtDNA genome will likely enhance the genetic resolution needed to detect these patterns. In addition, rapidly evolving microsatellite loci have been shown to elucidate finer scale dispersal patterns (Costello et al. 2003). Finally, bull trout's phylogenetic and demographic history should be examined in more detail to better understand the colonization history of the Columbia River through the use of new coalescence analyses that incorporate migration (Beerli and Felsenstein 2001). The dynamic geological history of northwestern North America has created complex biogeographic patterns. The longitudinal and latitudinal phylogeographic patterns displayed by bull trout and Dolly Varden, respectively, are concordant with those observed in numerous taxa 98  in northwestern North America (Soltis et al. 1997; Nielson 2001). These data, therefore, contribute to the growing body of regional phylogeographic studies. The ability of N C A to reconstruct the historical processes responsible for creating a hybrid zone between these two species was put to the test in this complex setting. Many of the ambiguous relationships amongst haplotypes raise the question of whether parsimony accurately reflected the evolution of the mtDNA control region in bull trout and Dolly Varden given the potential for homoplasy. In addition, the number of alternative inferences produced by N C A suggests that the inference key is not equipped to accommodate species with complex demographic histories, such as bull trout. Mosaic hybrid zones are thought to form as a result of environmental heterogeneity in which parental genotypes are adapted to alternative habitats with hybrids occurring in patches of transition areas (reviewed in Harrison and Rand 1989). For aquatic organisms in particular, this environmental heterogeneity may exist at different spatial scales. First the hybrid zone could exist within a specific watershed that contains patches of habitat suited to either species. On a larger scale, the mosaic pattern is contingent upon local historical processes such as watershed exchanges that allowed sympatry in some watersheds, but not others. At either scale, each patch is essentially a geographically isolated subzone with potentially different historical and/or ecological characteristics. These natural replicates provide the opportunity to study whether reproductive isolation evolves in different ways across subzones. M y analysis did infer that some areas of sympatry containing interior bull trout and northern Dolly Varden are the product of secondary contact, a result that is common in hybrid zone studies (Gava and Freitas 2003; Zamudio and Savage 2003). By contrast, phylogeographical substructure within the DB clade suggested that while Dolly Varden and bull trout may have a shared refugial history south of the ice sheet, this "shared history" does not necessarily imply that the species were always in continuous contact. In fact, Angers and 99  Bernatchez (1998) presented evidence of distinctive phylogeographic subgroups of brook trout (S. fontinalis) that all originated from the same refuge in eastern North America. Across the entire contact zone between Dolly Varden and bull trout, there are, consequently, three types of replicate hybrid zones: areas of secondary contact such as northcoast rivers, areas of presumed continuous contact on the Olympic Peninsula, and areas of where both processes may have resulted in extant sympatric populations (Southgate and Homathko rivers). Complex, interspecific mosaic hybrid zones have also been shown to form between species of toads (Bufo) (Green and Parent 2003) and grasshoppers (Chorthippus) (Bridle et al. 2001). The existence of these different kinds of char hybrid zones provides rich opportunities for further analysis. Initially it would be interesting to compare the sympatric populations on the Olympic Peninsula, W A to those on the northcoast of B.C. to evaluate ecological differences between the species in each area and to see if different periods of contact have resulted in different interactions between species. For instance, when a high degree of assortative mating has evolved, as seems to be the case for bull trout and Dolly Varden, natural selection is expected to act directly on mate choice to reduce the production of unfit hybrids, through the process of reinforcement (Noor 1995; Jiggins and Mallet 2000). In chapter 3,1 had predicted that areas of continuous contact may have had more time to build up reproductive isolating barriers and, therefore, reinforcement may be stronger in some subzones compared to others. This comparative approach outlined above provides an excellent opportunity to test this hypothesis. Under this same reasoning, I had predicted that southern Dolly Varden would be more likely to hybridize with interior bull trout than coastal bull trout because the latter and Dolly Varden have been sympatric for longer owing to their shared refugial history. Although preliminary analysis of the mtDNA of hybrids in the Southgate River indicated no bias in hybridization, more substantial analyses are required to test this idea rigorously. Alternatively, evidence of unidirectional hybridization suggests that small Dolly Varden males "sneak" spawn 100  with bull trout females. This type of parasitic mating strategy, common in many salmonids, may compromise selection for size-assortative mating, reducing the likelihood for reinforcement (Redenbach and Taylor 2003).  Phylogeny of Salvelinus  Reconstructing the evolutionary history and relationships among members of Salvelinus have long vexed evolutionary biologists and taxonomists (Behnke 1980; Jonsson and Jonsson 2001). In chapter 4 of my thesis, I investigated the phylogeny of Pacific basin char and significantly extended previous analyses by the inclusion of a geographically wide array both of Dolly Varden and bull trout. Monophyly of the four Arctic char (S. alpinus) lineages was rejected. Although the deep biogeographic events within the Arctic char complex are difficult to establish due to low internode values in the phylogenetic tree, the phylogeographic structuring of each of these lineages provides some insight into where species came into contact and possibly hybridized throughout the Pleistocene. Most notably, the bull trout and the Arctic lineage probably experienced secondary contact in the Mackenzie River drainage while Dolly Varden likely interacted with Arctic char along the Siberian coast. Conversely, lack of reciprocal monophyly may represent incomplete lineage sorting (Avise 2000) or indicate rapid morphological diversification (Savvaitova 1995). Comparative nuclear phylogenies incorporating the same species throughout their geographic ranges are required to further test these hypotheses. Using multiple nuclear markers is also useful for identifying genealogical concordance since selection could be acting differently across loci. Salvelinus exhibit a deep and shallow phylogenetic history similar to other northwestern species complexes such as Sorex (Demboski and Cook 2001), and illustrate how dynamic interactions across broad geographic area involving multiple events of introgressive hybridization can produce trans-species polymorphisms (Klein et al. 1998). Therefore, the effect 101  of introgressive hybridization on molecular and phenotypic evolution remains a topic for future research. Evolutionary significance of introgressive hybridization  Although once considered evolutionarily