"Science, Faculty of"@en . "Zoology, Department of"@en . "DSpace"@en . "UBCV"@en . "Hagen, John"@en . "2009-07-06T23:59:58Z"@en . "2000"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "Experimental evidence and theoretical arguments support the notion that ecological\r\nfactors can be important in divergence and speciation. Empirical evidence suggests that\r\necology plays a role in divergence, but field studies addressing the importance of ecology\r\nin the evolution of reproductive isolation are few. Here I suggest that divergent natural\r\nselection is contributing to reproductive isolation between the closely related salmonid\r\nfish species Dolly Varden {Salvelinus malma) and bull trout (S. confluentus) in areas of\r\ntheir sympatry. In the Thutade watershed of northcentral British Columbia, hybridization\r\noccurs between these species naturally and produces viable, fertile offspring, suggesting\r\nthe sympatry is not being maintained by genetic incompatibility. Both forms are found in\r\nthe tributary streams, but resource use differences are subtle and insufficient to reduce the\r\nintensity of interspecific competition, suggesting no obvious basis, in this habitat, for\r\nnatural selection against phenotypically intermediate hybrids. Life history differences,\r\nhowever, are obvious. Bull trout undergo a niche shift that the obligatorily streamresident\r\nDolly Varden do not, migrating to Thutade Lake and becoming piscivorous. The\r\nsubstantial level of pre-mating isolation is probably caused by the large differences in\r\nbody size (and correlated spawning behaviours) that result from the life history\r\ndifferences, while selection against intermediate, hybrid phenotypes can also be\r\nreasonably expected due to the completely different physical and biological environments\r\nin lake and stream habitats. Two potential ecological mechanisms for the life history\r\ndifferences are supported: i) divergent natural selection for alternative resource\r\nenvironments, which is suggested by morphological adaptations by each species to\r\ndifferent-sized prey, and ii) divergent natural selection to reduce inter-specific\r\ncompetition, which is suggested by evidence for character displacement in life history."@en . "https://circle.library.ubc.ca/rest/handle/2429/10305?expand=metadata"@en . "4619208 bytes"@en . "application/pdf"@en . "Reproductive isolation between Dolly Varden (Salvelinus malma) and bull trout (5. confluentus) in sympatry: the role of ecological factors by JOHN HAGEN B.Sc. University of British Columbia, 1990 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S ' F O R T H E D E G R E E OF M A S T E R OF S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of Zoology) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A February 2000 \u00C2\u00A9 John Hagen In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^ZooloQ^ The University of British Columbia Vancouver, Canada Date February )5. J 2000 DE-6 (2/88) ABSTRACT Experimental evidence and theoretical arguments support the notion that ecological factors can be important in divergence and speciation. Empirical evidence suggests that ecology plays a role in divergence, but field studies addressing the importance of ecology in the evolution of reproductive isolation are few. Here I suggest that divergent natural selection is contributing to reproductive isolation between the closely related salmonid fish species Dol ly Varden {Salvelinus malma) and bull trout (S. confluentus) in areas of their sympatry. In the Thutade watershed of northcentral British Columbia, hybridization occurs between these species naturally and produces viable, fertile offspring, suggesting the sympatry is not being maintained by genetic incompatibility. Both forms are found in the tributary streams, but resource use differences are subtle and insufficient to reduce the intensity of interspecific competition, suggesting no obvious basis, in this habitat, for natural selection against phenotypically intermediate hybrids. Life history differences, however, are obvious. B u l l trout undergo a niche shift that the obligatorily stream-resident Dol ly Varden do not, migrating to Thutade Lake and becoming piscivorous. The substantial level of pre-mating isolation is probably caused by the large differences in body size (and correlated spawning behaviours) that result from the life history differences, while selection against intermediate, hybrid phenotypes can also be reasonably expected due to the completely different physical and biological environments in lake and stream habitats. Two potential ecological mechanisms for the life history differences are supported: i) divergent natural selection for alternative resource environments, which is suggested by morphological adaptations by each species to different-sized prey, and ii) divergent natural selection to reduce interspecific competition, which is suggested by evidence for character displacement in life history. ii T A B L E O F C O N T E N T S ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES vi ACKNOWLEDGEMENTS vii INTRODUCTION Models of speciation 1 Basic allopatry model 1 Reinforcement model 2 Sympatric speciation 3 Bottleneck speciation 4 Ecology and speciation in northern temperate zone fish assemblages 5 Taxonomy and distribution of Dolly Varden and bull trout in North America 8 Study questions 10 MATERIALS AND METHODS Study area , 12 Study design 13 Resource environments 13 Interspecific competition 16 Data Collection 17 Resource environments - strict sympatry 17 Life history 19 iii Density compensation 21 Analyses 21 Resource environments - strict sympatry 21 Life history 24 Density compensation 24 RESULTS Resource environments - strict sympatry 26 Habitat 26 Food. 27 Time 29 Life history 30 Density compensation 33 DISCUSSION The evidence for pre- and postmating isolation 36 Adaptation to different resource environments and reproductive isolation 37 Tributary stream environment 37 Life history differences 41 Interspecific competition and reproductive isolation 47 Tributary stream environment - density compensation 47 Character displacement in life history 48 Alternative hypotheses 52 Conclusion 54 REFERENCES 57 iv L I S T O F T A B L E S Table 1. Interspecific comparisons of microhabitat observations between Dol ly Varden and bull trout in strict sympatry 69 Table 2. Interspecific comparisons of food resource use between Dol ly Varden and bull trout in strict sympatry 70 Table 3. LoglO-transformed measures of diel feeding activity versus log io body size: interspecific comparisons ( A N C O V A ) between Dol ly Varden and bull trout in sympatry 71 Table 4. Measurements of trophic morphology versus body size: interspecific comparisons ( A N C O V A ) between Dol ly Varden and bull trout in strict sympatry 72 Table 5. Interspecific comparisons of reproductive biology of Dol ly Varden and bull trout in sympatry 73 Table 6. Surveys of spawning activities for bull trout in North Kemess Creek, 1994-1998 74 Table 7. Surveys of spawning activities for Dol ly Varden in North Kemess Creek and seepage-fed tributaries, 1994-1998 75 Table 8. Thutade Lake tributary reaches used in analysis of density compensation of one species for missing individuals of the other species (by ordinary least squares regression) 76 LIST OF FIGURES Figure 1. Distribution of Dol ly Varden (Salvelinus malma) and bull trout (S. confluentus) in western Canada, and location of study site 78 Figure 2. Life histories of Dol ly Varden and bull trout in the Thutade watershed... 80 Figure 3. Interspecific comparisons of habitat type use during (a) daytime and (b) nighttime observation periods by sympatric bull trout and Dol ly Varden 82 Figure 4. Interspecific comparisons of stream depth at focal point during (a) daytime and (b) nighttime observation periods for sympatric bull trout and Dol ly Varden 84 Figure 5. Interspecific comparisons of mean current velocity at focal point during (a) daytime and (b) nighttime observation periods for sympatric bull trout and Dol ly Varden 86 Figure 6. Observations of foraging mode during (a) daytime and (b) nighttime observation periods for sympatric bull trout and Dol ly Varden 88 Figure 7. Interspecific comparison of average prey item biomass for sympatric bull trout and Dol ly Varden 90 Figure 8. Interspecific comparison of taxonomic composition of daytime stomach contents for sympatric (a) bull trout and (b) Dol ly Varden 92 Figure 9. Interspecific comparison of taxonomic composition of nighttime stomach contents for sympatric (a) bull trout and (b) Dol ly Varden 94 Figure 10. Interspecific comparisons of total prey biomass of (a) daytime and (b) nighttime stomach samples for sympatric bull trout and Dol ly Varden....96 Figure 11. Interspecific comparisons of trophic morphology for sympatric bull trout and Dol ly Varden in terms of (a) mouth width and (b) gillraker spacing 98 Figure 12. Tributary reaches within the Thutade watershed used for investigation of Dol ly Varden/bull trout density compensation 100 Figure 13. Dol ly Varden/bull trout density compensation within independent stream reaches of the Thutade Lake watershed, British Columbia 102 Figure 14. Comparison of average total parr densities for both species combined between reaches of local allopatry and sympatry 104 vi ACKNOWLEDGEMENTS M y thanks and acknowledgements go first to my primary academic supervisor, Dr. Eric Taylor. M y decision to attend graduate school was a direct result of conversations I had with Rick over a two-year period leading up to my eventual enrollment. What won me over was his enthusiasm, both for these char species themselves and also for the beautiful wilderness area that is the Thutade Lake watershed, the setting we jointly envisioned for our study. A s a supervisor Rick has been perfect, allowing me the freedom to make decisions about what I wanted to study, yet helping wherever possible to ensure the project actually worked out. The other two members of my supervisory committee are Dr.'s. J. Don McPha i l and Dolph Schluter. With respect to my way of thinking about fish and ecology generally, these two have been tremendously important and complementary influences. Don's influence was to broaden my thinking, allowing me to see a great number of possibilities for interesting study (many more than I could manage!). A l l of Dolph's comments and writings have been extremely valuable, and he probably saved me a year's worth of work by charitably discussing with me some of my earlier, unworkable ideas. Fellow students and members of the Department of Zoology have been important teachers as well . I am indebted to Dr. Ernest Keely and Gordon Haas particularly. Ernest helped me with almost every aspect of the project, from the study design through to the statistical analysis, and assisted me in the field during the Thutade Lake survey. Gordon also visited the site, and furthermore helped me by making his extensive collection of information available to me. Gordon's most important contribution, however, was to get me interested in char biology in the first place, long before I was considering studying them in graduate school. I have benefited greatly from being part of a close and helpful lab group, the members of which are: A l l an Costello, Jason Ladell , Steve Latham, Megan McKusker , Dave O'Brien, Zoe Redenbach, Mike Stamford, Josh Taylor, and Peter Troffe. Zoe's project was a sister to my own, and molecular genetic data from her study provided the essential context for my results. Peter assisted me during the most grueling vii period of fieldwork, working most of the day and all of the night, while providing plenty of helpful advice besides. James Baxter, Jessica Bratty, Joel Sawada, and Shannon Stotyn are close friends of mine, and assisted me both by reviewing my presentations and also by shoring up my spirits. The potential for this study was first revealed during environmental impact assessments associated with the Kemess Mine. The individuals associated with these assessments are important collaborators in this study - it simply could not have been as extensive or as interesting without their contributions. Ted Down and Don Cadden of the British Columbia Ministry of Environment facilitated the project in a number of important ways, and were generally enthusiastic and helpful. Dave Bustard, Rob Dams, Graham McLean, and Kate Portman are the staff of the environmental consulting company Dave Bustard and Associates of Smithers. They are close friends of mine, and are responsible for the long-term population assessment database and life history observations that provide crucial data for my study. Essential to the success of this study was the co-operation and assistance of the environmental staff (formerly) of Royal Oak Mines, Inc. - Harold Bent, Larry Connell, and Matt Robson. They provided me with helicopter access to my study sites, which was essential given the remote nature of the sites but would have been impossible for me to organize on my own. Financial support for the study came primarily from an N S E R C operating grant to my supervisor (E.B. Taylor), for which I am grateful. M y mother and father, Rena Myles of Nelson and Barry Hagen of Prince George, were big supporters as usual, and their help during lean times allowed me to complete the study debt-free. Thanks to all! viii INTRODUCTION Models of speciation Ecological processes, or interactions occurring between organisms and their physical and biological environments, have long been thought to be important in promoting diversity between biological forms, and in speciation. The notion that ecological factors can drive phenotypic divergence, in particular, is supported by certain theoretical arguments (Maynard Smith 1966, Wilson 1989), experimental evidence (reviewed in Rice and Hostert 1993), and empirical evidence (Schluter 1988, 1996b, Grant and Grant 1989, Robinson and Wilson 1994, Carroll et al. 1997, Smith et al. 1997, Orr and Smith 1998). The processes of phenotypic diversification and speciation, however, are not identical. Speciation, by the Biological Species Concept of Mayr (1942), is defined as the evolution of reproductive isolation between forms that formerly could exchange genes. The study of speciation is then the study of the specific processes that create barriers to gene exchange within a single species, thereby leading to two or more new species. What these specific processes are has been a highly contentious issue in evolutionary biology for most of this century, and my own research interest is in the role of ecological factors particularly. Four basic speciation models have been proposed, incorporating both ecological and genetic processes. Basic allopatry model (a.k.a. dichopatric, vicariant speciation). Certainly the most widely accepted model of speciation is the allopatric model summarized by Mayr (1942). The first step in the process occurs when a physical barrier to dispersal and migration (mountain range, river, etc.) is formed within the distribution of a single 1 species, which eliminates gene flow between populations on either side of the barrier. Because the homogenizing effects of gene flow and recombination have been halted between these populations, the specific processes of genetic drift and/or natural selection can then lead to genetic differentiation between them. Speciation is complete when this differentiation is sufficient to produce incidental reproductive isolation between the evolving forms. The basic allopatry model is uncontroversial in its theoretical content, and empirical support for it is extensive (summarized in Mayr 1942, Lynch 1989). Experimental evidence (reviewed in Rice and Hostert 1993) strongly suggests that divergent natural selection can cause pre- and postmating reproductive isolation for allopatric populations. However, for the case of pre-mating isolation the conditions under which this could occur were specific - the characters adapting the organisms to the environments had to either cause premating isolation themselves or else be strongly linked to other characters that did. The experimental evidence for genetic drift alone giving rise to reproductive isolation is scant, but the time scales for those experiments may simply have been inappropriate relative to speciation in nature. Reinforcement model. In the reinforcement model of Dobzhansky (1940), speciation is initiated during a period of geographic isolation, but in this instance is not completed until the barrier to gene flow breaks down, and the forms come into secondary contact. The specific mechanism proposed by Dobzhansky was natural selection for assortative mating to reduce the production of hybrid offspring, which were presumed to be of lower fitness. It is common to observe stronger premating isolation between closely related species from areas of sympatry than when the same species are compared from naturally allopatric areas (Butlin 1989, Coyne and Orr 1989, Otte 1989), a pattern 2 which supports the reinforcement model. Strong objections to speciation by reinforcement have also been raised, however, because i) alternative explanations exist for the observed patterns (Butlin 1989), ii) no repeatable experiments have demonstrated reinforcement (reviewed in Rice and Hostert 1989), and iii) simulation models require very restrictive conditions (very strong selection against hybrids, very close linkage between mating and selected traits, and low levels of gene flow into the hybrid zone) to achieve speciation (Felsenstein 1981, Butl in 1989). Sympatric speciation (a.k.a. divergence-with-gene-flow, parapatric speciation). In the sympatric model (summarized in Maynard Smith 1966), phenotypic divergence and speciation take place without a physical barrier to gene flow. The first step in the process is the creation of a stable polymorphism in a single species as a result of strong, divergent selection for adaptation to different locales, habitats, or niches. Speciation takes place incidentally during this divergence, when sufficient genetic differences have accumulated to result in pre- or postmating isolation between the forms. It is very difficult to assess the importance of sympatric speciation in nature, as the hypothesis of geographic isolation in the unknown past can rarely be rejected. Nonetheless, sympatric speciation appears to be theoretically plausible (Maynard Smith 1966, Wilson 1989, Bush 1994), and the model is supported by experimental evidence as well (reviewed in Rice and Hostert 1993). Taken together, the results of the simulations and experimental studies suggest that sympatric speciation is feasible when divergent selection is strong relative to gene flow, first of all, and secondly when the characters adapting the populations to the environments cause premating isolation