UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hybridization in western trout : spatial variation and the role of environmental factors Yau, Monica M. 2013

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

Item Metadata

Download

Media
24-ubc_2014_spring_yau_monica.pdf [ 1.75MB ]
Metadata
JSON: 24-1.0165667.json
JSON-LD: 24-1.0165667-ld.json
RDF/XML (Pretty): 24-1.0165667-rdf.xml
RDF/JSON: 24-1.0165667-rdf.json
Turtle: 24-1.0165667-turtle.txt
N-Triples: 24-1.0165667-rdf-ntriples.txt
Original Record: 24-1.0165667-source.json
Full Text
24-1.0165667-fulltext.txt
Citation
24-1.0165667.ris

Full Text

Hybridization in Western Trout:spatial variation and the role ofenvironmental factorsbyMonica M. YauB.Sc., The University of British Columbia, 2008A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinThe Faculty of Graduate and Postdoctoral Studies(Zoology)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)November 2013c? Monica M. Yau 2013AbstractHybridization and introgression with non-native salmonids is one of thegreatest factors threatening native cutthroat trout species. Westslope cut-throat trout (Oncorhynchus clarkii lewisi ; WSCT) were recently listed un-der the Canadian Species at Risk Act (SARA) as ?special concern? (BritishColumbia populations) and ?threatened? (Alberta populations). I employeda 10 locus-microsatellite DNA assay to investigate levels of hybridization be-tween westslope cutthroat trout and introduced rainbow trout (O. mykiss;RT) at 159 sampling locations in southwestern Alberta and parts of south-eastern British Columbia. My results revealed that hybridization is exten-sive across the region sampled. Admixture levels (qwsct of 0 = pure rainbowtrout, 1.0 = pure westslope cutthroat trout) at sampling locations rangedfrom 0.01 to 0.99. An average qwsct below 0.99 is a criterion that has beenused in previous work to designate a population as ?hybridized.? Landscapegenetic analysis using regression trees indicate that water temperature, el-evation, distance to the nearest stocking site and distance to the nearestrailway were significant components of a model that described 34% of thevariation in qwsct across 58 sites for which habitat variables were available.Building on this finding, I explored the role of water temperature, the bestpredictor of hybridization levels amongst the variables tested, in limitingthe spread of admixture by evaluating cold tolerance in both species usingcritical thermal methods (CTM). Analysis of variance revealed a statisti-cally significant difference between the critical thermal minima (CTMin) ofWSCT and RT acclimated to 15 ?C (1.0 ? 0.8 ?C and 1.4 ? 1.0 ?C, respec-tively). The heritability of cold tolerance observed in this study appears tobe complex and does not seem to behave in a simple additive manner. Theidentification of water temperature as a major factor influencing admixtureand subsequent test for physiological differences in cold tolerance provideevidence to support a hypothesis that cold water habitats act as a naturalbarrier to hybridization between WSCT and RT. This information providesinsight into the evolutionary history of WSCT and RT and will be usefulin assisting conservation efforts aimed at mitigating the wide-spread loss ofWSCT to genomic extinction.iiPrefaceThe experiments carried out in Chapter 3 were covered by Animal CareCertificate number A11-0009.Chapter 2. A version of this chapter has been published as [Yau M.M. and Taylor E. B.. Environmental and anthropogenic correlates of hy-bridization between westslope cutthroat trout (Oncorhynchus clarkii lewisi)and introduced rainbow trout (O. mykiss). Conservation Genetics 14:855-900, 2013]. I was responsible for data collection, data synthesis, conceptformation and manuscript composition. Tissue samples were collected byAlberta Sustainable Resource Development and Parks Canada. Jen Gowand Eric Taylor developed the genetic methods that I employed in thisstudy. Eric Taylor was the supervisory author on this project and was in-volved throughout the project in concept formation, statistical analysis andmanuscript composition.iiiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Westslope Cutthroat Trout . . . . . . . . . . . . . . . . . . 41.2 Hybridization with Rainbow Trout . . . . . . . . . . . . . . . 51.3 Conservation Status of Westslope Cutthroat Trout . . . . . . 101.4 Hybridization Studies . . . . . . . . . . . . . . . . . . . . . . 111.5 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . 132 Distribution of hybridization and environmental factors . 142.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . 162.2.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . 162.2.2 Genetic analyses . . . . . . . . . . . . . . . . . . . . . 162.2.3 Population genetic analyses . . . . . . . . . . . . . . . 182.2.4 Admixture analyses . . . . . . . . . . . . . . . . . . . 182.2.5 Stream characteristics and anthropogenic variables . 192.2.6 Statistical analysis . . . . . . . . . . . . . . . . . . . . 202.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.3.1 Genetic analyses . . . . . . . . . . . . . . . . . . . . . 232.3.2 Influences on admixture values . . . . . . . . . . . . . 24ivTable of Contents2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282.4.1 Spatial variation in admixture . . . . . . . . . . . . . 282.4.2 Environmental correlates of admixture . . . . . . . . 292.4.3 Anthropogenic correlates of admixture . . . . . . . . 322.4.4 Implications for recovery planning . . . . . . . . . . . 353 Cold tolerance limits hybridization between westslope cut-throat trout and rainbow trout . . . . . . . . . . . . . . . . . 373.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 373.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . 403.2.1 Trout populations . . . . . . . . . . . . . . . . . . . . 403.2.2 Rearing set-up . . . . . . . . . . . . . . . . . . . . . . 423.2.3 Acclimation . . . . . . . . . . . . . . . . . . . . . . . 433.2.4 Critical thermal minima determination . . . . . . . . 443.2.5 Statistical analysis . . . . . . . . . . . . . . . . . . . . 453.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453.3.1 Effect of acclimation temperature . . . . . . . . . . . 453.3.2 Effect of body size . . . . . . . . . . . . . . . . . . . . 453.3.3 Interspecific differences . . . . . . . . . . . . . . . . . 483.3.4 Intraspecific differences in rainbow trout . . . . . . . 483.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513.4.1 Acclimation temperature . . . . . . . . . . . . . . . . 523.4.2 Inter- and intraspecific differences . . . . . . . . . . . 533.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584.1 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . 584.2 Defining ?Genetic Purity? . . . . . . . . . . . . . . . . . . . . 594.3 Anthropogenic Hybridization . . . . . . . . . . . . . . . . . . 604.4 Concluding Thoughts and Future Directions . . . . . . . . . 61References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63AppendicesA Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78B Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93vList of Tables2.1 Definition of variables potentially explaining variation in ad-mixture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.2 Variable importance values and model improvement ratiosfor the full and final models determined from randomForestanalysis on admixture values . . . . . . . . . . . . . . . . . . 253.1 Critical thermal minima, length and mass of trout test groupsacclimated at 15 ?C and 18 ?C. . . . . . . . . . . . . . . . . . 493.2 Summary statistics for analysis of covariance (ANCOVA) be-tween critical thermal minima (CTMin) and body size introut test groups at 15 ?C and 18 ?C acclimation . . . . . . . 50A.1 Summary of genetic data for sampling locations in southwest-ern Alberta and southeastern British Columbia, Canada . . . 78A.2 Variable data of streams used in RANDOMFOREST model . 90B.1 Critical thermal minima, length and mass data for each trouttested . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93viList of Figures1.1 The percent of all listed endangered species threatened byhabitat loss, introduced species, overexploitation and pollu-tion in the United States and Canada . . . . . . . . . . . . . 21.2 The percent of endangered fish species (marine and freshwa-ter) threatened by habitat loss, introduced species, overex-ploitation and pollution in the United States and Canada . . 31.3 Native Range of westslope cutthroat trout (Oncorhynchusclarkii lewisi) . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.4 Native Range of rainbow trout (O. mykiss) in North America 71.5 Juvenile westslope cutthroat trout (O. c. lewisi), rainbowtrout (O. mykiss) and westslope cutthroat trout x rainbowtrout hybrid . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.1 Map of sampling locations in southwestern Alberta and south-eastern British Columbia, Canada . . . . . . . . . . . . . . . 172.2 Observed admixture values (qwsct) and predicted values de-rived from RANDOMFOREST . . . . . . . . . . . . . . . . . 262.3 Partial dependence plots for water temperature, elevation,distance from sample site to stocking site and distance fromsample site to nearest railway line . . . . . . . . . . . . . . . 273.1 Source locations of trout used in critical thermal minima study 413.2 Simple linear regressions of CTMin on acclimation tempera-ture for trout test groups . . . . . . . . . . . . . . . . . . . . 463.3 Critical thermal minima of trout test groups . . . . . . . . . . 47B.1 Spread of CTMin data of trout test groups acclimated at15 ?C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106B.2 Spread of CTMin data of trout test groups acclimated at18 ?C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107viiAcknowledgementsFirst, I would like to thank my supervisor, Dr. Eric Taylor for his guidance,support, academic input, and the freedom he provided to make this projecttruly ?my-own?. Thank you for giving me the reigns and allowing me tolearn what was too much and when to pull back; for the opportunities towork on research beyond my thesis, that challenged me and allowed me toperfect my craft. Your dedication to your work and fish conservation issomething to be admired.I would also like to thank my ?mentor?, Patrick Tamkee, who created avision of my future and is adamant that I achieve it. He was with me everystep of the way in this project and without his support none of this wouldhave been possible. I owe you.I would like to thank my committee members, Trish Schulte and JohnRichardson for their helpful input and suggestions.Thank you to Dr. Jennifer Gow for technical support and developingthe genetic assay used in this study. Immense thanks to Charlie Pacas(Parks Canada) and Jennifer Earle (Alberta Sustainable Resource Develop-ment) who helped coordinate the sampling of tissue used in this study andfor supplying environment data used in the GIS analysis. Big thanks toSara Northrup and everyone at the Freshwater Fisheries Society of BritishColumbia, who supplied fish used in this study. Sara, you were an invaluableresource in everything from transport logistics to spawning protocol.The Taylor Lab (past and present): Les Harris, Jon Mee, JS Moore, MattSiegle, Stefan Dick, Patricia Woodruff, Chad Ormond, Damon Nowasad,Carla Crossman, Shannan May-McNally, Jenn Ruskey, Amanda Moirera,thanks for all the encouragement and support.Many thanks to the UBC Department of Zoology, the professors, thestudents, and more specifically Vince and Bruce from the Shop, and ourgrad secretary, Alice Liou, for helping me through the submission process.I need to acknowledge Victor Chan, Jessica Lu and Haley Kenyon whoassisted with fish care when I was unable to do so. Sarhiar Kabir and AshleySmith who helped with spawning and transporting fish to UBC. Carita Chanand Rebecca Seifert for lab assistance and Andrew Thompson for help inviiiAcknowledgementsthe field.Ryan Goldade, thank you for your first-hand understanding of the trialsand tribulations of being a grad student and for always providing an honestperspective and opinion.Last, but not least, a huge thanks goes out to my family and friends, foralways cheering me on as I traversed a path that can be traced back to myearly childhood. Mom and Dad, thank you for encouraging me to pursuea career that I love and providing me the means to do so. I?d also like tothank my grandma, Yeat Lau Yau, for insisting that I focus on my studiesand for instilling a strong work ethic for fear of ever letting her down. Jodyand Steven, thank you for supporting my passion and taking more than justa passing interest. Finally, to my Aunt Kitty, who had long ago, envisionedreading an article written by her ?fish-loving niece?.Funding for this project was provided by the Canadian Wildlife Feder-ation Endangered Species Fund, grant # 22R64916, NSERC Discovery Re-search and Equipment Grants, Parks Canada, Alberta Sustainable ResourceDevelopment and the Department of Zoology Teaching Assistantship.ixDedicationThis work is dedicated to my grandfather, Yau Kwan Leung, as well as,Charlie Pacas and the biologists at Alberta Sustainable Resource Develop-ment and Parks Canada who caught the fish and collected tissue samplesused in this thesis.xChapter 1IntroductionThe introduction of exotic species is a critical issue in conservation biology(Rhymer and Simberloff 1996; Allendorf et al. 2001; Levin 2002). Withmarked growth in transoceanic and transcontinental transport and travel,non-native species are brought into new habitats and have the potentialto disrupt natural ecosystems. The outcomes of introductions are hard topredict. Some taxa may never colonize the new environment, while othersthat do, may have positive, negative or zero impact on the system. Thosethat spread and yield negative consequences are termed ?invasive? and cangenerate massive, unforeseen costs (Dawson 2002; Pimentel et al. 2005). Inthe United States, the number of exotic species is estimated to be 50,000and rising (Pimentel et al. 2005). This translates to almost $120 billiondollars (US) in environmental damages annually. In Canada, the same modelestimates a loss of $7.5 billion (CDN) per annum (Dawson 2002).Introductions may be deliberate, usually done in an attempt to enhanceeconomic gain through agriculture or aquaculture, or unintentional. Re-gardless of the motivation behind the transfer, invasive species pose a seri-ous threat to the environment and to biodiversity (Rhymer and Simberloff1996). Native taxa, which are often already threatened by other stressorssuch as pollution and habitat destruction, can be pushed to extinction whenan exotic species enters the system. Invasive taxa can impact an ecosystemthrough predation, habitat alteration, competition, infection and hybridiza-tion (Simberloff 2005). Introduced taxa are responsible for the ?endangered?status of nearly 49% of listed species in the United States and 22% of theendangered species in Canada (Fig. 1.1, Wilcove et al. 1998; Venter et al.2006). Amongst the listed fish species, introduced species threaten 53% offreshwater and marine fishes in the United States and 25% of the endan-gered fishes of Canada (Fig. 1.2, Wilcove et al. 1998; Venter et al. 2006).Unlike other stressors that can be mitigated or restored, rarely is it possibleto completely extirpate an invasive species (Allendorf et al. 2004).Hybridization with introduced species can seriously threaten naturallyevolved, regional taxa (Rhymer and Simberloff 1996; Allendorf et al. 2001;Levin 2002). Introductions and habitat modifications can bring previously1Chapter 1. IntroductionFigure 1.1: The percent of all listed endangered species threatened by habi-tat loss, introduced species, overexploitation and pollution in the UnitedStates (n=1880, Wilcove et al. 1998) and Canada (n=488, Venter et al.2006). These values are not exclusive which is why they add to a valuegreater than 1.2Chapter 1. IntroductionFigure 1.2: The percent of endangered fish species (marine and freshwa-ter) threatened by habitat loss, introduced species, overexploitation andpollution in the United States (n=213, Wilcove et al. 1998) and Canada(n=95,Venter et al. 2006) These values are not exclusive which is why theyadd to a value greater than 1.31.1. Westslope Cutthroat Troutallopatric populations into contact creating an opportunity to hybridize (Al-lendorf and Leary 1988). These groups generally lack evolved characteristicsthat prevent hybridization and the production of viable, hybrid offspring. Ifthe hybrids are fertile, they then act as a vector to introgressive hybridiza-tion by reproducing with each other or an individual from either parentalspecies, moving alleles from one genetically distinct group to another. Un-controlled hybridization can lead to the homogenization of locally adaptedgene complexes into a hybrid swarm and the loss of native genotypes (Al-lendorf et al. 2001).In freshwater habitats, fishes may be brought in as an additional foodsource, but more often, non-native fishes are introduced in an attempt toenhance recreational fisheries (Allendorf and Leary 1988). Salmonid fishes(salmon, trout, char, grayling and whitefishes) are amongst the most widelystocked worldwide. They are generally excellent sport fishes, a naturalsource of protein, and can be reared and transported with relative ease.Fishes within this group that do not naturally co-occur, generally lacksignificant divergence in behaviour and spawning habitats (Allendorf andLeary 1988; Behnke 1992). This, combined with external fertilization, makessalmonids especially prone to hybridization (reviewed by Taylor 2004).Over 65% of the invasive species affecting native fishes in Canada arethe result of intentional introductions (Dextrase and Mandrak 2006). Theearliest record of a species introduced outside their natural range in NorthAmerica dates back to the late 1600s (DeKay 1842). In much of westernNorth America, unexpected hybridization has become commonplace and atleast two subspecies of inland cutthroat trout (Salmonidae: Oncorhynchusclarkii) are now genomically extinct (Behnke 2002). The common cause inboth cases is hybridization with introduced rainbow trout (Oncorhynchusmykiss (Walbaum, 1792); RT) that are a staple in many hatcheries andare farmed worldwide. Rainbow trout are generally limited to freshwatersystems, although anadromous populations are also represented in O. mykiss(steelhead trout). This species is broadly stocked and can be found on allcontinents with the exception of Antarctica (Welcomme 1992).1.1 Westslope Cutthroat TroutThe westslope cutthroat trout (Oncorhynchus clarkii lewisi (Pratt and Gra-ham, 1884); WSCT) is one of at least ten and perhaps as many as fourteensubspecies of cutthroat trout (Behnke 1992). It is native to the interiordrainages of southeastern British Columbia, southwestern Alberta, Canada,41.2. Hybridization with Rainbow Troutand adjacent watersheds in Idaho and Montana scattered areas in westernWashington and Oregon in the United States (Fig. 1.3). The formation ofwaterfalls approximately 70,000 years ago in many of the large tributaries ofthe upper Columbia River appear to have shaped the natural distribution ofWSCT (Figure 1.3, Behnke 1992). They were likely able to colonize watersabove barrier falls due to high water levels produced by glacial melt, priorto isostatic rebound that follows glacial retreat (Behnke 1992). Rainbowtrout found in this region are absent above the barrier falls and are believedto have been restricted to the lower Columbia River