insignificant by many zoologists, the extent to which hybridization and introgression generates increased genetic variation for evolutionary forces to act upon is receiving increased attention (Arnold 1997; Rieseberg 1998; Grant and Grant 2000). Northern freshwater fish hybridize more than any other vertebrates (Hubbs 1955), and clearly char are no exception. It is, therefore, important to consider the role hybridization has played in generating adaptive genetic variability that may explain, in part, the often bewildering number of taxa of char that have been described (e.g. Behnke 1980). The life history and resource use polymorphisms exhibited by char are comparable to those observed in African cichlids and Galapogos finches that have experienced relatively recent adaptive radiations. The ability to shift to new resource environments requires heritable additive genetic variation for directional selection to act on or can result from a punctuated event whereby hybrids fill an unexploited niche. Hybridization as a historical evolutionary force is thought to provide a means by which populations can change ecological zones (DeMaris et al. 1992) because the production of novel recombinant genotypes through hybridization is expected to occur more rapidly than through mutation alone (Dowling et al. 1997). Grant and Grant (2000) quantified the production of phenotypic additive genetic variation through introgression in species of Darwin's finches and found that Geopeszia scandens is more variable phenotypically as a result of hybridizing with G. fortis than in the absence of hybridization. Likewise, hybridization had a larger effect on the phenotype of G. fortis because it receives genes from both G. scandens and G. fulginosa. In cichlids, experimental hybrids were shown to have a novel jaw morphology, distinct from either parent species, suggesting that interspecific hybridization can produce a phenotypic shift large enough to reduce competition with parental 102  taxa and potentially form a new evolutionary lineage (McElroy and Kornfield 1993). It is not known if hybridization is responsible for production of distinct evolutionary lineages, but it is likely that hybridization has acted as a potent evolutionary force in the diversification char producing a "nightmare for systematics, and paradise for the evolutionary biologist" (Mayr 1947; cited in Savvaitova 1995).  103  References Alexandrino, J., J. W. Arntzen, and N . Ferrand. 2002. Nested clade analysis and genetic evidence for population expansion in the phylogeography of the golden-striped salamander, Chioglossa lusitanica (Amphibia: Urodela). Heredity 88:66-74. Alves, M . J., H. Coelho, M . J. Collares-Pereira, and M . M . Coelho. 2001. Mitochondrial D N A variation in the highly endangered cyprinid fish Anaecypris hispanica: importance for conservation. Heredity 87:463-473. Arnold, M . L. 1997. Natural hybridization and evolution. Oxford University Press, New York. Avise, J. A . 1992. Molecular population structure and the biogeographic history of a regional fauna: A case history with lessons for conservation biology. Oikos 63:62-76. Avise, J. A . 1994. Molecular Markers, Natural History, and Evolution. Chapman & Hall, N Y . Avise, J. A . 2000. Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, M A . Ballard, J. W., and M . Kreitman. 1995. Is mitochondrial D N A a strictly neutral marker? Trends in Ecology and Evolution 10:485-488. Ballard, J. W., and M . Whitlock. 2003. The incomplete natural history of mitochondria. Molecular ecology, in press. Baxter, J. S., E. B. Taylor, R. H. Devlin, J. Hagen, and J. D. McPhail. 1997. Evidence for natural hybridization between Dolly Varden (Salvelinus malma) and bull trout (Salvelinus confluentus) in a northcentral British Columbia watershed. Canadian Journal of Fisheries and Aquatic Science 54:421-429. Beerli, P., and J. Felsenstein. 2001. Maximum likelihood estimation of migration rates and effective population numbers in two populations using a coalescent approach. Genetics 152:763-773. Behnke, R. J. 1972. The systematics of salmonid fishes of Recently glaciated lakes. Journal Fish. Res. Bd. Canada 29:639-670. Behnke, R. J. 1980. A systematic Review of the genus Salvelinus. Pp. 441-480 in E. K . Balon, ed. Charrs: salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague. Behnke, R. J. 1984. Organizing the diversity of the Arctic charr complex in L. J. a. B. L . Burns, ed. Biology o the Arctic charr, Proceedings of the International Symposium on Arctic Charr, Winnipeg, Manitoba, May 1981. Univ. Manitoba Press, Winnipeg.  Bermingham, E., and J. A . Avise. 1986. Molecular zoogeography of freshwater fishes of the southeastern United States. Genetics 113:939-965. 104  Bernatchez, L. 2001. The evolutionary histroy of brown trout (Salmo trutta L.) inferred from phylogeographic, nested clade and mismatch analyses of mitochondrial D N A variation. Evolution 55:351-379. Bernatchez, L., and B. Angers. 1998. Combined use of S M M and non-SMM methods to fine structure and evolutionary history of brook char (Salvelinus fontinalis, Salmonidae) populations from microsatellites. Molecular Biology and Evolution 15:143-159. Bernatchez, L., and R. G. Danzmann. 1993. Congruence in control-region sequence and restriction-site variation in mitochondrial D N A of brook charr (Salvelinus fontinalis Mitchill). Molecular Biology and Evolution 10:1002-1014. Bernatchez, L., and J. J. Dodson. 1990. Allopatric origin of sympatric populations of lake whitefish (Coregonus clupeaformis) as revealed by mitochondrial-DNA restriction analysis. Evolution 44: 1263-1271. Bernatchez, L., and J. J. Dodson. 1991. Phylogeographic structure in mitochondrial D N A of the lake whitefish (Coregonus clupeaformis) and its relation to Pleisotcene glaciations. Evolution 45:1016-1035. Bernatchez, L., H . Glemet, C. C. Wilson, and R. G. Danzmann. 1995. Introgression and fixation of Arctic char (Salvelinus alpinus) mitochondrial genome in an allopatric population of brook trout (Salvelinus fontinalis). Canadian Journal of Fisheries and Aquatic Sciences 52:179-185. Bernatchez, L., R. Guymonard, and F. Bonhomme. 1992. D N A sequence variation of the mitochondrial control region among geographically and morphologically remote European brown trout Salmo trutta populations. Molecular Ecology 1:161-173. Bernatchez, L., J. A . Vuorinen, R. A. Bodaly, and J. J. Dodson. 1996. Genetic evidence for reproductive isolation and multiple origins of sympatric trophic ecotypes of whitefish (Coregonus). Evolution 50:624-635. Bernatchez, L., and C. C. Wilson. 1998. Comparative phylogeography of Nearctic and Palearctic fishes. Molecular Ecology 7:431-452. Bridle, J. R., S. J. E. Baird, and R. K. Butlin. 2001. Spatial structure and habitat variation in a grasshopper hybrid zone. Evolution 55:1832-1843. Brunner, P. C , M . R. Douglas, A. Osinov, C. C. Wilson, and L. Bernatchez. 2001. Holarctic phylogeography of arctic charr (Salvelinus alpinus L.) inferred from mitochondrial D N A sequences. Evolution 55:573-586. Brunner, P. C , M . R. Douglas, A. Osinov, C. C. Wilson, and L. Bernatchez. 2001. Holarctic phylogeography of arctic charr (Salvelinus alpinus L.) inferred from mitochondrial D N A sequences. Evolution 55:573-586.  105  Brunsfeld, S. J., S. J., D. E. Soltis, and P. S. Soltis. 2002. Comparative phylogeography of northwestern North America: a synthesis. Pp. 319-339. Caan, R. L., M . Stoneking, and A. C. Wilson. 