themselves or are strongly linked to other characters that do. 3 Bottleneck (founder effect, peripatric) speciation. B y the bottleneck speciation model (reviewed in Carson and Templeton 1984), speciation occurs rapidly and as a result of a major reorganization of the genetic material. This reorganization is possible, theoretically, when a population passes through an extremely drastic reduction in population size. Carson and Templeton (1984) suggest by way of example that the geographic isolation of the Hawaiian Islands, not only from each other but also from the continents, implies that colonizing populations would be very small in size. The extraordinary number of endemic species on the islands, in their view, thus supports the founder effect model, although Barton and Charlesworth (1984) point out that other models, also, can be invoked to explain the radiations. Importantly, experiments simulating bottleneck speciation have led only to weak levels of pre- and postmating isolation, leading Rice and Hostert (1983) to conclude that population bottlenecks may facilitate but not cause speciation. Two primary areas of focus are apparent from this brief review. The first of these concerns the geographic mode of speciation (feasibility of speciation without physical barriers to gene flow), and the second concerns the relative importance of divergent natural selection versus genetic factors in the evolution of reproductive isolation (e.g. importance of bottleneck speciation). M y own research interest is in the role of ecological factors (divergent natural selection) particularly. Importantly, the notion that divergent natural selection can lead to the evolution of reproductive isolation has both theoretical and experimental support, irrespective of the geographic mode of speciation. However, recent reviewers have pointed out that in spite of the attention evolutionary biologists have focussed on the importance of natural selection in speciation, very few 4 empirical studies have ever been able to convincingly make this link (Coyne 1992, Schluter 1996a, Orr and Smith 1998). A need obviously exists, therefore, for field studies specifically evaluation the role of natural selection in the evolution of reproductive isolation. Ecology and speciation in northern temperate zone fish assemblages Closely related groups of fish taxa from post-glacial, northern temperate aquatic ecosystems are uniquely suited for field studies of phenotypic divergence and speciation generally, and more particularly of ecology's role in these processes (examples: Behnke 1972, Ryman et al. 1979, Lindsey 1981, McPhai l 1984, 1992, Hindar et al. 1986, Malmquist et al. 1992, Schluter and McPhai l 1993, Taylor and Bentzen 1993, Bernatchez et al. 1996, Skulason et al. 1996, Taylor et al. 1996, Wood and Foote 1996, L u and Bernatchez 1998). Schluter (1996b) reviewed recent studies of divergence in fish taxa that live in post-glacial lakes, and suggested that four attributes of these assemblages are consistent with the idea that divergent natural selection plays an important role in the speciation process: 1. First of all , reproductive isolation has evolved rapidly between the forms, which have likely diverged since the last glaciation (less than 15,000 years ago). Examples of pairs of divergent forms (which Schluter considers good species despite a lack of taxonomic recognition) are presented from seven different taxa, each of which exhibits: i) very close genetic relationship between the forms, ii) assortative mating amongst the forms, and iii) a genetic basis (rather than 5 environmentally induced) for the observed phenotypic differences. Pairs of species in some cases appear to have evolved independently at more than one location, and such parallel evolution argues against genetic drift as a cause of the observed patterns. 2. In cases for which molecular genetic data are available, the forms are persisting despite known gene flow. Natural selection against hybrid phenotypes (post-mating isolation) is implied in these instances, otherwise the homogenizing effects of gene flow and recombination would have decayed premating isolation over time. 3. A high degree of ecological niche differentiation between the forms is typical, and shows remarkably consistent patterns across diverse taxa. In the majority of cases one form wi l l specialize on zooplankton in the pelagic zone of the lake, with the other specializing on larger invertebrates in benthic and littoral areas. 4. Finally, results from the small number of studies that have investigated hybridization suggest that the intrinsic viability and fertility of the interspecific hybrids is high. Genetic incompatibility between the forms, therefore, does not appear to be the basis for the observed postmating isolation. The suitability of northern temperate, polymorphic species (or closely related species pairs) for the study of ecological forces in speciation is in part due to the fact that 6 they constitute relatively young radiations into novel, post-glacial environments, and genetic incompatibility between the forms is unlikely (but see L u and Bernatchez 1998). However, the significance of the ecological factors under study at maintaining reproductive isolation over the longer term, and not just promoting it in the early stages of speciation, are unknown, because, as Schluter (1996b) pointed out, non-ecological factors related to genetic incompatibility can be invoked as explanations. The taxonomist Savvaitova (1980), in recommending a conservative approach in assigning taxonomic labels to divergent phenotypes in northern environments, suggested that not only reproductive isolation at the time of observation be considered when assessing whether speciation had occurred, but also the possibility of its reversal. She felt that the existence of a high number of closely related, endemic species in northern environments was unlikely over the longer term, due to the instability of these environments and the high degree of inherent phenotypic plasticity within the groups (examples: Lindsey 1981, Nordeng 1983). If it is typical that environmental change leads to convergence and introgression, as Savvaitova believed, then intraspecific groups formed rapidly in response to ecological opportunity may not be stable. Support for this notion is provided by the case of the closely related Gasterosteus (stickleback) species in coastal British Columbia, which, as McPhai l (1994) pointed out, is a paradox. First of all , Gasterosteus in this region is prone to isolation and rapid adaptive divergence (e.g. into benthic and pelagic species within lakes, and stream-resident and migratory species in streams), and reproductive isolation in areas of secondary contact/sympatry has evolved in less than 12,000 years. Because of this apparent tendency for evolutionary divergence the lineage should contain a large number of species both young and old, as Gasterosteus has been 7 present on the Pacific coast of North America since at least the Miocene. Yet the opposite is true, and there are only two well-delineated old species in the genus. What this example suggests, when taken together with Savvaitova's cautionary point, is that the general importance of ecological factors in speciation may be overestimated by studying only taxonomic divergence that is in its early stage. A need exists, therefore, for field studies of a complementary nature, which address the importance of ecology in maintaining reproductive isolation over the longer term. Ideally, the study species for such an inquiry would have proven themselves as good species over broader spatial and temporal scales, yet still be closely enough related that hybrids were viable and fertile. The sympatry of two northern temperate zone salmonid species from the genus Salvelinus (chars), the Dol ly Varden (S. malma) and the bull trout (S. confluentus), provides a study system which meets these criteria. Taxonomy and distribution of Dolly Varden and bull trout in North America Prior to 1978, the bull trout was taxonomically included in the Dol ly Varden species complex, which had a distribution around the entire arc of the North Pacific from the Sea of Japan to Northern California, and also north in Alaska to its border with the Arctic Ocean (Cavender 1980). This group had received considerable attention up to that time (McPhail 1961, Morton 1970, Behnke 1972), but the bull trout form was not distinguished because of its similarities to the Dol ly Varden both in appearance and also in the morphological characters typically used in char taxonomy: number of gi l l rakers, pyloric caecae, and vertebrae (Cavender 1978). Cavender (1978), however, was able to distinguish the species throughout their respective ranges, which he accomplished by 8 including in his study an extensive set of additional morphological characters. He felt that the morphological differences between the species were best summarized as adaptations, in terms of mouth morphology particularly, to 'piscivory specialist' and 'generalist' feeding niches by bull trout and Dol ly Varden, respectively (see also Haas and McPha i l 1991). In western North America, the portion of the geographic range formerly ascribed to Dol ly Varden but that is actually occupied by bull trout encompasses areas south of 47\u00C2\u00B0 N , and interior regions between 47\u00C2\u00B0 N and approximately 60\u00C2\u00B0 N . Dol ly Varden are found in coastal areas north of 47\u00C2\u00B0 N (Haas and McPhai l 1991), although the distributions of the two species do overlap significantly in British Columbia (Fig. 1). Molecular genetic evidence has now confirmed that Dol ly Varden and bull trout are distinct, and furthermore suggests that divergence between the form predates the last glaciation period (Grewe et al. 1990, Phillips et al. 1992, 1994, 1995, Pleyte et al. 1992, Crane et al. 1994). Understanding of the species' precise taxonomic relationship, however, has been hampered because of suggestions of extensive, past hybridization between the lineages (Phillips et al. 1994, 1995, Taylor et al. 1999). Hybridization between the species also takes place where they are currently in sympatry, although the present levels are relatively low in these areas (Cavender 1978, McPhai l and Taylor 1995, Baxter et al. 1997; Z .R. Redenbach and E . B . Taylor, U . B . C . Dept. of Zoology, unpubl. data). Despite this hybridization, populations of the two species appear to be maintaining distinct gene pools throughout the zone of overlap (Cavender 1978, Haas and McPha i l 1991, Baxter et al. 1997, Leary and Allendorf 1997), although one area is a possible exception. The exception may be the Skagit River watershed, located at the 9 southern extent of the zone of overlap, where molecular genetic and morphological evidence suggest the possibility of introgression (McPhail and Taylor, 1995). The potentially exceptional case of the Skagit populations notwithstanding, the two forms do appear to have proven themselves as good biological species (groups of potentially interbreeding populations reproductively isolated from other such groups - Mayr 1942), yet the hybridization evidence suggests that in this case the process of speciation is not past the point of irreversibility. The study system is suitable, therefore, for the investigation of ecology's role in preserving reproductive isolation between species in which intrinsic isolating mechanisms are slow to evolve. Study questions Because of the strong theoretical and experimental support for its potential importance, I am specifically interested in the role of divergent natural selection in maintaining reproductive isolation between Dolly Varden and bull trout. Schluter (1996a) has proposed that divergent natural selection in two forms, for adaptation for different resource environments, first of all, and secondly to reduce the intensity of interspecific competition during sympatry, may be important in speciation. To investigate the potential occurrence of these ecological modes of speciation in my study system I posed the following, two-part questions: 1. Are Dolly Varden and bull trout in sympatry using different resource environments (this is consistent with having adapted to different resource environments during either i) sympatry, or ii) a previous period of allopatry)? If 10 such differences exist, the question then becomes: are the resource environment differences contributing to reproductive isolation? 2. Is interspecific competition between Dolly Varden and bull trout resulting in phenotypic divergence? If so, this question then becomes: is this divergence contributing to reproductive isolation? M y investigation thus required two components, which I w i l l refer to hereafter as the 'resource environment' study and the 'interspecific competition' study. This thesis reports the results of my investigation into these ecological factors, which took place in the pristine Thutade Lake watershed in northcentral British Columbia. 11 M A T E R I A L S AND M E T H O D S Study A r e a The Thutade Lake watershed is situated near the Pacific / Arctic continental divide in north central British Columbia, at approximately 57.0\u00C2\u00B0 N latitude and 126.7\u00C2\u00B0 W longitude (Figure 1). Thutade Lake is the headwater of the Finlay River, a tributary of the Peace River and ultimately the Mackenzie, which flows into the Arctic Ocean. For a number of reasons, the basin provided the most suitable site for a Dol ly Varden / bull trout sympatry study that had yet been identified. First of all , the fish community is relatively simple. The tributary stream reaches are dominated by the two char species, meaning that at most sites the potentially confounding effects of other competitors were not an issue. The fish populations are also thought to be at levels that are close to the environment's natural carrying capacity, which was important for the assessment of the effects of interspecific competition. The reason for this belief is that there is almost no recreational use of Thutade Lake, which does not have developed access. Furthermore, there is very little fish population impact from habitat loss, as most of the drainage basin is pristine (the only obvious loss to date came in 1997 during the construction of the Kemess gold and copper mine in the Kemess Creek watershed, when the entire valley of one of the tributaries, South Kemess Creek, was permanently lost under the tailings impoundment). A particularly valuable attribute of the location was that the fisheries impact assessment associated with the mine development had been extensive, and provided me with a great deal of usable information that would have been impossible to collect on my own. Finally, and importantly, hybridization between Dol ly Varden and 12 bull trout and hybrid fertility in the watershed had been documented (Baxter et al. 1997). Dol ly Varden/bull trout hybridization in the Thutade watershed was also investigated, in a much more thorough manner, in a study concurrent with my own (Redenbach 2000). Study design Resource environments. M y first goal for the resource environment study was to investigate whether the ecological niches for the two species were in fact distinct. Evidence of such differences is consistent with the idea that the species have adapted to different resource-environments, either in sympatry or during a previous period of allopatry. The second goal, then, was to be able to speculate about what these differences meant for reproductive isolation. The life history patterns for Dol ly Varden and bull trout in the Thutade watershed are different (Figure 2), and this affected the study's design. Dol ly Varden remain residents of relatively small streams for their entire life cycle. B u l l trout spend up to four years (Baxter 1994) in these same streams at the beginning of their lives, but then undergo a niche shift, migrating to Thutade Lake and becoming piscivorous. There are, therefore, two broad categories of resource environment to consider: lake and tributary stream. Because I was looking for differences within each of these broad categories, the complete exclusion of Dol ly Varden from the lake environment simplified fieldwork priorities. In the lake, I needed only to confirm that Dol ly Varden were not present there and, therefore, that the life histories of the two species were in fact different. This is not to suggest, however, that the life histories were not an important focus of the investigation. Differences could potentially affect aspects of each species' biology that 13 may be related to premating isolation, such as body size and its potential correlates spawning location, spawning timing, and spawning site hydraulic and bed material characteristics. Even though premating isolation or even interspecific differences had not been considered before, a record of the above attributes nonetheless existed already in the reporting of the annual population monitoring study related to the mine development, and was available for analysis (Bustard and Royea 1995, Bustard 1996, 1997, 1998, 1999 in prep.). The primary focus of my actual field work in the resource environment study could then be the strict sympatry of the tributary reaches, because only there could the species' resource use be compared under identical conditions. Schoener (1974) reviewed existing ecological information and identified three general niche dimensions along which species may segregate: food, habitat, and time. Ross (1986) followed this approach for fish assemblages and confirmed its utility. Hypothetically, niche overlap could be greatly reduced by substantial resource use differentiation along even just one of these dimensions. A l l , therefore, were included in my study of the species' resource environments in strict sympatry (the tributary streams). I considered there to be several logical subdivisions of each of these dimensions, some which have been identified already for Dol ly Varden and bull trout, and my goal was to treat each with a consistent level of effort as to not prejudge its importance. For salmonids generally, habitat separation has been frequently documented for sympatric, stream dwelling species, both in terms of hydraulic characteristics, (e.g. depth and velocity: Hartman 1965, Lister and Genoe 1970, Everest and Chapman 1972, Dol lof f and Reeves 1990, Heggenes et al. 1996), and also hydraulic habitat type (pool versus riffle: Hartman 1965, Glova 1986, Taylor 1991). Food and feeding behaviour separation has 14 been treated for stream-dwelling char specifically, in sympatry both with other salmonid genera and also congeners, and food resource sub-dimensions already identified are prey type (Nakano et al. 1992, Nakano and Kaeriyama 1995) and foraging mode (drift vs. benthic: Nakano and