during the last glacialperiod, allowing WSCT to become established in large in-land regions ofNorth America, geographically isolated from RT (Figure 1.4, Behnke 1992).Westslope cutthroat trout evolved with few other fish species and theirbiology appears to be driven more by abiotic environmental factors ratherthan interspecific interactions (Griffith 1988). Fluvial and resident formsare common, while adfluvial populations are less so (Cleator et al. 2009).They are considered an indicator species for pristine environments, havinga relatively narrow range of suitable living and spawning conditions. Adultsmake seasonal movements into spawning habitats that are characterized bysilt-free, well-oxygenated water and clean gravel. Once developed, juvenilefish seek shallow pools with low water velocity. Adults require cool, clearwater with in-stream structural complexity and riparian cover. In the sum-mer, they are typically found in water temperatures between 9 and 12 ?C.The adults are generalist predators, foraging on terrestrial and aquatic in-vertebrates available in cold and sometimes turbid water. Deep pools orareas with ground water discharge are needed to overwinter. Westslope cut-throat trout are typically not piscivorous even when forage fish are available(Cleator et al. 2009).Westslope cutthroat trout typically spawn in late spring between Mayand June when water temperatures reach 10 ?C (Nelson and Paetz 1992;Cleator et al. 2009). Females reach sexual maturity after 3-5 years and malesbetween 2-4 years. Fry emerge from the gravel between early July and lateAugust and can remain in their rearing habitats for 1-4 years depending onthe productivity of the stream.1.2 Hybridization with Rainbow TroutCutthroat trout and rainbow trout shared a common ancestor between 3.5-6 million years ago (McKay et al. 1996; Smith et al. 2002; Loxterman andKeeley 2012). Salmonids evolved approximately 100 million years ago, after51.2. Hybridization with Rainbow TroutFigure 1.3: Native range of westslope cutthroat trout (Oncorhynchus clarkiilewisi). Modified from Behnke (2002).61.2. Hybridization with Rainbow TroutFigure 1.4: Native range of rainbow trout (O. mykiss) in North America.Modified from Behnke (2002).71.2. Hybridization with Rainbow Troutchromosome duplication (tetraploidy) separated them from most other fishspecies (Behnke 2002). Delineating the evolutionary history of WSCT andRT has proven difficult, due to one or more hybridization events that haveoccurred since their initial split (Allendorf and Waples 1996; Behnke 2002).Divergence in morphology, ecology and life history has also hampered effortsto define taxonomic groups. Both rainbow trout and cutthroat have signif-icant intraspecific variation and local adaptation, represented by numerousproposed subspecies within each group (Behnke 2002).Introduced RT are the greatest peril for WSCT (Allendorf and Leary1988; COSEWIC 2006). They have restricted WSCT to the upper extremesof their range and will hybridize, and produce viable offspring (Fig. 1.5).Despite having different chromosome numbers (58-60 in RT, 66 in WSCT),WSCT and RT have the same number of chromosome arms (104) (Behnke2002). This allows chromosome pairing to carry on as normal, resulting inhybrid offspring without major developmental deficiencies. Hybrids tend tobe less fit (Allendorf and Leary 1988). Lab-spawned hybrids exhibit slowergrowth and post hatching survival. Recent work has also found that a hy-brid individual will experience a 50% reduction in fitness with as little astwenty percent RT admixture (Muhlfeld et al. 2009a). First generation hy-brids that survive to maturity, however, are highly fertile . They experiencegreater fertilization and hatching success than pure WSCT (Muhlfeld et al.2009a). Under these circumstances, hybrid swarms will form as any progenyproduced by a hybrid will be a hybrid and all subsequent offspring will alsobe hybrids, carrying some degree of RT admixture.Introgression of RT genes into the WSCT genome can break up unique,locally adapted, gene complexes. Westslope cutthroat trout are charac-terized by their adaptation to local conditions (Behnke 1992). Vast streamnetworks subdivide populations and the genetic composition of a single pop-ulation can be significantly different from adjacent populations (e.g., Tayloret al. 2003). Levels of genetic differentiation amongst WSCT tend to behigher within a watershed than between. Thus, the loss of a single popu-lation could eliminate a significant portion of the genetic diversity withina system. In addition, the presence of RT restricts pure populations ofWSCT to high elevation, cold-water habitats, while RT and hybrids domi-nate downstream (Hitt et al. 2003; Weigel et al. 2003; Muhlfeld et al. 2009c;Rasmussen et al. 2010). This removes habitat connectivity and isolates pureWSCT populations from one another. These isolated populations tend tobe small and vulnerable to inbreeding and stochastic events such as drift(Mayhood and Taylor 2011). If a population is lost, the lack of connectivityprevents WSCT from recolonizing the habitat.81.2. Hybridization with Rainbow TroutFigure 1.5: a) juvenile westslope cutthroat trout (Oncorhynchus clarkiilewisi ; note the orange slash on the underside of the mandible, body spotsconcentrated in the posterior region of the body), b) juvenile rainbow trout(O. mykiss; absence of orange slash, spots are uniform across the lengthof the body), c) westslope cutthroat trout x rainbow trout hybrid (orangeslash on mandible present, spots have uniform distribution across the lengthof the body; the hybrid characters are clearly visible in this individual, butcan be much harder to distinguish in other specimens. Photo credit: a) USNational Park Service, b) FISHBIO, c) Austin McPherson.91.3. Conservation Status of Westslope Cutthroat Trout1.3 Conservation Status of Westslope CutthroatTroutWidespread hybridization with RT is pervasive across the entire range ofWSCT. In the United States, WSCT in Montana, Idaho, Washington andOregon are recognized as a single population for conservation purposes andcarry a status of ?special concern?, occupying 59% of their historical distribu-tion (Shepard et al. 2005). Between 65-85% of the remaining area no longercontain ?genetically-pure? WSCT populations. They are the state fish ofMontana, but occupy only 72% of their historical range and 69% show somelevel of hybridization with RT. Westslope cutthroat trout fishing is regulatedby catch limits and in some places is restricted to only catch-and-release an-gling. Westslope cutthroat trout also enjoy a level of protection as many?strong-holds? tend to be found in national parks. In addition, some of theirdistribution coincides with habitats of endangered species such as the bulltrout (Salvelinus confluentus) and steelhead trout (O. mykiss), which areprotected under the Endangered Species Act (US Fish and Wildlife Service.1999). As a result, WSCT were denied listing under the ESA.In Canada, the Committee on the Status of Endangered Wildilfe inCanada (COSEWIC) recognizes two designatable units (DUs) of WSCT,one in British Columbia and another in Alberta (COSEWIC 2006). The twoDUs occupy separate ecozones and lack dispersal between them due to sep-aration by the Rocky Mountains (COSEWIC 2006; The Alberta WestslopeCutthroat Trout Recovery Team 2013). There are also marked differencesin the conservation status of each DU. In BC, WSCT maintain most of theirhistorical distribution, but there is evidence of extensive hybridization withRT as illustrated in studies by Rubidge et al. (2001) and Bennett et al.(2010). In the upper Kootenay River drainage, however, WSCT were foundin only 22% of their historical range (Rubidge et al. 2001). In 2010, theBC population of WSCT was federally listed as ?special concern? under theCanadian Species at Risk Act (SARA).In southwestern Alberta, WSCT historically occurred, in abundance,in streams and rivers from the Bow River to the Alberta-Montana border(Fig. 1.3). Numbers began to decline following the construction of theCanadian Pacific Railway in 1883, which opened up access to the region.Overexploitation by early European settlers in the late 1880s and 1900sremoved substantial numbers of fish from the system and likely caused thelocal extinction of many WSCT populations (Mayhood and Taylor 2011).The introduction of brook trout (Salvelinus fontinalis), rainbow trout and101.4. Hybridization Studiesbrown trout (Salmo trutta) in large numbers further displaced native WSCT.These species are now common across the historical distribution of WSCT,threatening native populations with competition, hybridization, predationand disease (Mayhood and Taylor 2011).Westslope cutthroat trout are the only native subspecies of cutthroatin Alberta (Behnke 1992). This DU, however, appears to be at an ele-vated level of risk and was designated as ?threatened? by COSEWIC in 2006(COSEWIC 2006). Current estimates suggest WSCT occupy 5% of theirhistorical distribution in the province and only 50 of approximately twohundred and seventy-four populations are considered ?genetically pure? (un-hybridized) populations (Mayhood and Taylor 2011). These populations aresaid to have less than 5000 mature adults and 16% have a low chance of re-covery. As of 2009, Alberta?s populations are protected provincially underthe Wildlife Act and in 2013, listed as ?threatened? under SARA. In Marchof the same year, the Alberta Species at Risk Program released a recoveryplan for Alberta WSCT (The Alberta Westslope Cutthroat Trout RecoveryTeam 2013). The recovery plan highlights the importance of identifying ge-netically pure populations of WSCT with continual monitoring as well as?an evaluation of environmental and biological factors that promote and/orlimit hybridization between westslope cutthroat trout and rainbow trout?as part of their recovery strategy (The Alberta Westslope Cutthroat TroutRecovery Team 2013).1.4 Hybridization StudiesStudies to identify variables that influence levels of hybridization in salmonidshave arisen in response to conservation concerns. Combinations of stockinghistory, habitat disturbance, water temperature, habitat connectivity (pres-ence of barriers) and elevation appear to impact levels of hybridization;however, the relative influences of each vary across studies and geographicregion. Rubidge and Taylor (2005) found that water temperature appearedto influence levels of admixture and that the Koocanusa Reservoir acted asa source of RT and, that hybridization decreased with increasing distancefrom the reservoir in the upper Kootenay River, British Columbia. Hittet al. (2003) concluded that hybridization was spreading in an upstreammanner in the Flathead River system in Montana, and agreed with Rubidgeand Taylor (2005) that the presence of physical barriers was likely the onlyobstacle constraining the spread of RT and hybridization in the system. Incontrast, hybridization was absent near areas of known stocking in the Clear-111.4. Hybridization Studieswater River Basin, Idaho, and locales in the Upper Oldman River, Alberta(Weigel et al. 2003; Rasmussen et al. 2010). In these systems, there were nophysical barriers to movement and these authors hypothesized that the envi-ronment was exerting extrinsic control, primarily through changes in watertemperature. Subsequently, Rasmussen et al. (2010) found evidence thatparental habitat choice, likely driven by water temperature, and differentiallife history strategies contributed to hybrid zone structure. While in BritishColumbia and Montana, Muhlfeld et al. (2009c) determined that a mixtureof stocking history, connectivity, anthropogenic disturbance, and water tem-perature all played major roles in predicting levels of hybridization in theupper Flathead River. These studies describe hybridization between thesame species, but reveal high variability in individual conclusions. Conse-quently, the studies to date highlight the necessity of assessing local habitatconditions for effective conservation.Despite geographic variation, one pattern appears to persist across moststudies. A genotypic gradient is commonly reported with changes in ele-vation; genetically pure populations of WSCT are consistently reported athigh elevations with increasing levels of admixture with RT downstream(Hitt et al. 2003; Weigel et al. 2003; Rubidge and Taylor 2005; Rasmussenet al. 2010). This pattern persists through time even when exotic fish areintroduced to high elevation habitats (Weigel et al. 2003). Cold headwa-ters appear to be the only habitat in which genetically pure, indigenoussalmonids experience apparent immunity from competition and hybridiza-tion with introduced taxa (Paul and Post 2001). This observation has leadto the ?elevation refuge hypothesis?, where temperature-mediated compe-tition dictates species dominance along an elevational gradient (Paul andPost 2001). Under this hypothesis, the invasive taxon appears to be a bet-ter competitor in warmer waters and the reverse is true at high elevations,where the native species is competitively superior.Laboratory studies on a number of trout species support a shift in com-petitive ability (growth, survival and behaviour) as temperatures are altered(Taniguchi et al. 1998; Selong et al. 2001; Bear et al. 2007). Westslope cut-throat trout and rainbow trout, however, exhibit virtually identical optimalgrowth temperatures (WSCT: 13.6 ?C; RT: 13.1 ?C, Bear et al. 2007). Re-cent research by McHugh and Budy (2005), suggest clinal zonation may bedriven not only by interspecific competition, but also physiological limita-tions. Specifically, at cold temperatures, typical of high elevation headwa-ters, where a reversal in competitive success favouring the native species isnot observed. Work by Rasmussen et al. (2012) revealed metabolic differ-ences between WSCT, RT and their hybrids. The authors suggest that cold,121.5. Research Objectivesheadwater habitats are likely unable to support the energetic demands ofpure RT and admixed offspring. These data suggest that temperature maybe limiting the spread of hybridization in a way that is not solely based oncompetitive interaction, as suggested by the elevation refuge hypothesis, butalso by way of physiological performance.1.5 Research ObjectivesThere is an urgent need to mitigate the impacts of hybridization and protectspecies that are becoming threatened by it. A proper analysis to determinethe levels of hybridization as well as the factors influencing hybridization isintegral for an effective conservation programme. The aim of this researchwas to document the extent of hybridization between WSCT and introducedRT at sampling locations along the British Columbia-Alberta border, as wellas to identify and further understand factors that may be driving or limitinghybridization between these species.To do this, I conducted the following studies:1. I evaluated the distribution of interspecific hybridization amongst pop-ulations of WSCT and introduced RT through DNA analysis of tis-sues collected across southwestern Alberta and southeastern BritishColumbia.2. I used landscape data obtained from Alberta Sustainable ResourceDevelopment and other resources visualized in ArcGIS, to identifyenvironmental factors that may be influencing rates of hybridizationacross the region.3. I tested the elevation refuge hypothesis by conducting a laboratorystudy evaluating the cold tolerance of WSCT and RT. Here, I ex-plored the possibility that cold-water temperatures may be limitingthe spread of RT upstream to high elevation headwaters, where ge-netically pure populations of WSCT are characteristically found inwatersheds with admixed populations further downstream.13Chapter 2Distribution of hybridizationand environmental factors2.1 IntroductionHybridization between native and introduced species is an on-going con-servation issue (Rhymer and Simberloff 1996; Allendorf et al. 2001; Levin2002). Uncontrolled gene flow between previously isolated groups disruptslocal genotypes and can result in the genomic extinction of indigenous taxa(Epifanio and Philipp 2000; Allendorf et al. 2004; Muhlfeld et al. 2009a). Infreshwater fishes, this problem has arisen in large part due to human effortsto enhance recreational fisheries (Larson and Moore 1985; Allendorf andLeary 1988) or by habitat alteration (e.g., Hubbs 1955; Vonlanthen et al.2012). Many situations include fishes that have evolved in allopatry suchthat reproductive isolation is often incomplete (Allendorf and Leary 1988)and following introduction of non-indigenous species, widespread hybridiza-tion results (Rhymer and Simberloff 1996).Salmonid fishes (trout, salmon and their relatives) are particularly sus-ceptible to hybridization (Taylor 2004). Like most fishes, salmonids fer-tilize their eggs externally, however, they often lack significant divergencein spawning habitats and behaviour that results in incomplete reproductiveisolation. Introgressive hybridization is so pervasive in freshwater salmonids,it has been called the most important factor responsible for the loss of nativetrout species (Allendorf and Leary 1988). Hybridization with stocked rain-bow trout (Oncorhynchus mykiss; RT) has already claimed two subspeciesof inland cutthroat trout (O. clarkii) by genomic extinction (Miller et al.1989) and threatens many other species, including westslope cutthroat trout(O. clarkii lewisi ; WSCT).The abundance of WSCT has severely declined across its entire range.Westslope cutthroat trout are native to southeastern BC and southwesternAlberta, Canada and parts of the United States (Schmetterling 2001; Weigelet al. 2003; Rubidge and Taylor 2005; Mayhood and Taylor 2011). Westslope142.1. Introductioncutthroat trout are the only trout native to Alberta and currently inhabitless than 20% of their historical range owing to overexploitation, habitat lossand habitat degradation (COSEWIC 2006). They are thought to numberfewer than 5,000 adults (Mayhood and Taylor 2011). Populations of WSCTpersist primarily in the headwater tributaries of the Oldman and Bow Rivers?drainage systems, which form part of the western headwaters of the SouthSaskatchewan River system (Mayhood and Taylor 2011). Early records ofRT stocking date back to the mid-1920?s, and some sixty million fish havebeen stocked since that time (Alberta Sustainable Resource Development,Airdree, Alberta, unpublished data). Hybridization with RT is considered tobe widespread across extant populations (DFO 2009; Mayhood and Taylor2011).Although hybridization in freshwater fishes has a long history of study(e.g., Hubbs 1955; Taylor 2004; Hansen and Mensberg 2009), the biological,environmental, and anthropogenic factors that influence the extent and spa-tial distribution of hybridization are not well known. Heath et al. (2010) pre-sented evidence that a combination of anthropogenic habitat alteration (e.g.,logging and urbanization) and stocking intensity were important drivers ofspatial variation in hybridization between rainbow trout and coastal cut-throat trout (O. c. clarkii) in southwestern BC. At the intraspecific level,Marie et al. (2012) demonstrated an effect of stocking intensity, habitatsize, dissolved oxygen levels and pH among wild lake-dwelling populationsof brook char (Salvelinus fontinalis) in Que?bec, Canada. Previous workon RT and WSCT hybridization has indicated that some environmentalfeatures, especially water temperature and elevation, seem to co-vary withadmixture levels (e.g., Rubidge and Taylor 2005; Muhlfeld et al. 2009c; Ras-mussen et al. 2010), while in