1987. Mitochondrial D N A and human evolution. Nature 325:31-36. Castelloe, J., and A. R. Templeton. 1994. Root Probabilities for intraspecific gene trees under neutral coaslescent theory. Molecular Phylogenetics and Evolution 3:102-113. Cavalli-Sforza, L. L., P. Menozzi, and A. Piazza. 1994. The history and geography of human genes. Princeton University Press, Princeton. Cavender, T. M . 1978. Taxonomy and distribution of the bull trout, Salvelinus confluentus (Suckley), from the American Northwest. California Fish and Game 64:139-174. Cavender, T. M . 1980. Systematics of Salvelinus from the North Pacific Basin in E. K . Balon, ed. Charrs, salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague, the Netherlands. Cavender, T. M . 1997. Morphological distinction of bull trout from the McCloud River system to northern California. Friends of the Bull Trout Conference Proceedings, Calgary , Canada 271-282. Cavender, T. M . , and S. Kimura. 1989. Cytotaxonomy and interrelationships of Pacific basin Salvelinus. Physiol. Ecol. Jap. Spec. Vol. 1:49-68. Clague, J. J. 1981. Late Quaternary geology and geochronology of British Columbia. Part 2: Summary and discussion of radiocarbondated Quaternary history. Geological Survey of Canada, Paper 80-35, 41 pp. Clague, J. J. 1991. Quaternary glaciation and sedimentation, Chapt. 12. Pp. 419-434 in G. Gabrielse and Y. C.J., eds. Geology of the Cordilleran Orogen in Canada. Geological Survery of Canda, Geology of Canada. Clement, M . , D. Posada, and K. A . Crandall. 2000. TCS: a computer program to estimate gene geneologies. Molecular Ecology 9:1657-1659. Costello, A. B., T. E. Down, S. M . Pollard, C. J. Pacas, and E. B . Taylor. 2003. The influence of history and contemporary stream hydrology on the evolution of genetic diversity withing species: an examination of microsatellite D N A variation in bull trout, Salvelinus confluentus (Pisces: Salmonidae). Evolution 57:328-344. Crandell, K. A . 1994. Intraspecific cladogram estimation: Accuracy at higher levels of divergence. Systematic biology 43:222-235. Crandell, K. A., and A. R. Templeton. 1993. Empirical test of some predication from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics 134:959-969.  106  Crane, P. A., L. W. Seeb, and J. E. Seeb. 1994. Genetic relationships among Salvelinus species inferred from allozyme data. Canadian Journal of Fisheries and Aquatic Sciences 51:182197. Cruzan, M . B., and A. R. Templeton. 2000. Paleoecology and coalescence: phylogeographic analysis of hypotheses from the fossil record. Trends in Ecology and Evolution 15:491495. Deagle, B . E., T. E. Reimchen, and D. B . Levin. 1996. Origins of endemic stickleback from the Queen Charlotte Islands: Mitochondrial and morphological evidence. Canadian Journal of Zoology 74:1045-1056. DeMaris, B. D., T. E. Dowling, M . E. Douglas, W. L . Minckley, and P. Marsh, C. 1992. Origin of Gila seminuda (Teleostei: Cyprinidae) through introgressive hybridization: Implications for evolution and conservation. Proceedings National Academy of Science 89:2747-2751. Demboski, J. R., and J. A. Cook. 2001. Phylogeography of the dusky shrew, Sorex monticolus (Insectivora, Soricidae): insights into deep and shallow history in northwestern North America. Molecular Ecology 10:1227-1240. Dowling, T. E., R. E. Broughton, and B. D. DeMaris. 1997. Significant role for historical effects in the evolution of reproductive isolation: Evidence from patterns of introgression between the cyprinid fishes Luxilus cornutus and Luxilus chrysocephalus. Evolution 51:1574-1583. Durand, J. D., A . R. Templeton, B. Guinand, A . Imsiridou, and Y . Bouvet. 1999. Nested clade and phylogeographic analyses of the chub Leucisus cephalus (Teleostei, Cyprinidae), in Greece: Implications for Balkan Peninsula biogeography. Molecular Phylogenetics and Evolution 13:566-580. Efron, B., E. Halloran, and S. Holmes. 1996. Boostrap confidence level for phylogenetic trees. Proceedings National Academy of Science 93:13429-13434. Emerson, B. C , E. Paradis, and C. Thebaud. 2001. Revealing the demographic histories of species using D N A sequences. Trends in Ecology and Evolution 16:707-716. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791. Felsenstein, J. 1995. PHYLIP (Phylogeny Inference Package). Department of Genetics, University of Washington, Seattle, W A . Frolov, S. V . 2001. Karyological differences between northern Dolly Varden {Salvelinus malma) and white char (S. albus) from the Kamchatka River basin. Russian Journal of Genetics 37:269-275. Gava, A., and T. R. O. Freitas. 2003. Inter and intra-specific hybridization in tuco-tucos (Ctenomys) from Brazilian coastal plains (Rodentia : Ctenomyidae). Genetica 119:11-17. 107  Givet, D., and R. J. Petit. 2002. Phylogeography of the common ivy (Hedera sp.) in Europe: genetic differentiation through space and time. Molecular Ecology 11: 1351-1362 Glemet, H., P. Blier, and L . Bernatchez. 1998. Geographical extent of Arctic char (Salvelinus alpinus) mtDNA introgression in brook char populations (S. fontinalis) from eastern Quebec, Canada. Molecular Ecology 7:1655-1662. Goldman, N . , J. P. Anderson, and A. Rodrigo. 2000. Likelihood-based test of topologies in phylogenetics. Systematic biology 49:652-670. Gomez-Zurita, J., E. Petitpirre, and C. Juan. 2000. Nested cladistic analysis, phylogeography, and speciation in the Timarcha goettingensis complex (Coleoptera, Chrysomelidae). Molecular Ecology 9:557-570. Grant, P. R., and B. R. Grant. 2000. Quantitative genetic variation in populations of Darwin's finches. Pp. 3-40 in T. A . Mousseau, B. Sinervo and J. A. Endler, eds. Adaptive Genetic Variation in the Wild. Oxford University Press, Oxford. Green, D. M . , and C. Parent. 2003. Variable and asymmetric introgression in a hybrid zone in the toads, Bufo americanus and Bufo fowleri. Copeia:34-43. Grewe, P. M . , N . Billington, and P. D. N . Herbert. 1990. Phylogenetic relationships among members of Salvelinus inferred from mitochondrial D N A divergence. Canadian Journal of Fisheries and Aquatic Science 47:984-991. Haas, G. R., and J. D. McPhail. 1991. Systematics and distribution of Dolly Varden (Salvelinus malma) and bull trout (Salvelinus confluentus) in North America. Canadian Journal of Fisheries and Aquatic Science 48:2191-2211. Haas, G. R., and J. D. McPhail. 2001. The post-Wisconsinan glacial biogeography of bull trout (Salvelinus confluentus): a multivariate morphometric approach for conservation biology and management. Canadian Journal of Fisheries and Aquatic Sciences 58:2189-2203. Hagen, J., and E. B . Taylor. 2001. Habitat partitioning as a factor limiting gene flow in hybridizing populations of Dolly Varden char (Salvelinus malma) and bull trout (S. confluentus). Canadian Journal of Fisheries and Aquatic Science 58:2037-2047. Hammer, M . F., T. Karafet, A. Rasanayagam, E. T. Wood, T. K . Altheide, T. Jenkins, R. C. Griffiths, A. R. Templeton, and S. L. Zegura. 1998. Out of Africa and back again: Nested cladistic analysis of hyman Y chromosome variation. Molecular Biology and Evolution 15:427-441. Hare, M . P. 2001. Prospects for nuclear gene phylogeography. Trends in Ecology and Evolution 16:700-706. Harpending, H. 1994. Signature of ancient population growth in a low resolution mitochondrial D N A mismatch distribution. Human Biology 66:591-600. 