Furukawa-Tanaka 1994, Fausch et al. 1997). I also considered separation along another potential food dimension, prey size, although it has not yet been investigated for stream-dwelling chars. Cavender (1978) identified adaptations to piscivory in the bull trout trophic structures, including a larger mouth than that of the Dol ly Varden. I was interested in the possibility of prey size differences because the morphology of the mouth and related structures can affect the size of prey that fish are capable of handling (Werner 1977, Wankowski 1979, Wainwright 1996). Indeed, the result of larger prey size for bull trout was found during a preliminary survey of the watershed in 1997 (Appendix 1), which I wished to confirm under more conservative conditions. Henderson and Northcote (1985) discovered that Dol ly Varden were capable of reacting to prey at nighttime light levels, suggesting that separation along the time dimension could happen either by diel or seasonal activity partitioning. I was constrained to doing resource use observations during a single study period in mid-summer, so I studied the time dimension by replicating all interspecific comparisons of food and habitat for two time periods: one immediately surrounding mid-day, and the other surrounding mid-night. Overall, then my design for interspecific comparison in the tributary stream environment included the three main dimensions of Schoener (1974) subdivided into the following: 15 1. The 'habitat' dimensions water depth, mean current velocity, and hydraulic habitat type 2. The 'food' dimensions foraging mode (two categories), mean prey item biomass, and taxonomic classification of prey 3. The 'time' dimension as a replication of the 'habitat' and 'food' comparisons for day and night time periods, with the addition of the diel feeding activity measures total prey biomass and total foraging rate (for both day and night). Interspecific competition. M y first goal in studying interspecific competition was to investigate whether or not it was occurring at all. The second goal was to be able to speculate about whether the effects of this competition, should competition be demonstrated, were contributing to reproductive isolation between the species. The existence or importance of interspecific competition cannot be inferred from resource use differences alone. Alternative hypotheses for causes of the differences, such as phylogeny (Dobson 1985), or adaptation to different ecological niches, cannot be ruled out. Demonstrating the presence of interspecific competition in freshwater ecosystems generally requires evidence that interactions cause a niche shift in one or both species (e.g. Fausch and White 1981; Hearn 1987 for review). Demonstrating the extent, importance and dynamics of competition, however, is more difficult and less explored, and requires, over appropriate temporal and spatial scales, evidence that interactions cause changes in i) parameters affecting the fitness of individuals (survival, growth), or 16 ii) changes in the carrying capacity of the environment (density) for one or both species (Schoener 1983, Werner 1986 for reviews). Since 1995, with the assistance of University of British Columbia students and faculty, the consultant to the mine developer has conducted annual population monitoring of both Dolly Varden and bull trout, with sites located throughout the Thutade watershed. My plan was to utilize the data produced to investigate density compensation in seven independent reaches, as a study of this scale and expense would be simply impossible for me to mount on my own. I also planned to investigate interspecific competition in a second manner, by looking for niche shifts in sympatry. I did not feel that sympatric niche shifts could be measured in the Thutade watershed itself, even though I had identified locally allopatric and sympatric stream reaches for each species. The reason for this was that the required assumption, that I could measure in the field resource uses for the two species under identical environmental conditions, was unrealistic. Instead I planned to compare the resource environments of the two species in sympatry, as characterized by this study and a small number of unrelated investigations, to what has been recorded for each in allopatry. Data collection Resource environments - strict sympatry. Resource use measurements in the tributary stream environment, for both species concurrently, were made between July 18 and August 6, 1998. The study reach was a 2.0 km section of North Kemess Creek, a third-order tributary of the Attichika River (which in turn flows into Thutade Lake), and an area of known, strict sympatry. We used snorkel surveys to locate and observe individual fish of each species. If the snorkeler deemed the fish undisturbed, he observed 17 it for five minutes and recorded the number of surface, drift, and benthic foraging attempts as well as the species identification and visually estimated length. He then attempted to capture the fish using two large aquarium nets, and i f he was successful the fish was measured, its adipose fin clipped for later molecular genetic analysis (see Redenbach 2000), and it was then preserved in a solution of 10% formaldehyde for laboratory gut content analysis. If, however, the observer was unsuccessful at capturing the fish, the visually estimated length was later corrected according to the bias evident from comparing visually estimated and actual lengths for the fish that were captured, so that the observational data could be retained. A t the focal point location held by each observed fish, we measured stream depth and mean current velocity, and also noted hydraulic habitat type (riffle or pool, main channel or side channel). In habitats too shallow for snorkel observations, we used a battery-powered electro fisher and short pulses of electricity to locate and capture individual fish. Depth, velocity, and hydraulic habitat type were recorded and the fish sampled as above. We replicated underwater observations at nighttime using a diver's light shining through a translucent red filter, and replicated electrofishing with high-powered lighting from a number of sources. The use of visible light to observe fish at night is obviously contentious. Our method of using the translucent red filter was arrived at after trials at the beginning of the study, in which we attempted to find a method that simultaneously allowed us to observe the fish yet not disturb their activity. Heggenes et al. (1993) reported on the use of a red filter for nighttime observations, and in our trials it appeared to reduce disturbance of the fish. When the red filter was sanded enough to reduce light transmission to the minimum required for effective observations, fish that were approached carefully showed no 18 obvious signs of disturbance (moving away from the observed focal point, suspension of feeding activity). We attempted as much as possible to collect similar numbers of each species for each size class and for each collection method and time period, to avoid biasing interspecific comparisons. This was achieved by setting targets for the sample sizes for each of the categories. It should be noted that young-of-the-year char were not included in the study because i) it was difficult to determine their identity as to species, and ii) areas of sympatry do not correspond with those for older juveniles (Z.R. Redenbach, M . S c . candidate, U . B . C . Department of Zoology, unpubl. data.). Later, in the laboratory I conducted stomach content analysis on the entire preserved sample. Intact prey items removed from the fish stomachs were identified, counted, and measured by using a dissecting microscope connected through a digitizing tablet to a computer. A t this time also I measured mouth width and gi l l raker spacing, because Wankowski (1979) has suggested that these are morphological constraints that influence the size range of prey that salmonids can manage. I obtained the total biomass of intact prey items for each stomach by weighing the gut contents after they had been dried in a dessicator, and regular measurements had indicated that the weights had stabilized (typically after two days). Life history. Thutade Lake was surveyed from June 16 to June 21, 1998. We captured fish with small-mesh, monofilament tangle nets set both on the bottom and suspended from the surface, and at two different locations on the lake. From June 16 to June 18 we fished an open-water, deeper location adjacent to a large shoal area, and from June 19 to June 21 we fished a shallower, more complex shoreline near the lake outlet. 19 A l l fish captured were anaesthetized, identified as to species, measured, and released live i f possible. Additionally, from all char we removed a pelvic fin ray for aging and an adipose fin clip for molecular genetic analysis. We sampled char stomach contents using the pulsed gastric lavage method (Foster 1977) and preserved them in a solution of 10% formaldehyde for later measurements. Aspects of the species' biology related potentially both to the life history and premating isolation had already been recorded during fish population monitoring studies by the mining company's consultant (Bustard and Royea 1995; Bustard 1996, 1997, 1998, 1999 in prep.). For adult bull trout, body size data were collected by angling at locations along the Attichika River and along its tributary Kemess Creek. Captured specimens were identified as to sex, measured for fork length, had sections of their adipose and pelvic fins removed for molecular genetic analysis and aging, respectively, tagged for population monitoring purposes and then released. Adult Dol ly Varden, because they were difficult to distinguish from immature fish, were identified from a single sample of Dol ly Varden captured and sacrificed from tributaries throughout the Thutade watershed. Specimens were identified as to sex and maturity, measured for fork length, had a scale sample removed for aging, and were then preserved in alcohol for later molecular genetic analysis. Spawning timing was investigated in the Kemess Creek watershed for both species in all years of the population monitoring (1994-1999). A l l known spawning areas in the watershed were surveyed approximately weekly over the entire period of spawning for both species combined, from early August until early October of each year. For each species, the numbers of spawners present in spawning reaches were recorded (sex was recorded only for the larger, adult bull trout) as were the cumulative total number of 20 redds (spawning sites excavated into the stream bed material) on each survey date. A subsample from the total number of redds, from three independent reaches for both species, were assessed as to hydraulic and bed material characteristics. Mean stream depth and current velocity were recorded as the average of two measures, each over undisturbed bed material on opposite sides of the mid-point of the redd. The measure of bed material size was the diameter of the piece for which smaller (and larger) particles made up 50% of the redd area (D50). Spawning locations were recorded within 400 m marked sections of the surveyed reaches, and also in smaller or seepage-fed tributaries located along these marked sections. Density compensation. For all four years (1995-1998) of the population monitoring surveys field work was conducted during August, typically in the final two weeks by two crews working simultaneously. There are 27 sites for which multiple years' data exist, most of which require helicopter access, and these could typically be surveyed at a rate of 2 to 3 sites per crew per day. We sampled fish populations by the two-pass removal method (Seber and LeCren 1967) using electrofishing equipment at sites completely enclosed by stopnets. All fish captured were anaesthetized, identified as to species, measured, weighed, and then released. At sites for which aging and molecular genetic analyses were planned (Redenbach 2000), scales and tissue samples were also taken. Analyses Resource environments - strict sympatry. The Dolly Varden and bull trout populations from which the samples were taken include several age categories, and my 21 initial examination of resource use data showed generally that the values of these traits were positively correlated with body size. A s Werner and Gi l l iam (1984) pointed out, body size imposes important constraints on an organism's resource use, and for such size-structured populations it is not unusual for niche differences between size classes to be as great or greater than between sympatric species. I dealt with the potentially problematic effect of body size differences in the field, first of all, by collecting samples of both species from across the size range in similar proportions where possible. During analyses I attempted to remove the affect of body size differences statistically with the technique of analysis of covariance ( A N C O V A : Tabchnick and Fidell 1996), which I could apply to resource use traits with continuous distributions. The distribution of variances generally appeared log-normal, as is typical for allometric data (Harvey 1982), so interspecific comparisons were made with log-transformed data for the following resource utilization measures: i) 'habitat' traits water depth and mean current velocity; ii) 'food' traits drift foraging rate, benthic foraging rate, and mean prey item biomass; and iii) 'time' traits total foraging rate and total gut content biomass. I also used the analysis of covariance to compare mouth width and gill raker spacing for the species. Hydraulic habitat type data (riffle/pool, main channel/side channel) were in discrete categories, so I tested for independence from species category using contingency table analysis and the chi-square statistic (Zar 1996). Uncomplicated and powerful statistical tests for diet differences, which can be applied without computer programming, do not exist (Wallace 1981, Mueller and Altenberg 1985, Sevenster and Bouton 1998). The reason for this problematic state is that the pattern of resource utilization consists of several, non-independent variables, which take extreme distributions (Sevenster and Bouton 1998). 22 The Schoener (1970) index of proportional overlap, for the average diet proportion of each prey taxonomic category, is recommended by Wallace (1981) and is given by the formula: a = l - 0 . 5 ( Z | P \u00C2\u00AB - P ^ ( - | ) Where: a = Schoener index of proportional overlap P*, = Proportion of food category i in the diet of species x Pyi = Proportion of food category /' in the diet of species y Overlap indices such as this cannot be tested for statistical significance in a straightforward manner, relying instead upon the judgment of the researcher and tradition among related studies. The Spearman Rank Correlation Coefficient (Zar 1996) does provide a statistical test for overlap, but its use in diet analysis has been criticized by Wallace (1981) for being insensitive to proportions. I made interspecific diet comparisons using both methods, which suits the exploratory nature of the study. However, because the results from each must be interpreted cautiously, for the above reasons, I judged them not fit for overall hypothesis testing within the food resource dimension. A s several authors have pointed out (Rice 1989, Wright 1992, Chandler 1995), multiple statistical comparisons within one study inflate the risk of rejecting null hypotheses when they are in fact true (Type I errors). To treat all of the potentially important resource dimensions I had identified, and their subdivisions, I needed to make a minimum of 18 interspecific statistical comparisons. Adjustment of the significance levels for individual tests, which I conducted using the sequential Bonferroni adjustment 23 (Holm 1979, as cited in Rice 1989), was therefore necessary to maintain an acceptable overall Type I error rate. My goal for the resource partitioning analysis was sufficient power to conclude which, i f any, of the three main niche dimensions (habitat, food, time) were important, while ensuring conclusions at this level were not spurious. A l l of the interspecific comparisons belonging to each of these dimensions, then, form a logical \"table\" of statistical tests for control of Type I error rates. The overall Type I error rate for the resource partitioning study was then no greater than 15%, within the 10 to 15% guidelines for exploratory surveys suggested by Chandler (1995). Life history. A number of biological attributes influenced by the life histories, and which may be related to premating isolation, were tested for interspecific differences. I calculated the means of adult body size, spawning site depth, spawning site current velocity, and spawning site bed material size for each species, and then made the interspecific comparison using two-sample /-tests. I also calculated the means of adult body size for both bull trout and bull trout/Dolly Varden hybrids that had been identified by molecular genetic analysis. These too were compared using the two-sample /-test. Density compensation. The abundance data from the individual sites are not independent from other sites within the same reach because they belong to the same population (reach). The average of sites within a reach for a given year also is not independent from other years, as individuals may stay in the reach for multiple years. The proper replicate for comparison is thus the average for the reach of all site densities averaged over all the years. I calculated the regression of the average reach densities of each species upon the other using the ordinary least squares technique with untransformed data, which enabled me to compare the resulting slope with null 24 hypotheses slopes of zero and minus one using /-tests. Mean total densities of sympatric and allopatric reaches were compared using the two-sample /-test (Zar 1996). 