other cases spatial arrangement of populationsalso seems to be important (e.g., Hitt et al. 2003; Boyer et al. 2008).In this chapter, I report the results of an extensive survey of admixturelevels across more than 150 localities with WSCT in southwestern, Alberta.I then use these data to investigate the key habitat and stocking variablesthat may influence spatial variation in observed levels of hybridization. Byconducting this analysis, I hope to independently test the idea that certainenvironmental and anthropogenic factors (e.g., water temperature, humanhabitat disturbance) are important in influencing the degree of hybridizationbetween RT and WSCT as has been suggested in other areas and species.The identification of natural variables influencing hybridization may also berelevant to understanding factors important in the evolution of reproductiveisolation between these species (e.g., Culumber et al. 2012), while the identi-fication both of natural and anthropogenic-related variables could be useful152.2. Materials and Methodswhen designing programs to limit the spread of hybridization or mitigate itseffects.2.2 Materials and Methods2.2.1 SamplingTissue samples consisted of either fins clips stored in 95% ethanol or driedand stored in paper envelopes. All samples were obtained from populationsof WSCT across localities in southwestern Alberta and a small adjacent re-gion of British Columbia (BC) largely between 2006-2009, although a smallnumber of samples dated to 1999 (Fig. 2.1, Appendix A Table. A.1). Mostof my sampling localities spanned two major watersheds: the Bow and Old-man rivers and included samples from four national parks (Banff, Jasper,Kootenay and Yoho National Parks). The water that forms the Bow andOldman rivers originate from glaciers in the Rocky Mountains near the BCborder and the confluence of these rivers form the South Saskatchewan River(Hudson Bay drainage) representing the northern limit of the natural rangeof WSCT east of the Continental Divide. I also examined samples fromseveral tributaries of the upper Kootenay River (Columbia River drainage)in BC. Altogether, I analyzed 159 sampling locations with a minimum of 15fish sampled per site.2.2.2 Genetic analysesGenomic DNA was extracted from the fin samples using standard phenol-chloroform methods. Individuals were characterized for allelic variation us-ing 10 microsatellite based markers: Ssa85 and Ssa197 (O?Reilly et al. 1996)Ssa456 (Slettan et al. 1995), Ots3, Ots4 (Olsen et al. 1996), Ots104, andOts107 (Nelson and Beacham 1999), Oki3a (P. Bentzen, Dept.of Biology,Dalhousie University, Halifax, NS, unpublished data), Omy77 (Morris et al.1996), and Occ16 (Ostberg and Rodriguez 2002). Only Occ16 appears tobe strictly diagnostic between the two species (Ostberg and Rodriguez 2002and Taylor unpubl. data), but all other loci (except Ssa197) showed ma-jor differences in allele size ranges and frequencies. For instance, in myfinal learning dataset (see below under Admixture analyses) FST (?, Weirand Cockerham 1984) between the species averaged 0.50 (SD = 0.08) andranged from 0.17 (Ssa197) to 1.0 (Occ16). These loci were also scored in150 ?learning samples? of allopatric RT and 75 WSCT that previous analy-ses indicated had no detectable admixture and that represented a range of162.2. Materials and MethodsFigure 2.1: Map of 159 localities in southwestern Alberta and southeasternBritish Columbia, Canada, which were sampled for westslope cutthroat trout(Oncorhynchus clarkii lewisi) and subject to microsatellite DNA analyses.172.2. Materials and Methodspopulations from BC and Alberta (Taylor et al. 2003, 2007; Tamkee et al.2010).Polymerase chain reactions were performed in 20 ?l total volumes usingthe Qiagen Multiplex PCR Kit following the manufacturer?s instruction.An average of 30 individuals were assayed per sampling location and PCRproducts were evaluated using fluorescently labeled primers and assayed ona Beckman-Coulter CEQ 8000 automated genotyper.2.2.3 Population genetic analysesI used MICRO-CHECKER (van Oosterhout et al. 2004) to check for thepresence of null alleles or PCR artifacts that could compromise subsequentanalyses. Thereafter, basic descriptive statistics of sample size (N), num-ber of alleles (NA), observed (HO) and expected (HE) heterozygosity werecompiled using FSTAT ver 2.9 (Goudet 1995). Tests for deviations fromHardy-Weinberg equilibrium were performed for each locus-population com-bination using an exact test in which probability values were estimated usinga Markov chain method using GENEPOP ver. 3.3 (Rousset 2008). Tests forgenotypic linkage disequilibrium for all combinations of locus pairs within apopulation were also made using a Markov chain method with GENEPOPdefault values.2.2.4 Admixture analysesIndividual admixture values (q-values) and posterior probability intervalswere estimated for each locality using STRUCTURE (Pritchard et al. 2000).I set K (number of genetic populations) to K = 2 to represent the twospecies which were clearly distinct from one another in preliminary analyses(e.g., ordination analyses, assignment tests). Models were run under theadmixture model and assuming correlated allele frequencies with a burn-inof 100,000 steps and subsequent runs of 450,000 steps. I also calculatedadmixture values for a sample of 100 simulated hybrids between the twospecies. The simulated hybrid genotypes were generated by using the HY-BRIDLAB (Nielsen et al. 2006) program by randomly selecting alleles foreach locus from the allopatric, non-admixed populations of RT and WSCT.I performed admixture analyses in two steps. First, I ran simulations usingthe allopatric, non-admixed populations (?learning samples?) and all otherpopulation samples using five replicated analyses. Here, the non-admixedsamples were used as priors in the model, i.e., the program was forced toconsider these individuals as known RT and WSCT by invoking the USE-182.2. Materials and MethodsPOPINFO model in STRUCTURE (Pritchard et al. 2000). The admixturevalue, q, was expressed as the proportion of the genome estimated to stemfrom WSCT, qwsct (0 =pure rainbow trout, 1.0 =pure westslope cutthroattrout). The mean qwsct value across the five replicate analyses was thencalculated for all of the non-learning samples. From this analysis, threepopulations of WSCT for which all individuals had qwsct values of ? 0.99were added to the learning sample group and the admixture analyses wasrerun across five further replicates. I conducted the admixture analysesin two steps because: (i) the initial learning sample of WSCT was muchsmaller (75) than that of the RT, and (ii) the initial learning sample con-tained no non-admixed populations from Alberta and I wanted to accountfor possible genetic differentiation between BC and Alberta populations ofWSCT which are separated by the continental divide. The final learningsample set consisted of 150 RT and 165 WSCT. The final values of qwsct forAlberta populations of WSCT represent the averages calculated across thefive replicate analyses.2.2.5 Stream characteristics and anthropogenic variablesVariables were organized into three broad categories: stream geomorphol-ogy/environments, stocking history, and variables representing aspects ofanthropogenic-based habitat disturbance (Table 2.1). Environmental vari-ables were chosen based on current indicators considered to influence hy-bridization in western trout (Hitt et al. 2003; Weigel et al. 2003; Rubidgeand Taylor 2005; Muhlfeld et al. 2009c; Heath et al. 2010). Also, the numberof stocking events and the total number of RT introduced both representaspects of stocking intensity that may influence the potential for admixture(e.g., Ruzzante et al. 2004; Hansen and Mensberg 2009; Marie et al. 2010).Finally, habitat disturbance from human development has long been consid-ered an important variable influencing interspecific hybridization in fishes(Hubbs 1955). More recently, Heath et al. (2010) found evidence of a positiveassociation between logging activity and urban development and hybridiza-tion between naturally sympatric rainbow trout and coastal cutthroat trout(O. c. clarkii) in southwestern BC.Locality and stocking data were available for a total of 58 streams (Al-berta Sustainable Resource Development, Airdrie, AB; Appendix A Ta-ble. A.2). These data were mapped using ArcGIS 10.0 (ESRI, Redlands,CA, USA). Elevation data were extracted using a 25-metre Digital Ele-vation Model and stream order, a measure of stream branching, was as-sessed in ArcGIS using the Strahler method (Horton 1945; Strahler 1952).192.2. Materials and MethodsSpot water temperature was recorded with hand-held thermometers at thetime of fish sampling. To assess how well such instantaneous measuresmight reflect longer term, relative differences between localities, I obtainedlonger-term measures of air temperature near each locality through World-Clim (www.worldclim.org). These data layers are generated by interpolat-ing monthly climate data from the Global Historical Climatology Network(GHCN), the Food and Agriculture Organization of the United Nations, andthe World Meteorological Organization, and several other sources of climatedata for the period 1950-2000 (see Hijmans et al. 2005). The locality watertemperatures obtained during daytime fish sampling were positively corre-lated with mean annual air temperature (r = 0.31, df = 57, P = 0.002),maximum air temperature of the warmest month (r = 0.26, P = 0.01) aswell as average air temperature over the warmest three month period (r =0.40, P < 0.0001). Further, elevation and water temperature are expectedto be negatively correlated with one another in natural systems (e.g., Pauland Post 2001; Rasmussen et al. 2010) as was observed in my data (r =-0.51, P < 0.001). These associations with elevation and longer term tem-perature trends suggest that locality water temperatures recorded at time offish sampling represent a reasonable proxy for relative differences betweenlocalities. Human impact variables (e.g., distance of sample site to nearestroad) were measured as Euclidean distances and distance from the samplesite to the stocking location (stockD) was calculated as the fluvial distancein ArcGIS (Table 2.1). In some streams, stocking of RT took place in trib-utaries other than that in which genetic samples were obtained. In others,apparent upstream migration barriers (mapped in ArcGIS) separated stock-ing and genetic sample localities within tributaries. Preliminary analyses,however, indicated no difference in average qwsct values between localitieswhere stocking took place in the same or different tributaries or separatedby potential upstream migration barriers (Yau and Taylor, unpubl. data)so I did not include these potential effects in my analysis.Stocking intensity was represented by the number of stocking events atthe stocking location nearest to the genetic sample locality. Other variables,included the total number of RT stocked at each locality, and the year oflast stocking, which was assessed to investigate how time scale may influenceadmixture levels.2.2.6 Statistical analysisDifferences in the mean values (calculated across samples with a sample sizeof ? 10) of qwsct were tested for significance using using variants of one-way202.2. Materials and MethodsTable 2.1: Definition of variables potentially explaining variation in admix-ture levels between native westslope cutthroat trout and introduced rainbowtrout.Variable DefinitionStream VariablesDepth1 Maximum water depth at sample siteDepth2 Mean water depth at sample siteWater temperature Water temperature at time of samplingOrder Strahler stream order of sample site (1-5)Elevation Elevation in meters at sample siteAnthropogenic disturbance variablesRoad (RoadD) Euclidean distance from sample site to nearestroadPipeline (PipeD) Euclidean distance from sample site to nearestgas/oil pipelineRailway (RailD) Euclidean distance from sample site to nearestrail linePower (PowerD) Euclidean distance from sample site to nearestpower lineStocking variablesDistance (StockD) Fluvial distance to nearest stocking siteStocking intensity(StockI)Total number of stocking events at the stock-ing siteNumber of fish stocked(StockN)Total number of rainbow trout stocked at thestocking siteYear of last stocking(StockY)Total years between 2010 and year of laststocking at stocking site212.2. Materials and MethodsANOVA and subsequent post-hoc tests suitable for samples with unequalvariances using PAST (version 2.12), a general spreadsheet-based statisticalpackage (Hammer et al. 2001).To identify habitat and human impact variables that might influenceadmixture levels, I used a nonlinear, regression tree approach implementedin the Random Forest algorithm described by Breiman (2001). Regressiontrees use decision tree analysis to help to resolve relationships between aresponse variable (in my case mean admixture level) and several potentialpredictor variables that may behave in a non-linear fashion and have multi-ple interactions amongst them. Regression trees and their use in ecologicaldata analyses have been described by De?ath and Fabricius (2000) and Moi-sen (2008) have seen recent application in landscape genetics (e.g., Murphyet al. 2010; Hether and Hoffman 2012). Random Forest analysis uses boot-strapped, learning datasets to build an assemblage of regression trees eachof which generates predictions of the dependent variable (qwsct) based onthe independent variables. These predictions are then averaged across allbootstrapped iterations to yield a final prediction. Examining many boot-strapped samples coupled with random sampling of subsets of predictor vari-ables during tree construction helps to reduce the variance amongst differentregression tree results. Random forests also have the desirable qualities ofbeing insensitive to autocorrelation, are distribution free, do not requiretransformation of original variables, and they can assess complex interac-tions amongst many variables (De?ath and Fabricius 2000; Moisen 2008). Icombined the approaches of Murphy et al. (2010) and Hether and Hoffman(2012) by first analyzing a full model incorporating all nine potential pre-dictor variables, evaluating sub-models with subsets of predictor variables,running a final ?best? sub-model and assessing the relative importance ofeach predictor variable and overall model significance using the RANDOM-FOREST package v 4.5-28 with R (Liaw and Wiener 2002; R DevelopmentTeam 2012) under the regression mode with 10,000 trees.I converted the measures of variable importance (Ip) that are provided inRANDOMFOREST to model improvement ratios (MIRs) by dividing eachIp by the maximum Ip observed (Hether and Hoffman 2012). I then usedthe MIR to iteratively select the best sub-models from the full model. First,I ran a sub-model that incorporated only those variables that had MIR ofat least 0.3 (i.e., they improved the model to a degree of at least 30% ofthe best variable). Next, I ran 20 independent RANDOMFOREST analysesfor each of a series of sub-models that had one of the predictor variables re-moved, starting with the variable with the lowest MIR. I calculated the meanand 95% confidence intervals for the resulting pseudo-R2 (Liaw and Wiener222.3. Results2002) values for each such sub-model. I then chose a final sub-model thathad the fewest predictor variables, but whose mean pseudo-R2 had a 95%confidence interval overlapping that of the best sub-model (Hether and Hoff-man 2012). I used partial dependence plots to assess the marginal effects ofeach retained predictor variable to admixture (i.e., the effect not explainedby other predictor variables, Liaw and Wiener 2002; Cutler et al. 2007).Finally, I tested the significance of the derived sub-model following the ran-domization (of admixture values, N = 1,000) procedure and examining therandomized distribution of the simulated pseudo-R2 values relative to theobserved value as detailed in Hether and Hoffman (2012). I also exploredalternative methods of analysis of the data (e.g., standard multiple regres-sion on the original data or on scores from a principal components analysesof correlation matrices), but the RANDOMFOREST procedure consistentlyproduced lower mean square errors, higher R2, simpler models, identified thesame major variables as predictors, and had the added advantage of beingable to analyze the original variables without the need for transformations.This analyses employed the mean q-value for each stream which is unin-formative concerning the variation amongst individuals within streams inadmixture level. I also analysed the data using the median q-value (whichas expected showed a correlation r of > 0.9 with the mean q-value) andvery similar results were obtained (i.e., overall R2 value of 0.32, same fivevariables chosen in same relative ranking of explanatory power see below).Where appropriate, adjustments for multiple simultaneous statisticaltests incorporated the false discovery rate procedure of Narum (2006).2.3 Results2.3.1 Genetic analysesAnalysis of genetic data for all localities with at least 15 fish sampled indi-cated that evidence for null alleles was rare and scattered across individualloci and populations. Of 1,590 analyses (159 samples x 10 loci), five sug-gested that null alleles might be present at Ots107 in some samples. I found,however, no individuals that were null-null homozygotes at Ots107 suggest-ing that null alleles would not be a significant factor in subsequent analyses(cf. Taylor et al. 2003; Tamkee et al. 2010).Across localities and loci, observed and expected heterozygosities, andthe number of alleles averaged (SD) 0.32 (0.15), 0.34 (0.16), and 3.2 (1.4),respectively, and FIS averaged 0.072 (0.14) and 29 of 159 permutation testsindicated FIS values significantly greater than 0 (Appendix A Table. A.1).232.3. ResultsOf the 159 samples examined, there was a broad range of estimatedadmixture levels ranging from qwsct = 0.010 to 0.994. There was a strongright skew to the data with most localities having qwsct values > 0.90. PureRT (i.e., qwsct < 0.05), however, were found at 12 localities, six localities hadaverage qwsct values of < 0.10 and 34 of the localities exhibited extensivehybridization with q-values below 0.90 (Appendix A Table. A.1). Threelocalities were identified from admixture analysis and stocking records asYellowstone cutthroat trout (Oncorhynchus clarkii bouvieri) or Yellowstonecutthroat trout x westslope cutthroat hybrids and removed from furtheranalyses.Of those localities with qwsct values between 0.10 and 0.90 (N = 29), 24had FIS values not significantly different from 0, while five had significantlypositive FIS values associated with heterozygote deficiencies (e.g., HaidukLake, Smuts Creek, Fisher Creek, Spotted Wolf Creek, Appendix A Table.A.1). The estimated value of qwsct was negatively correlated with the av-erage (across loci) number of alleles observed; localities with high levels ofadmixture with RT (low qwsct) tended to have a higher average number of al-leles (rs = -0.73, P < 0.0001). The simulated hybrids (N = 100 individuals)had a mean qwsct value of 0.51 (SD = 0.053).2.3.2 Influences on admixture valuesIncorporation of all predictor variables into the RANDOMFOREST regres-sion tree analysis resulted in a pseudo-R2 value of 23.7% and a mean squareerror (MSE) of 0.059. The variable importances (Ip) for admixture pre-diction ranged from a low of 0.037 for stream order to a high of 1.019 forwater temperature. Model improvement ratios relative to the best predictor(1.0 = water