108  Harrison, R., G. 1993. Hybrids and hybrid zones: historical perspective. Pp. 3-12 in R. Harrison, G., ed. Hybrid Zones and the Evolutionary Process. Oxford University Press, Oxford. Harrison, R., G., and D. M . Rand. 1989. Mosaic hybrid zones and the nature of species boundaries. Pp. 111-133 in D. Otte and J. A. Endler, eds. Speciation and its Consequences. Sinauer Assoc., Sunderland, M A . Hatfield Consulting Ltd. 2001. Overview 1:50,000 fish and fish habitat inventory of 5 mainland coast watersheds. Ministry of Environment, Lands and Parks, Lower mainland region., Surrey. 70 pages. 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. 2000. The genetic legacy of the Quaternary ice ages. Nature 405:907-913. Higgins, D. D., A. J. Bleasby, and R. Fuchs. 1992. C L U S T A L V: Improved software for multiple sequence alignment. Comput. Appl. Biosci. 8:189-191. Hocutt, C , and E. O. Wiley. 1986. The zoogeography of North American freshwater fishes. John Wiley and Sons, New York. Howell, N . , I. Kubacka, and D. A. Mackey, 1996. How rapidly does the human mitochondrial genome evolve? American Journal of Human Genetics 59:501-509. Hubbs, C. L . 1955. Hybridization between fish species in nature. Systematic zoology 4:1-20. Hundertmark, K . J., G. F. Sheilds, I. G . Udina, R. T. Bowyer, A . A . Danilkin, and C. C. Schwartz. 2002. Mitochondrial Phylogeography of Moose (Alces alces): Late Pleistocen divergence and population expansion. Molecular Phylogenetics and Evolution 22:375387. Hurwood, D. A., and R. M . Hughes. 1998. Phylogeography of the freshwater fish Mogurnda adspera, in streams of northeastern Queensland, Australia: evidence for altered drainage patterns. Molecular Ecology 7:1507-1517. Johnson, J. B. 2002. Evolution after the flood: Phylogeography of the desert fish Utah chub. Evolution 56:948-960. Jiggins, C. D., and J. Mallet. 2000. Bimodal hybrid zones and speciation. Trends in Ecology and Evolution 15:250-255. Jonsson, B., and N . Jonsson. 2001. Polymorphism and speciation in Arctic charr. Journal of Fish Biology 58:605-638. Klein, J., A. Sato, S. Nagl, and C. O'hUgin. 1998. Molecular trans-species polymorphism. Annual Review of Ecology and Systematics 29:1-21.  109  Knowles, L. L., and W. P. Maddison. 2002. Statistical phylogeography. Molecular Ecology 11:2623-2635. Law, J. H., and B. J. Crespi. 2002. The evolution of geographic parthenogenesis in Timema walking-sticks. Molecular Ecology 11:1471-1489. Lindsey, C. C , and J. D. McPhail. 1986. Zoogeography of fishes of the Yukon and Mackenzie Basins. Pp. 639-674 in C. Hocutt and E. O. Wiley, eds. Zoogeography of North American freshwater fishes. John Wiley and Sons, New York. Lu, G., D. J. Basley, and L. Bernatchez. 2001. Contrasting patterns of mitochondrial D N A and microsatellite introgresssive hybridization between lineages of lake whitefish (Coregonus clupeaformis); relevance for speciation. Molecular Ecology 10:965-985. Maddison, W. P. 1997. Gene trees in species trees. Systematic biology 46:523-536. Malyarchuk, B. A . 2002. Problems of molecular systematics of the genus Salvelinus based on nucleotide sequence variability in the major non-coding region of mitochondrial D N A . Russian Journal of Genetics 38:1148-1154. Maskas, S. D., and M . B. Cruzan. 2000. Patterns of intraspecific diversification in the Piriqueta caroliniana complex in southeastern North America and the Bahamas. Evolution 54:815827. Masta, S. E., N . M . Laurent, and E. Routman. 2003. Population genetic structure of the toad Bufo woodhousii: an empirical asessment of the effects of haplotype extinction on nested clade analysis. Molecular Ecology 12:1541-1554. Matos, J. A., and B. A . Schaal. 2000. Choroplast evolution in the Pinus montezumae complex: A coalescent approach to hybridization. Evolution 54:1218-1233. Mayr, E. 1963. Animal Species and Evolution. Harvard University Press, Cambridge, M A . McCart, P., and P. Craig. 1980. Meristic differences between anadromous and freshwaterresident Arctic char (Salvelinus alpinus) in the Sagavanirktok River drainage, Alaska. Journal of the Fisheries Research Board of Canada 28:115-118. McCusker, M . R., E. Parkison, and E. B . Taylor. 2000. Mitochondrial D N A variation in rainbow trout (Oncorhynchus mykiss) across its native range: testing biogeographical hypotheses and their relevance to conservation. Molecular Ecology 9:2089-2108. McElroy, D . M . , and I. Kornfield. 1993. Novel jaw morphology in hybrids between pseudotropheus zebra and Labeotropheus fuelleborni (Teleosteixichlidae) from Lake Malawi, Africa. Copeia 4:933-945. McPhail, J. D . 1961. Study of the Salvelinus alpinus complex in North America. Journal Fish. Res. Bd. Canada 173:1- 381.  110  McPhail, J. D. 1967. Distribution of freshwater fishes in western Washington. Northwest Science 41:1-11. McPhail, J. D., and C. C. Lindsey. 1970. Freshwater fishes of northwest Canada and Alaska. Journal Fish. Res. Bd. Canada: 1-381. McPhail, J. D . 1993. Ecology and evolution of sympatric sticklebacks (Gasterosteus): origin of the species pairs. Canadian Journal of Zoology 71:515-523. McPhail, J. D. 1997. Status of the Nooksack Dace. The Canadian Field-Naturalist 111:258-262. McPhail, J. D., and C. C. Lindsey. 1986. Zoogeography of the freshwater fishes of Cascadia (the Columbia system and rivers north to the Stikine). Pp. 615-637 in C. W. Hocutt, E.O, ed. Zoogeography of North American Freshwater Fishes. Wiley and Sons, New York. McPhail, J. D., and E. B . Taylor. 1995. The Final Report to the Skagit Environmental Endowment Commission: Skagit Char Project (Project 94-1). McPhail, J. D., and E. B . Taylor. 1999. Morphological and genetic variation in Northwestern longnose suckers Catostomus catostomus: The Salish sucker problem. Copeia 4:884-893. Montgomery, D. R., and Ff. M . Greenberg. 2000. Local relief and the height of Mount Olympus. Earth Surface Processes and Landforms 25:385-396. Morrow, J. E. 1980. Analysis of the Dolly Varden char Salvelinus malma, of northwestern North America and northeastern Siberia. Pp. 323-338 in E. K . Balon, ed. Charrs, Salmonid Fishes of the Genus Salvelinus. Dr. W. Junk bv Publishers, The Netherlands. Nei, M . 1987. Molecular Evolutionary Genetics. Columbia University Press, New York. Nichols, R. 2001. Gene trees ad species trees are not the same. Trends in Ecology and Evolution 16:358-364. Noor, M . A . 1995. Speciation by natural selection in Drosophila. Nature 375:674-675. Park, L. K., M . A . Brainward, D.A. Dightman, G. A . Winans. 1993. Low levels of intraspecific variation in the mitochondrial D N A of chum salmon (Oncorhynchus keta). Marine Biology and Biotechnology 2:362-370. Paulo, O. S., W. C. Jordan, M . W. Bruford, and R. A. Nichols. 2002. Using nested clade analysis to assess the history of colonization and the persistence of populations of an Iberian Lizard. Molecular Ecology 11:809-819. Petren, K., B . R. Grant, and P. R. Grant. 1999. A phylogeny of Darwin's finches based on microsatellite D N A length variation. Proceedings of the Royal Society of London Series B-Biological Sciences 266:321-329.  Ill  Pfenninger, M . , and D. Posada. 2002. Phylogeographic history of the land snail Candidula unifasiata (Helicellinae, Stylommatophora): fragmentation, corridor migration, and secondary contact. Evolution 56:1776-1788. Phillips, R. B., L. I. Gudex, K. M . Westrich, and A. L. DeCicco. 1999. Combined phylogenetic analysis of ribosomal ITS1 sequences and new chromosome data supports three subgroups of Dolly Varden char (Salvelinus malma). Canadian Journal of Fisheries and Aquatic Science 56:1504-1511. Phillips, R., S. L. Sajdak, and M . J. Domanico. 