25 RESULTS Resource environments - strict sympatry Habitat. My comparisons of habitat utilization for Dolly Varden and bull trout in strict sympatry revealed a great deal of overlap. I made interspecific comparisons for three traits in two time periods within the habitat niche dimension (Table 1), totaling six statistical tests. I therefore sequentially adjusted the significance level (a,\u00E2\u0080\u009E where n = rank in ascending order of P-value) to hold the overall Type I error rate for the six tests at 5%. The two species in strict sympatry used the hydraulic habitat categories mainstem pool, mainstem riffle, sidechannel pool, and sidechannel riffle (Figure 3) in similar proportions, both during the day {chi square = 5.76, P < 0.25) and night {chi square = 3.10, P < 0.50) observation periods. The comparison of focal point water depths (Figure 4) between Dolly Varden and bull trout was also not significant for the daytime observation period (F= 1.99, P = 0.16). The nighttime difference was more evident, with Dolly Varden being found at locations in deeper water. After the critical significance level was adjusted for multiple tests, however, the comparison was found to be not significant (F= 6.91, P - 0.010, a, = 0.008), although only marginally so. The interspecific comparison of daytime, mean focal point current velocity revealed different relationship slopes for the two species (Figure 5a), with Dolly Varden showing a greater difference relative to bull trout between small and large body size categories. Again, the comparison was found to be not significant (although close) after the correction for multiple tests (Slopes: F = 6.32, P = 0.014, a2= 0.010). Both Dolly Varden and bull trout were found at locations of reduced current velocity at night, and the interspecific 26 comparison (Figure 5b) revealed no significant difference (F = 2.83, P = 0.097). Despite the lack of statistical significance in the velocity comparisons, the scatterplots when visually inspected do seem to indicate consistent though subtle preferences of the largest Dol ly Varden for deeper and faster water relative to bull trout. A l l fish observed were closely associated with the stream bottom. Food. I made five statistical comparisons to assess differences in food resource utilization (Table 2). The underwater observations of feeding behaviour suggested both species had similar diel shifts in feeding behaviour, but interspecific differences were still apparent. A s observations were made using visible light, nighttime results must be interpreted cautiously. The assumption of the method is that introduced error w i l l affect both species similarly. Although I had no way of assessing how valid this assumption was, I felt that including interspecific comparisons in the analysis was warranted, given the exploratory nature of the study. During the day foraging from the drift was more important than benthic foraging for both species (Figure 6), but the relative importance of benthic foraging increased at night. Interspecific comparison, however, revealed that benthic foraging was marginally (statistically speaking) more important for bull trout than for Dol ly Varden, during both daytime (F= 7.82, P = 0.008, a, = 0.010) and nighttime (F = 6.72, P = 0.013, a 2 =0.013) observation periods. Dol ly Varden had a marginally higher (but not statistically significant) drift foraging rate by day (F= 3.65, P = 0.062). Conversely, at night the bull trout foraging rate was higher. This difference was pronounced only for larger individuals, resulting in a difference in the relationship slopes. The difference was not sufficient to attain statistical significance (F = 5.01, P = 0.031, a 2 =0.025), however, after the correction for the multiple tests in the food resource 27 dimension. Surface foraging was not important to either species during the period of the study, and thus this category was dropped from the analysis. Gut content analysis also revealed interspecific differences, chiefly that bull trout consumed larger prey items (Figure 7) than did Dol ly Varden (F= 5.89, P = 0.017, a 3 = 0.017). This result confirmed the finding of larger prey for bull trout during a preliminary survey of the watershed in 1997, which had been conducted under less conservative conditions (Appendix 1). Note that the day and night average prey biomass data were combined for each species (Table 2), this having been done because they were not significantly different and the fit of the model was improved. Furthermore, this step improved the power of the overall analysis by reducing the total number of statistical tests in the food resource dimension, which meant that significance levels for individual tests were less conservative. Interpretations of prey type differences were more ambiguous. For the daytime stomach samples, the taxonomic composition of the average Dol ly Varden diet appeared different from that of the bull trout (Figure 8), having much higher proportions of winged insects. Adult Dipterans and Hymenopterans made up 33.5% and 7.5%, respectively, of the Dol ly Varden diet, as opposed to 11.3% and 0.6% of the bull trout diet. The value of the Schoener (1970) index of proportional overlap for this comparison was 0.71, however, which implies significant overlap by exceeding the commonly used \"biological significance\" level of 0.60 (Wallace 1981). Average diet proportions at night appeared more similar (Figure 9), and the value of the Schoener index was relatively high (0.89). The analysis of diet proportion overlap using the Spearman rank correlation coefficient was highly significant for the nighttime sample (/\"s(.05),7 = 0.964, P < 0.005), agreeing with the Schoener index, but was not significant for 28 the daytime sample (rS(.05),7 = 0.429, P < 0.50). Neither of these two methods is considered suitable for statistical inference in diet studies, however (Wallace 1981; Sevenster and Bouton 1998), so diet composition comparisons were not included at all in the overall statistical test for differences within the food resource dimension. Time. With respect to diel activity partitioning, the day and night comparisons of habitat and food resource utilization patterns have so far not revealed much statistically, only that Dol ly Varden are in deeper water than are bull trout at night but not during the day. I made four interspecific comparisons of total feeding activity to assess potential differences - these are summarized in Table 3. Daytime overall foraging rates (Figure 6) for the two species were very similar (F= 0.12, P = 0.73). The night time rates were significantly different (F= 8.77, P = 0.0048, a, = 0.013), however, suggesting that Dol ly Varden made a nighttime reduction in overall foraging rate relative to bull trout. The shift in Dol ly Varden stomach fullness relative to bull trout, from the gut content analysis, was in the same direction (Figure 10), but Dol ly Varden had fuller stomachs than the feeding observations would predict. Interspecific comparisons were not significant (day: F- 1.31, P - 0.26; night: F=0.\0,P = 0.76). Apart from experimental error, an explanation for this discrepancy may be that the gut contents included prey captured over a much longer time period than the feeding observations encompassed. The period of nighttime darkness during the study was less than 4 hours, and prey captured at the end of the daylight period may still have been intact in stomachs during the observation period, which began only one hour later (2300 PST). In summary, for the resource environment study in the tributary streams, it appears that support for the hypotheses of interspecific differences along the resource 29 dimensions habitat, food, and time exists but is somewhat marginal, especially for the resource dimensions habitat and time. Wankowski (1979) has suggested that gillraker spacing and mouth width are morphological constraints on the size of prey (minimum and maximum, respectively) that salmonids can handle, so these measures also were taken during the resource partitioning study (Table 4, Figure 11). B u l l trout had larger mouths than Dol ly Varden (Slopes: F = 4.8, P = 0.03) and wider gillraker spacing (Slopes: F = 7.14, P = 0.009), consistent with the finding of relatively larger prey in bull trout diets. Even though the slope differences precluded testing differences in these measures using the analysis of covariance, the separation between them was clear (Figure 11). A single drift sample was also taken for each of the day and night study periods, and the comparison revealed that the total biomass of drifting invertebrates was far - I T greater at night (8662 mg/m /hr vs. 552 mg/m /hr). Life history Sixteen char, ranging in length from 199 to 850 mm, were captured in Thutade Lake from approximately 80 hours of tangle-netting effort in both near-shore and open water habitats. A l l were bull trout. Other fish species captured during netting were kokanee salmon {Oncorhynchus nerka), mountain whitefish (Prosopium williamsoni), rainbow trout {Oncorhynchus mykiss), and longnose suckers {Catostomus catostomus). This result is consistent with my assertion that bull trout undergo a niche shift that Dol ly Varden do not, which is also supported by my observations in the tributaries of Dol ly Varden within the size range known for sexually mature adults (Appendix 2): 113-235 30 millimeters. Piscivory as an ecological specialization was also confirmed for bull trout in the lake. Two whitefish of 132 and 123 mm were recovered from the stomach of a 410 mm bull trout, and kokanee of 197 and 184 mm were recovered from bull trout of 800 and 600 mm, respectively. A l l other char stomachs were empty, although the stomachs of three bull trout of less than 230 mm fork length were not sampled. The two species were different in their reproductive biology, a fact that was potentially related directly to the life history differences. Dol ly Varden and bull trout differed in: i) adult body size (Table 5; appendix 2); ii) physical habitats for spawning, in terms of site depth, current velocity, and bed material size (Table 5; Appendix 3); and iii) spawning timing and spawning site distribution (Tables 7, 8). Only 12 adult Dol ly Varden made up the sample for body size comparisons, as fish had to be sacrificed in order to confirm that they were in fact mature (Appendix 2). The interspecific differences were nonetheless obvious and dramatic. The mature Dol ly Varden ranged in size from 113-235 mm fork length, averaging 148 mm (S.E. = 11 mm) whereas the 34 bull trout had a size range from 440-890 mm and averaged 741 mm (S.E. = 16 mm). Statistical comparison of the means confirmed the obvious difference (t = 11.92, P \u00C2\u00AB 0.001). Tissue samples from the migratory, adult fish presumed to be bull trout have been analyzed using molecular genetic techniques (Z.R. Redenbach, U . B . C . Dept. of Zoology, M . S c . candidate, unpubl. data), and a small number (n = 3) of Dol ly Varden/bull trout hybrids (Appendix 2) have been identified within this group. The hybrids averaged 403 mm (S.E. = 33 mm), and were therefore intermediate in size between adult Dol ly Varden and bull trout. Very little overlap in body size existed between hybrids and the parental forms, and these differences were indeed statistically 31 significant in both directions (Dolly Varden versus hybrid: .^05(2)35 = 10.23, P \u00C2\u00AB 0.001; bull trout versus hybrid: .^05(2)35 - 6.15, P < 0.001). None of the large migrants were found to be Dol ly Varden. A study of whether hybrids occurred within the adult Dol ly Varden spawning group could not occur because of the difficulty in distinguishing immature, potentially migratory fish from mature fish within the sampled stream resident population. Not surprisingly, given the body size differences, spawning sites for the two species differed significantly in terms of their physical characteristics (Table 5; Appendix 3). The larger bull trout spawned at sites that were in deeper water than those of Dol ly Varden (34.7 cm versus 9.3 cm: /.05(2)56 = 10.80, P \u00C2\u00AB 0.001), and which also had higher current velocities (36.8 cm/s versus 20.7 cm/s: ^.05(2)56 = 5.51, P < 0.001). The bull trout sites were also found in much larger bed material (D50 = 4.9 cm versus 1.6 cm: /.os(2)56 = 12.59, P \u00C2\u00AB 0.001). N o spawning observations were made for interspecific hybrids. Interspecific comparisons of spawning timing and spawning site distribution can only be made for a single watershed, that of North Kemess Creek. Because of the once-weekly (at best) timing of surveys the spawning period for each species can only be approximated. However, from these surveys and other, incidental observations of spawning activity it appears that the maximum range of activity is roughly from August 22 to September 12 for bull trout, and from September 3 to October 8 for Dol ly Varden. Dol ly Varden spawning observations have been taken in a more sporadic manner than those for bull trout, but despite missing surveys the peak of Dol ly Varden spawning activity is clearly later (late September versus early September: Tables 6,7). For the four years that Dol ly Varden spawning data is reasonably complete (1994, 1996-1998), 32 activity is only 4-50% complete by Sept. 12, the end of the bull trout spawning period (Table 7). The spatial distributions of spawning sites show even less overlap. Within the North Kemess watershed virtually all Dol ly Varden spawning is limited to the reach comprised of the top 4.0 km of North Kemess Creek and the small, mostly seepage-fed tributaries that feed this section. No bull trout spawners at all were recorded in this section prior to 1997 (Table 6), and only 4.9% of the total adult bull trout population using North Kemess Creek used it during 1998. Only in 1997 was there any appreciable overlap in the spawning distributions of the two species. In that year, 27% of North Kemess Creek bull trout spawners, representing 19% of the total Kemess Creek population, used this upper reach. Displacement of spawners from other areas in the Kemess watershed was the likely cause of the 1997 anomaly. The construction of the mine's tailings dam in that year on another Kemess Creek tributary, South Kemess Creek, had the effects of simultaneously eliminating spawning habitat for approximately 20% of the total Kemess watershed population and also greatly increasing turbidity and sedimentation in the lower reaches of Kemess Creek. No bull trout whatsoever spawned in the seepage tributaries that are the most important spawning habitats for Dol ly Varden in North Kemess Creek. Density compensation To investigate the nature of the relationship of the two species' densities in sympatry and local allopatry I used data from seven reaches that I considered to be independent (n = 2l sites total), which are presented in Table 8. A l l were sections of small to medium-sized tributaries of either Thutade Lake or of two larger rivers that drain 33 into the lake (Figure 12). Sites existed along the two larger mainstems themselves (n = 6), but regularly contained only very low densities of both Dol ly Varden and bull trout age one year and older (Table 8). The species assemblages of these two reaches were different also, being dominated by the sculpin Cottus asper, which was rare or absent in the tributary reaches. Consequently, I discarded the two mainstem reaches as outliers, my reasoning being that their habitat capability for char was not comparable to smaller systems. I looked at the density relationships by regressing the mean abundance per reach of Dol ly Varden on that of bull trout. There was a linear relationship between the densities (Figure 13), and the strength of association was moderately high (r 2 = 0.61). The obvious first null hypothesis was that of no negative relationship, implying no effect on Dol ly Varden carrying capacity from the presence of bull trout and vice versa. B y t-testing the observed regression coefficient (J3 = -0.94) against that of the null hypothesis (j30 = 0), I determined that the relationship was negative and significant (/.05(i),6 = 2.79, P < 0.025), a result consistent with adverse competitive effects between the species. The specific prediction of greatest interest, however, and one that is crucial to competition theory, was that resource use differences between species should lessen the intensity of interspecific competition relative to intraspecific competition (Schoener 1983). The second null hypothesis, therefore, was that of complete compensation by one species for missing individuals of the other (J30 = -1.0), implying no reduction of competition between individuals belonging to different species categories relative to competition between individuals belonging to the same category. Figure 13 suggests that the null model explains the observed relationship rather well , and indeed a /-test did not reject the 34 null hypothesis of 60 = -1.0 (/.05(i),6 = 0.19, P > 0.50). This null hypothesis was also investigated in a second manner, by comparing the mean total densities (both species combined) of locally allopatric and sympatric reaches (Figure 14). The average total density of the four sympatric reaches (11.43 fish A 00 m 2 ) was not significantly different 0.05(2),6 = 0.32, P > 0.50) from that of the three allopatric reaches (12.99 fish /100 m 2 ) . 