temperature) ranged from 0.036 (stream order) to 0.654 (ele-vation). Five variables, water temperature, elevation, stockD, powerD andrailD were selected for sub-model analyses based on MIRs of at least 0.3. Amodel incorporating just these variables produced a pseudo-R2 = 33.2% anda MSE = 0.052 (P < 0.001). Analyses of the various sub-models (Table.2.2) indicated that a four variable sub-model incorporating water temper-ature (MIR = 1.0), elevation (MIR = 0.824), railD (MIR = 0.717) andstockD (MIR = 0.612) (but removing powerD, MIR = 0.632) had a meanpseudo-R2 (34.4%, MSE = 0.051, P < 0.001 from randomization test) thatwas not significantly worse than the full five variable sub-model, but thatremoving any other single variable resulted in significantly lower pseudo-R2values (minimum value = 26.3% after removing temperature). This final,four variable model was used to predict admixture levels based on 10,000242.3. ResultsTable 2.2: Variable importance values (Ip) and model improvement ratios(MIR) for the full and final, four variable (in boldface) models determinedfrom randomForest analysis of the relationship between admixture values(qwsct) between westslope cutthroat trout and rainbow trout and streamvariables for 58 localities, in southwestern Alberta.Variable Ip-full MIR-full Ip-final MIR-finalMean depth 0.1378 0.1351 NA NAMaximum depth 0.1511 0.1482 NA NATemperature 1.0294 1.0000 1.0730 1.000Elevation 0.6672 0.6545 0.8848 0.8247Stream order 0.0373 0.0365 NA NARoadD 0.2567 0.2498 NA NAPipeD 0.2000 0.1964 NA NAPowerD 0.3364 0.3300 NA NARailD 0.3762 0.3691 0.7697 0.7172StockI 0.0833 0.0788 NA NAStockN 0.1751 0.1717 NA NAStockY 0.2130 0.2089 NA NAStockD 0.3297 0.3234 0.6121 0.6122bootstrapped regression trees.The correlation between observed admixture values and those predictedby the randomForest model was 0.61 (df = 58, P < 0.001). The distributionsof observed and predicted admixture levels mirrored each other reasonablyclosely, but the greatest deviations were observed at very low and very highqwsct levels; about 21% of the samples had an observed qwsct of > 0.99, butthe RANDOMFOREST model predicted that the highest qwsct was 0.96(Fig. 2.2). Partial dependence plots indicated the effect of each retainedpredictor variable after averaging out the effects of all other predictor vari-ables. The response of qwsct to changes in water temperature suggested astep-like response as qwsct decreased abruptly between about 11 ?C and 13 ?C(Fig. 2.3). Conversely, qwsct increased steadily as elevations rose above 1300m (Fig. 2.3). The qwsct also tended to increase with increasing stockD, al-though somewhat more irregularly and leveled off after about 4000 m (Fig.2.3). Finally, distance to the nearest railway (RailD) was the most irregularshowing variable responses across different distances (Fig. 2.3).252.3. ResultsFigure 2.2: Observed (upper) and predicted (from RANDOMFOREST de-rived relationships, lower) admixture values (qwsct, 0 = rainbow trout, 1= westslope cutthroat trout) as a function of variation in water temper-ature, elevation, distance from sample site to stocking site and distancefrom sample site to the nearest railway line for westslope cutthroat trout(Oncorhynchus clarkii lewisi) sampled across 58 localities in southwesternAlberta, Canada, and assayed at 10 microsatellite DNA loci.262.3. ResultsFigure 2.3: Partial dependence plots showing the relationship between a)water temperature, b) elevation, c) distance from sample site to stockingsite, and d) distance from sample site to nearest railway line and admixturevalues (qwsct, 0 = rainbow trout, 1 = westslope cutthroat trout) for westslopecutthroat trout (Oncorhynchus clarkii lewisi) sampled from 58 localities insouthwestern Alberta, Canada, and assayed at 10 microsatellite DNA loci.272.4. Discussion2.4 DiscussionHybridization and introgression have become critical concerns in the con-servation of biodiversity as genomic extinction and admixture have led todemographic declines (Rhymer and Simberloff 1996; Huxel 1999; Muhlfeldet al. 2009a). The study of admixture in cutthroat trout and other taxa hasbeen widespread throughout its native range and dates back to the early1980s (e.g., Leary et al. 1984; Allendorf and Leary 1988). All together, ithas been estimated that genetically non-admixed WSCT may exist in lessthan 10% of the total historical native range (Trotter 2008). In western Al-berta, WSCT and RT hybrids had been first recognized, morphologically, asearly as 1947 and were considered widespread by 1950 (Mayhood and Taylor2011). It is now estimated that fewer than 10% of populations of WSCT inwestern Alberta have no detectable introgression with RT and that many ofthese populations are relatively isolated in headwater reaches (Mayhood andTaylor 2011). The perilous state of these populations over the medium tolong term was a major factor leading to their assessment as ?threatened? bythe Committee on the Status of Endangered Wildlife in Canada (COSEWIC2006) and later listing under the federal Species at Risk Act (SARA). Bycontrast, there have been very few studies that have sought to understandthe fate of hybrids or general explanations for spatial variation in admixturebetween native WSCT and introduced RT. Some laboratory-based (Learyet al. 1984; Ferguson et al. 1985) and, more recently, field-based studies(Muhlfeld et al. 2009a; Rasmussen et al. 2012) have shed light on the per-formance and fate of hybrids and in some cases have demonstrated thateven small amounts of admixture from RT can influence physiology or re-duce survival of hybrid offspring. The demonstration of real fitness costs toadmixture stresses the importance of understanding what influences spatialvariation in admixture in nature. Understanding the factors that explainthis variation can aid our understanding of the biology of the hybridizingtaxa, if and how admixture spreads, and what strategies might be best todeal with admixture.2.4.1 Spatial variation in admixtureMy survey indicated that while some localities showed considerable admix-ture with RT (i.e., qwsct < 0.75), most areas showed qwsct levels of between0.85 and 0.99. I am unaware of any study of admixture between RT andWSCT trout that is as geographically extensive as the present study, buteven on smaller geographic scales a similar pattern of spatially variable ad-282.4. Discussionmixture is usually found. For instance, Hitt et al. (2003) documented a fewheavily admixed populations (i.e., > 85% genetic contribution from RT) anda majority of localities with 1-27% admixture from RT across 42 samplesfrom the upper Flathead River drainage of Montana. Such variation mightbe expected given differences in various aspects of introduction intensity andhistory across areas. For instance, both Rubidge et al. (2001) and Hitt et al.(2003) reported that admixture with RT decreased upstream from sourcesof introduction of RT. By contrast, Bettles et al. (2005) and Heath et al.(2010) reported extensive, but less structured spatial variation in the ex-tent of natural admixture between native populations of rainbow trout andcoastal cutthroat trout (O. c. clarkii). The finding that spatial variationin admixture does not appear to be governed simply by idiosyncrasies ofspecific artificial introduction programs suggests that local conditions alsoappear to play an important role (Aldridge and Campbell 2008). Regard-less of the geographic pattern of variation in admixture levels, my resultsdemonstrate that admixture with RT is widespread in southwestern Alberta.There has been no comparably extensive survey in British Columbia (BC),but surveys by Rubidge et al. (2001), Taylor et al. (2003) and Rubidge andTaylor (2005) indicate that many geographic localities in the upper Koote-nay River system in BC are similarly significantly admixed.2.4.2 Environmental correlates of admixtureThe landscape genetic approach to studies of the spatial distribution of ge-netic variability (e.g., Manel et al. 2003; Van Houdt et al. 2005; Holdereggerand Wagner 2006; Tamkee et al. 2010) can also clearly apply to under-standing how admixture between species is distributed and spreads acrossa landscape. My analyses were conducted over a series of watersheds andthus add generality to the observations made in smaller scale studies acrosswestern North America suggesting that a consistent set of physical factorsinfluence the extent of admixture between WSCT and RT (e.g., Weigel et al.2003; Rubidge and Taylor 2005; Muhlfeld et al. 2009b). Specifically, watertemperature and elevation are consistently associated with admixture levels;admixture between the species is consistently lowest in high elevation, loworder and cooler streams. In particular, all three of these variables tend to beintercorrelated, but water temperature was the most important factor asso-ciated with admixture in my regression tree models. The influence of watertemperature observed in my study is consistent with previous studies of hy-bridization in WSCT and other species of salmonids that typically reporta gradient of admixture with genetically pure, native trout at the highest292.4. Discussionelevations, increasing levels of admixture downstream and pure genotypes ofthe introduced species in the lowest reaches (Larson and Moore 1985; Pauland Post 2001; Hitt et al. 2003; Rubidge and Taylor 2005; Muhlfeld et al.2009c; Rasmussen et al. 2010, 2012). Westslope cutthroat trout and rain-bow trout exhibit very similar optimal temperatures (13.6 ?C and 13.1 ?C,respectively), at least as inferred by limited laboratory study, but RT dis-played better growth at higher water temperatures and a broader range ofgrowth at warmer temperatures (Bear et al. 2007). Rainbow trout may alsobe better adapted to warmer temperatures as suggested by having an upperincipient lethal temperature that is 4.7 ?C greater than in WSCT (Bear et al.2007). These trends in the elevational distribution of admixture and obser-vations of different thermal characteristics of hybridizing species have led tothe development of the ?elevation refuge hypothesis? (Fausch 1989; Paul andPost 2001; Rasmussen et al. 2010). Here, habitats at high elevations mayprovide a refuge for cold water-adapted, native trout, at temperatures thatintroduced species are unable to exploit as effectively.A number of laboratory studies support this hypothesis, demonstrating ashift in growth rates and survival as water temperatures change (e.g., Reeseand Harvey 2002; McMahon et al. 2007). Recent work in bull trout (Salveli-nus confluentus) and brook trout (S. fontinalis) showed a competitive ad-vantage of introduced brook trout at warm temperatures reared in sympatrywith bull trout (McMahon et al. 2007), but no significant differences wereobserved when these species were grown in allopatry. Further, Rasmussenet al. (2010, 2012) found significant associations among admixture gradient,elevation, and difference in life history characters and metabolism betweenRT and WSCT and hybrids in the upper Oldman River, Alberta. Westslopecutthroat trout with low levels of admixture with RT tended to predominatein headwater reaches of the upper river and were also older and grew moreslowly than hybrids and pure RT which tended to predominate in lower el-evation and warmer reaches. Hybrids were found in mid-elevation reachesand exhibited intermediate metabolic rates that seem well-suited to suchecotonal habitats (Rasmussen et al. 2012).Weigel et al. (2003) proposed one mechanism by which cold summer tem-peratures may impede the establishment of RT alleles. Cold summer tem-peratures can delay egg production and prolong incubation of early springspawning RT compared to WSCT that spawn later in the season (Hubertet al. 1994; Stonecypher et al. 1994). Such a developmental delay may act tocompound a decrease in growth rate at cooler water temperatures and alsoreduce overwinter survival of RT and hybrid fry. Indeed, Culumber et al.(2012) presented evidence that physiological adaptation to different ther-302.4. Discussionmal niches could explain, in large part, the elevational pattern of genotypicdistribution of two species of swordtails (Xiphophorus) and their hybrids insoutheastern Mexico.My analysis suggested, from the partial dependence plots, that the ef-fect of water temperature was non linear and that a temperature of be-tween 11-13 ?C might represent a threshold at which the likelihood of ad-mixture changes abruptly. This temperature range as a specific thresholdmust clearly be interpreted with caution because my temperature samplingwas superficial and denser sampling, spatially and temporally, must be com-pleted to see if such a threshold has any veracity. Interestingly, however, Ras-mussen et al. (2010) proposed that a temperature threshold of about 7.3 ?Cmarks a point at or below which admixture was low to absent within one trib-utary of the Oldman River system, Alberta. Of course, even with such moreintensive sampling of water temperature, Rasmussen et al. (2010) cautionedthat its extension to other streams was problematic given the many featuresthat can vary between streams and influence admixture (e.g., barriers). Fur-ther, there is evidence that stream bed temperatures can vary significantlyacross a stream channel within a site (Webb et al. 2008) and that smalltemperature changes of 1 ?C can alter salmonid distributions (Fausch et al.1994). If these temperature thresholds to admixture exist, recent advancesin the ability to accurately measure stream and river temperatures (Webbet al. 2008) can perhaps help to predict responses to environmental changes(e.g., long-term climate change, land use changes) that can influence streamwater temperatures in terms of changes in the levels of genetic admixture,its geographic distribution, and the speed at which such changes might oc-cur (cf. Rasmussen et al. 2010; Isaak et al. 2012). There is some evidenceof spatial segregation of RT and WSCT and differences in growth poten-tial as a function of water temperature (e.g., Bozek and Hubert 1992; Sloatet al. 2005; Bear et al. 2007), but McHugh and Budy (2005) reviewed sev-eral studies and suggested that, overall, condition (temperature)-dependentcompetition is an unlikely explanation for all salmonid zonation patterns. Abetter understanding of the physiological performance of these species andtheir hybrids across multiple life stages, particularly at low water tempera-tures is needed and I address this need in the next chapter of my thesis.My analyses also suggest that elevation plays a role in influencing ad-mixture levels (cf. Weigel et al. 2003; Rasmussen et al. 2012). There isa clear negative association between water temperature and elevation (thisstudy, see also Rasmussen et al. 2010; Culumber et al. 2012), but the partialdependence plots suggest an effect of elevation in excess of that which can beaccounted for by its association with water temperature. Rainbow trout tend312.4. Discussionto be more abundant in wider, lower elevation, lower gradient stream reaches(MacCrimmon 1971; Gard and Flittner 1974; Bozek and Hubert 1992; Pauland Post 2001). Henderson et al. (2000) and Muhlfeld et al. (2009b) foundthat upstream-downstream segregation in WSCT and RT, respectively, mayalso include spawning habitats, but that they also overlapped in some tribu-taries. Given that gradient increases and stream width generally declines aselevation increases, the physically smaller and more energetic higher eleva-tion habitats may be less preferred by RT regardless of water temperatureand reduce the incidence of interactions and hybridization with WSCT. Inaddition, given that stocking of RT was characteristically in the lower por-tions of my study streams, the negative association between admixture leveland elevation (and its association with water temperature) may result fromnon-equilibrium conditions in which invasive RT and admixture are still inthe process of extending in upstream directions (Rubidge et al. 2001; Hittet al. 2003; Boyer et al. 2008; Muhlfeld et al. 2009c). My data can serveas a useful baseline to monitor changes in admixture with time and/or inresponse to habitat disturbances such as fires or floods (e.g., Isaak et al.2012).2.4.3 Anthropogenic correlates of admixtureA number of anthropogenic factors related to stocking operations and en-vironmental disturbance were found to be related to admixture with RTamongst my samples. For instance, the level of admixture has been of-ten observed to decline with distance from the point(s) of stocking (Ru-bidge et al. 2001; Hitt et al. 2003; Rubidge and Taylor 2005; Muhlfeld et al.2009c). My results are in agreement with these findings and emerged de-spite the likelihood that all such comparisons suffer to varying degrees fromincomplete stocking histories, undocumented movement of trout by the pub-lic, failure to establish spawning populations, and migration of stocked fishfrom sampling location (Moring 1993; Weigel et al. 2003). I expected thatthe number of fish stocked and stocking intensity would increase propagulepressure of RT and be associated with increasing admixture levels as hasbeen reported in other salmonid systems (Lockwood et al. 2009; Muhlfeldet al. 2009b; Marie et al. 2010, 2012). I also expected an increase in numberof years elapsed since the last stocking and increasing isolation (a combina-tion of stream distance and presence of upstream migration barriers) fromthe stocking site would be associated with low levels of admixture (e.g., Ru-bidge et al. 2001; Ruzzante et al. 2004). Only the latter variable, however,was identified as an important variable in the RANDOMFOREST analysis.322.4. DiscussionTwo streams, Loomis Creek in the Bow River and South Castle Creek inthe Oldman River had unusually high numbers of fish stocked and stock-ing intensities. These streams had between five and 30 times as many fishstocked (up to 3 million in South Castle Creek) and between six and 10times the stocking intensity (up to 64 years in South Castle Creek) as theaverage values for all other localities (Appendix A Table. A.2). Despitesuch high stocking levels, these two streams had levels of admixture withRT that were negligible in Loomis Creek (qwsct = 0.97) to about the aver-age (qwsct = 0.88) in South Castle Creek (qwsct = 0.87). Loomis Creek hadone of the coolest water temperatures (7 ?C which was in the lowest 15thpercentile of all streams) and was almost 5 km upstream of the stocking sitewhich was below a migration barrier and in a different tributary. Further,the water temperature of South Castle Creek (9.2 ?C) was below the averagevalue of 10 ?C for all streams. Thus if the spot water temperatures that Iused do accurately reflect cooler than average conditions in these streams,this factor combined with the location of stocking sites in terms of distanceand in relation to migration barriers, and other aspects of habitat that maybe unsuitable for RT may help to explain the relatively low admixture lev-els despite high levels of stocking of RT in some systems. More generally,these results suggest that the numbers and intensity of stocking interactwith local environmental conditions to influence admixture levels (Weigelet al. 2003; Taylor et al. 2007; Muhlfeld et al. 2009c). Certainly, other casesof intraspecific salmonid supplementation programs have illustrated the id-iosyncratic nature of the outcome of stocking that is dependent on more thanjust the numbers and intensity of non-native fish stocked (e.g., Krueger andMenzel 1979; Largiade`r and Scholl 1995; Taylor et al. 2007; Halbisen andWilson 2009; Marie et al. 2012). My analysis reinforces the importance ofconnectivity between streams (as influenced by fluvial distance and migra-tion barriers) as a critical factor influencing admixture in many situations(Rubidge et al. 2001; Gunnell et al. 2008; Muhlfeld et al. 2009c).The effects of human disturbance factors on the spread of non-nativespecies and hybridization in stream fishes may be direct, from increasing ac-cess to streams through road or railway construction, or indirect, from land-scape developments that influence key features such as water temperaturewhich themselves influence dispersal of non-native species and hybridization(e.g., Dunham et al. 2002; McMahon et al. 2007; Heath et al. 2010). Forinstance, Muhlfeld et al. (2009c) found that the number of upstream roadcrossings was positively correlated with levels of admixture between nativeWSCT and invasive RT in the upper Flathead River in British Columbiaand Montana. In my study systems, only the distance to the nearest rail-332.4. Discussionway appeared to influence admixture levels, but in an erratic manner. Theapparent lack of an influence of roads, pipelines, and powerlines should beinterpreted cautiously as they may not have represented measures of hu-man disturbance per se particularly well. For instance, Muhlfeld et al.