1995. Relationships among charrs based on D N A sequences. Nordic J. Freshwater Res. 71:378-391. Pielou, E. C. 1991. After the Ice Age: The Return of Life to Glaciated North America. University of Chicago Press, Chicago. Pleyte, K. A., S. D. Duncan, and R. B . Phillips. 1992. Evolutionary relationships of the salmonid fish genus Salvelinus inferred from D N A sequences of the first internal transcribed spacer (ITS 1) of ribosomal D N A . Molecular Phylogenetics and Evolution 1:223-230. Posada, D., and K. A . Crandall. 1998. MODELTEST: testing the model of D N A substituion. Bioinformatics 14:817-818. Posada, D., and K. A . Crandall. 2001. Intraspecific gene genealogies: trees grafting into networks. Trends in Ecology and Evolution 16:37-45. Posada, D., K . A. Crandell, and A. R. Templeton. 2000. GeoDis: a program for the cladistic nested analysis of the geographical distribution of henetic haplotypes. Molecular Ecology 9:487-488. Radchenko, O. A. 2002. Genetic differentiation inferred from data on restriction analysis of mitochondrial D N A in the northern and southern forms of the Dolly Varden char. Russian Journal of Genetics 38:421-428. Rand, D. M . 2001. Mitochondrial genomics flies high. Trends in Ecology and Evolution 16:2-4. Redenbach, Z. 2000. Molecular evidence of current and historical introgressive hybridization between bull trout (Salvelinus confluentus) and Dolly Varden (Salvelinus malma). Pp. 140. Zoology. University of British Columbia, Vancouver. Redenbach, Z., and E. B . Taylor. 2002. Evidence for historical introgression along a contact zone between two species of char (Pisces:Salmonidae) in northwestern North America. Evolution 56:1021-1035. Redenbach, Z., and E. B. Taylor. 2003. Evidence for bimodal hybrid zones between two species of char (Pisces: Salmonidae: Salvelinus) in northwester North America. Journal of Evolutionary Biology, in press.  112  Reist, J. D., D. Johnson, and T. J. Carmichael. 1991. Morphological variation in charr from northwestern Canada- Salvelinus alpinus and S. malma. Proceedings of the Sixth ISACF workshop on Arctic char, Drottningholm, Sweden 135-145. Reist, J. D., D. Johnson, and T. J. Carmichael. 1997. Variation and specific identity of char from northwestern Arctic Canada and Alaska. American Fisheries Society Symp. Spec. Publ. 19:250-261. Reist, J. D., G. Low, J. D. Johnson, and D. McDowell. 2002. Range extension of bull trout, Salvelinus confluentus, to the central Northwest Territories, with notes on identification and distribution of Dolly Varden, Salvelinus malma, in the western Canadian Arctic. Arctic 55:70-76. Remington, C. L . 1968. Suture-zones of hybrid interactions between recently joined biotas. Evolutionary Biology 2:321-428. Rieseberg, L. H . 1998. Molecular ecology of hybridization. Pp. 243-265 in G. R. Carvalho, ed. Advances in Molecular Ecology. IOS Press. Robinson, B. W., and D. Schluter. 2000. Natural selection and the evolution of adaptive genetic variation in northern freshwater fishes. Pp. 65-94 in T. A . Mousseau, B. Sinervo and J. A . Endler, eds. Adaptive Genetic Variation in the Wild. Oxford University Press, Oxford. Rodriguez, F., J. F. Oliver, A. Mable, and M . J.R. 1990. The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142:485-501. Rogers, A. R. 1995. Genetic evidence for a Pleistocene population explosion. Evolution 49:608615. Rogers, A . R., and H. Harpending. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9:552-569. Rosas, J., and R. Rozas. 2000. DnaSP Version 3.5: a novel software package for extensive molecular population genetics analysis. Computer Applications in Bioscience 13:307311. Rundle, H. R., L. Nagel, J. W. Boughman, and D. Schluter. 2000. Natural selection and parallel speciation in sympatric sticklebacks. Science 287:306-308. Salmenkova, E. A., V . T. Omelchenko, A. A . Kolesnikov, and T. V . Malinina. 2000. Genetic differentiation of charrs in the Russian north and far east. Journal of Fish Biology 57:136157. Savvaitova, K . A . 1980. Comments to the "systematic review of the genus Salvelinus". Pp. 480481 in E. K . Balon, ed. Charrs: the salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague.  113  Savvaitova, K. A . 1980. Comments to the "systematic review of the genus Salvelinus". Pp. 480481 in E. K . Balon, ed. Charrs: the salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague. Savvaitova, K . A . 1995. Patterns of diversity and processes of speciation in Arctic char. Nordic J. Freshwater Res. 71:81-91. Schluter, D. 1996. Ecological speciation in postglacial fishes. Proceedings of the Royal Society of London B 351:807-814. Schneider, S., D. Roessli, and L. Excoffier. 2001. Arlequima software package for population genetics data analysis. Department of Antropology, University of Geneva, Switzerland, Geneva. Schneider-Broussard, R., D. L. Felder, C. A . Chlan, and J. Neigel. 1998. Tests of phylogeographic models with nuclear and mitochondrial D N A sequence variation in the stone crabs, Menippe adina and Menippe mercenaria. Evolution 52:1671-1678. Seddon, J. M . , F. Santucci, J. Reeve, and G. M . Hewitt. 2001. D N A footprints of European hedgehogs, Erubaceus europaeus and E. concolor. Pleistocene refugia, postglacial expansion and colonization routes. Molecular Ecology 10:2187-2198. Shaw, K. L . 2002. Conflict between nuclear and mitochondrial D N A phylogenies of a recent species radiation: What mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proceedings National Academy of Science 99:16122-16127. Shimodaira, H . 2002. A n approximately unbiased test of phylogenetic tree selection. Systematic Biology 51:492-508. Shimodaira, H., and M . Hasegawa. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16:1114-1116. Shimodaira, H., and M . Hasegawa. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17:1246-1247. Slatkin, M . , and R. Hudson. 1991. Pairwise comparisons of mitochondrial D N A sequences in stable and exponentially growing populations. Genetics 129:555-562. Smith, G. R. 1992. Introgression in fishes: significance for paleontology, cladistics, and evolutionary rate. Systematic Biology 41:41-57. Smith, C. T., R. J. Nelson, C. C. Wood, and B. F. Koop. 2001. Glacial biogeography of North American coho salmon (Oncorhynchus kisutch). Molecular Ecology 10:2775-2785. Soltis, D. E., M . A . Gitzendanner, D. D. Strenge, and S. P.S. 1997. Chloroplast D N A intrspecific phylogeography form the Pacific Northwest of North America. Plant Systematics and Evolution 206:353-373.  114  Sota, T., R. Ishikawa, M . Ujiie, F. Kusumoto, and P. Vogler. 2001. Extensive trans-species mitochondrial polymorphisms in the carabid beetles Carabus subgenus Ohomopterus caused by repeated introgressive hybridization. Molecular Ecology 10:2833-2847. Stamford, M . D . and E.B. Taylor. 2003. Evidence for multiple phylogeographic lineages of Arctic grayling (Thymallus arcticus) in North America. Molecular Ecology, In review. Stone, K. D., and J. A . Cook. 2000. Phylogeography of black bears (Ursus americanus) of the Pacific Northwest. Canadian Journal of Zoology 78:1218-1223. Swofford, D . L. 2000. PAUP*. Phylogenetic Analysis using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, M A . Swofford, D. L., G. J. Olsen, P. J. Waddell, and D. M . Hillis. 1996. Phylogenetic inferences. Pp. 407-514 in D. M . Hillis, C. Moritz and B. K . Mable, eds. Molecular Systematics. Sinauer Associates, Sunderland, M A . Taberlet, P., L. Fumagalli, A. G. Wust-Saucey, and J. F. Cosson. 1997. Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology:453564. Taylor, E. B. 1999. Species pairs of north temperate freshwater fishes: Evolution, taxonomy, and conservation. Reviews in Fish Biology and Fisheries 9:1-26. Taylor, E. B. 2003. Evolution in mixed company: evolutionary inferences from studies of natural hybridization in Salmonidae. Pp. in press in A. P. Hendry and S. Stearns, eds. Evolution in Salmonids. Oxford University Press, Oxford. Taylor, E. B., and P. Bentzen. 1993. Evidence for mulitple origins and sympatric divergence of trophic ecotypes of smelt (Osmerus) in Northeastern North America. Evolution 47:813832. Taylor, E. B., C. J. Foote, and C. C. Wood. 1996. Molecular genetic evidence for parallel lifehistory evolution within a Pacific salmon (sockeye salmon and kokanee, Oncorhynchus nerka). Evolution 50:401-416. Taylor, E. B., and J. D. McPhail. 2000. Historical contingency and ecological determinism interact to prime speciation in sticklebacks (Gasterosteus). Proceedings of the Royal Society of London B 267:2375-2384. Taylor, E. B., S. M . Pollard, and D. Louie. 1999. Mitochondrial D N A variation in bull trout (Salvelinus confluentus) from northwestern North America: implications for zoogeography and conservation. Molecular Ecology 8:1155-1170. Taylor, E. B., Z. Redenbach, A. B. Costello, S. M . Pollard, and C. J. Pacas. 2001. Nested analysis of genetic diversity in northwestern North American char, Dolly Varden (Salvelinus malma) and bull trout (Salvelinus confluentus). Canadian Journal of Fisheries and Aquatic Science 58:406-420. 115  Teel, D. M . , GB; Winans, G A ; Grant, WS. 2000. Genetic population structure and origin of life history types in chinook salmon in British Columbia, Canada. Trans. Am. Fish. Soc. 129:194-209. Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7:381-398. Templeton, A. R. 2001. Using phylogeographic analyses of gene trees to test species status and processes. Molecular Ecology 10:779-791. Templeton, A. R., E. Routman, and C. Phillips. 1995. Separating population structure from population history: A cladistic analysis of the geographical distribution of mitochondrial D N A haplotypes in the tiger salamander, Ambystoma tigrinum. Genetics 140:767-782. Templeton, A . R., and C. F. Sing. 1993. A cladistic analysis of phenotypic association with haplotypes inferred from restriction endonuclease mapping. IV. Nested analyses with cladogram uncertainty and recombination. Genetics 134:659-669. Thorson, R. M . 1980. Ice-sheet glaciation of the Puget lowland, Washington, during the Vashon state (late Pleistocene). Quaternary Research 13:303-321. Travisano, M . , J. A. Mongold, A. F. Benett, and R. Lenski. 1995. Experimental test of the roles of adaptation, chance, and history in evolution. Science Wash. 267:87-90. Turgeon, J., and L. Bernatchez. 2001. Mitochondrial D N A phylogeography of lake cisco {Coregonus artedi): evidence supporting extensive secondary contacts between two glacal races. Molecular Ecology 10:987-1001. Turner, T. F., J. C. Trexler, D. N . Kuhn, and H. W. Robison. 1996. Life-history variation and comparative phylogeography of darters (Pisces: Percidae) from the North American central highlands. Evolution 50:2023-2036. Vila, C , I. R. Amorim, D. Leonard, D. Posada, F. Castroviejo, F. Petrucci-Fonseca, A. Crandall, H. Ellegren, and R. K . Wayne. 1999. Mitochondrial D N A phylogeography and population history of the grey wolf Canis lupus. Molecular Ecology 8:2089-2103. Waters, J. M . 1994. Mitochondrial D N A variation suggest river capture as a source of vicariance in Gadopsis bispinosus (Pisces: Gadopsidae). Journal of Fish Biology 44:549-551. Waters, J. M . , D. Craw, J. H . Youngson, and G. P. Wallis. 2001. Genes meet geology: Fish phylogeographic pattern reflects ancient, rather than modern, drainage connections. Evolution 55:1844-1851. Waters, J. M . , and G. P. Wallis. 2000. Across the Southern Alps by river capture? Freshwater fish phylogeography in South Island, New Zealand. Molecular Ecology 9:1577-1582. Westrich, K. M . , N . R. Konkol, M . P. Matsuoka, and R. B. Phillips. 2002. Interspecific relationships among chairs based on phylogenetic analysis of nuclear growth hormone intron sequences. Environmental Biology of Fishes 64:217-222. 116  Wilson, A. C , R. L. Cann, and S. M . Carr. 1985. Mitochondrial D N A and two perspectives on evolutionary genetics. Biological Journal of the Linnean Society 26:375-400. Wilson, C. C , and P. D. N . Herbert. 1993. Natural hybridization between Arctic char (Salvelinus alpinus) and lake trout (Salvelinus namaycush) in the Canadian Arctic. Canadian Journal of Fisheries and Aquatic Science 50:2652-2658. Wilson, C. C , and P. D. N . Herbert. 1998. Phylogeography and postglacial dispersal of lake trout (Salvelinus namaycush) in North America. Canadian Journal of Fisheries and Aquatic Science 55:1010-1024. Wilson, C. C , P. D . N . Herbet, J. D. Reist, and J. B. Dempson. 1996. Phylogeography and postglacial dispersal of arctic char Salvelinus alpinus in North America. Molecular Ecology 5:198-197. Yang, Z. 1996. Among site rate variation and its impact on phylogentic analysis. Trends in Ecology and Evolution 11:367-372. Zamudio, K . R., and W. K . Savage. 2003. Historical isolation, range expansion, and secondary contact of two highly divergent mitochondrial lineages in spotted salamanders (Ambystoma maculatum). Evolution 57:1631-1652. Zink, R. M . 1996. Comparative phylogeography in North American birds. Evolution 50:308-330.  117  Appendix 3.1: Sample localities and number code (see Fig. 3.1), sample sizes, names give to mtDNA haplotypes resolved in phylogenetic analysis including haplotypes that are shared. Also given are the coordinates used for the nested geographical distance analysis with latitude north and longitude west unless otherwise noted. D V - N haplotypes notated BER# are from Brunner et al. (2001).  Bull trout Sample  n  haplotype  BT-I Keele R. N W T (1) Funeral Cr., N W T (2) Daugherty L., Yukon (3) Wolverine Lake, N . central B C (4) Chowika R., Upper Peace R. (5) Hotel Cr., Upper Liard R. (6) Nakina R. north coast B.C. (7) Naas tributaries, north coast B.C. (8) Telkwa R., Skeena R., B C (9) E. Buckley R., Skeena R. B C (10) W. Bulkley R., Skeena R. B C (11) Klinaklini R., southcoast B.C (12) Homathko R., southcoast, B.C. (13) Southgate River, southcoast B.C. (14) Upper Anderson, mid Fraser R., B C (15) Upper Kootenay R. (16) Salmo R. southeast B.C. (17) CallCr., U . Pine, BC(18) Flatbed L.., Murray R., B C (19) North Thomspson R., B C (20) Belly Cr. S. Saskatchewan R. A B (21) Mystic Cr. southwest A . B . (22) AthabascaR. A . B . (23) M i l l Cr., Walla Walla, W A (24) Phelps Cr., Wenatchee, W A (25) Early Winters Cr. Methow, W A (26) Tuccannon R. W A (27) Indian Cr, Yakima, W A (28) Flathead R. Basin M T (29)  1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1 BT-1  118  Coordinates Longitude Latitude 126 26 64 15 124 44 61 36 130 57 60 05 128 35 58 21 124 45 56 45 126 00 58 00 133 08 58 54 128 50 55 49 127 40 54 30 125 10 54 10 127 40 55 15 125 30 51 30 125 05 51 15 124 20 50 50 121 14 49 36 115 06 49 15 49 11 55 40 54 58 52 30 49 43 51 27 54 00 46 05 47 57 48 34 46 36 46 42 49 00  117 09 122 15 120 52 119 20 113 49 115 42 117 00 118 25 120 58 120 37 118 10 121 20 114 50  Jarbridge R, Snake R., N V (30) Boise Cr., Snake R. ID (31) Bear Cr., Snake R. W A (32)  41 50 43 49 45 50  117 40 117 00 118 06  1 BT-9 1 BT-9 1 BT-9 1 BT-7 2 BT-9 1 BT-9 1 BT-6 =DV-G 1 BT-6 1 BT-11 2 BT-10 =DV-J 2 BT-8 1 