35 DISCUSSION The evidence for pre- and postmating isolation The fact of hybridization between Dol ly Varden and bull trout in the Thutade watershed demonstrates that reproductive isolation between them is imperfect, but the nature of the hybridization evidence suggests that pre- and postmating isolation are substantial. Because F l hybrids are known to be completely viable after fertilization (Haas and McPha i l 1991), the index of premating isolation between 'pure' parental forms is the rate of formation of F l hybrids. Such hybridization occurs but is infrequent, with the rate being measured to be 0.61% in a total sample of 661 young-of the year (Redenbach 2000). It should be noted that this measure is likely an underestimate, as part of the sample was taken from locally allopatric (parapatric) areas. Not surprisingly, given the scarcity of F l progeny, young-of-the-year from backcross matings (4.8% of total) were more common than were F2's (0.15%). A n additional 1.7% were 'Dolly Varden' that had bull trout mitochondrial D N A , presumably indicating a hybridization event long enough ago to be no longer detectable in the nuclear D N A (because mitochondrial D N A is maternally inherited it is not subject to recombination and can therefore be preserved, unlike nuclear D N A ) . This last also suggests a likely direction of hybridization: Dol ly Varden male with bull trout female. The hybridization evidence suggests strong postmating isolation also, albeit indirectly, otherwise the homogenizing effects of gene flow and recombination would have led to introgression and decayed premating isolation over time. Even extremely low rates of natural hybridization and hybrid fertility appear to have led to introgression in 36 natural populations (Arnold et al. 1999). What follows is a discussion of whether and how results of my study implicate divergent natural selection (as opposed to intrinsic isolating mechanisms), either during adaptation to different resource environments or to reduce the intensity of interspecific competition, as a cause of this pre- and postmating reproductive isolation. A d a p t a t i o n to di f ferent resource env i ronments a n d reproduc t ive isolat ion Tributary stream environment. In the strict sympatry of Thutade Lake's tributary streams, the resource environments of the two char species could be distinguished statistically, but these differences were subtle. Even when the comparisons were statistically significant distributions of all of the resource use variables showed extensive interspecific overlap. Because the overall distributions of each variable for both species combined appear to be continuums, with no evidence of discrete niches, the results did not show much support for the notion that the species are adapted to different resource environments within the tributary streams. There appears to be no obvious basis for natural selection against intermediate, hybrid phenotypes, therefore, during the stream resident phase of their life history. Interspecific comparisons for other salmonid taxa have typically found more obvious resource partitioning. I found no obvious differences between Dol ly Varden and bull trout in their use of microhabitat categories riffle versus pool (Table 1, Figure 3), or in their orientation to the bottom. Within the genus Oncorhynchus, more clearly defined microhabitat partitioning has been documented (examples: Hartman 1965, Johnson and Ringler 1980, Glova 1986, Taylor 1991). Comparisons between members of Oncorhynchus versus Salvelinus have also revealed 37 more obvious differences, with Dol ly Varden (Dolloff and Reeves 1990, Nakano and Kaeriyama 1995) and bull trout (Nakano et al. 1992) both being found closely associated with the stream bottom relative to their Oncorhynchus counterparts. Depth and velocity differences (examples: Lister and Genoe 1970, Everest and Chapman 1972, Dol lof f and Reeves 1990, Taylor 1991, Heggenes et al. 1996) are reported also for sympatric pairs of stream-dwelling salmonid species, yet the only interspecific difference I observed was a marginally significant tendency of Dol ly Varden to occupy deeper habitats relative to bull trout at night (Table 1, Figures 4,5). Differences in feeding behaviour are thought to be an important mechanism allowing co-existence between Salvelinus species and other salmonids, and have typically been linked to diet differences. Both Dol ly Varden (Nakano and Furukawa-Tanaka 1994, Nakano and Kaeriyama 1995, Fausch et al. 1997) and bull trout (Nakano et al. 1992) exhibited a strong orientation toward benthic foraging relative to Oncorhynchus species. In my comparison, however (Table 2, Figure 6), benthic foraging was significantly more important to bull trout than to Dol ly Varden, but this did not have the effect of reducing drift foraging overlap (more important to both species during the day). Stomach contents of the two forms were statistically distinguishable only in terms of prey size (Table 2, Figure 7). Prey size comparisons have generally not been a part of salmonid resource partitioning studies, but prey size is theoretically an important diet resource axis because of its link with morphology (Werner 1977, Wankowski 1979, Wainwright 1986). Indeed, that link can be made for my study system - bull trout, which consumed larger prey on average, had both larger mouths and greater gillraker spacing for a given body size (Table 4, Figure 11). 38 Die l activity partitioning has been reported during interspecific comparisons between co-existing salmonids (Johnson and Ringler 1980) and between salmonids and other taxa (Glova and Sagar 1991), but has not been investigated for stream-dwelling Salvelinus species. M y observations of reduced nighttime feeding activity of Dol ly Varden relative to bull trout (Table 3, Figure 6) has to be interpreted cautiously, as they were made using visible light and not supported to the same degree by the gut content analysis (Table 3, Figure 10). However, even i f feeding efficiency was greatly impaired at night the potential importance of diel activity partitioning should be conceded, as drift abundance was greater by an order of magnitude relative to the day. This difference in drift abundance is typical in streams used by salmonids for rearing (Rader 1997). The above salmonid literature comparisons have suggested that clear dichotomies in resource use are possible for co-existing species in streams. Resource partitioning in animal communities generally is well documented, with a high degree of resource use separation typical between species along at least one major niche dimension (Schoener 1974, Ross 1986 for reviews). The sympatry in streams of Dol ly Varden and another closely related congener, the Asian white-spotted char {Salvelinus leucomaenis), provides an important contrast to the Dol ly Varden/bull trout sympatry. Clear resource partitioning between these species has been demonstrated in Japanese streams, in terms of feeding behaviour, foraging microhabitat, and diet (Fausch et al. 1994, 1997, Nakano and Furukawa-Tanaka 1994). The more aggressive white-spotted char feed almost exclusively from the drift, hold mid-water focal points, and dominate Dol ly Varden of the same size. Conversely, benthic foraging is much more important to the non-aggressive Dol ly Varden, and individuals are closely associated with the stream bottom. These 39 behavioural differences result in substantial diet partitioning. Interestingly, both Dol ly Varden and bull trout follow this pattern of non-aggressive, benthic oriented feeding when in sympatry with salmonid species from the genus Oncorhynchus (Dolloff and Reeves 1990, Nakano et al. 1992, Nakano and Kaeriyama 1995). In my study, then, it is perhaps not surprising that both Dol ly Varden and bull trout in the tributary streams were non-aggressive and closely associated with the bottom - this is the feeding niche they appear adapted to in allopatry. I did record higher drift foraging rates for both species than are known from the above studies, but I suspect the swift current of the steep, mountain streams of the Thutade watershed may favour this more stationary foraging mode. The anomalous apparent lack of resource partitioning by Dol ly Varden and bull trout within the tributary stream environment of the Thutade watershed may have other explanations, as well . Adaptive divergence between the species in this environment may be constrained by: i) historical effects, in terms of either phylogenetic constraints (Ross 1986) due to similar phenotypes, or a lack of time since secondary contact for adaptation in sympatry to proceed very far; ii) a lack of alternative ecological niches in the steep, mountain streams of the Thutade watershed, and iii) natural selection for alternative niches that is weak relative to gene flow, a condition which is not likely to favour adaptive divergence in sympatry (Maynard Smith 1966, Felsenstein 1981, But l in 1989, Rice and Hostert 1993). The observed niche differences in strict sympatry do not suggest an obvious ecological basis for postmating isolation (note: premating isolation was not considered here because bull trout do not mature in this environment), but what this primarily 40 illustrates is the importance of considering the entire life cycle before making inferences about the action of ecological processes. When considered over the entire life history and not just the strict sympatry of the tributary streams, distinctions in the resource environments are obvious. Life history differences. The fact that bull trout undergo a niche shift that Dol ly Varden do not, by migrating from the tributary streams to Thutade Lake and becoming piscivorous (Figure 2), means that the two species are exposed to completely different resource environments for a portion of their lives. In the lake environment, bull trout are exposed to a different predation environment, different food resources (bull trout prey upon kokanee salmon Oncorhynchus nerka and mountain whitefish Prosopium williamsoni in the lake, while the food resource in the streams is primarily invertebrates), and potential interspecific competition from rainbow trout. It appears that adaptation by each of these species has taken place to these different environments. In fact, the formal separation of the species from one another taxonomically was based on morphological differences, in terms of mouth morphology particularly, which were considered to be adaptations to 'piscivore specialist' and 'generalist' feeding niches by bull trout and Dol ly Varden, respectively (Cavender 1978). These differences were consistent throughout the respective ranges of these species, so adaptation to these resource environments appears to have taken place prior to secondary contact. The greater gi l l raker spacing and larger mouths for bull trout that I recorded in my study are consistent with the notion that they are better adapted to the large prey in the lake, and that Dol ly Varden are better adapted to the smaller prey of the streams. B u l l trout selected larger prey than Dol ly Varden even in strict sympatry, which may reflect morphological adaptations to piscivory later in life 41 as well . Presumably, these adaptations come at a cost when feeding on invertebrates, and the smaller mouths and finer gi l l raker spacing of Dol ly Varden are better adapted to the stream environment. Indeed, Wankowski (1979) has suggested that mouth size and gi l l raker spacing set limits to the maximum and minimum size, respectively, of prey that salmonids can handle. Even i f Dol ly Varden and bull trout in the Thutade watershed are experiencing different resource environments, it still remains to link these differences to reproductive isolation between the species. To begin with, the discontinuous nature of the Thutade watershed, with completely different physical and biological environments in lake and stream habitats, suggest a potential mechanism of selection against interspecific hybrids (post-mating isolation). Hybrids between Dol ly Varden and bull trout are known to be intermediate in morphology (Haas and McPhai l 1991), so hybrids in the Thutade watershed may not be adapted well to either lake or stream environments. The fate of hybrids that migrate to the lake can be speculated upon, as small number of adults have been identified from within the bull trout spawning population (Redenbach 2000). The survival rate of these fish relative to bull trout is unknown, but what is apparent is that they do not achieve the large body size of the bull trout adults (Appendix 2, but see Taylor and McPha i l 1995). The obvious potential mechanism for this disparity is that the hybrids, being morphologically intermediate, are less proficient at exploiting the prey fish resource in Thutade Lake. It is known, for instance, that predators that have larger mouth gapes become piscivorous at younger ages and smaller sizes (Mittelbach and Persson 1998). If the smaller body size of hybrids is a result of a lower growth rate then these fish may be at a disadvantage in terms of survival, also. Mortality in fish is typically size-42 dependent (Werner and Hal l 1988, Emerson et al. 1994, Damsgard 1995), so hybrids may take longer to emerge from the size-dependent period of maximum vulnerability to gape-limited predators (adult bull trout). Body size is related to fecundity in fishes, so the smaller body size of the hybrids may also represent a reproductive disadvantage relative to bull trout (Gross 1987, Werner 1988). The evolution of postmating isolation as a result of divergent natural selection has been demonstrated convincingly only in the laboratory (reviewed in Rice and Hostert 1993). However, results from field observations and experiments using other taxa offer support for the notion that the intermediate phenotypes of interspecific hybrids are at a fitness disadvantage in parental niches. For example, it is possible that divergent natural selection for alternative life histories has contributed to post-mating isolation between the grasshopper species Melanopus saguinipes, which is found at higher elevations, and M. devastator, which is found lower and drier locations. Life history traits for populations studied by Orr (1996) changed along a climatic gradient associated with altitude, and, importantly, levels of life history divergence between high- and low- altitude populations were correlated with levels of hybrid inviability in experimental crosses. Another example comes from southwestern British Columbia, where recently evolved, sympatric pairs of stickleback species (Gasterosteus aculeatus) adapted to foraging in benthic/littoral ('benthic' forms) and open water areas ('limnetic' forms) have been described (McPhai l 1984, McPhai l 1992). Hybrids between the species are intermediate in morphology, and in a field experiment have been found to grow more slowly than parental forms in each of the respective parental niches (Hatfield and Schluter 1999). Furthermore, hybrids were less successful than the parental forms at feeding on the prey 43 source that each is adapted to, zooplankton and benthic invertebrates for limnetic and benthic species, respectively (Schluter 1993). Another example is provided by the sympatry of kokanee and sockeye salmon, which are similarly very closely related (Taylor et al. 1996). Kokanee salmon are freshwater residents for their entire lives, while sockeye are anadromous, spending a portion of their life cycle feeding in the ocean. Small kokanee males 'sneak' matings with sockeye females (Foote and Larkin 1988), yet despite the existence of this potential for gene flow the two species are maintaining distinct gene pools (Taylor et al. 1996), suggesting substantial postmating selection against hybrids. Experimental evidence exists for reduced performance of F l hybrids for a variety of traits associated with anadromous and freshwater resident life histories: i) salinity tolerance, ii) swimming performance, iii) growth and development, and iv) age of maturity (reviewed in Wood and Foote 1996). The adaptation of Dol ly Varden and bull trout in the Thutade watershed to different life histories also provides a potential mechanism for strong premating isolation between the species. For animal species generally body size is known to be an important determinant of success in intrasexual competition for mates (examples: Light 1976, Howard and Kluge 1985, M c L a i n and Boromisa 1987). It has been well demonstrated that in fish taxa (including salmonids), in particular, individuals of both sexes select pairing based on body size, both of themselves and prospective mates (Kodric-Brown 1977, Rowland 1982, Jonsson and Hindar 1982, Sargent et al. 1986, Foote 1988, Foote and Larkin 1988, Foote 1989). In the Thutade watershed, then, the great differences in body size alone for the Dol ly Varden, bull trout, and hybrid adults (Appendix 2 - no overlap) allow a prediction of assortative mating among the forms. 44 In the Thutade watershed body size differences also appear to segregate the species in terms of physical habitats for spawning and spawning site locations. B u l l trout spawn in deeper and faster water than do Dol ly Varden and in coarser bed material (Table 5; Appendix 3), and such sites are found in the mainstem reaches of small- to medium-sized Thutade Lake tributaries. Spawning habitats suitable for Dol ly Varden are found in smaller channels, which are formed either by headwater reaches or in very small tributaries. It is even possible that the water quality characteristics of these smaller channels modify the selection pressures on spawning timing and are responsible for the observed differences (Tables 6, 7). The seepage fed tributaries that are most important to spawning Dol ly Varden receive a greater input of groundwater than the larger tributaries do, which means they are warmer over the winter, and are more protected from the spring freshet (Bustard and Royea 1995). Premating isolation is not completely effective, however, as the fact of hybridization obviously attests. The mechanism by which this gene flow occurs may be 'sneak' spawning of the much smaller Dol ly Varden males with bull trout pairs, a behaviour that has been documented for other salmonids (Maekawa and Onozato 1986, Hutchings and Myers 1988, Wood and Foote 1996), and which is consistent with genetic evidence (bull trout mitochondrial D N A in Dol ly Varden -Redenbach 2000). Strong theoretical and experimental support exists for the notion that premating isolation can evolve or be maintained as a consequence of adaptation (summarized in Maynard Smith 1966, Wilson 1989, Rice and Hostert 1993, Bush 1994), but specific conditions are required. These are: i) divergent selection for alternative resource environments is strong relative to gene flow (in sympatry only) and ii) the characters 45 adapting the populations to the environments cause premating isolation themselves or else are strongly linked to other characters that do. One potential mechanism for the crucial second condition is habitat-specific mating, whereby mate choice takes place on or within a preferred habitat or host (Bush 1994). Perhaps the most convincing demonstration of this mechanism, providing some of the best empirical evidence for the importance of divergent natural selection in speciation generally, is the case of the apple maggot fly, Rhagoletetis pomonella. Adults from two sympatric races, hawthorn- and apple-infesting respectively, showed a strong tendency to reproduce on the fruits they developed in as larvae, thereby restricting gene flow between them to approximately 6% per year (Feder et al. 1994). Assortative mating without habitat-specific mating can occur i f the characters adapting the populations to the different environments also form the basis of premating isolation. This mechanism has also been investigated by a small number of field studies, and possibly the best-known example is the sympatry of the benthic and limnetic stickleback species pairs in southwestern British Columbia, which I have discussed previously with respect to postmating isolation. Strong premating isolation exists between the forms, and assortative mating has been experimentally demonstrated to be based on body size. The considerable difference in body size between the benthic and limnetic forms is linked to their exploitation of open water, zooplankton and benthic/littoral invertebrate food resources, respectively (Nagel and Schluter 1998). Sockeye and kokanee salmon, which spawn in the same areas in sympatry, also mate assortatively based on body size (Foote 1988, 1989, Foote and Larkin 1988). The sockeye/kokanee salmon pair is in many respects a parallel example to the sympatry of Dol ly Varden and bull trout. Sockeye salmon achieve their greater size at maturity by 46 migrating to the ocean after a period in freshwater, thereby undergoing a niche shift that kokanee do not. These fish examples suggest that link between body size and sexual selection may be an important mechanism of promoting or maintaining reproductive isolation generally, for the circumstance where ecological