(2009c) found that road density did not have a significant effect in theirstudy, but that number of upstream road crossings did. In this case, thelatter variable is perhaps a more direct (and hence more sensitive) mea-sure of potential habitat disturbance for streams. Similarly, my measuresof disturbance were expressed only as distance to the nearest anthropogenicstructure and did not incorporate density nor actual crossings. Interestingly,railways were historically the source of much stocking pressure (Mayhood1999), but my analysis revealed a complex interaction between distance tothe nearest railway (RailD) and admixture. This result may stem from thehighly bimodal distribution of RailD values I obtained; one mode was foundat about 16 km and the other at about 46 km (Appendix A Table. A.2).Still, I could not resolve any consistent directionality to the influence RailDand admixture. Finally, other human disturbance factors such as loggingactivity, recreational land use including angling activity or natural factorssuch as seasonal flood dynamics, and their interactions, may play roles inadmixture and should be investigated (e.g., Fausch et al. 2001; Heath et al.2010).In summary, the analysis of variation in admixture levels between WSCTand RT in western Alberta further emphasize what appears to be a basicspatial pattern; admixture levels tend to be low in cooler, high order, highelevation streams (cf. Paul and Post 2001; Rubidge et al. 2001; Weigelet al. 2003; Rubidge and Taylor 2005; Muhlfeld et al. 2009b). This patternhas led to the temperature/elevation refuge hypothesis that suggests thatnative population of salmonid fishes in mountainous areas may be less sus-ceptible to invasion of non-natives if native fishes have a physiological orbehavioural advantage over non-native and hybrids at cooler water temper-atures (Paul and Post 2001; McMahon et al. 2007; Rasmussen et al. 2010).In addition, the characterization of WSCT and RT as relatively cool-waterand warm-water adapted, respectively, may be considered consistent withtheir evolutionary and biogeographic history given the concentration of theformer in high elevation areas of the Rocky Mountains and adjacent moun-tain ranges. Still, the relative performance of either species at low watertemperatures, a key component of the temperature refuge hypothesis, hasbeen little explored, as has the potential role of thermal adaptation in spe-ciation in fishes (cf. Culumber et al. 2012; Keller and Seehausen 2012). Myresults and those of others (e.g., Rubidge et al. 2001; Fausch 2008; Muhlfeld342.4. Discussionet al. 2009b; Marie et al. 2012) clearly indicate, however, that factors otherthan water temperature (e.g., habitat area, stocking practices, migrationbarriers) can also influence admixture levels and that the influence of hu-man disturbance factors (resource extraction, road and pipeline density androutes) will often be contingent on the local regulatory regime.2.4.4 Implications for recovery planningMy comprehensive survey of Alberta populations provides insights into theinteraction between hybridizing species and their environment that are com-plementary to previous, more localized studies, and can focus future researchto assist with recovery planning for threatened WSCT (COSEWIC 2006).For instance, my results can help to prioritize populations for conservationduring the assessment of recovery potential. Populations with little to nodetectable admixture may be the highest priority for conservation and per-haps as sources of fish for recovery in more affected streams (Muhlfeld et al.2009c; Mayhood and Taylor 2011). Second, the RANDOMFOREST anal-ysis performed reasonably well in terms of predicting admixtures levels inbootstrapped samples and could used as an initial ?triage? procedure tochoose amongst a series of localities with unknown admixture levels as towhich may be more or less likely to exhibit admixture with RT. The use ofthis model would, however, prove conservative because it was unable to accu-rately predict qwsct values of > 0.96. This would be problematic for correctlyidentifying genetically non-admixed populations of WSCT under a proposedthreshold value of qwsct > 0.99 (Allendorf et al. 2004). Finally, more inten-sive sampling of stream water temperatures through time could help betterassess the role of low water temperature on relative performance of geno-types and its effects on reproductive isolation and hybridization betweenspecies. My results also support the idea that higher order, lower elevation,and warmer streams appear to be favourable environments for hybridizationand provide sources for the spread of admixture between WSCT and RT. Inaddition, physical stream characteristics, stocking practices, anthropogenichabitat modifications, and the local regulatory regime (e.g., within or out-side protected areas) all interact to influence admixture levels in salmonidfishes in diverse ways which emphasizes the need for context-specific solu-tions.The growing body of work both on intraspecific and interspecific hy-bridization in salmonid fishes (e.g., Rubidge and Taylor 2005; Halbisen andWilson 2009; Muhlfeld et al. 2009b; Heath et al. 2010; Marie et al. 2012)suggests that admixture levels are somewhat predictable although much of352.4. Discussionthe variation remains unexplained. Even some small understanding of whatenvironmental or human factors influence the probability and extent of ad-mixture should improve conservation efforts for native fishes. For instance,managers could be alerted to situations where stocking should not occur,e.g., in streams with conditions that might favour non-natives and/or hy-brids especially if such areas show high levels of interconnectivity with otherstreams) or, if stocking does occur or did occur in the past, its likely con-sequences to native fish gene pools (e.g., relative probability of admixture).Finally, understanding what influences admixture can help suggest potentialremedial actions, e.g., streams that have marginal conditions for non-nativesor that were stocked only lightly may present the best cases for recovery ofnative fishes.36Chapter 3Cold tolerance limitshybridization betweenwestslope cutthroat troutand rainbow trout3.1 IntroductionSuitable conditions for survival play a large role in defining a species? dis-tribution in nature. These conditions include biotic interactions that aremediated by abiotic factors that vary across a landscape (Baltz et al. 1982;Rahel 1984; Sih 1987; Dunson and Travis 1991; Warner et al. 1993; De Stasoand Rahel 1994). In freshwater fishes for example, zonal patterns of speciesdominance are often observed in rivers and streams that flow along an eleva-tional gradient (e.g., Paul and Post 2001). Factors that vary with elevationsuch as water temperature, water velocity and substrate, alter the abioticconditions to a state that can favor one species over another up or down-stream (Vannote et al. 1980; Rahel 1984; Fausch et al. 1994; Taniguchi et al.1998). Water temperature in particular, appears to shape the distributionof salmonid fishes, where small changes of only 1 ?C have been linked to thepresence or absence of certain species (Fausch et al. 1994).The temperature of the water alters metabolism and behaviour in ec-tothermic fishes, which in turn affects their competitive ability (Taniguchiet al. 1998; Selong et al. 2001; Bear et al. 2007). A common distributionpattern in habitats that have been stocked with non-native salmonids isa restriction of the native species to high elevation headwaters and tribu-taries, followed by interspecific hybrids and complete replacement by theintroduced species downstream (Rahel and Hubert 1991; Fausch et al. 1994;Hitt et al. 2003; Weigel et al. 2003; Rubidge and Taylor 2005; Rasmussenet al. 2010). Cold waters at high elevations appear to provide a refuge forthe native species allowing them to persist despite the absence of physical373.1. Introductionbarriers that would stop the movement of introduced trout and hybrids up-stream (Paul and Post 2001; Weigel et al. 2003; Rubidge and Taylor 2005;Rasmussen et al. 2010). According to the ?elevation refuge hypothesis? firstintroduced by Paul and Post (2001), cold temperatures impart a competitiveadvantage to native trout at high elevation and the reverse is true at low ele-vations where introduced taxa typically dominate. Hybrids, if viable, exhibitintermediate behaviours and are found at intermediate habitats (Rasmussenet al. 2010).Laboratory and controlled field studies have yielded mixed results insupport of the elevation refuge hypothesis (see review in McHugh and Budy2005). Some studies, such as that carried out by De Staso and Rahel (1994)on Colorado River cutthroat trout (Oncorhynchus clarkii pleuriticus) andbrook trout (Salvelinus fontinalis), supported a shift in competitive abil-ity as water temperature changed. By contrast, work by McMahon et al.(2007) failed to detect a reversal in competitive advantage favouring thenative species at cold temperatures. McHugh and Budy (2005) were alsounable to demonstrate that Bonneville cutthroat trout (O. c. utah) weremore successful than exotic brown trout (Salmo trutta) at field sites locatedat high elevation. In cases where a reversal in competitive ability is notobserved, McHugh and Budy (2005) argued that physiological limitationsexerted a stronger influence at cold temperatures. Based on their data, theauthors suggested that interspecific competition was likely driving speciesdominance at low elevation and that non-native fishes would displace indige-nous taxa upstream to a point where their metabolic needs could no longerbe met (McHugh and Budy 2005).The elevation refuge hypothesis has been used to describe the distri-bution of westslope cutthroat trout (Oncorhynchus clarkii lewisi ; WSCT),which are currently threatened by hybridization with introduced rainbowtrout (O. mykiss; RT) (Rasmussen et al. 2010). Westslope cutthroat troutinhabit approximately 20% of their historical range in southwestern Alberta(COSEWIC 2006). Fragmented populations of genetically pure WSCT areat risk due to extensive stocking of hatchery RT to enhance recreationalfisheries (Allendorf and Leary 1988). Introduced RT have become natural-ized in some locales and will hybridize with WSCT. The hybrids are ableto reproduce and typically form a hybrid swarm at mid-elevation habitats.Although the two parental taxa have virtually identical optimal growth tem-peratures (WSCT: 13.6 ?C; RT: 13.1 ?C, Bear et al. 2007), studies by Pauland Post (2001) and Hitt et al. (2003) have found that RT tend to disperseto low elevations and warmer waters regardless of the elevation at whichthey were introduced. Observational differences in upper thermal tolerance383.1. Introductionhave also been established between these species, suggesting they occupydifferent, but overlapping thermal regimes (Bear et al. 2007). Geneticallypure WSCT require cool, clear water and continue to persist only in theupper headwaters of their range, suggesting that the elevation refuge hy-pothesis may be applicable to their interactions with introduced salmonids(Paul and Post 2001; Rasmussen et al. 2010).The inability of work by McMahon et al. (2007) to demonstrate a com-petitive advantage for the native taxa at high elevations has lead to furtherexploration of the theory put forth by McHugh and Budy (2005). Researchby Rasmussen et al. (2010) identified a life history gradient associated withthe clinal distribution of RT alleles. Fish with RT characters tended togrow more quickly and mature faster than genetically pure WSCT. Habi-tats at low elevations are more productive and appear to be better suited tosupport the energetic demands of RT. Cold-water habitats, being less pro-ductive, are able to support slow-growing WSCT. Rasmussen et al. (2010)suggested that a tradeoff exists between metabolic scope (RT) and growthefficiency (WSCT) within the system. In a more recent study (Rasmussenet al. 2012), differences in metabolism were evaluated by looking at ratesof oxygen consumption, and the activity of lactate dehydrogenase and cit-rate synthase. The authors concluded that fish with RT alleles tended tohave higher metabolic demands than fish comprised primarily of WSCTbackground. Hybrids generally had intermediate metabolic traits and werebetter able to balance the tradeoff between energetic scope and growth ef-ficiency, allowing them to be successful further upstream than RT. Thatnon-admixed WSCT persist in high elevation tributaries, likely reflects thephysiological and metabolic limitations of RT and interspecific hybrids (Ras-mussen et al. 2010, 2012).Understanding the role of thermal limits can help pinpoint species leveldifferences that shape the zonal distribution observed in so many studies.Despite having similar optimal growth temperatures, RT can grow over abroader range of temperatures and exhibit an upper lethal temperature4.7 ?C higher than WSCT (Bear et al. 2007). As suggested by McHughand Budy (2005), competition is more likely to explain species dominanceat low elevations that reach maximum temperatures significantly lower thanphysiologically defined, upper lethal limits. McHugh and Budy (2005) alsohighlighted a need to better understand the physiological performance ofboth species and their hybrids at low thermal limits.To date, the majority of studies on freshwater fish have focused on ther-mal maxima, the upper threshold of temperature tolerance (see review inBeitinger et al. 2000). Typically, these studies assay the critical maximum393.2. Materials and Methodstemperature (CTMax) defined as the maximum temperature at which a fishcan maintain a normal, upright swimming position (?equilibrium?). Theminimum temperature at which a fish can maintain such equilibrium istermed CTMin and has been much less commonly studied (but see Bar-rett et al. 2011; Darveau et al. 2012). For a comparison, upper thermallimits have been estimated in 72 studies for 103 different freshwater fishspecies and only fifteen studies report a lower limit across 29 species (re-viewed by Beitinger et al. 2000). This disparity exists for several reasons.Maximum limits tend to be easier to measure and interspecific variabilityat thermal maxima is much greater. Endpoints for thermal minima canbe harder to define and tend to be near 0 ?C for many freshwater species.There is also a heightened interest in exploring upper thermal limits withthe onset of global warming and increases in thermal stress, where fish aremore likely to experience temperatures much higher than their evolutionarypast (Beitinger et al. 2000).Understanding the underlying mechanisms driving the distribution ofintroduced species, native taxa, and their hybrids is integral to proper re-covery management. Introgressive hybridization threatens WSCT across theentirety of its historical range (Allendorf and Leary 1988; Shepard et al. 2005;COSEWIC 2006; Trotter 2008). High elevation isolates of WSCT, representfew remaining strongholds for these coldwater species, that at one time,had much more extensive distributions (Mayhood and Taylor 2011). In thischapter, I measure CTMin for WSCT and RT to test whether cold temper-atures present a greater physiological constraint for RT. This may explain,in part, why rainbow trout and hybrids tend to be distributed in warmerwaters at low elevations. This study presents a ?first-step? in exploring thephysiological limits of cold tolerance in WSCT, divergent populations of RTand their hybrids.3.2 Materials and Methods3.2.1 Trout populationsWestslope cutthroat trout used in this study were obtained from a sourcepopulation in Connor Lake, British Columbia (Fig. 3.1). Fry were rearedwith the help of the Kootenay Hatchery in Cranbrook, BC, before transferto the UBC Aquatics Facility in April 2012. Fish from this brood stock havebeen used in all WSCT stocking events in BC over the last three decadesand represent native WSCT genotypes.Rainbow trout from Blackwater River (BW) and Tzenzaicut Lake (TZ),403.2. Materials and MethodsFigure 3.1: Map of source locations for westslope cutthroat trout (On-corhynchus clarkii lewisi, Connor Lake) and rainbow trout (O. mykiss,Blackwater River (BW) and Tzenzaicut Lake (TZ)) used in critical ther-mal minima trials.413.2. Materials and MethodsBritish Columbia, were spawned with the assistance of the Freshwater Fish-eries Society of British Columbia (Fig. 3.1). These wild populations werechosen based on ease of access and divergence in specific characters. Forexample, BW rainbow trout are characterized by fast-growth and recom-mended for stocking in competitive habitats, while TZ rainbow trout aresuited to colder habitats with low productivity and survival is significantlybetter in stocked TZ fish than BW fish (Clarke et al. 2008; Northrup andGodin 2009). Based on these characteristics, the TZ fish appear to be moresimilar to WSCT (i.e. slow growing, utilize habitats with low productiv-ity), while BW fish possess attributes typical of the RT described in studiesof hybridized populations (i.e. fast-growth, poor survival and highly com-petitive) (Rasmussen et al. 2010). The availability of sexually mature F1hybrids between BW and TZ, presented an opportunity to assess interpop-ulation and interpopulation hybrid differences. I evaluated cold tolerance inthe F2 generation and backcrosses to each parental population (Blackwaterbackcross, BWB; Tzenzaicut backcross, TZB) in an attempt to assess howhigher-order hybrids between a WSCT-like trout (TZ) and an ?average? RT(BW) perform in CTMin. Although these are not true WSCT x RT hybridsand backcrosses, they were used here to illustrate possible outcomes hadWSCT x RT hybrids been obtainable for this study.Here, I evaluated the critical thermal minima (CTMin) for 431 troutrepresenting groups of WSCT, RT, F2 hybrids of the RT populations andhybrid backcrosses to each parental RT population acclimated to 15 ?C and18 ?C.3.2.2 Rearing set-upFish were kept in an environmental chamber (manufactured by Environ-mental Growth Chambers) in the Biological Sciences Building at the UBCcampus. Constant conditions were maintained at 12:12 L:D, and an ambi-ent temperature of 15 ?C with water supply at 10 ?C. Fertilized eggs wereplaced into baskets constructed