BT-8 1 BT-9 1 BT-12 3 BT-5  51 15 50 48 49 41 49 37 49 20 49 04 47 15 48 00 47 28 48 00 48 00 47 32 46 15 42 24 44 39  124 05 124 25 123 10 122 40 122 05 121 10 122 30 122 12 123 10 124 28 123 37 124 05 121 58 120 53 121 32  1 DV-E 1 DV-E 1 DV-E 1 DV-E 1 DV-E 1 DV-E 1 DV-F 1 D V - G =BT-6 1 DV-G 1 DV-G 1 DV-G 1 DV-H 1 DV-I 1 D V - J =BT-10  56 00 55 35 51 10 50 50 49 40 49 44 48 38 49 22 48 20 49 14 49 10 50 10 50 30 48 00  124 30 129 10 124 05 124 22 123 15 123 09 122 12 121 28 123 07 121 10 121 12 126 47 124 21 124 36  2 1 1  BT-1 BT-1 BT-2  BT-C  Homathko R., south coast, B.C. (33) Southgate R., south coast B.C. (34) Squamish R., south coast, B.C. (35) Upper Pitt R., Fraser R., B C (36) Chilliwack R., Fraser R., B C (37) Skagit R. south coast/Puget Sound (38) Puyallup R., Puget Sound, W A (39) Snohomish R., Puget Sound, W A (40) Skokomish R., Puget sound, W A (41) Soleduck R. Olympic Peninsula, W A (42) Elwah R. Olympic Peninsula W A (43) Queets R. Olympic Peninsula W A (44) Swift R., Lewis W A (45) Long Cr., Klamath R. OR (46) Metolius R., OR (47)  Dolly Varden  DV-S Omineca R., Upper Peace R., B C (48) Zolap R., Nass R., B C (49) Homathko R., south coast, B C (50) Southgate R., south coast, B C (51) M i l l Cr., South coast, B C (52) Mamquam R., south coast, B C (53) Nooksack R., Puget Sound, W A (54) Silverhope R., soutcoast, B C (55) Dungeness R., Olympic Peninsula W A (56) Sumallo R., Skagit R., B C (57) Klesikwa R., Skagit R., B C (58) Zeballos R. Vancouver Is., B C (59) Toba R., south coast, B C (60) Soleduck R., Olympic Peninsula, W A (61)  119  DV-N Aero R., Queen Charoltte Is. B C (62) Brent Cr. Queen Charlotte Is. B C (63) Klutina R., Copper R., A K . (64) Taku R., north coast, B C (65) Nooseneck River, mid-coast, B C (66) Tahlatan R., Stikine R., B C (67) Seutakan R., Chukotka Peninsula (68) Togiak R., Bristol Bay, A K (69) Ingenika R. north-central, B C (70) C o w R . Vancouver Is., B C (71) Ayton Cr., Skeena R., B C (72) Noyse Is., Skeena R., B C (73) Ogden Ch., Skeena R., B C (74) Kuril Islands, Russia (75) Pedro Ponds, Alaska (76) Paratuhka R., Kamchatka Peninsula, (77) Achen Lake, Chukotka Peninsula (78) Kamchatka R., Kamchatka Peninsula (79) Kamchatka R., Kamchatka Peninsula (79) Kuma R., E. Siberia Paramushir Isle (80) Kakhmauri R., Paramushir Isle (81) Clear Hatchery, Alaska (82) Prince William Sound, Alaska (83) Prince William Sound, Alaska (84) Auke Cr., Alaska (85)  1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2 1 2 2 2 1 1 2  DV-A DV-A DV-A DV-A DV-A DV-A D V - A =BER3 DV-B DV-C DV-C DV-C DV-C DV-C DV-C DV-D DV-Y DV-Z BER1 BER2 BER4 BER5 BER6 BER7 BER8 BER9  120  53 05 53 01 61 45 58 35 52 20 58 00 65 40 59 25 54 47 49 16 54 12 54 07 53 50 46 30 59 47 53 00 64 50 56 15 67 26 50 25 50 43 64 30 60 24 60 24 62 15  131 56 132 00 145 32 133 39 127 00 131 17 176 56 161 45 125 51 126 10 129 44 124 14 130 17 152 30 154 06 158 23 E 174 48 162 30 E 146 46 E 155 50 E 156 10 E 160 30 147 00 147 00 159 40  Ui  o o  CJ O  s c Q  c•  CJ  *H  Ui  Q  03  •a . 5 ^T3 S <u .2 CO ^ o « S 03 F  CO oci  1  o  CO o3 CU  X <D  P  o  o  oo  CJ  c 03  ob o P M  C  CO "D  •  .S? o 22  a O  S  s  Ui -"  i 00 CN 03  T-  f*l . oo _CJ 53 b •is 3 "a o3 O P 03 oo 03 .5. <*> C OH O 53 3 F^ 5 O C oo 4_.  OJ CVj  1  in  ^  rt  4—> , 03 D O  ;-H  -a c S 03  03  CJ  |  OH  r .  U  i  T3 oo  C C 03 o CJ  c j X£ o3 > scj P U  c3  T3  «  00 o3  P  oo  O CJ  CCS  co  I  o cn  3 g  V—  ob T  CS  «J  O  ~u  o c o -  M  ~  00  •T3 -t-  1  c\i cri CO CM  o  coni r^cn  Ui CV1 in  co CO OJ  CO CO <N CO  ,_  O  oi •cao —  •z.  C\J CO "O JO  0)  co ro  o  o co  S  T^-  CD  •a ra  LU OC  oj O  O  C u 03  00  c  g  oo  -d  CJ  00  o  (30 CJ  C  CJ  £  I OO  CJ  fc  1 2  73  CO  73  5  >>  V-  •C  03  03  X  1  c 2 O  o •a ra  CJ  .22  "c3  H  pO  I  CJ  w  ;3 Cu CJ  CJ  2 >  g- cl 53 t3 cl OH 03 2 -a 3  1  o cs b oo  CO  CO  OJ  CO  oo 03  OH  oo  or  w  Ui in r-~ _l _i in in ci .— o in co cri cncn Ui Ui -4 c riO2 o c r i coV ? Q cmnocnm co q oi 5 OJ  6 Pi. & " T3 F 00 -gCJ C 00 ii" o 'oo  o co ui  ~  ,  "Z  ^ cc OJ  "  LU  CO "D  CO "O  ra O  ra o  O  a s Oi  C^  W CO  cri oi o cd CO  ^j- CD CO CM  ^  i  co co iri o co cri T oo co CM T— CO  co  I -  cn  O)  T-  N  S  1  O  OI ^ O CO  O  O  00 T -  i-  CO ^1-  cr  M  «3  oo  CNJ  CO CO OJ  00  = 3o O^ cj cd PH C  M  x  p  CO  iri m co •  CO  00 --j  in  CO  CM  s "I C cj iscj 53C  S Q  — T co  OJ  CJ  k-H  o  oi  CD  Ul Ui  CO  33  CJ  -*-» OO  DC  co  m  CO <N  in <o  CJ  w C1 O FcS 53 j •3 .5 00 ^ c£0 C « b oo 03 F < c j -C OF3 ooov § ^ C 03 ^ T3 03 <*>  ro  JS  < = >2  s  O  ,CJ  •D  | ^ CD  CO  ra  •S  cu  OJ CO CO "D  Ui w  73  0 o •s c  co  O  C cj  O . -a cj oo  •  •>-> T3 oo o3  a-° T3  CJ  >,.SP p  O  . —  m  as-  oo  ^ ov CJ  O <N  co  CO  C? °  O  <g n S CM ^  UJ —  -p  i  oo  co  CO  CO  cn cn cn co CO  o _L •<*• ~Z O  " o ^j D ra _  -  CO  1 o  H  CN  CO  o 1_l _l O) 0 0  c Q  CN  m r~C\J  od T  ~  _l _i •<* CD CM CO CD oo C J r--  o Q  co  CN  4 CU h-  CU  TJ  co  TJ _!_  g c CO  E  O) CO  CO CJ CO  co  c Q  CM  CM CO CM  l< CM  r~ CO  _ l »  CO  o Q  CO  4  S So O z  o  1  co  CU  CM  I  re  ^  CD  T  "a — JS O  CO  a>  CM  TJ  JS  o CO CM TJ-  CO  o  u Q  _l <"  co  i*co  c Q  CM O CO CM  o i-  CO in  co oo  co  g  d co  cn co —  co  O  t--  CD  O  9  _l  J °  5 V  co „ CD  Q  a Z op  V  CO CM CO  CM  TJ  v  ro O  TJ  O  CO  •  Z  4  TJ  re O  CO CM  co  c Q  CO  o Q  in  TT  . cn co m co o cn *co ' 1  o cn m _l Q  Tf  i-  CO  T—  d  V  CO  TJ  LL  re o  O  c Q  cn  T-  N  d co s  (O O) O)  4  o  CM  — t  DC  CD  —i  "S c3  ro o Qj  oo  CD S in •<* oo <°  4  CM N  CM CO CO  LO  oo  cr o  i  V  5r  TJ  ro O  co o CD  N < CQ Q N  00  -'  CM LOC M  i o V ^ r;  CO  d »o o un o  TJ  "tf  CM  CM  O  Q  CU  JS  CO  ^ o oo oo in o ^ '  in cu  CM  TT  T>  4  ri  d d°  cu  cn cn  CD  in A  CJJ  V  CO  co •§  LL  LU h- >  CO — "O  f~  SJ 8S9  - CD cc co  TJ  CO  oo  CN u. CD  CN CU  CV  o co CO  C\l  h  OC CC —  LU LU o CO CQ Z  O  >  cn  T-  h  CC CC —  LU LU o CQ  mz  t <? CM  TJ  re  co —  "D J S O  CN CN  Appendix 3.3: Templeton's revised inference key 24 October 2001  Inference Key for the Nested Haplotype Tree Analysis of Geographical Distances Start with haplotypes nested within a 1-step clade: 1. Are there any significant values for Dc, Dn, or I-T within the clade? • NO - the null hypothesis of no geographical association of haplotypes cannot be rejected (either panmixia in sexual populations, extensive dispersal in non-sexual populations, small sample size, or inadequate geographical sampling). Move on to another clade at the same or higher level. • YES - Go to step 2. 2. Is at least one of the following conditions satisfied? a. The Dc 's for one or more tips are significantly small and the Dc's for one or more of the interiors are significantly large or non-significant. b. The Dc's for one or more tips are significantly small or non-significant and the Dc's for some but not all of the interiors are significantly small. c. The I-T Dc is significantly large. • N O - G o to step 11. • YES - Go to step 3. • Tip/Interior Status Cannot be Determined - Inconclusive Outcome. 3. Is at least one of the following conditions satisfied? a. Are any Dn and/or I-T Dn values significantly reversed from the Dc values? b. Do one or more tip clades show significantly large Dn 's with the corresponding Dc values being non-significant? c. Do one or more interior clades show significantly small Dn's with the corresponding Dc values being non-significant? d. Does I-T have a significantly small Dn with the corresponding Dc values being non-significant? • NO - Go to step 4. • YES - Go to step 5. 