specialization leads to large body size differences. Importantly, as corroborating examples they also lend support for my suggestion that assortative mating between Dol ly Varden and bull trout in sympatry may be due to ecological specialization. Interspecific competition and reproductive isolation Tributary stream environment - density compensation. M y analysis indicated that Dol ly Varden and bull trout affected each other's abundances in the tributary stream reaches of the Thutade watershed, suggesting that interspecific competition occurs between them. Importantly, reaches in which both species were present had no greater total carrying capacity (for all fish irrespective of species) than did reaches containing only one species, suggesting that the relative strengths of intraspecific and interspecific competition were similar despite the subtle resource use differences. A crucial assumption of competition theory is that the greater the resource use overlap between species, the greater w i l l be the intensity of interspecific relative to intraspecific competition (MacArthur and Levins 1967, Schoener 1974, 1983, Grant 1994, Robinson and Wilson 1995). Therefore, i f interspecific competition was an important factor in the differentiation of the species in terms of resource use, then these differences should have reduced the intensity of interspecific relative to intraspecific competition, which is not the case. A s has been discussed already, the resource use differences in the tributary stream 47 environments do not seem to be of such a nature as to be important in reproductive isolation between the species. Therefore, regardless of whether or not interspecific competition was playing a role in the divergence of Dol ly Varden and bull trout in terms of resource use in the tributaries, the effects of this competition cannot be linked convincingly to reproductive isolation either. It is possible that interspecific competition in the tributary streams is actually important in maintaining reproductive isolation between the species, yet this study is unable to show it. For species that have complex life histories, our understanding of the importance of an ecological interaction to that species population dynamics requires that we integrate the effects of the interaction over the entire life cycle (Werner and Gi l l i am 1984, Werner 1988, Persson 1988, Olson et al. 1995). I have suggested that the life history differences between the species, which result from a niche shift bull trout make but Dol ly Varden do not, may play a role in assortative mating. Relative mortality:growth ratios in alternative environments have been shown to influence whether or not niche shifts occur in fish (reviewed in Werner 1988). These parameters are affected by interspecific competition, providing a plausible mechanism for competition in the tributaries to contribute to the life history differences. Character displacement in life history. Known almost exclusively to have stream-resident and migratory life histories, respectively, Dol ly Varden and bull trout in sympatry appear to be restricted in their range of life history options relative to allopatry. Three life history strategies have been described for Dol ly Varden in allopatry: i) stream resident for the entire life cycle (as in the case of Thutade fish), ii) adfluvial (migration to a lake for part of the life cycle), and iii) anadromous (Armstrong and Morrow 1980 for 48 review). B u l l trout also exhibit three life history strategies in allopatry: i) adfluvial (as in the case of Thutade Lake fish), ii) fluvial (migration to a larger river for part of the life cycle), and ii i) stream resident (Goetz 1989, McPhai l and Baxter 1996 for reviews). Anadromy is not thought to be important to bull trout, although individuals are known to make short forays at sea (Haas and McPhai l 1991). The possibility that the life histories of the two species are more recognizably different when they are found together in the Thutade watershed than when they are alone in allopatry suggests the possibility of ecological character displacement (Grant 1972, 1994, Robinson and Wilson 1995 for reviews). Dol ly Varden in particular appear to have been excluded from a preferred habitat in the Thutade watershed. In southeast Alaska, nearly all landlocked lakes with native fish faunal assemblages contain Dol ly Varden, although a relatively small proportion of these lakes contain fish that mature at body sizes greater than 400 mm. In Humpback Lake near Ketchikan, however, which contains a kokanee salmon {Oncorhynchus nerka) prey fish base (analogous to Thutade Lake), Dol ly Varden are known to reach 2.7 kg (Armstrong and Morrow 1980). Relatively large, adfluvial Dol ly Varden are known from lakes on the Queen Charlotte Islands (G.R. Haas, British Columbia Ministry of Fisheries, pers. comm.), and mature fish from Buttle Lake on Vancouver Island, which also has kokanee, range in size from 350 to 600 mm (S. Rimmer, British Columbia Ministry of Environment, pers. comm.). Stream resident populations of bull trout are relatively common in allopatry, but what is unclear from reviews (Goetz 1989, McPhai l and Baxter 1996) is whether such populations occur when barriers to migration downstream do not exist or are ephemeral, the situation that exists for Dol ly Varden in the Thutade watershed. In other words: are bull trout always 49 migratory in allopatry when they have the opportunity? One case of intraspecific sympatry of adfluvial and resident forms has been described (McPhail and Murray 1979), but it is unknown i f such sympatry is common (Goetz 1989, McPhai l and Baxter 1996). Conversely, stream-resident populations of bull trout appear to be exceedingly rare in the zone of overlap with Dol ly Varden in northern British Columbia, and none are known from areas of strict sympatry (D. Bustard, Smithers, British Columbia environmental consultant, pers. comm.). Very little reliable information is available for comparative purposes from other areas of Dol ly Varden and bull trout sympatry. The observations from three studies conducted in the Skeena watershed of northwestern British Columbia, however, suggest that the life history patterns observed in the Thutade watershed may occur generally in sympatry (Bustard et al. 1997a, 1997b, 1999). In the watersheds of Goathorn Creek (tributary to the Bulkley River), Gosnell Creek, and Thautil River (tributaries to the Morice River), bull trout are fluvial, using the Bulkley and Morice Rivers respectively for a relatively large portion of their lives. These fish reach an adult body size of 330 to 500 mm, whereas Dol ly Varden are stream resident, and only attain body sizes from 110 to 230 mm at maturity. The spawning locations for the two species are segregated by stream channel size, just as in the Thutade watershed, and the peak of Dol ly Varden spawning is similarly later than that for bull trout. Exceptions to this pattern in the Skeena watershed have not been identified (G.R. Haas, British Columbia Ministry of Environment, pers. comm.; D . Bustard, Smithers, British Columbia fisheries consultant, pers. comm.), although anadromous Dolly Varden are found in the lower reach of the drainage near the downstream extent of the bull trout distribution. 50 I have already discussed the possibility that the life history differences between Dol ly Varden and bull trout in sympatry are linked to pre- and postmating isolation. It also appears possible that these differences are at least partially a result of ecological character displacement. The results of the study are consistent therefore, with the notion that divergent natural selection to reduce the intensity of interspecific competition (ecological character displacement) is contributing to reproductive isolation between the species. It does not appear that the evolution of reproductive isolation because of character displacement has ever been demonstrated directly (Grant 1994, Schluter 1996a), or i f has, examples are few. However, evidence that competition plays a role in speciation comes from studies of adaptive radiations in response to ecological opportunity, which have looked at taxa such as finches and lizards on islands (Schluter 1988, Losos et al. 1997), and fish in recently glaciated lakes (Robinson and Wilson 1994 for review). Divergence in these cases is thought to be driven by competition between colonizing forms, and facilitated by a lack of competition from other taxa (Schluter and McPhai l 1993 for review). The evidence for competition's role in speciation would be more convincing i f the diverging adaptive characters either caused reproductive isolation themselves or were closely linked to characters that did, a condition which experimental evidence and theoretical arguments suggest is required for speciation by natural selection (Maynard Smith 1966, Butl in 1989, Rice and Hostert 1993, Bush 1994). These links are suggested by results from the first experimental test of the character displacement hypothesis, which again concerns the benthic/limnetic stickleback species pairs of southwestern British Columbia. Limnetics were added to a population of a morphologically diverse 51 stickleback species intermediate between the benthics and limetics yet closely related, which then experienced directional natural selection for morphotypes most different from the introduced competitors (Schluter 1994). This experiment recreates the double colonization of lakes that is thought to have led to the benthic/limnetic pairs. Because the adaptations of these forms to different habitats is thought to contribute to pre- and postmating isolation between them (Nagel and Schluter 1998, Hatfield and Schluter 1999), as I've already discussed, a link between character displacement and reproductive isolation is suggested. Beyond this example, however, such a paucity of information exists from field studies relating competition to reproductive isolation that the results of my study are potentially important corroborative evidence. Alternative hypotheses. Although the pattern of life history differences between Dol ly Varden and bull trout in sympatry is consistent with character displacement and adaptation to different resource environments, other explanations for the pattern cannot be ruled out. This situation is typical for correlative studies of the phenomenon of character displacement in particular, and has prevented widespread acceptance of its prevalence in nature (Grant 1972, 1994, Walter et al. 1984, Schluter 1988, Schluter and McPhai l 1992, 1993, Abrams 1990). Each of my two explanations (adaptation to different resource environments alone, character displacement) obviously presents a viable alternative to the other. Body size in salmonids is a sexually selected character (Foote 1988, 1989, Foote and Larkin 1988), so divergence in this character is also consistent with the hypothesis of reinforcement of premating isolation (Dobzhansky 1940) in response to natural selection 52 against hybrids. Alternatively, the observed displacement may be solely a phenotypic response, and not the result of evolution. Yet another potential explanation is that the sympatric populations may have been pre-adapted in allopatry, and the observed pattern is a result of competitive exclusion of incompatible populations rather than character displacement (competition-driven evolution). It also possible that predation contributes to the exclusion of Dolly Varden from migratory life history options in sympatry with bull trout, given the highly piscivorous nature of the bull trout. The information I have gathered during this study is insufficient to discriminate between these various hypotheses, and it may be very difficult to do this at all. However, a consistent element among most of them is that the life history differences in sympatry are a result of divergent natural selection. The one exception in this regard, phenotypic plasticity, is probably amenable to an experimental test. My goal was to investigate the potential role of divergent natural selection in maintaining reproductive isolation between Dolly Varden and bull trout, so perhaps further investigations may be more profitably directed down other avenues than the subdividing of this mechanism of speciation. Crucial to my explanation of how the two species are maintaining sympatry are the following assumptions: i) first of all, that body size and correlated factors are responsible for premating isolation, and ii) postmating isolation by ecological selection against hybrids is occurring, otherwise premating isolation would have broken down over time. Investigating these assumptions should perhaps be the highest priority for further work on this study system. Experimental tests are possible. Bull trout and Dolly Varden adults of similar body sizes can be collected from areas of allopatry, and assortative mating investigated. The importance of ecological factors in postmating isolation can perhaps be 53 best investigated by ruling out the alternative - genetic incompatibility. Experimental hybridization is being conducted currently (E.B. Taylor, University of British Columbia, pers. comm.). If F l and subsequent generations of hybrids are available for ecologically-relevant experiments, then insight as to the mechanisms of postmating isolation may be gained. Conclusion. This investigation was exploratory by design - very little ecological information of any kind existed with respect to the co-existence of Dolly Varden and bull trout in sympatry, let alone the mechanisms of reproductive isolation. Hence, many of the possibilities that are detailed here are just that: possibilities, which are in need of confirmation and clarification. The study was necessarily limited in its temporal and geographic scope, and therefore confirmation of the patterns I observed is needed from other areas of sympatry between these species, and more detailed comparisons required with areas of allopatry. The two crucial assumptions that I have made in order to state that divergent natural selection may be maintaining reproductive isolation also require more direct investigation. As suggested above, these assumptions are that body size and correlated factors are responsible for premating isolation, first of all, and secondly that postmating isolation is not being caused by factors related to genetic incompatibility. Nonetheless, I feel that the results of this study suggest mechanisms whereby ecological factors maintain, and perhaps even promote, pre- and postmating isolation between genetically compatible forms. The results from my study, furthermore, are complementary to those of the other studies of divergence in northern temperate, 54 freshwater environments, in that they treat the potential importance of ecological relative to genetic factors in reproductive isolation for taxa that are later in the speciation process. The first such mechanism of reproductive isolation is divergent natural selection for different life histories that affect body size, a character important in premating isolation. Ecologically-caused body size differences also seem to be an important basis for reproductive isolation in other aquatic study systems also, suggesting that this mechanism may be common. A condition suggested by theoretical arguments and experimental studies (Rice and Hostert 1993, Bush 1994 for reviews) for the evolution of reproductive isolation caused by natural selection, which is that the adapting characters themselves cause or are closely linked to premating isolation, therefore appears to be attainable in nature. When complete premating isolation does not occur, which may be common among taxa that are closely related, postmating isolation due to ecological inferiority of hybrids may act in concert with premating isolation to maintain the sympatry (postmating isolation is theoretically implausible without some kind of mechanism for first reducing gene flow - Rice and Hostert 1993, Bush 1994). The very differences in life history or lifestyle that give rise to body size differences and premating isolation hypothetically also provide a basis for postmating isolation. For Dol ly Varden and bull trout in sympatry, the small body size of hybrids relative to bull trout in the lake environment do suggest the possibility of hybrid inferiority. However, despite how intuitively obvious the notion of postmating isolation because of adaptation is, a compelling demonstration of this process cannot be drawn anywhere from the literature of field studies of animal speciation. 55 With respect to how divergent natural selection in the Thutade watershed has led to the life history differences themselves, alternative theories cannot at this point be discriminated. However, the results of this study are consistent with two such mechanisms: 1. Divergent natural selection has occurred for adaptation to different resource environments. The larger mouths of bull trout and the finer gill raker spacing of Dolly Varden are potentially advantageous for each of these species in their respective life history specializations. 2. Divergent natural selection has occurred to reduce the intensity of interspecific competition. Dolly Varden and bull trout in sympatry have restricted life history options relative to allopatry, a pattern consistent with ecological character displacement. This study therefore is suggestive of future research opportunities concerning these taxa. Understanding the role of ecology in maintaining the sympatry between Dolly Varden and bull trout will require more powerful investigations of the mechanisms of selection leading to the life history differences, and also of the precise nature of postmating isolation. 56 REFERENCES Abrams, P . A . 1990. Ecological vs. evolutionary consequences of competition. Oikos 57: 147-151. Armstrong, R . H . and J. E . Morrow. 1980. The Dol ly Varden charr, Salvelinus malma, p. 99-140. In E . K . Balon (ed.) Charrs: salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague, The Netherlands. Arnold, M . L . , M . R . Bulger, J . M . Burke, A . L . Hempel, and J .H. Will iams. 1999. Natural hybridization: How long can you go and still be important? Ecology 80: 371-381. Barton, N . H . , and B . Charlesworth. 1984. Genetic revolutions, founder effects, and speciation. Ann. Rev. Ecol . Syst. 15: 133-164. Baxter, J.S. 1994. Morphometric analysis and species identification of char from the Kemess Creek watershed. Unpub. memo. Dept. of Zoology, University of B . C . 10 p. Baxter, J.S., E . B . Taylor, R . H . Devlin, J. Hagen, and J.D. McPhai l . 1997. Evidence for natural hybridization between Dol ly Varden and bull trout in a northcentral British Columbia watershed. Can. J. Fish. Aquat. Sci. 54: 421-429. Behnke, R.J . 