from 2-inch, PVC pipe and mesh nettingsealed with aquarium grade silicone. The baskets were housed in a verticalincubator, with a continuous flow of freshwater. Individual baskets werelabeled by family. After approximately thirty days post-fertilization, theeggs ?eyed-up? and following another twenty days, alevins (larvae) began tohatch. One week, post-hatch, the alevins were transferred to plastic, con-tainers, submerged in a fiberglass, rearing trough. One and a half-inch holeswere drilled in the sides of the container, covered in mesh netting, and sealedwith aquarium grade silicone to allow for water flow. The alevins remained423.2. Materials and Methodsin these containers until their yolk sacs absorbed and they were fed crushedtrout chow (approximately 3 weeks post-hatch). Fry were then transferredto wooden troughs that had been painted and sealed with fiberglass coat-ing. The troughs were housed in a three-level shelving unit and each troughhad a one-inch, PVC, water wand that supplied a continuous flow of cleanfreshwater. The troughs drained to a common 1.5-inch pipe which flowed toa floor drain. Trout were fed 1.2 mm Biovita trout chow.3.2.3 AcclimationEach test group was acclimated for a minimum of two weeks at 15 ?C or18 ?C. Fish were housed in four, 195 L, fiberglass tanks that were connectedto a common sump. Both the acclimation tanks and CTMin apparatuswere set-up in the same environmental chamber held at 5 ?C, to controlwater cooling and eliminate temperature fluctuations during the trials. Toachieve the appropriate acclimation temperature, Odyssey Heatpro aquar-ium heaters were placed in the common sump, which pumped water heatedto 25 ?C into each tank. A nozzle controlled the flow of this water suchthat only a trickle of the heated water would enter the tank at any giventime. Fine control of the inflow of warm water, in combination with tankwater being cooled by the ambient 5 ?C air, allowed me to obtain a constantacclimation temperature of 15 ?C or 18 ?C.The acclimation tanks were initially set-up as a recirculating system, butwere later modified to a flow-through system when a fungus killed 60% ofthe WSCT individuals acclimating to 18 ?C. No mortalities were observedin the WSCT acclimating to 15 ?C or any of the RT test groups at either ac-climation temperature. These groups showed no physical signs of the fungalinfection even though some were exposed to the same water as the infectedWSCT in the recirculating set-up. I believe the development of fungus wasbrought on by heat-stress experienced by WSCT while acclimating to 18 ?C(see discussion). The spread of the infection was curbed by dropping thewater temperature to 10 ?C and allowing fish to stabilize for one week. Aheader tank was added so that cool, clean water could continuously be addedto the system, and the tanks were gradually brought back up to the acclima-tion temperatures. After the addition of the header tank and conversion toa flow-through system, WSCT acclimating to 18 ?C did not show any signsof fungal infection and were able to complete the acclimation procedure.During acclimation, fish were fed 1.2 mm Biovita trout chow every day.433.2. Materials and Methods3.2.4 Critical thermal minima determinationCTMin experiments on trout test groups were carried out at different timesover an eight month period in 2013. This was because some groups haddifferent growth rates and I wanted to minimize the size variation amongstfish used in the study. WSCT grew much slower than RT and even at thetime of trial, were smaller than the RT being tested. The two week accli-mation period and number of acclimation tanks available also contributedthe the duration of the study. In addition, unforeseen, technical issues, (i.e.chamber compressor failing) also caused trials to suspend for a short periodof time adding to the length of the test period.The experimental apparatus consisted of a single, large antifreeze bathin which twelve, individual containers were floated and held in place by alarge Styrofoam sheet, fitted with holes to hold each container. A large,20-gallon, Rubbermaid container was placed on a table and elevated withcement blocks above the antifreeze bath. A flap was cut into the lid of thiscontainer to facilitate the addition of dry ice to the system, allow CO2 toescape, and to prevent uncontrolled bubbling and splatter. The containerwas outfitted with a brass nozzle to control the flow of cooled antifreezeentering the antifreeze bath. Below the antifreeze bath, an identical Rub-bermaid container collected the antifreeze that had already flowed throughthe apparatus. A Little Giant, 115V-60Hz, pump in this container pumpedthe used antifreeze back up to the elevated Rubbermaid container to becooled again and re-circulated through the system during the trial.Before the trial, 25 ?C water, from the sump of the acclimation set-up,was added to the twelve containers in the antifreeze bath. Twelve fish werecollected from the acclimation tanks and transferred into a bucket. A singlefish was placed into each of the twelve containers once the water cooledto meet the acclimation temperature. The fish were given fifteen minutesto acclimate to the test container, during which dry ice was added to theelevated Rubbermaid container. After the fifteen minute acclimation, thenozzle of the Rubbermaid container was opened and the cooled antifreezebegan to flow into the large bath. The pump in the collection container wasturned on to circulate the antifreeze through the set-up. Each of the twelvecontainers within the antifreeze bath was connected to a digital thermometerand had an air stone, to ensure the water was saturated with oxygen andto allow for homogenous cooling. The rate of cooling was monitored usingthe digital thermometers and kept at a rate of -0.3 ?C per minute, by theaddition of dry ice to the system. The temperature at which a fish lostequilibrium was recorded as its critical thermal minima (CTMin). Once all443.3. Resultstwelve fish lost equilibrium, the trial ended and the fish were euthanized inMS222, weighed and measured. The fish were then labeled and preservedin 95% ethanol for future potential DNA analysis. For each acclimationtemperature, three CTMin trials of 12 fish were carried out for a total of 36individuals per test group.3.2.5 Statistical analysisCold tolerance data were analyzed first using simple linear regression (SLR)to test for effects of length (cm) and mass (g) on CTMin at both acclimationtemperatures. Differences in means of the test groups were analyzed usingANOVA on the data collected at 15 ?C acclimation (SLR yielded no effectof length or mass) and ANCOVA on the 18 ?C dataset incorporating lengthand mass as covariates. Values corrected for differences in length and massdid not differ from actual values by more than 0.1 to 0.2 ?C (see below).3.3 Results3.3.1 Effect of acclimation temperatureCritical thermal minima increased with acclimation temperature in all testgroups (i.e., the temperature at which fish lost equilibrium increased withacclimation temperature). Simple linear regressions of CTMin on acclima-tion temperature were significant in both RT and WSCT test groups (RT:r2 = 0.059; p < 0.001; WSCT: r2 = 0.256, p < 0.001; Fig. 3.2). Significantdifferences in CTMin within each test group were observed in BWB, WSCT,F2 and TZ at 15 ?C and 18 ?C acclimation (Fig. 3.2). Regression analysisrevealed that RT experience a 0.16 ?C increase in CTMin for every 1 ?C in-crease in acclimation temperature. The rate of change was nearly double inWSCT, 0.29 ?C increase in CTMin with every 1 ?C increase in acclimationtemperature.3.3.2 Effect of body sizeThere were significant differences in fork length (cm) and mass (g) amongsttest groups (ANOVA length: p < 0.001, mass: p < 0.001). Rainbow troutwere generally larger than WSCT ( Table. 3.1). Simple linear regressionrevealed no significant effect of fork length and mass on CTMin at an ac-climation temperature of 15 ?C (p = 0.729). At 18 ?C acclimation, however,453.3. ResultsFigure 3.2: Simple linear regressions of CTMin on acclimation temperaturefor trout test groups. Blackwater River rainbow trout (BW; r2 = 0.034, p =0.121), Blackwater River backcross (BWB; r2 = 0.067, p = 0.030), westslopecutthroat trout (WSCT; r2 = 0.256, p < 0.001), Blackwater River x Tzen-zaicut Lake F1 hybrid x Blackwater River x Tzenzaicut Lake F1 hybrid (F2;r2 = 0.180, p < 0.001), Tzenzaicut Lake rainbow trout (TZ; r2 = 0.053, p =0.050), Tzenzaicut Lake backcross (TZB; r2 = 0.003, p = 0.661).463.3. ResultsFigure 3.3: Critical thermal minima values for each trout test group at 15 ?C(blue, solid line) and 18 ?C (green, dashed line) acclimation with 95% con-fidence intervals. Blackwater River rainbow trout (BW), Blackwater Riverbackcross (BWB), westslope cutthroat trout (WSCT), Blackwater River xTzenzaicut Lake F1 hybrid x Blackwater River x Tzenzaicut Lake F1 hy-brid (F2), Tzenzaicut Lake rainbow trout (TZ), Tzenzaicut Lake backcross(TZB).473.3. Resultslength and mass did exert a significant effect (p = 0.028). Analysis of covari-ance, controlling for the effect of fork length and mass, revealed differencesin CTMin amongst test groups acclimated at 15 ?C, but failed to detect anystatistically significant differences for fish acclimated to 18 ?C (p = 0.006and p = 0.110, respectively; Table. 3.2). Size-corrected means generatedfrom the ANCOVA for test fish acclimated to 15 ?C did not differ from theactual values by more than 0.2 ?C or 0.1 ?C in 18 ?C acclimated fish (Table.3.1).3.3.3 Interspecific differencesRainbow trout had significantly higher CTMin than WSCT at 15 ?C (1.4 ?Cand 1.0 ?C, respectively, p = 0.01; Table. 3.1). At 18 ?C, interspecific differ-ences in critical thermal minima were not observed (p = 0.929; Table. 3.1).Here, mass seemed to be a better predictor of cold tolerance (p = 0.022),but only explained 3% of the variation in CTMin.3.3.4 Intraspecific differences in rainbow troutAverage CTMin for RT test groups acclimated at 15 ?C ranged from 1 ?Cfor the F2 fish to 1.7 ?C for TZB (Table. 3.1). Differences between thetest groups proved to be subtle, but statistically significant (p = 0.006; Ta-ble. 3.2). Levene?s test of equality of error variances amongst test groupsapproached significance (p = 0.085), however, robust tests of equality ofmeans for groups with unequal variances continued to yield significant dif-ferences between test groups (Welch p = 0.003; Brown-Forsythe p = 0.007).One-way ANOVA followed by a post-hoc test revealed that the only signif-icant differences in CTMin amongst RT test groups were between the F2population and TZB (Tukey, p = 0.02; Fig. 3.3).As predicted (see Materials and methods), BW fish had a higher aver-age CTMin than TZ fish, although these differences were not statisticallysignficant (1.6 ?C and 1.3 ?C, respectively, Tukey, p = 0.934) and suggeststhat these results are not more than what would be obtained by chance.At an acclimation temperature of 18 ?C, CTMin across all test groupsranged from a low of 1.8 ?C in Tzenzaicut Lake backcross (TZB) individuals,to 2.1 ?C in Blackwater River backcross (BWB) (Table. 3.1). There wereno statistically significant differences in means amongst the RT test groupsacclimated at 18 ?C when controlling for differences in fork length and mass(ANCOVA, p = 0.647; Table. 3.2).483.3. ResultsTable 3.1: Critical thermal minima (CTMin), length, mass and CTMin cor-rected for length and mass of trout test groups acclimated at 15 ?C and18 ?C. Values are mean ?s.d.; N is the number of fish; Rainbow trout(RT; all populations); westslope cutthroat trout (WSCT); RT populations:Blackwater River rainbow trout (BW); Blackwater River backcross (BWB);Blackwater River x Tzenzaicut Lake F1 hybrid x Blackwater River x Tzenza-icut Lake F1 hybrid (F2); Tzenzaicut Lake rainbow trout (TZ); TzenzaicutLake backcross (TZB).15 ?C AcclimationTest group Fork length Mass Mean CTMin Size-corrected N(cm) (g) ( ?C) CTMin ( ?C)RT 9.5?1.3 9.2?3.6 1.4?1.0 1.5 179WSCT 7.3?0.8 4.0?1.7 1.0?0.8 0.8 36BW 8.6?1.4 7.1?3.3 1.6?0.9 1.5 36BWB 9.3?1.3 8.9?3.7 1.5?1.2 1.6 35F2 9.9?1.2 9.7?3.3 1.0?1.2 1.1 36TZ 10.2?1.1 10.5?3.2 1.3?0.7 1.5 36TZB 9.6?1.0 10.0?3.4 1.7?0.7 1.8 3618 ?C AcclimationRT 9.6?1.3 8.4?3.1 1.9?0.9 1.9 180WSCT 7.4?0.8 4.7?1.9 1.9?0.7 2.0 36BW 9.5?1.3 8.5?3.3 1.9?1.0 1.8 36BWB 9.3?0.8 8.5?2.4 2.1?0.8 2.1 36F2 9.8?0.9 9.5?2.8 2.0?1.0 2.0 36TZ 10.0?1.0 9.5?3.0 1.8?0.8 1.7 36TZB 9.6?0.8 9.6?2.2 1.9?0.9 1.9 36493.3. ResultsTable 3.2: Summary statistics for analysis of covariance (ANCOVA) betweencritical thermal minima (CTMin) and body size in trout test groups at 15 ?Cand 18 ?C acclimation using Type III sum of squares.15 ?C AcclimationSource Sum of Squares df Mean Square F SigCorrect Model 17.667 7 2.524 2.905 0.006Intercept 3.782 1 3.782 4.353 0.038Length 0.408 1 0.408 0.470 0.494Mass 0.013 1 0.013 0.015 0.903Test group 17.079 5 3.416 3.932 0.002Error 179.846 207 0.869Total 600.910 215Corrected Total 197.513 21418 ?C AcclimationCorrect Model 9.678 7 1.383 1.701 0.110Intercept 0.117 1 0.117 0.144 0.704Length 3.775 1 3.775 4.644 0.032Mass 5.757 1 5.757 7.082 0.008Test Group 3.803 5 0.761 0.936 0.459Error 169.082 208 0.813Total 972.260 216Corrected Total 178.760 215503.4. Discussion3.4 DiscussionThe results of my critical thermal minima experiments reveal that thereare subtle, but significant differences between species acclimated at 15 ?C.Despite the variation in average CTMin amongst RT test groups, therewas a statistically significant difference between average values obtained forWSCT and RT acclimated at 15 ?C (1.0 ?C and 1.4 ?C, respectively). Whenacclimation temperature was increased to 18 ?C, there were no significantinterspecific differences observed (Table. 3.1). The loss of species leveldifferences at high acclimation temperatures could be the result of heatstress experienced by WSCT during acclimation which may have inhibitedtheir cold tolerance performance (see below).The duration of the CTMin trials extended longer than anticipated anddue to the length of time, raises the concern of differences in seasonal per-formance amongst test groups. Maintaining constant lab conditions duringrearing and acclimation, as done in my study, should control for any effectof seasonality. In all cases, the timing of the three CTMin trials for each testgroup at 15 ?C or 18 ?C, were chosen based on the size of the individuals (tocontrol for size variation amongst fish tested) and were never carried out ina single month, but spanned a minimum of two months. Further, significantdifferences in CTMin (such as that between WSCT and TZB) overlappedin the timeframe in which they were tested.Despite sharing similar thermal optima, WSCT and RT appear to havesignificant differences in other aspects of their respective thermal regimes(Bear et al. 2007). Field studies have repeatedly described a gradient ofspecies dominance that follows a gradient in elevation and water temperaturewhen WSCT and RT exist in sympatry (Weigel et al. 2003; Hitt et al. 2003;Rubidge and Taylor 2005; Muhlfeld et al. 2009b; Rasmussen et al. 2010).In laboratory studies, interspecific differences exist when both species weretested at the upper extremes of their thermal scope (Bear et al. 2007). Rain-bow trout grow over a broader range and continue to grow at temperaturesbeyond 20 ?C, which are lethal to juvenile WSCT. When tested for upperlethal limits, RT can survive temperatures 4.7 ?C above WSCT thresholds(Bear et al. 2007). My work, however, is the first to evaluate species differ-ences between WSCT and RT at the lower limits of their thermal range.As a whole, literature on thermal tolerance is biased towards experi-ments of upper thermal tolerance (UTT, Beitinger et al. 2000). This islikely because definitive endpoints are easier to observe, there is a greaterdegree of interspecific variation, and heat tolerance is more likely to limitspecies distribution in nature (discussed in Beitinger et al. 2000). I used513.4. Discussiona non-lethal procedure that employs a gradual decline in water tempera-ture (critical thermal methods; CTM). This method can be contrasted withincipient lethal temperature (ILT) or Fry method that evaluates thermaltolerance based on abrupt transfers to temperatures above or below accli-mation until death is reported in 50% of the sample population (Fry 1947).In contrast, critical thermal methodology, utilizes a dynamic change in tem-perature, which is arguably more representative of conditions in the wild(see discussion in Beitinger et al. 2000). Further, CTM requires fewer accli-mation temperatures and smaller sample sizes (for a detailed discussion ondifferences between CTM and ILT see Beitinger et al. 2000).3.4.1 Acclimation temperatureTo date, two studies have evaluated CTMin in RT acclimated to 10, 15and 20 ?C and there are no reported CTMin values for WSCT (Becker andGenoway 1979; Currie et al. 1998). The acclimation temperatures used inmy study, were chosen such that a quantifiable CTMin could be obtained forboth species. An acclimation temperature of 15 ?C with a linear decline of0.3 ?C min?1 was used in my experiments so that a direct comparison couldbe made with previous work on RT (Becker and Genoway 1979; Currieet al. 1998). A second acclimation temperature of 18 ?C with the same rateof decline was chosen in an attempt to capture species level differences thatmay be understated at 15 ?C. An acclimation temperature above 18 ?C wasnot employed in my study as these temperatures are likely too warm forWSCT (Bear et al. 2007), and an acclimation temperature below 15 ?C wasnot tested as RT held at 10 ?C in the previous studies, did not exhibit lossof equilibrium when the water began to freeze at 0 ?C (Becker and Genoway1979; Currie et al. 1998).In my experiments, there was a significant effect of acclimation temper-ature on cold tolerance performance in WSCT and RT. Regression analysisrevealed an increase of 0.29 ?C in CTMin with every 1 ?C increase in ac-climation temperature for WSCT. This value is similar to those reportedin killifish (Fundulus heteroclitus, Fangue et al. 2006), sheepshead minnow(Cyprinodon variegatus, Bennett and Beitinger 1997) and other studies ofRT (Becker and Genoway 1979; Currie et al. 1998). The rainbow trout eval-uated in my work, however, only exhibited a 0.16 ?C increase in CTMin withevery 1 ?C increase in acclimation temperature. This is considerably lowerthan the value of 0.36 ?C per 1 ?C rise in acclimation reported by Beckerand Genoway (1979) and Currie et al. (1998). Differences between thesevalues could be the result of testing only two acclimation temperatures in523.4. Discussionmy study that differed by only 3 ?C. Becker and Genoway (1979) and Currieet al. (1998) evaluated three acclimation temperatures at 5 ?C increments.Thus, the values drawn from my study are applicable to a smaller rangeof temperatures and thus may not have elicited as large as response asthese previous studies. Further, the RT used in my work represent differentpopulations and crosses, which increases the potential genetic variability incold tolerance and may translate to higher variability in cold performance.Acclimation temperature will generally explain a significant portion of thevariation observed in CTMin experiments (Becker and Genoway 1979; Ben-nett and Beitinger 1997; Currie et al. 1998; Fangue et al. 2006). Similarwork generally reported r2 values of CTMin on acclimation temperatureabove 0.95. My study produced dramatically lower r2 values likely a resultof high levels of intraspecific variation (Fig. 3.2).3.4.2 Inter- and intraspecific differences15 ?C acclimationMy results reveal a statistically significant difference in average CTMin be-tween RT and WSCT acclimated to 15 ?C (Table. 