4. Are both of the following conditions satisfied (when no lower level nested clades have significant effects, answer only part a.)? a. Do the clades (or 2 or more subsets of them) with restricted geographical distributions have ranges that are completely or mostly non-overlapping with the other clades in the nested group (particularly interiors)? b. Does the pattern of restricted ranges represent a break or reversal from lower level trends within the nested series (applicable to higher-level clades only)? • N O - Restricted Gene Flow with Isolation by Distance (Restricted Dispersal by Distance in Non-sexual species). This inference is strengthened if the clades with restricted distributions are found in diverse locations, if the union of their ranges roughly corresponds to the range of one or more clades (usually interiors) within the same nested group (applicable only to nesting clades with many clade members or to the highest level clades regardless of number), and if the Dc values increase and become more geographically widespread with increasing clade level within a nested series (applicable to lower level clades only). • YES - Go to step 9. 5. Do the clades (or 2 or more subsets of them) with restricted geographical distributions have 123  ranges that are completely or mostly non-overlapping with the other clades in the nested group (particularly interiors), and does the pattern of restricted ranges represent a break or reversal from lower level trends within the nested series (applicable to higher-level clades only)? • N O - Go to step 6. • YES - Go to step 15. 6. Do clades (or haplotypes within them) with significant reversals or significant D n values without significant D c values (identified in step 3) satisfy one or more of the following conditions: a. They define two or more geographically concordant subsets (that is, they have similar geographic distributions). b. They are geographically concordant with other haplotypes/clades showing similar distance patterns. • No - Go to step 7. • YES - Go to step 13. • TOO F E W C L A D E S (< 2) TO D E T E R M I N E C O N C O R D A N C E - Insufficient Genetic Resolution to Discriminate between Range Expansion/Colinization and Restricted Dispersal/Gene Flow - Proceed to step 7 to determine if the geographical sampling is sufficient to discriminate between short versus long distance movement. 7. Are the clades with significantly large D n 's (or tip clades in general when D n for I-T is significantly small) separated from the other clades by intermediate geographical areas that were sampled? • N O - Go to step 8. • Y E S - Restricted Gene Flow/Dispersal but with some Long Distance Dispersal. 8. Is the species absent in the non-sampled areas? • NO - Sampling Design Inadequate to Discriminate between Isolation by Distance (Short Distance Movements) versus Long Distance Dispersal • YES - Restricted Gene Flow/Dispersal but with some Long Distance Dispersal over Intermediate Areas not Occupied by the Species. 9. Are the different geographical clade ranges identified in step 4 separated by areas that have not been sampled? • N O - Past Fragmentation. (If inferred at a high clade level, additional confirmation occurs if the clades displaying restricted by at least partially non-overlapping distributions are mutationally connected to one another by a larger than average number of steps.) • YES - Go to step 10. 10. Is the species absent in the non-sampled areas? • N O - Geographical Sampling Scheme Inadequate to Discriminate Between Fragmentation and Isolation By Distance. • YES - Allopatric Fragmentation. (If inferred at a high clade level, additional confirmation occurs if the clades displaying restricted by at least partially non-overlapping distributions are mutationally connected to one another by a larger than average number of steps.) 11. Is at least one of the following conditions satisfied? 124  a. The Dc value(s) for some tip clade(s) is/are significantly large. b. The Dc value(s) for all interior(s) is/are significantly small. c. The I-T Dc is significantly small. • N O - Go to step 17 • YES - Range Expansion, go to step 12. 12. Are any of the Dn and/or I-T Dn values significantly reversed from the Dc values? • N O - Contiguous Range Expansion. • YES - Go to step 13. 13. Are the clades with significantly large Dn 's (or tip clades in general when Dn for I-T is significantly small) separated from the geographical center of the other clades by intermediate geographical areas that were sampled? • NO - Go to step 14. • YES - Long Distance Colonization. 14. Is the species absent in the non-sampled areas? • N O - Sampling Design Inadequate to Discriminate between Contiguous Range Expansion and Long Distance Colonization. • YES - Long Distance Colonization. 15. Are the different geographical clade ranges identified in step 5 separated by areas that have not been sampled? • NO - Past Fragmentation. (If inferred at a high clade level, additional confirmation occurs if the clades displaying restricted by at least partially non-overlapping distributions are mutationally connected to one another by a larger than average number of steps.) • YES - Go to step 16. 16. Is the species absent in the non-sampled areas? • N O - G o to step 18. • YES - Allopatric Fragmentation. (If inferred at a high clade level, additional confirmation occurs if the clades displaying restricted by at least partially non-overlapping distributions are mutationally connected to one another by a larger than average number of steps.) 17. Is at least one of the following conditions satisfied? a. Are the Dn values for tip or some (but not all) interior clades significantly small? b. Are the Dn for one or more interior clades significantly large? c. Is the I-T Dn value significantly large? • N O - Inconclusive Outcome. • YES - Go to step 4. 18. Are the clades found in the different geographical locations separated by a branch length with a larger than average number of mutational steps. • NO - Geographical Sampling Scheme Inadequate to Discriminate Between Fragmentation, Range Expansion, and Isolation By Distance. • YES - Geographical Sampling Scheme Inadequate to Discriminate Between Fragmentation and Isolation By Distance. 125  Appendix 3.4: Calculation for time since population expansion.  In the expression x =2ut we need to estimate u which is the mutation rate of the entire region of sequence under study, not the mutation rate per nucleotide. To do this we use the formula of Nei and Tajima (1981 eq. 5 ) expressed as u=2\ik where k is the number of nucleotides (550) and 2(i is the nucleotide divergence rate estimated between 1% /Myr and 2% /Myr. With a generation time for these fish being approximately 5 years, then the that implies that atl%,  2u=0.01/10"  at 2%, 2p= 0.02/10"  6  6  u=5 X 10" /generation X 5  u=l X 10~ /generation X 5  (1=2.5 X 10" / generation  u=5.0 X 10" / generation  9  8  8  next solve for u:l%,  8  u = (2) 2.5 X IO" / gen.X (550 bp) = 2.75 X 10" 8  5  2%, u = (2) 5.0 X 10" / gen X (550 bp) = 5.5 X IO" 8  5  insert into equation T = 2ut  for T=1.4 (DV-S) the time since expansion t for a divergence rate of 1% /Myr is t = 1.4/ (2) (2.75 X 10" ) = 12,727 generations X 5 years/ generation =63,636 years 5  126  

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.831.1-0091184/manifest

Comment

Related Items