1972. The systematics of salmonid fishes of recently glaciated lakes. J. Fish. Res. Bd . Can. 29: 639-671. Bernatchez, L . , J. 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. Bush, G . L . , 1994. Sympatric speciation in animals - new wine in old bottles. Trends Ecol . Evo l . 9: 285-288. Bustard, D . 1999 (in prep.). Kemess South Project: 1998 fish monitoring studies. Consult, rept. prep, by D. Bustard and Assoc. for Kemess Mines, Inc. Bustard, D . 1998. Kemess South Project: 1997 fish monitoring studies. Consult, rept. prep, by D . Bustard and Assoc. for Kemess Mines, Inc., Seattle, W A . Bustard, D . 1997. Kemess South Project: 1996 fish monitoring studies. Consult, rept. prep, by D . Bustard and Assoc. for Kemess Mines, Inc., Seattle, W A . 57 Bustard, D . 1996. Kemess South Project: 1995 fisheries studies final report. Consult, rept. prep, by D . Bustard and Assoc. for E l Condor Resources, Ltd. and Royal Oak Mines, Inc., Vancouver, B C . Bustard, D . and D.J . Royea. 1995. Kemess South Gold-Copper Project fisheries studies. Consult, rept. by D. Bustard and Assoc. and Hallam, Knight, Piesold Ltd. for E l Condor Resources Ltd. and St. Phillips Resources Inc., Vancouver, B C . Bustard, D . , R. Dams, and C. Portman. 1999. 1998 stream inventory report, Gosnell Creek watershed. Consult, rept. by D. Bustard and Assoc. for Northwood, Inc. and Forest Renewal B . C . Bustard, D . , R. Dams, and C. Portman. 1997. Stream inventory, Thautil River watershed 1996. Consult, rept. by D . Bustard and Assoc. for Houston Forest Products Ltd. and Forest Renewal B . C . Bustard, D. , R. Dams, and C. Portman. 1998. Aquatic resource baseline studies: Telkwa Coal Project. Draft consult, rept. by D . Bustard and Assoc. for Manalta Coal , Ltd. Butl in, R. 1989. Reinforcement of premating isolation. Pp. 158-179 in D . Otte and J .A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass. Carroll, S.P., H . Dingle, and S.P. Klassen. 1997. Genetic differentioation of fitness-associated traits among rapidly evolving populations of the soapberry bug. Evolution 51: 1182-1188. Carson, H . L . , and A . R . Templeton. 1984. Genetic revolutions in relation to speciation phenomena: the founding of new populations. Ann. Rev. Ecol . Syst. 15: 97-131. Cavender, T . M . 1978. Taxonomy and distribution of the bull trout, Salvelinus confluentus (Suckley) from the American Northwest. Calif. Fish. Game 64: 139-174. Cavender, T . M . 1980. Systematics of Salvelinus from the North Pacific Basin, p. 295-322. In E . K . Balon (ed.) Charrs: salmonid fishes of the genus Salvelinus. Dr. W. Junk, The Hague, The Netherlands. Chandler, C R . 1995. Practical considerations in the use of simultaneous inference for multiple tests. An im. Behav. 49: 524-527. Coyne, J .A. 1992. Genetics and speciation. Nature 355: 511-515. Coyne, J. A . , and H . A . Orr. 1989. Patterns of speciation in Drosophila. Evolution 43: 362-381. 58 Crane, P .A . , L . W . Seeb, and J.E. Seeb. 1994. Genetic relationships among Salvelinus species inferred from allozyme data. Can. J. Fish. Aquat. Sci. 51(Suppl. 1): 182-197. Damsgard, B . 1995. Arctic charr, Salvelinus alpinus (L.), as prey for piscivorous fish - a model to predict prey vulnerabilities and prey size refuges. Nordic J. Freshw. Res. 71: 190-196. Dobson, F.S. 1985. The use of phylogeny in behavior and ecology. Evolution 39: 1384--1388. Dobzhansky, T. 1940. Speciation as a stage in evolutionary divergence. A m . Nat. 74: 312-321. Dolloff, C A , and G . H . Reeves. 1990. Microhabitat partitioning among stream-dwelling juvenile coho salmon, Oncorhynchus kisutch, and Dol ly Varden, Salvelinus malma. Can. J. Fish. Aquat. Sci. 47: 2297-2306. Emerson, S.B., H . W . Greene, and E . L . Charnov. Allometric aspects of predator-prey interactions, p. 123-139. In P .C. Wainwright and S . M . Reil ly, eds. Ecological Morphology. The University of Chicago Press, Chicago, Illinois. Everest, F . H . , and D . W . Chapman. 1972. Habitat selection and spatial interaction by juvenile chinook salmon and steelhead trout in two Idaho streams. J. Fish. Res. B d . Canada 29: 91-100. Fausch, K . D . , and R.J . White. 1981. Competition between brook trout and brown trout for positions in a Michigan stream. Can. J. Fish. Aquat. Sci. 38: 1220-1227. Fausch, K . D . , S. Nakano, K . Ishigaki. 1994. Distribution of two congeneric charrs in streams of Hokkaido Island, Japan: considering multiple factors across scales. Oecologia 100: 1-12. Fausch, K . D . , S. Nakano, and S. Kitano. 1997. Experimentally induced foraging mode shift by sympatric charrs in a Japanese mountain stream. Behav. Ecol . 8: 414-420. Feder, J .L. , S.B. Opp, B.Wlazlo, K . Reynolds, and W. Go. Host fidelity is an effective premating barrier between sympatric races of the apple maggot fly. Proc. Natl . Acad. Sci. 91: 7990-7994. Felsenstein, J. 1981. Skepticism towards Santa Rosalia, or why are there so few kinds of animals? Evolution 35: 124-138. Foote, C.J. 1988. Male mate choice based on male size in salmon. Behaviour 106: 63-80. 59 Foote, C.J . 1989. Female mate choice in Pacific salmon. An im. Behav. 38: 721-723. Foote, C.J. , and P. A . Larkin. 1988. The role of male choice in assortative mating of anadromous and non-anadromous sockeye salmon. Behaviour 106: 43-62. Foster, J.R. 1977. Pulsed gastric lavage: an efficient method of removing the stomach contents of live fish. The Progressive Fish Culturist 39: 166-169. Glova, G.J . 1986. Interaction for food and space between experimental populations of juvenile coho salmon (Oncorhynchus kisutch) and coastal cutthroat trout (Salmo clarki) in a laboratory stream. Hydrobiologia 131: 155-168. Glova, G.J. , and P . M . Sagar. 1991. Dietary and spatial overlap between stream population of a native and two introduced fish species in New Zealand. Aust. J. Mar. Freshwater Res. 42: 423-433. Goetz, F. 1989. Biology of the bull trout, Salvelinus confluentus: a literature review. U . S . D . A . , Willamette National Forest, Eugene, Oregon, 53p. Grant, P.R. 1972. Convergent and divergent character displacement. B i o l . J. L inn . Soc. 4: 39-68. Grant, P.R. 1994. Ecological character displacement. Science 266: 746-747. Grant, P.R., and B . R . Grant. 1989. Sympatric speciation in Darwin's Finches. Pp. 433-457 in D . Otte and J .A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass. Grewe, P . M . , N . Billington, and P . D . N . Hebert. 1990. Phylogenetic relationships among members of Salvelinus inferred from mitochondrial D N A divergence. Can. J. Fish. Aquat. Sci. 47: 984-991. Gross, M . R . 1987. Evolution of diadromy in fishes. American Fisheries Society Symposium 1: 14-25. Haas, G.R. and J.D. McPhai l . 1991. The systematics, zoogeography, and evolution of Dol ly Varden and bull trout in North America. Can. J. Fish. Aquat. Sci. 48: 2191-2211. Hartman, G.F. 1965. The role of behavior in the ecology and interaction of underyearling coho salmon (Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Bd . Canada 22: 1035-1081. Harvey, P . H . 1982. On rethinking allometry. J. Theor. B i o l . 95: 37-41. 60 Hatfield, T., and D . Schluter. 1999. Ecological speciation in sticklebacks: Environment-dependent hybrid fitness. Evolution 53: 866-873. Hearn, W . E . 1987. Interspecific competition and habitat segregation among stream-dwelling trout and salmon: a review. Fisheries 12: 24-31. Heggenes, J., O . M . W . Krog, O.R. Lindas, J.G. Dokk, and T. Bremnes. 1993. Homeostatic behavioural responses in a changing environment: brown trout {Salmo trutta) become nocturnal during winter. J. An im. Ecol . 62: 295-308. Heggenes, J., S.J. Saltveit, and O. Lingaas. 1996. Predicting fish habitat use to changes in water flow: modelling critical minimum flows for Atlantic salmon, Salmo salar, and brown trout, S. trutta. Regulated Rivers-Research and Management 12:331-344. Henderson, M . A . , and T .G . Northcote. Visual prey detection and foraging in sympatry cutthroat trout {Salmo clarki clarki) and Dol ly Varden {Salvelinus malma). Can. J. Fish. Aquat. Sci. 42: 785-790. Hindar, K . , N . Ryman, and G . Stahl. 1986. Genetic differentiation among local populations and morphotypes of Arctic charr, Salvelinus alpinus. B i o l . J. L inn . Soc. 27: 269-285. Howard, R .D . , and A . G . Kluge. 1985. Proximate mechanisms of sexual selection in wood frogs. Evolution 39: 260-277. Hutchings, J .A. , and R . A Myers. 1988. Mateing success of alternative maturation phenotypes in male Atlantic salmon, Salmo salar. Oecologia 75: 169-174. Johnson, J .H. , and N . H . Ringler. 1980. Diets of juvenile coho salmon {Oncorhynchus kisutch) and steelhead trout {Salmo gairdneri) relative to prey availability. Can. J. Zool . 58: 553-558. Johnson, J .H. , and E . Z . Johnson. 1981. Feeding periodicity and diel variation in diet composition of subyearling coho salmon, Oncorhynchus kisutch, and steelhead, Salmo gairderi, in a small stream during summer. Fishery Bulletin 79: 370-376. Jonsson, B . , and K . Hindar. 1982. Reproductive strategy of dwarf and normal Arctic charr {Salvelinus alpinus) from Vangsvatnet Lake, western Norway. Can. J. Fish. Aquat. Sci. 39: 1404-1413. Kodric-Brown, A . 1977. Reproductive success and the evolution of breeding territories in pupfish {Cyprinidon). Evolution 31: 750-766. Leary, R.F . , and F .W. Allendorf. 1997. Genetic confirmation of sympatric bull trout and Dol ly Varden in western Washington. Trans. A m . Fish. Soc. 126: 715-720. 61 Light, L . E . 1976. Sexual selection in toads (Bufo americanus). Can J. Zool . 54: 1277-1284. Lindsey, C C . 1981. Stocks are chameleons: plasticity in gi l l rakers of coregonid fishes. Can. J. Fish. Aquat. Sci. 38: 1497-1506. Lister, D . B . , and H.S. Genoe. 1970. Stream habitat utilization by cohabiting underyearlings of chinook (Oncorhynchus tshawytscha) and coho (O. kisutch) salmon in the B i g Qualicum River, British Columbia. J. Fish. Res. B d . Canada 27: 1215-1224. Losos, J .B. , K . I . Warheit, and T.W. Schoener. 1997. Adaptive differentiation following experimental island colonization mAnolis lizards. Nature 387: 70-73. L u , G . , and L . Bernatchez. 1998. Experimental evidence for reduced hybrid viability between dwarf and normal ecotypes of lake whitefish (Coregonus clupeaformis). Proc. Roy. Soc. Lond. B 265: 1025-1030. Lynch, J.D. 1989. The guage of speciation: on the frequencies of modes of speciation. Pp. 527-553 in D . Otte and J.A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass. MacArthur, R . H . , and R. Levins. 1967. The limiting similarity, convergence, and divergence of co-existing species. A m . Nat. 101: 377-385. M c L a i n , R., and Boromisa, R . D . 1987. Male choice, fighting ability, assortative mating and the intensity of sexual selection in the milkweed beetle, Chauliognathus pennsylvanicus. Evolution 36: 1227-1235. McPhai l , J .D. 1961. A systematic study of the Salvelinus alpinus complex in North America. J. Fish. Res. Bd . Can. 18: 793-815. McPha i l , J.D. 1984. Ecology and evolution of sympatric sticklebacks (Gasterosteus): morphological and genetic evidence for a species pair in Enos Lake, British Columbia. Can. J. Zool . 62: 1402-1408. McPhai l , J.D. 1992. Ecology and evolution of sympatric sticklebacks (Gasterosteus): morphological and genetic evidence for a species pair in Paxton Lake, British Columbia. Can. J. Zool . 70: 361-369. McPhai l , J.D. 1994. Speciation and the evolution of reproductive isolation in the sticklebacks (Gasterosteus) of south-western British Columbia. Pp. 399-437 in M . A . Be l l and S.A. Foster, eds. The Evolutionary Biology of the Threespine Stickleback. Oxford University Press, Oxford. 62 McPhai l , J.D., and C . B . Murray. 1979. The early life history and ecology of Dol ly Varden {Salvelinus malma) in the upper Arrow Lakes. Report to B C Hydro and British Columbia Ministry of Environment, Fisheries Branch, Nelson. 113 p. McPhai l , J.D., and E . B . Taylor. 1995. Skagit char project (Project 94-1). Final Rept. to Skagit Environmental Endowment Commission. 39 p. McPhai l , J.D., and J.S. Baxter. 1996. A review of bull trout {Salvelinus confluentus) life-history and habitat use in relation to compensation and improvement opportunities. Province of British Columbia Fisheries Management Report N o . 104, 35 p. Maekawa, K . , and H . Onozato. 1986. Reproductive tactics and fertilization success of mature Miyabe charr, Salvelinus malma miyabei. Environ. B i o l . Fish. 15: 119-129. Malmquist, H.J . , S.S. Snorrason, S. Skulason, B . Jonsson, O.T. Sandlund, and P . M . Jonasson. 1992. Diet differentiation in polymorphic Arctic charr in Thingvallavatn, Iceland. J. An im. Ecol . 61: 21-35. Maynard Smith, J. 1966. Sympatric speciation. A m . Nat. 100: 637-650. Mayr, E . 1942. Systematics and the origin of species. Columbia University Press. N e w York. Mittelbach, G . G . , and L . Persson. The ontogeny of piscivory and its ecological consequences. Can. J. Fish. Aquat. Sci. 55: 1454-1465. f Morton, W . M . 1970. On the validity of all subspecific descriptions of North American Salvelinus malma (Walbaum). Copeia 1970: 581-587. Mueller, L . D . , and L . Altenberg. 1985. Statistical inference on measures of niche overlap. Ecology 66: 1204-1210. Nagel, L . , and D . Schluter. 1998. Body size, natural selection, and speciation in sticklebacks. Evolution 52: 209-218. Nakano, S., K . D . Fausch, T. Furukawa-Tanaka, K . Maekawa, and H . Kawanabe. 1992. Resource utilization by bull char and cutthroat trout in a mountain stream in Montana, U . S . A . Japanese Journal of Icthyology 39: (3)211-216. Nakano, S., and T. Furukawa-Tanaka. 1994. Intra- and interspecific dominance hierarchies and variation in foraging tactics of two species of stream-dwelling chars. Ecological Research 9: 9-20. 63 Nakano, S., and M . Kaeriyama. 1995. Summer microhabitat use and diet of four sympatric stream-dwelling salmonids in a Kamchatkan stream. Fisheries Science 61: 926-930. Nordeng, H . 1983. Solution to the \"char problem\" based on Arctic char (Salvelinus alpinus) in Norway. Can. J. Fish. Aquat. Sci. 40: 1372-1387. Olson, M . H . , G . G . Mittelbach, and C W . Osenberg. 1995. Competition between predator and prey: resource-based mechanisms and implications for stage structured dynamics. Ecology 76: 1758-1771. Orr, M . R . 1996. Life history adaptation and reproductive isolation in a grasshopper hybrid zone. Evolution 50: 704-716. Orr, M . R . , and T .B . Smith. 1998. Ecology and speciation. Trends Ecol . Evo l . 13: (12) 502-506. Otte, D . 1989. Speciation in Hawaiian crickets. Pp. 482-526 in D . Otte and J .A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass. Persson, L . 1988. Asymmetries in Competitive and Predatory Interactions in fish populations. Pages 203-218 in B . Ebenman and L . Persson, eds. Size-structured populations: ecology and evolution. Springer-Verlag, Berlin. Phillips, R . B . , K . A . Pleyte, and M . R . Brown. 1992. Salmonid phylogeny inferred from ribosomal D N A restriction maps. Can. J. Fish. Aquat. Sci. 49: 2345-2353. Phillips, R . B . , S.L. Sajdak, and M . J . Domanico. 1995. Relationships among charrs based on D N A sequences. Nordic J. Freshw. Res. 71: 378-391 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 . M o l . Phyl . Evol . 1: 223-230. Rader, R . B . 1997. A functional classification of the drift: traits that influence invertebrate availability to salmonids. Can. J. Fish. Aquat. Sci. 54: 1211-1234. Redenbach, Z .R. 2000. Patterns of hybridization between Dol ly Varden (Salvelinus malma) and bull trout (S. confluentus) in nature. M.Sc . Thesis, University of British Columbia. Rice, W.R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Rice, W.R. , and E .E . Hostert. 1993. Laboratory experiments on speciation: what have we learned in 40 years? Evolution 47: 1637-1653. 64 Robinson, B . W . , and D.S. Wilson. Character release and displacement in fishes: a neglected literature. A m . Nat. 144: 596-627. Ross, S.T. 1986. Resource partitioning in fish assemblages: a review of field studies. Copeia 1986(2): 352-388. Rowland, W.J . 1982. Mate choice by male sticklebacks, Gasterosteus aculeatis. A n i m . Behav. 30: 1093-1098. Ryman, N . , F .W. Allendorf, and G . Stahl. 1979. Reproductive isolation with little genetic divergence in sympatric populations of brown trout {Salmo trutta). Genetics 92: 247-262 Savvaitova, K . A . 1980. Taxonomy and biogeography of charrs in the Palearctic, p. 281-481. In E . K . Balon (ed.) Charrs: salmonid fishes of the genus Salvelinus. Dr. W . Junk, The Hague, The Netherlands. Sargent, R . C . , M . R . Gross, and E.P. Van den Berghe. 1986. Male mate choice in fishes. A n i m . Behav. 34: 545-550. Schluter, D . 1988. Character displacement and the adaptive divergence of finches on islands and continents. A m . Nat. 131: 799-824. Schluter, D . 1993. Adaptive radiation in sticklebacks: size, shape, and habitat use efficiency. Ecology 74: 699-709. Schluter, D . 1994. Experimental evidence that competition promotes divergence in adaptive radiation. Science 244: 798-801. Schluter, D . 1996a. Ecological causes of adaptive radiation. A m . Nat. 148 (Suppl. 1): S40-S64. Schluter, D . 1996b. Ecological speciation in postglacial fishes. Proceedings of the Royal Society of London B 351: 807-814. Schluter, D . , and J.D. McPhai l . 1992. Ecological character displacement and speciation in sticklebacks. A m . Nat. 140: 85-108. Schluter, D . , and J.D. McPhai l . 1993. Character displacement and replicate adaptive radiation. Trends Ecol . Evol . 8: 197-200. Schoener, T .W. 1970. Non-synchronous spatial overlap of lizards in patchy habitats. Ecology 51: 408-418. Schoener, T .W. 1974. Resource partitioning in ecological communities. Science 185: 27-39. 65 Schoener, T .W. 1983. Field experiments on interspecific competition. A m . Nat. 122: 240-285. Seber, G . A . F . , and E . D . LeCren, 1967. Estimating population parameters from catches large relative to the population. J. An im. Ecol . 36: 631-643. Sevenster, J .G., and N . Bouton. 1998. The statistical significance of diets and other resource utilization patterns. Netherlands Journal of Zoology. 48: 267-272. Skulasson, S., S.S. Snorrason, D . L . G . Noakes, and M . M . Ferguson. 1996. Genetic basis of life history variations among sympatric morphs of Arctic char, Salvelinus alpinus. Can. J. Fish. Aquat. Sci. 53: 1807-1813. Smith, T .B . , R . K Wayne, D.J. Girman, M . W . Bruford. 1997. A role for ecotones in generating rainforest biodiversity. Science 276: 1855-1857. Tabachnick, B . G . , and L . S . Fidell . 1996. Using Multivariate Statistics, Third Edition. Harper Collins, N e w York. Taylor, E . B . 1991. Behavioural interaction and habitat use in juvenile chinook, Oncorhynchus tshawytscha, and coho, O. kisutch, salmon. An im. Behav. 42: 729-744. Taylor, E . B . , and P. Bentzen. 