3.1). Some RT individualswere capable of performing just as well, if not better than certain WSCT,and thus, a definitve conclusion that WSCT are more cold tolerant than RTcannot be made (Appendix B Fig. B.1). Further tests of intraspecies vari-ation in WSCT (testing other populations) would be necessary to obtainmore compelling support for this hypothesis. These results do, however,agree with data that suggest the distribution of RT alleles at lower eleva-tions may be the result of habitat preference for warmer waters (Hitt et al.2003; Rasmussen et al. 2010, 2012). Differences in physiological response towater temperature between the two species, may be responsible for the zonalpattern observed in hybridizing populations and as a result, may be limitingthe spread of RT alleles into cold headwaters (Rasmussen et al. 2010, 2012).The F2 crosses of RT had a mean CTMin value equivalent to WSCT at15 ?C acclimation. These results suggest that it may be possible for RT todisplace WSCT at higher elevations given the appropriate combination ofgenes. It may simply be a matter of time and the lack of competitive pres-sure at lower elevations that prevent RT from seeking out and establishingpopulations in the upper headwaters. It is unknown whether RT individualswith high cold tolerance are competitively superior in all aspects, even atlow elevations, and are therefore, never forced to seek less favourable habi-tats in cold extremes. Further tests are needed to determine whether RT533.4. Discussionwith high cold tolerance are competitively superior in low and high elevationhabitats.There were high levels of variation in cold performance between differ-ent populations of RT; however, statistical significance was only observedbetween two test groups (F2 and TZB; Fig. 3.3). Second-generation (F2)hybrids between ?warm? (BW) and ?cold? (TZ) water-adapted RT wereable to perform just as well as WSCT when acclimated at 15 ?C and hadthe highest degree of variation amongst all the groups tested. Assumingadditive genetic variability for any trait, the F2 hybrid offspring should becharacterized by high levels of genetic variation as genotypic combinationsof both parental genotypes as well as hybrid combinations between the twopopulations are represented in this group. The low CTMin obtained for thisgroup is difficult to explain, as simple additive heritability would predict anaverage CTMin performance (a value intermediate to the parental popula-tions (BW, TZ)). Unfortunately the genetic architecture of cold tolerancein salmonids is poorly understood. Quantitative trait locus (QTL) analyseswould need to be carried out to identify stretches of DNA that control or arelinked to cold tolerance performance in trout. This would increase our un-derstanding of the underlying genetic controls for this trait (e.g. Xiao et al.1995). Without this information, it is difficult to know why the backcrossgroups (BWB, TZB) and the F2 hybrids performed the way they did. Thepatterns observed suggest that the genes controlling cold tolerance do notbehave in a simple, additive manner.Tzenzaicut fish had an average CTMin lower than BW fish (higher coldtolerance, Table. 3.1). This result is supported by reports from Clarkeet al. (2008) and Northrup and Godin (2009) that TZ fish perform better ingrowth and survival than BW fish when stocked in colder habitats. Further,it lends support to the rationale used in my study; that wild fish populationsthat evolved in colder habitats perform better in cold tolerance trials. Thisresult agrees with my predictions and suggests that the results of my studyhave relevance in nature.As a whole, the CTMin values reported in my study are higher thanprevious reports in RT (Becker and Genoway 1979; Currie et al. 1998). Atan acclimation temperature of 15 ?C and a rate of temperature decline of0.3 ?C min?1, Currie et al. (1998) report a CTMin of only 0.2 ?C. Beckerand Genoway (1979) report a similar CTMin of 0.7 ?C under similar con-ditions. In contrast, the average CTMin of RT in my study was 1.4 ?Ctested under the same conditions. This discrepancy amongst individuals ofthe same species could be the result of minor but significant variations inmethodology, experimental set-up, population, age and size of fish, etc. For543.4. Discussioninstance, RT used by Becker and Genoway (1979) were collected straightfrom the Columbia River, while fish used in the study by Currie et al.(1998) had been purchased from a hatchery in Missouri. My fish originatedfrom experimental crosses using fish collected in the wild and subsequentlyraised under laboratory conditions. Currie et al. (1998) also used a CTMinendpoint that deviated from my methods. The temperature they reportedwas taken one minute after the fish exhibited an initial loss of equilibrium(LOE). I recorded my temperatures as the initial onset of LOE and did notwait one minute to record CTMin values. Becker and Genoway (1979) re-ported CTMin as LOE50 (the temperature at which 50% of a sample lostequilibrium). Further, both Becker and Genoway (1979) and Currie et al.(1998) introduced cold water directly into their system, while I used anantifreeze bath to gradually cool the test water. The fish used by Currieet al. (1998) were only four centimetres long and only 6 weeks old. My fishwere considerably larger and older, and my experiments demonstrated thatfork length and mass were significant predictors of cold performance in fishacclimated at 18 ?C (Table. 3.2, 3.1, Appendix B Table. B.1). Analysisof variance (ANOVA) revealed that length had a positive relationship withCTMin, where longer fish tended to lose equilibrium at higher CTMin values(regression coefficient = 0.325). The opposite was true for mass. Fish thatweighed more tended to have lower CTMin (regression coefficien t= -0.161).These two factors are likely to play a role in how well a fish can maintainits equilibrium.Despite the discrepancy in the reported critical thermal minima of RTin previous studies and my own, my work does reveal that RT and WSCTexhibit marked differences in thermal minima, when tested under a commonexperimental procedure at least when acclimated to 15 ?C.18 ?C acclimationWhen acclimation temperature was increased to 18 ?C, there were no signif-icant intra- and interspecific differences in cold tolerance (Table. 3.2). Atthis temperature, body length and mass were better predictors of CTMin.This suggests that the genetically controlled, physiological differences thatexist when fish are held at temperatures in the mid-range of their tolerancethresholds are no longer relevant when one or both species is beginning toexperience some degree of heat-stress (Beitinger et al. 2000; Bear et al. 2007).A temperature of 18 ?C represents above average summer temperatures forWSCT and approaches the upper extreme at which juvenile fish can sur-vive (survival drops dramatically above 20 ?C, Bear et al. 2007). Westslope553.5. Conclusionscutthroat trout held at this temperature may already be experiencing somelevel of heat stress. In my study, the spread of a fungal infection in WSCTseemed to be triggered by the high acclimation temperature. Westslopecutthroat trout acclimated to 15 ?C showed no symptoms of the infection,despite exposure to the water of infected fish in the recirculating acclimationset-up. Similarly, RT showed no symptoms of the fungus at either 15 ?C or18 ?C acclimation.At 18 ?C, F2 fish loss the cold tolerance advantage observed at 15 ?Cacclimation (Table. 3.1). The performance of this group is similar to theresults obtained for WSCT at both temperatures. It is possible that thisgroup may have also been experiencing some degree of temperature stress at18 ?C, but the symptoms were not as conspicuous. These fish were tested inthe autumn of 2012, before any experiments were done on WSCT. Becausethey were tested earlier, the fungus that infected the WSCT in my studymay not have been in the system while the F2 were undergoing acclimation.3.5 ConclusionsWith the exception of two test groups (F2 and TZB), trends in cold tol-erance performance met my predictions. Blackwater River (BW) fish hadan average CTMin higher than Tzenzaicut Lake (TZ) fish and WSCT hada lower CTMin than pure RT genotypes and backcross individuals (Table.3.1). While I did not resolve measurable differences at 18 ?C acclimation,this may have been expected given that I was pushing the thermal limits ofWSCT.I was able to show that there are measurable differences in CTMin be-tween WSCT and RT. This finding lends support to the theory that phys-iological limitations may be preventing RT from invading cooler, higherelevation habitats as suggested by work by McHugh and Budy (2005) andRasmussen et al. (2010, 2012). This may explain, at least in part, thespecies gradient reported in numerous field studies that report zonal pat-terns of salmonid distribution (Rahel and Hubert 1991; Fausch et al. 1994;Hitt et al. 2003; Weigel et al. 2003; Rubidge and Taylor 2005; Rasmussenet al. 2010).Some RT individuals exhibited cold tolerance comparable to WSCTwhich suggests that they may do just as well in cold water, however, mystudy still remains limited by its scope (Fig. 3.3, Appendix B Fig. B.1,B.2). Futher analysis of additional WSCT populations may reveal high lev-els of interspecific variation in CTMin as observed in the RT groups tested.563.5. ConclusionsIt is possible that RT may not survive long-term exposure to cool, oftenunproductive habitats. As proposed by Rasmussen et al. (2012), cold-waterhabitats may not meet the metabolic needs of RT and prevent their es-tablishment in headwaters where we find genetically pure WSCT. A studyevaluating the thermal requirements of RT (perhaps different populations)through all life-stages (egg to adult) would be necessary to test this idea.The Alberta Westslope Cutthroat Trout Recovery Plan outlines effortsto restock genetically pure WSCT in habitats across their historical range(The Alberta Westslope Cutthroat Trout Recovery Team 2013). Conserva-tion efforts for WSCT should focus on rehabilitating populations at highelevations, that provide natural refuge from RT invasion via cold water andlow productivity as suggested by my study and previous work (McHugh andBudy 2005; Rasmussen et al. 2010, 2012). A better understanding of thegenetics that underlie cold tolerance may also be valuable, particularly ifcertain populations of WSCT show a heightened tolerance to cool tempera-tures. If we could identify such populations, they could be useful in effortsto restock high elevation locales as their cold tolerance could provide anextra means of resistance to RT invasion and subsequent hybridization withnative WSCT.57Chapter 4Conclusion4.1 Summary of FindingsMy thesis produced three general findings:1. Hybridization between westslope cutthroat trout (Oncorhynchus clarkiilewisi ; WSCT) and rainbow trout (O. mykiss; RT) is extensive acrosssouthwestern Alberta. Over 150 sites were tested and most popula-tions had qwsct between 0.85 and 0.99. A population with qwsct valuebelow 0.99 is considered ?hybridized?.2. Amongst the locales tested, water temperature, elevation, distance tonearest stocking site and distance to nearest railway were significantpredictors of levels of admixture. ?Genetically pure? populations ofWSCT are more common in cool, high elevation habitats that are farfrom stocking sites. These results are consistent with previous workdone over smaller spatial scales. The complex interaction with distanceto railway is difficult to interpret; however, railways were historicallyused to stock RT in southwestern Alberta.3. Rainbow trout and WSCT exhibit subtle, but significant differences inaverage CTMin when acclimated to 15 ?C. At this acclimation tem-perature, WSCT were able to withstand an average water temperaturecooler than RT. This result provides additional evidence that RT andhybrids may prefer warmer water and are limited in their upstreamspread by physiological demands that cannot be met in cold temper-ature habitats. These constraints may be preventing RT and hybridsfrom overtaking the final WSCT ?strongholds? at high elevations whenphysical barriers are not present.My analyses provide evidence that hybridization is indeed widespread insouthwestern Alberta and efforts to mitigate the loss of WSCT to genomicextinction are urgently needed. The identification of landscape variablesthat appear to influence levels of hybridization may assist recovery plans.584.2. Defining ?Genetic Purity?For instance, low elevation sites near former RT stocking locales are likelyto show a high degree of hybridization and may contain genetically pure RT.The likelihood of recovering pure WSCT in these types of habitats may bevery low. Efforts to restock WSCT in these sites without the eliminationof existing hybridized individuals and RT may be futile, if conditions inthese higher stream order habitats indeed favor RT and hybrids over WSCT.Work by Henderson et al. (2000) and Hitt et al. (2003) suggest that hybridsdisperse more than the parental taxa and are the agent of spread. Targetsites for recovery will require elements that suppress RT movement, such asthermal or physical barriers.4.2 Defining ?Genetic Purity?The standards set to classify a population as ?hybridized? or ?geneticallypure? for conservation purposes have been widely debated (Allendorf et al.2005; Campton and Kaeding 2005). Allendorf et al. (2001) recommend strin-gent guidelines, setting a 1% threshold of allowable admixture, such that?pure? populations are those with qwsct ? 0.99. They argued that becausehybrids with low levels of RT admixture experience a significant reductionin fitness in laboratory studies, any level of admixture in WSCT populationswould be undesirable. Further, protecting WSCT under such a conservativelimit would act to preserve the evolutionary legacy of the WSCT gene pool,while acknowledging that polymorphisms may still be shared between thetwo species or that historical, natural hybridization between them may haveoccurred (resulting in admixture of no more than 1%). These guidelineswould also reduce the likelihood of protection of populations with higherlevels of admixture that may serve as a means for introgression into purepopulations. Campton and Kaeding (2005) contested the ?1% admixturethreshold, and stated that admixture is present in naturally sympatric pop-ulations of WSCT and RT. These authors also added that in the long term,failing to protect populations with as low as 10% introgression would dra-matically reduce our capacity to recover WSCT. For instance, if only pop-ulations with ?1% admixture qualified for protection and the remainingpopulations were subject to eradication, and if eradication was not 100%,as is often the case, the remaining ?pure? populations could continue to mixwith residual admixed individuals and show evidence of introgression afterseveral generations and may no longer be subject to protection under strictconservation guidelines. This could continue on until we have depleted ourreserve of WSCT populations and no longer have any populations to pro-594.3. Anthropogenic Hybridizationtect that meet the ?1% admixture standard. Muhlfeld et al. (2009a) showedthat fitness of WSCT in nature may decline by up to 50% with as little as20% admixture with RT, although the mechanisms of such fitness declinesare unknown. More such studies under different environmental contexts arerequired before we can make broad generalizations about the fitness con-sequences of introgression given a starting level of admixture. Many whowork in this field agree that until the fitness consequences of introgressionin nature are better understood, case-by-case assessments concerning con-servation of individual populations is the best policy (see discussion in TheAlberta Westslope Cutthroat Trout Recovery Team 2013).Many of the sites measured in my study were found to have admixturelevels (qwsct) significantly lower than 0.99 (the amount of hybridizaton issignificant). The recovery plan outlined for WSCT in Alberta, adopt thestandards set by Allendorf et al. (2001) to identify the populations that aretop priorities for conservation, citing the same rationale; that slight levelsof admixture significantly reduces the fitness of female and male trout asdemonstrated by Muhlfeld et al. (2009a), and that 1% admixture accountsfor historical hybridization between species. The recovery team also rec-ognize that ?populations with low levels of hybridization (qwsct ? 