1993. Evidence for multiple origins and sympatric divergence of trophic ecotypes of smelt (Osmerus) in northeastern North America. Evolution 47: 813-832. Taylor, E . B . , C J . Foote, and C C . Wood. 1996. Molecular genetic evidence for parallel life-history evolution within a Pacific salmon (sockeye salmon and kokanee, Oncorhynchus nerka). Evolution 50: 401-416. Taylor, E . B . , A . Costello, Z .R. Redenbach, S. Pollard, and C. Pacas. 1999. Nested analysis of genetic diversity in bull trout (Salvelinus confluentus): biogeography, speciation, and conservation. Under Review, Can. J. Fish. Aquat. Sci. Wainwright, P . C 1996. Ecological explanations through functional morphology: the feeding biology of sunfishes. Ecology 77: 1336-1343. Wallace, R . K . 1981. A n assessment of diet-overlap indexes. Trans. A m . Fish. Soc. 110: 72-76. Walter, G . H . , P .E. Hulley, A . J . F . K . Craig. 1984. Speciation, adaptation, and interspecific competition. Oikos 43: 246-248. 66 Wankowski. J .W.J. 1979. Morphological limitations, prey size selectivity, and growth response of juvenile Atlantic salmon, Salmo salar. J. Fish. B i o l . 14: 89-100. Werner, E . E . 1977. Species packing and niche complementarity in three sunfishes. A m . Nat. I l l : 553-579. Werner, E . E . 1986. Species interactions in freshwater fish communities. Pages 344-358 in J. Diamond and T. Case, eds. Community ecology. Harper & Row, N e w York. Werner, E . E . 1988. Size, scaling, and the evolution of complex life cycles. Pages 60-81 in B . Ebenman and L . Persson, eds. Size-structured populations: ecology and evolution. Springer-Verlag, Berlin. Werner, E .E . , and J.F. Gi l l iam. 1984. The ontogenetic niche and species interactions in size-structured populations. Ann. Rev. Ecol . Syst. 15: 393-425. Werner, E . E . and D.J . Hal l . 1988. Ontogenetic habitat shifts in bluegill: the foraging rate-predation risk trade-off. Ecology 69: 1352-1366. Wilson, D.S. 1989. The diversification of single gene pools. Pp. 366-385 in D . Otte and J .A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass. Wood, C C , and C J . Foote. 1996. Evidence for sympatric genetic divergence of anadromous and nonanadromous morphs of sockeye salmon (Oncorhynchus nerka). Evolution 50: 1265-1279. Wright, S.P. 1992. Adjusted P-values for simultaneous inference. Biometrics 48: 1005-1013. Zar, J .H. 1996. Biostatistical Analysis. Prentice Hal l , Inc., Upper Saddle River, N e w Jersey. 67 TABLES 1 - 8 68 0 u to c '23 CJ >, M \"3 e C cd O O D TJ ea a cr P-OQ 1 S chi CO c o CO \u00E2\u0080\u00A2c o u o o OJ. as a, o 00 oj 1 \"S > ( 3 , OH , 4\u00E2\u0080\u0094\u00C2\u00BB \u00E2\u0080\u00A24\u00E2\u0080\u0094\u00C2\u00BB -4\u00E2\u0080\u0094\u00C2\u00BB cd * * cd C CN O A 0, A O m \u00C2\u00A9 * cd e O cd c cd c o ^ ^ o cd cd 1=1 C o c-~ r- vo TT' \u00C2\u00A9 cd cd O i n 00 \u00C2\u00A9 \u00C2\u00A9 00 o o o i - H o o CN \u00C2\u00A9 \u00C2\u00A9 00 m \u00E2\u0080\u00A2n m o O VO 00 m o o 0 * * cd c * o d O N o d 2 8 m 2 P o 0 0 d CN OO CN OO I m - H I CN vo ^ o cn cn 11 \u00C2\u00AB OJ -t-\u00C2\u00BB , \ 4-* r\u00E2\u0080\u0094 cd B o -8 cd Q H CQ > Q is 00 o II H ON cn BH CN P -T cn cn 03 II Ji. > Q Q \u00C2\u00AB ffl Q S &, OJ TJ e ' o &, \"cd CJ .2 -4\u00E2\u0080\u0094\u00C2\u00BB (Z> CJ o > o II H O N cn II c > BH CS \u00C2\u00A3 13 II ji o ^ 00 cj -4\u00E2\u0080\u0094< o _o 13 > c ' o Id ^ cn| \u00C2\u00AB 11 H1 <2 e. ,t=.| -T H K > J 3 H > a ffl O .2? ffl fl Q Z 69 < > 2 O u \u00E2\u0080\u00A2S 5 <+-< CJ 3^ S (11 03 C5 O c c o \u00E2\u0080\u00A2c o o o o O-OS a, o co 03 I c CM Q * * o3 CN o d CN O oo o VO VO o 03 XJ CN o cn o r-CN o Os o cn o CN O CN o o d cn o d oo o o l-H o \u00C2\u00A9 00 cn 1^ o o ^ CN 00 cn VO o cn o oo P d o 9 p Q 03 c PQ 60 o d o 2 \u00C2\u00B0 CN 1^ o \u00C2\u00B0 d O >/-> o d CN r-i\u00E2\u0080\u0094I oo 0O *~H d d C/3 03 O IS ll PQ CN > Q e PQ VO CN II H PQ T1-CN II > Q 0\ I 3 T?i Hi A I 60 03 \u00C2\u00BB-i <0 > H > PQ Q 70 Ci Ci oJ ai OH di Ci \u00C2\u00BB C Ci. o GO Ci 1 c Ci T j c Ci & Ci Q * 03 e co o in VO O o O o o oo CN co <^ \u00E2\u0080\u0094; o o d IT) o o VO CN CN 00 o CN T f O i-H o o d d oo CN CN *-H o d vo CN O CO VO CN o o o o CO CN * * c VO CN ON CN r d o I r - | CO r CN CN I CJ) g '5b T j O H oo TT H PQ CN II > Q in 60 c \"5b 03 \"S II CN <^> >> 03 X J e>0 1/3 s O IS > c Ci ll S o H H PQ > c o o 3 3 PQ Q oo 00 B O S c Ci 4\u00E2\u0080\u0094\u00C2\u00BB c o o O CN II H PQ CN II > Q 71 ao co 3 X> T3 C cd C CJ \u00E2\u0080\u00A2s cd 2 \u00C2\u00B0 ^ Q SJ CJ 3 & c o co \u00E2\u0080\u00A2a i * o cj as (3 -8 e >n r- vo o o d d O ON II E-i pq r-cn > Q ON ON cn II c H > PQ Q 72 t-l o co CJ \"cd I1 O CJ CJ \u00C2\u00AB->. OH o o o o \u00C2\u00A9 \u00C2\u00A9 V V o o \u00C2\u00A9 V cd 3 O H 3 00 CJ cn 00 II c CN CN OO CN cn in CN oo CN II 00 vb cn cn d oo CN II ON OH OJ 00 I CN GO 3 < i CN CN CJ 1 o Q CO CJ fc! oo CJ OV d CN oo OV d o cn cn as wn o cn d CN o cn II vo o O \u00E2\u0080\u00A2 oo o OH aj 00 I cn o OJ a cd \u00E2\u0080\u0094 CO co O o 5 \u00E2\u0080\u00A2e \u00C2\u00AB l l CJ X3 cj | l OH S 8 OJ & CO .1) OJ 3 \" f i cd XI .\u00C2\u00ABJ &H OJ TJ OJ (50 c 'S cd OH 00 1 CJ cj o > c e 3 cj OJ ao c \u00E2\u0080\u00A2 i-H cd OH 00 B CJ * O in Q \"cd \"S a TJ OJ X) aj 00 C '3 cd DH 00 * * TJ O \"B OH bO g cd OH 00 73 c \u00C2\u00A9 o fl u co fl \"V \u00C2\u00AB ON O e -4\u00E2\u0080\u0094> O H 00 X J -a C N co C N o o o o o o o o o o o o o o o o 1 cf1 x] I rt \u00C2\u00AB \u00C2\u00AB J fl tu ,o cu TJ 2 > S \"o '\u00C2\u00AB> Q SP ^ CU QQ CU CU CU 4\u00E2\u0080\u0094\u00C2\u00BB 03 T3 >. CU > 3 o & 03 E 3 o oo XJ XJ cu e o o ' - 4 \u00E2\u0080\u0094 < 03 3 cj 00 XJ XJ CU CU fl o - 4 \u00E2\u0080\u0094 * o (U \"3 E 3 CJ 00 X J XJ a cu c o 03 o cu E 3 o XJ XJ (U CU c o 03 \u00E2\u0080\u00A2 4 \u00E2\u0080\u0094 \u00C2\u00BB o CU _> E 3 O o ;-CU oo Os os vo Os Os Os o\ Os 74 TJ TJ CU i-i O 6 c 13 o H m o TI-O N 00 00 o CN o CN in o o o o o m o o c3 S3 03 c cn m o o CU TJ CO 3 \u00C2\u00ABJ 6 o p. D . a (U C O i 00 CN Cu CU C O I l\u00E2\u0080\u0094H CN D-cu C O I CN oo 3 < I O cn 03 O c m 03 O S3 m m o m in o o o Tt\" oo 03 00 C cn vo oo cn cn o o cn CN T f r CN 0O o o 00 o O 03 c 03 C 03 03 C o3 S3 03 Tf vo vo cn 03 C in 03 S3 03 (3 00 3 < i cn CN o o o o o o o3 e 03 S3 o3 C 03 S3 00 TJ TJ CU c o o3 O cu ^> B 3 O on TJ TJ cu I - I CU c 4 - 1 o 03 \u00E2\u0080\u00A2*\u00E2\u0080\u0094\u00C2\u00BB O CU e 3 cj TJ TJ S CU S3 o 03 4\u00E2\u0080\u0094\u00C2\u00BB o CJ 03 3 o TJ TJ Si CU S3 4w O * 03 \u00C2\u00BB O -*-> <+-< o CU _> '-*\u00E2\u0080\u0094< 03 3 a 3 o 00 TJ TJ cu c o _03 O CU _> '-\u00C2\u00BB-> 03 3 e 3 o 03 CU oo Os os O N O N V O Os m Os Os T i -ers O N 75 oo ON as r-H I >n as as CO CO C CD X ) X= o cd B CD W) cd V-CD > < Id E 2 0 . E Q \u00C2\u00A9 pq p CD O o cj td U X ! o cd CD o o SO \u00E2\u0080\u00941 o m 00 CM C N vo m r~; cn Ov OO ON SO T-H ON CN d 00 so o o 00 CO. ON ON C N in * # in cn ON r\u00E2\u0080\u0094I ON ^ \u00C2\u00ABn cn CN xr ^ CN cn vo . 1\u00E2\u0080\u00941 oi 0 h 2 \u00C2\u00AB 00 cn i-H 00 CN \u00C2\u00A7 00 o . Ov h 0 CN d xr CN cn CN CN \u00E2\u0080\u00941 m > Q H CQ > Q & & & & & & & & 0-1 E & E ^ E E J S E E t o c d c o c d t / 3 t o c d c o c o CD CD u CD u CD E CD X ! ti o to CD CD x : t: o CD & OH CD g u CO CO CD E CD CD > cd \u00E2\u0080\u00A2 ^ XT < cd CD CH 3 X) CD CD 1-1 u cd CH X ! H\u00E2\u0080\u0094\u00C2\u00BB 3 O CO CD CD 1-u 13 o fe 3 O CD > s cd IS CD < < Z 76 FIGURES 1 -14 77 Figure 1. Distribution of Dol ly Varden (Salvelinus malma) and bull trout (S. confluentus) in western Canada, showing in particular areas where the two species' distributions are thought to overlap' (dark shading). The study site is the Thutade Lake watershed, located at approximately 57.0\u00C2\u00B0 N latitude and 126.7\u00C2\u00B0 W longitude at the source of the Finlay River. 78 79 Figure 2. Life histories of Dol ly Varden and bull trout in the Thutade watershed. Dol ly Varden remain residents of relatively small tributary streams for their entire life cycle. B u l l trout spend up to four years in these same streams at the beginning of their lives, but then undergo a niche shift, migrating to Thutade Lake and becoming piscivorous. 80 Life History Differences adult DV juvenile BT, DV (invertebrate feeders) tributary streams Thutade Lake adult, sub-adult BT (piscivorous) 81 Figure 3. Interspecific comparisons of habitat type use during (a) daytime (chi square 5.76, P < 0.25) and (b) nighttime (chi square = 3.10, P < 0.50) by sympatric bull trout (dark shading; daytime n = 32, nighttime n = 40) and Dolly Varden (light shading; daytime n = 37, nighttime n = 39). 82 0.8 (a) day BT f| DV 0.6 0.4 o o w o Q_ o J - 0.6 (b) night 0.4 0.2 main ch. pool main ch. riffle side ch. pool side ch. riffle Hydraulic Habitat T ype 83 Figure 4. Interspecific comparisons of stream depth at focal point during (a) daytime (F = 1.99, P = 0.16) and (b) nighttime (F= 6.91, P = 0.010, adjusted a = 0.008) for sympatric bull trout (closed circles; daytime n = 32, nighttime n = 40) and Dolly Varden (open circles; daytime n = 37, nighttime n = 39). 84 100 50 (a) day o o o 20 E 10 o CD c3 1 0 0 CO 50 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 t o \u00E2\u0080\u00A2 - \u00E2\u0080\u00A2 o \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 o O o o o o \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 3 ; o BT DV J L (b) night o o o o o \u00C2\u00B0 o. o o \u00C2\u00ABp . * o o o 20 10 BT DV j i 50 70 90 110 130 150 170 190 Fork Length (mm) 85 F i g u r e 5. Interspecific comparisons of mean current velocity at focal point during (a) daytime (Slopes: F= 6.32, P = 0.014) and (b) nighttime (F= 2.83, P = 0.097) for sympatric bull trout (closed circles; daytime n - 32, nighttime n = 40) and Dolly Varden (open circles; daytime n = 37, nighttime n = 39). Differences were not statistically significant after sequential Bonferroni adjustment for multiple tests. 86 100 50 20 10 (a) day E o 8 2 CD > 1 C 100 CD 03 CD 20 10 BT DV \u00E2\u0080\u0094 \u00E2\u0080\u0094 \u00E2\u0080\u00A2 -\u00C2\u00A9\u00E2\u0080\u00A2 J L (b) night o o o o o o o \u00E2\u0080\u00A2 o 2 1 BT DV 50 70 90 O -e\u00E2\u0080\u00946 o 110 130 150 170 190 Fork Length (mm) 87 Figure 6. Observations of foraging mode during (a) daytime and (b) nighttime for sympatric bull trout (dark shading; daytime n = 18, nighttime n = 26) and Dol ly Varden (light shading; daytime n = 24, nighttime n = 24). Figures show adjusted means, but interspecific comparisons were made by the analysis of covariance technique (see results). 88 12 10 8 (a) day c \"E LO (/) -t\u00E2\u0080\u0094' Q . E CD CD 10 C CD 03 0 (b) night BT DV 0 Drift Benthic Foraging Mode Total 89 Figure 7. Interspecific comparison of average prey item biomass (F = 5.89, P = 0.017, adjusted a = 0.017) for sympatric bull trout (closed circles; n = 49) and Dolly Varden (open circles; n = 39). 90 20 D V 1 (mg) 10 BT \u00E2\u0080\u00A2 \"ey Biomass + 5 3 \u00E2\u0080\u00A2 i o \u00E2\u0080\u00A2 o \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 o \u00E2\u0080\u00A2 CL cb 2 \u00E2\u0080\u00A2 w ^ ^ ^ ^ ^ o ^ ^ 2 ^ 0 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 Av 1 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 * | \u00E2\u0080\u00A2 , o 8 o i i i 50 70 90 110 130 150 170 190 Fork Length (mm) 91 Figure 8. Interspecific comparison of taxonomic composition of daytime stomach contents (Schoener Index of Proportional Overlap = 0.71) for sympatric (a) bull trout (n 24) and (b) Dolly Varden (n = 21). 92 a) bull trout Diptera Adult 11.3% Hymenoptera 0.6% Diptera Larvae 14.3% b) Dolly Varden Diptera Adult 33.5% Ephemeroptera 46.2% Tricoptera 11.4% Plecoptera 10.6% Coleoptera 5.5% Ephemeroptera 33.6% Hymenoptera 7.5% Tricoptera 3.2% Plecoptera 8.4% Coleoptera 5.7% Diptera Larvae 8.0% 93 Figure 9. Interspecific comparison of taxonomic composition of nighttime stomach contents (Schoener Index of Proportional Overlap = 0.89) for sympatric (a) bull trout (n 27) and (b) Dolly Varden (n = 21). 94 a) bull trout Diptera Adult 11.9% Hymenoptera 0.7% Diptera Larvae 13.7% b) Dolly Varden Diptera Adult 12.0% Hymenoptera 2.8% Ephemeroptera 52.7% Fish 0.5% Tricoptera 11.4% Plecoptera 6.9% Coleoptera 2.2% Ephemeroptera 43.0% Diptera Larvae 22.6% Fish 0.4% Tricoptera 11.0% k Plecoptera 6.2% Coleoptera 1.9% 95 Figure 10. Interspecific comparisons of total prey biomass of (a) daytime (F= 1.31, P 0.26) and (b) nighttime (F= 0.10, P = 0.76) stomach samples for sympatric bull trout (closed circles; daytime n = 24, nighttime n = 27) and Dolly Varden (open circles; daytime n = 21, nighttime n = 21). 96 100 (a) day o o o o \u00C2\u00B0 . 10 C O E Q 0.1 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 B T \u00C2\u00A3 1 0 0 c o O (b) night CD 10 o o - * o 0.1 J L DV \u00C2\u00AB 3 -J l I I I 50 70 90 110 130 150 170 190 Fork Length (mm) 97 Figure 11. Interspecific comparisons of trophic morphology for sympatric bull trout (closed circles; n = 49) and Dol ly Varden (open circles; n = 37), in terms of (a) mouth width (Slopes: F= 4.8, P = 0.03) and (b) gillraker spacing (Slopes: F= 7.14, P = 0.009). 98 99 Figure 12. Tributary reaches within the Thutade Lake watershed used for the investigation of density compensation. Reaches of local allopatry are indicated by the symbols [ B T ] and [ D V ] for bull trout and Dolly Varden, respectively. Reaches of strict sympatry are indicated by the symbol [ SP ]. Thutade Lake is drained at its north end by the Findlay River, a tributary of the Peace River. 100 Thutade Lake watershed study reaches Figure 13. Density compensation within independent stream reaches of the Thutade Lake watershed, British Columbia. The slope of the regression of Dol ly Varden densities on bull trout densities (solid line; n = 7, (3 = -0.94) was significantly different from the null hypothesis slope of zero (t = 2.79, P < 0.025), but was not different statistically from the null hypothesis slope of minus-one (dashed line; t = 0.19, P > 0.50). 102 Observed 0 5 10 15 BT Density (parr/100m*m) 103 Figure 14. Comparison of average total parr densities for both species combined between reaches of local allopatry (light shading; n = 3) and sympatry (dark shading; 4). The comparison was not statistically significant (t = 0.32, P > 0.50). 104 105 APPENDICES 1 - 3 106 Appendix 1, Table Al . Average prey size data collected during preliminary field work in 1997. Preserved stomach contents analyzed in the laboratory by L . Sulek, University of British Columbia, Sept. 1997. No. Species Standard length (mm) Average prey length (mm) 1 dv 111 7.2 2 dv 92 2.9 3 dv 79 5.4 4 dv 77 2.7 5 dv 71 5.8 6 dv 65 6.8 7 dv 74 2.7 8 dv 63 3.2 . 9 dv 72 5 10 dv 75 4 11 dv 90 5.8 12 dv 75 5.5 13 dv 77 6.2 14 dv 72 4.2 15 dv 68 6 16 dv 79 4.1 17 dv 85 3.3 18 dv 92 4.7 19 dv 99 3.5 20 dv 83 4.5 21 dv 63 4.5 1 bt 107 6.3 2 bt 81 5.5 3 bt 76 3.8 4 bt 84 4.1 5 bt 86 6.3 6 bt 83 5.3 7 bt 81 4.6 8 bt 79 4.3 9 bt 62 6.4 10 bt 73 6.8 11 bt 77 8.5 12 bt 77 7 13 bt 73 5.6 14 bt 55 4.7 15 bt 95 4.7 16 bt 60 4.7 17 bt 130 17.3 107 Appendix 1, Figure Al . Average prey item lengths for Thutade watershed bull trout (closed circles; n - 17) and Dol ly Varden (open circles; n = 21). Samples were collected from throughout the watershed in the summer of 1997, during preliminary studies of the watershed. 108 109 Appendix 2. Fork lengths for mature, Thutade watershed char. Dol ly Varden species identity was confirmed by morphometric analysis (Baxter 1994, Haas and McPha i l 1991), while bull trout and hybrid species identities were confirmed using molecular genetic tech-niques (Z.R. Redenbach and E . B . Taylor, University of British Columbia Department of Zoology, unpubl. data). Species Fork length (mm) Sex ^ 0 ^ Species Fork length (mm) Sex 1 D V 113 m 14 B T 760 f 2 D V 115 m 15 B T 720 f 3 D V 117 m 16 B T 850 m 4 D V 120 m 17 B T 650 f 5 D V 122 f 18 B T 840 m 6 D V 137 m 19 B T 850 m 7 D V 142 f 20 B T 700 m 8 D V 143 f 21 B T 840 m 9 D V 169 f 22 B T 760 m 10 D V 172 f 23 B T 850 m 11 D V 194 m 24 B T 560 f 12 D V 235 f 25 B T 670 f 1 B T 700 f 26 B T 440 m 2 B T 740 f 27 B T 700 m 3 B T 750 f 28 B T 760 m 4 B T 730 f 29 B T 800 f 5 B T 650 f 30 B T 740 m 6 B T 700 f 31 B T 710 m 7 B T 740 f 32 B T 890 m 8 B T 760 f 33 B T 750 m 9 B T 800 f 34 B T 740 f 10 B T 580 f 1 Hybrid 370 na 11 B T 810 f 2 Hybrid 470 f 12 B T 820 f 3 Hybrid 370 m 13 B T 830 m 110 Appendix 3. Hydraulic and bed material characteristics at spawning sites located through-out the Thutade watershed (Bustard 1996). Dolly Varden Site Velocity Depth D50* left right mean left right mean Seep.B 31 16 23.5 6 4 5 0.8 Seep.B 19 16 17.5 4 6 5 1.5 Seep.B 16 25 20.5 4 3 3.5 1.5 Seep.B 17 8 12.5 5 4 4.5 1.5 Seep.B 24 5 14.5 6 9 7.5 2 Seep.B 28 4 16 6 3 4.5 1.5 N K 2 1 16 40 28 9 9 9 1.5 N K 2 1 32 22 27 7 10 8.5 2 N K 2 1 19 36 27.5 17 20 18.5 2.5 ' N K 2 1 31 3 17 10 12 11 2.5 N K 2 1 42 7 24.5 8 12 10 2.5 N K 2 1 32 27 29.5 10 14 12 2.5 N K 2 1 20 22 21 14 9 11.5 2 N K 2 1 15 32 23.5 5 8 6.5 2 T r i b H 39 32 35.5 11 11 11 1.5 T r i b l 4 13 29 21 16 32 24 1.5 T r i b H 18 15 16.5 17 15 16 1.5 T r i b l 4 27 13 20 19 14 16.5 1 T r i b H 23 41 32 5 14 9.5 1.5 T r i b H 36 29 32.5 14 14 14 1 T r i b H 36 38 37 9 11 10 1 T r i b H 11 28 19.5 4 6 5 1.5 T r i b H 16 29 22.5 5 8 6.5 1 T r i b H 28 17 22.5 12 8 10 0.8 Attycell. 4 3 3.5 11 11 11 2 Attycell . 19 3 11 2 4 3 1 Attycell. 8 8 8 6 7 6.5 2 Attycell. 5 1 3 7 7 7 1 Attycell. 14 15 14.5 6 7 6.5 1 Attycell . 21 19 20 5 4 4.5 1 Defined in methods 111 Appendix 3, continued. Hydraulic and bed material characteristics at spawning sites located throughout the Thutade watershed (Bustard 1996). Bull trout Site Velocity Depth D50* left right mean left right mean N . K e m . 63 25 44 29 28 28.5 6 N . K e m . 41 29 35 31 31 31 7 N . K e m . 46 11 28.5 24 27 25.5 6 N . K e m . 43 16 29.5 25 22 23.5 4 N . K e m . 2 19 10.5 15 21 18 5 N . K e m . 68 3 35.5 21 28 24.5 5 N . K e m . 10 17 13.5 46 47 46.5 4 N . K e m . 36 5 20.5 30 42 36 5 N . K e m . 14 51 32.5 28 24 26 5 N . K e m . 52 0 26 29 40 34.5 5 Attichik. 51 30 40.5 52 19 35.5 5 Attichik. 51 19 35 23 21 22 4 Attichik. 28 42 35 57 67 62 6 Attichik. 52 49 50.5 39 48 43.5 4 Attichik. 66 59 62.5 43 48 45.5 7 Attichik. 50 40 45 53 61 57 5 Attichik. 14 28 21 66 54 60 6 Attichik. 29 43 36 40 26 33 5 Attichik. 50 61 55.5 48 25 36.5 7 Attichik. 8 37 22.5 40 47 43.5 5 Niven . 27 48 37.5 39 34 36.5 2 Niven 41 55 48 24 25 24.5 2 Niven 43 21 32 18 13 15.5 4 Niven 39 48 43.5 33 26 29.5 6 Niven 34 65 49.5 39 38 38.5 5 Niven 37 51 44 28 36 32 5 Niven 24 37 30.5 39 35 37 6 Niven 62 68 65 25 25 25 2 Defined in methods 112 "@en . "Thesis/Dissertation"@en . "2000-05"@en . "10.14288/1.0099497"@en . "eng"@en . "Zoology"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Reproductive isolation between Dolly Varden (salvelinus malma) and bull trout (s. confluentus) in sympatry : the role of ecological factors"@en . "Text"@en . "http://hdl.handle.net/2429/10305"@en .