0.95 but< 0.99) may be important for species conservation and recovery?, as thesepopulations may contain migratory or adfluvial life history forms, may beadapted to unique environments, or be the least introgressed populationwithin a geogrphic area. They may also have unique behaviours or phe-notypes that researchers and experts consider important to conserve (TheAlberta Westslope Cutthroat Trout Recovery Team 2013).4.3 Anthropogenic HybridizationIt is important not to ignore the role that hybridization plays in evolu-tion. Hybridization introduces genetic variation that can act as a sourcefor beneficial adaptations (Verspoor and Hammart 1991). Hybridizationwith a closely related taxon may also rescue small populations experienc-ing inbreeding depression (Willi et al. 2007). Hybridization may even play arole in speciation; examples of allopolyploid speciation exist in plants, whereparental species that differ in chromosome number hybridize and produce anoffspring that is unable to reproduce with either parental species because itcarries double the number of chromosomes of the parental species (Widmerand Baltisberger 1999). Some species of fish have also arisen as a result ofintrogressive hybridization (Demarais et al. 1992; reviewed in Taylor 2004).604.4. Concluding Thoughts and Future DirectionsBy contrast, hybrids that arise from anthropogenic activity generally neg-atively impact an ecosystem by threatening native species (Allendorf andLeary 1988; Rhymer and Simberloff 1996). ?Anthropogenic hybridization?as described by Allendorf et al. (2001) needs to be reduced and ?hybridsthat originate as a result of human activity should only be protected if thehybrids contain the only remaining genetic information from a species thathas otherwise been lost by genetic mixing or when their origin is unclear.?4.4 Concluding Thoughts and Future DirectionsMy research builds upon previous work on hybridization between WSCT andRT in North America. I employed a larger suite of genetic markers to betterestimate levels of admixture and extend my analysis of landscape variablesacross a larger spatial scale. The congruence between environmental factorshighlighted in my model and that of previous studies brings generality to thefindings. Our ability to detect hybrids has increased with growth in genetictechnology. Increasing sample sizes and developing genome-wide methodsto identify specific genes that differentiate WSCT, RT and their hybridswill provide more precise estimates of admixture and population structure(Hohenlohe et al. 2011). Fine-scale, geographic mapping and measurementof abiotic and anthropogenic variables that exist in WSCT habitats are alsoneeded to obtain a better understanding of the complex interactions whichlead to spatial variability in hybridization levels across the range of WSCT.Temporal sampling and genetic analysis will give us a better understand-ing of how admixture values change over time and an evaluation of the lifehistory requirements of WSCT will also be vital to develop effective conser-vation strategies.My data revealed a measurable difference in average CTMin betweenWSCT and RT at 15 ?C acclimation (1.0 ?C and 1.4 ?C, respectively). Thissupports previous laboratory work, which suggests that RT are better adaptedto warmer temperatures, being able to tolerate temperatures significantlyhigher than WSCT (Bear et al. 2007). My results also agree with fieldresearch that propose a habitat preference, supported by species level differ-ences in metabolism and life history strategy (Rasmussen et al. 2010, 2012).Together, these studies provide a foundation for the hypothesis put forthby McHugh and Budy (2005), that physiological constraints to cold waterat high elevations prevent the spread of RT into the headwaters, and thattemperature-mediated competition with RT is likely excluding WSCT fromlow elevations.614.4. Concluding Thoughts and Future DirectionsCarrying forward, molecular research into the genetic basis and architec-ture of thermal tolerance will provide further insight into the evolutionaryhistory of physiological adaptation in each taxon. If lower thermal limitsare physiologically constraining the movement and establishment of RT up-stream, determining the genes responsible for thermal tolerance may alsohelp us identify RT populations that are unlikely to overtake remainingWSCT populations if stocking continues (i.e., one management solution tocontinued demands for RT could be to stock genotypes that exhibit poorcold tolerance). In conclusion, the results of my thesis agree with and ex-tend previous research that identify water temperature as a major elementinfluencing levels of hybridization in WSCT, and contributes to the under-standing of how cold water in high elevation habitats may provide refuge forindigenous taxa threatened by hybridization with non-native species.62ReferencesAldridge, G. and Campbell, D. (2008). Genetic and morphological pat-terns show variation in frequency of hybrids between Ipomopsis (Polemo-niaceae) zones of sympatry. Heredity, 102(3):257?265.Allendorf, F. W. and Leary, R. F. (1988). Conservation and distribution ofgenetic variation in a polytypic species, the cutthroat trout. ConservationBiology, 2(2):170?184.Allendorf, F. W., Leary, R. F., Hitt, N. P., Knudsen, K. L., Boyer, M. C.,and Spruell, P. (2005). Cutthroat trout hybridization and the U. S. En-dangered Species Act: one species, two policies. Conservation Biology,19(4):1326?1328.Allendorf, F. W., Leary, R. F., Hitt, N. P., Knudsen, K. L., Lundquist, L. L.,and Spruell, P. (2004). Intercrosses and the US Endangered Species Act:should hybridized populations be included as westslope cutthroat trout?Conservation Biology, 18(5):1203?1213.Allendorf, F. W., Leary, R. F., Spruell, P., and Wenburg, J. K. (2001). Theproblems with hybrids: setting conservation guidelines. Trends in Ecology& Evolution, 16(11):613?622.Allendorf, F. W. and Waples, R. S. (1996). Conservation and genetics ofsalmonid fishes. Conservation genetics: case histories from nature, pages238?280.Baltz, D. M., Moyle, P. B., and Knight, N. J. (1982). Competitive inter-actions between benthic stream fishes, riffle sculpin, Cottus gulosus, andspeckled dace, Rhinichthys osculus. Canadian Journal of Fisheries andAquatic Sciences, 39(11):1502?1511.Barrett, R. D., Paccard, A., Healy, T. M., Bergek, S., Schulte, P. M.,Schluter, D., and Rogers, S. M. (2011). Rapid evolution of cold tolerancein stickleback. Proceedings of the Royal Society B: Biological Sciences,278(1703):233?238.63Chapter 4. ReferencesBear, E. A., McMahon, T. E., and Zale, A. V. (2007). Comparative thermalrequirements of westslope cutthroat trout and rainbow trout: implicationsfor species interactions and development of thermal protection standards.Transactions of the American Fisheries Society, 136(4):1113?1121.Becker, C. D. and Genoway, R. G. (1979). Evaluation of the critical thermalmaximum for determining thermal tolerance of freshwater fish. Environ-mental Biology of Fishes, 4(3):245?256.Behnke, R. (2002). Trout and Salmon of North America. The Free Press,New York.Behnke, R. J. (1992). Native trout of western North America. AmericanFisheries Society Monograph.Beitinger, T. L., Bennett, W. A., and McCauley, R. W. (2000). Tempera-ture tolerances of North American freshwater fishes exposed to dynamicchanges in temperature. Environmental Biology of Fishes, 58(3):237?275.Bennett, S. N., Olson, J. R., Kershner, J. L., and Corbett, P. (2010).Propagule pressure and stream characteristics influence introgression: cut-throat and rainbow trout in British Columbia. Ecological Applications,20(1):263?277.Bennett, W. A. and Beitinger, T. L. (1997). Temperature tolerance of thesheepshead minnow, Cyprinodon variegatus. Copeia, pages 77?87.Bettles, C., Docker, M., Dufour, B., and Heath, D. (2005). Hybridization dy-namics between sympatric species of trout: loss of reproductive isolation.Journal of Evolutionary Biology, 18(5):1220?1233.Boyer, M. C., Muhlfeld, C. C., and Allendorf, F. W. (2008). Rainbow trout(Oncorhynchus mykiss) invasion and the spread of hybridization with na-tive westslope cutthroat trout (Oncorhynchus clarkii lewisi). CanadianJournal of Fisheries and Aquatic Sciences, 65(4):658?669.Bozek, M. A. and Hubert, W. A. (1992). Segregation of resident trout instreams as predicted by three habitat dimensions. Canadian Journal ofZoology, 70(5):886?890.Breiman, L. (2001). Random Forest. Machine Learning, pages 5?32.Campton, D. E. and Kaeding, L. R. (2005). Westslope cutthroat trout,hybridization, and the US Endangered Species Act. Conservation Biology,19(4):1323?1325.64Chapter 4. ReferencesClarke, A., Godin, T., Yesaki, T., and Cena, C. (2008). Assessment of 2nand 3n Fraser Valley rainbow trout strains for stocking coastal oligotrophiclakes. Freshwater Fisheries Society of British Columbia. 10pp.Cleator, H., Earle, J., Fitch, L., Humphries, S., Koops, M., Martin, K.,Mayhood, D., Petry, S., Pacas, C., Stelfox, J., et al. (2009). Informationrelevant to a recovery potential assessment of pure native westslope cut-throat trout, Alberta population. Fisheries and Oceans Canada, CanadianScience Advisory Secretariat Research Document 2009036, iv+ 24 p.COSEWIC (2006). COSEWIC assessment and update status report on thewestslope cuttthroat trout Oncorhynchus clarkii lewisi (British Columbiapopulation and Alberta population) in Canada. Committee on the Statusof Endangered Wildlife in Canada, Ottawa, Ontario.Culumber, Z., Shepard, D., Coleman, S., Rosenthal, G., and Tobler, M.(2012). Physiological adaptation along environmental gradients and repli-cated hybrid zone structure in swordtails (teleostei: Xiphophorus). Jour-nal of Evolutionary Biology, 25(9):1800?1814.Currie, R. J., Bennett, W. A., and Beitinger, T. L. (1998). Critical thermalminima and maxima of three freshwater game-fish species acclimated toconstant temperatures. Environmental Biology of Fishes, 51(2):187?200.Cutler, D. R., Edwards Jr, T. C., Beard, K. H., Cutler, A., Hess, K. T.,Gibson, J., and Lawler, J. J. (2007). Random forests for classification inecology. Ecology, 88(11):2783?2792.Darveau, C. A., Taylor, E. B., and Schulte, P. M. (2012). Thermal physi-ology of warm-spring colonists: Variation among lake chub (Cyprinidae:Couesius plumbeus) populations*. Physiological and Biochemical Zoology,85(6):607?617.Dawson, M. (2002). Plant quarantine: a tool for preventing the introductionand spread of alien species harmful to plants. Alien Invaders in Canada?sWaters, Wetlands and Forests. Edited by R. Claudi, P. Nantel, and E.Muckle-Jeffs. Canadian Forest Service, Ottawa, Ont, pages 243?251.De Staso, J. and Rahel, F. J. (1994). Influence of water temperature oninteractions between juvenile Colorado River cutthroat trout and brooktrout in a laboratory stream. Transactions of the American FisheriesSociety, 123(3):289?297.65Chapter 4. ReferencesDe?ath, G. and Fabricius, K. E. (2000). Classification and regression trees:a powerful yet simple technique for ecological data analysis. Ecology,81(11):3178?3192.DeKay, J. (1842). Zoology of New York-IV: Fishes. W. and A. Albany, NewYork.: White and J. Visscher.Demarais, B. D., Dowling, T. E., Douglas, M. E., Minckley, W., and Marsh,P. C. (1992). Origin of Gila seminuda (Teleostei: Cyprinidae) throughintrogressive hybridization: implications for evolution and conservation.Proceedings of the National Academy of Sciences, 89(7):2747?2751.Dextrase, A. J. and Mandrak, N. E. (2006). Impacts of alien invasive specieson freshwater fauna at risk in Canada. Biological Invasions, 8(1):13?24.DFO (2009). Recovery potential assessment of pure native westslope cut-throat trout, Alberta population. DFO Can Sci Advis Sec Sci Advis Rep2009/050.Dunham, J. B., Adams, S. B., Schroeter, R. E., and Novinger, D. C. (2002).Alien invasions in aquatic ecosystems: toward an understanding of brooktrout invasions and potential impacts on inland cutthroat trout in westernNorth America. Reviews in Fish Biology and Fisheries, 12(4):373?391.Dunson, W. A. and Travis, J. (1991). The role of abiotic factors in commu-nity organization. American Naturalist, pages 1067?1091.Epifanio, J. and Philipp, D. (2000). Simulating the extinction of parentallineages from introgressive hybridization: the effects of fitness, initial pro-portions of parental taxa, and mate choice. Reviews in Fish Biology andFisheries, 10(3):339?354.Fangue, N. A., Hofmeister, M., and Schulte, P. M. (2006). Intraspecificvariation in thermal tolerance and heat shock protein gene expression incommon killifish, Fundulus heteroclitus. Journal of Experimental Biology,209(15):2859?2872.Fausch, K. D. (1989). Do gradient and temperature affect distributionsof and interactions between brook char (Salvelinus fontinalis) and otherresident salmonids in streams? Special vol 1, Physiology and EcologyJapan, Kyoto, pages 303?322.Fausch, K. D. (2008). A paradox of trout invasions in North America.Biological Invasions, 10(5):685?701.66Chapter 4. ReferencesFausch, K. D., Nakano, S., and Ishigaki, K. (1994). Distribution of two con-generic charrs in streams of Hokkaido Island, Japan: considering multiplefactors across scales. Oecologia, 100(1-2):1?12.Fausch, K. D., Taniguchi, Y., Nakano, S., Grossman, G. D., and Townsend,C. R. (2001). Flood disturbance regimes influence rainbow trout invasionsuccess among five holarctic regions. Ecological Applications, 11(5):1438?1455.Ferguson, M. M., Danzmann, R. G., and Allendorf, F. W. (1985). Absenceof developmental incompatibility in hybrids between rainbow trout andtwo subspecies of cutthroat trout. Biochemical Genetics, 23(7-8):557?570.Fry, F. E. J. (1947). Effects of the environment on animal activity. Biolog-ical Series 55 (Publication of the Ontario Fisheries Research Laboratory68:162) University of Toronto Studies.Gard, R. and Flittner, G. A. (1974). Distribution and abundance of fishes inSagehen Creek, California. The Journal of Wildlife Management, pages347?358.Goudet, J. (1995). FSTAT version 2.9.3.1 Updated from Goudet J (1995).Journal of Heredity, 86(6):485?486.Griffith, J. (1988). Review of competition between cutthroat trout andother salmonids. In American Fisheries Society Symposium, volume 4,pages 134?140.Gunnell, K., Tada, M. K., Hawthorne, F. A., Keeley, E. R., and Ptacek,M. B. (2008). Geographic patterns of introgressive hybridization betweennative Yellowstone cutthroat trout (Oncorhynchus clarkii bouvieri) andintroduced rainbow trout (O. mykiss) in the South Fork of the SnakeRiver watershed, Idaho. Conservation Genetics, 9(1):49?64.Halbisen, M. A. and Wilson, C. C. (2009). Variable introgression from sup-plemental stocking in Southern Ontario populations of lake trout. Trans-actions of the American Fisheries Society, 138(4):699?719.Hammer, ?., Harper, D., and Ryan, P. (2001). Past: Paleontological statis-tics software package for education and data analysis Palaeontol. Elec-tronica 4: 1?9.67Chapter 4. ReferencesHansen, M. M. and Mensberg, K.-L. D. (2009). Admixture analysis ofstocked brown trout populations using mapped microsatellite DNA mark-ers: indigenous trout persist in introgressed populations. Biology Letters,5(5):656?659.Heath, D., Bettles, C. M., and Roff, D. (2010). Environmental factors associ-ated with reproductive barrier breakdown in sympatric trout populationson Vancouver Island. Evolutionary Applications, 3(1):77?90.Henderson, R., Kershner, J. L., and Toline, C. A. (2000). Timing andlocation of spawning by nonnative wild rainbow trout and native cut-throat trout in the South Fork Snake River, Idaho, with implicationsfor hybridization. North American Journal of Fisheries Management,20(3):584?596.Hether, T. and Hoffman, E. (2012). Machine learning identifies specifichabitats associated with genetic connectivity in Hyla squirella. Journal ofEvolutionary Biology, 25(6):1039?1052.Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G., and Jarvis, A.(2005). Very high resolution interpolated climate surfaces for global landareas. International Journal of Climatology, 25(15):1965?1978.Hitt, N. P., Frissell, C. A., Muhlfeld, C. C., and Allendorf, F. W. (2003).Spread of hybridization between native westslope cutthroat trout, On-corhynchus clarki lewisi, and nonnative rainbow trout, Oncorhynchusmykiss. Canadian Journal of Fisheries and Aquatic Sciences, 60(12):1440?1451.Hohenlohe, P. A., Amish, S. J., Catchen, J. M., Allendorf, F. W., andLuikart, G. (2011). Next-generation RAD sequencing identifies thousandsof SNPs for assessing hybridization between rainbow and westslope cut-throat trout. Molecular Ecology Resources, 11(s1):117?122.Holderegger, R. and Wagner, H. H. (2006). A brief guide to landscapegenetics. Landscape Ecology, 21(6):793?796.Horton, R. E. (1945). Erosional development of streams and their drainagebasins: hydro-physical approach to quantitative morphology. GeologicalSociety of America Bulletin, 56:275?370.Hubbs, C. L. (1955). Hybridization between fish species in nature. System-atic Zoology, 4(1):1?20.68Chapter 4. ReferencesHubert, W. A., Stonecypher Jr, R. W., Gern, W. A., and Bobbit, J. (1994).Communications: Response of cutthroat trout embryos to reduced incu-bation temperatures at different developmental stages. The ProgressiveFish-Culturist, 56(3):185?187.Huxel, G. R. (1999). Rapid displacement of native species by invasivespecies: effects of hybridization. Biological Conservation, 89(2):143?152.Isaak, D. J., Muhlfeld, C. C., Todd, A. S., Al-Chokhachy, R., Roberts, J.,Kershner, J. L., Fausch, K. D., and Hostetler, S. W. (2012). The pastas prelude to the future for understanding 21st-century climate effects onRocky Mountain trout. Fisheries, 37(12):542?556.Keller, I. and Seehausen, O. (2012). Thermal adaptation and ecologicalspeciation. Molecular ecology, 21(4):782?799.Krueger, C. C. and Menzel, B. W. (1979). Effects of stocking on geneticsof wild brook trout populations. Transactions of the American FisheriesSociety, 108(3):277?287.Largiade`r, C. and Scholl, A. (1995). Effects of stocking on the genetic diver-sity of brown trout populations of the Adriatic and Danubian drainagesin Switzerland. Journal of Fish Biology, 47(sA):209?225.Larson, G. L. and Moore, S. E. (1985). Encroachment of exotic rainbow troutinto stream populations of native brook trout in the southern AppalachianMountains. Transactions of the American Fisheries Society, 114(2):195?203.Leary, R. F., Allendorf, F. W., and Knudsen, K. L. (1984). Introgressionbetween westslope cutthroat trout and rainbow trout in Clark Fork Riverdrainage, Montana. Proc. Mont. Acad. Sci, 43:1?18.Levin, D. (2002). Hybridization and extinction in protecting rare species,conservationists should consider the dangers of interbreeding, which com-pound the more well-known threats to wildlife. American Scientist,90(3):254?261.Liaw, A. and Wiener, M. (2002). Classification and Regression by random-Forest. R news, 2(3):18?22.Lockwood, J. L., Cassey, P., and Blackburn, T. M. (2009). The more you in-troduce the more you get: the role of colonization pressure and propagulepressure in invasion ecology. Diversity and Distributions, 15(5):904?910.69Chapter 4. ReferencesLoxterman, J. L. and Keeley, E. R. (2012). Watershed boundaries andgeographic isolation: patterns of diversification in cutthroat trout fromwestern North America. BMC Evolutionary Biology, 12(1):38.MacCrimmon, H. R. (1971). World distribution of rainbow trout (Salmogairdneri). Journal of the Fisheries Board of Canada, 28(5):663?704.Manel, S., Schwartz, M. K., Luikart, G., and Taberlet, P. (2003). Landscapegenetics: combining landscape ecology and population genetics. Trendsin Ecology & Evolution, 18(4):189?197.Marie, A., Bernatchez, L., and Garant, D. (2010). Loss of genetic integritycorrelates with stocking intensity in brook charr (Salvelinus fontinalis).Molecular Ecology, 19(10):2025?2037.Marie, A. D., Bernatchez, L., Garant, D., and Taylor, E. (2012). En-vironmental factors correlate with hybridization in stocked brook charr(Salvelinus fontinalis). Canadian Journal of Fisheries and Aquatic Sci-ences, 69(5):884?893.Mayhood, D. W. (1999). Provisional evaluation of the status of westslopecutthroat trout in Canada. Proceedings of the Biological and Managementof Species and Habitats at Risk, Kamloops, BC.Mayhood, D. W. and Taylor, E. B. (2011). Contributions to a recoveryplan for westslope cutthroat trout (Oncorhynchus clarkii lewisi ) in Al-berta: distribution, population size and trends, report prepared by FWRFreshwater Research Limited, Calgary, AB, for Alberta Fish and WildlifeDivision, Cochrane, AB.McHugh, P. and Budy, P. (2005). An experimental evaluation of compet-itive and thermal effects on brown trout (Salmo trutta) and bonnevillecutthroat trout (Oncorhynchus clarkii utah) performance along an alti-tudinal gradient. Canadian Journal of Fisheries and Aquatic Sciences,62(12):2784?2795.McKay, S. J., Devlin, R. H., and Smith, M. J. (1996). Phylogeny of Pacificsalmon and trout based on growth hormone type-2 and mitochondrialNADH dehydrogenase subunit 3 DNA sequences. Canadian Journal ofFisheries and Aquatic Sciences, 53(5):1165?1176.McMahon, T. E., Zale, A. V., Barrows, F. T., Selong, J. H., and Danehy,R. J. (2007). Temperature and competition between bull trout and brook70Chapter 4. Referencestrout: a test of the elevation refuge hypothesis. Transactions of the Amer-ican Fisheries Society, 136(5):1313?1326.Miller, R. R., Williams, J. D., and Williams, J. E. (1989). Extinctions ofNorth American fishes during the past century. Fisheries, 14(6):22?38.Moisen, G. (2008). Classification and regression trees. pages 582?588.Moring, J. R. (1993). Records of long-range, downstream movementsof stocked rainbow trout (Oncorhynchus mykiss). Fisheries Research,16(2):195?199.Morris, D. B., Richard, K. R., and Wright, J. M. (1996). Microsatel-lites from rainbow trout (Oncorhynchus mykiss) and their use for geneticstudy of salmonids. Canadian Journal of Fisheries and Aquatic Sciences,53(1):120?126.Muhlfeld, C. C., Kalinowski, S. T., McMahon, T. E., Taper, M. L., Painter,S., Leary, R. F., and Allendorf, F. W. (2009a). Hybridization rapidlyreduces fitness of a native trout in the wild. Biology Letters, 5(3):328?331.Muhlfeld, C. C., McMahon, T. E., Belcer, D., and Kershner, J. L.(2009b). Spatial and temporal spawning dynamics of native westslopecutthroat trout, Oncorhynchus clarkii lewisi, introduced rainbow trout,Oncorhynchus mykiss, and their hybrids. Canadian Journal of Fisheriesand Aquatic Sciences, 66(7):1153?1168.Muhlfeld, C. C., McMahon, T. E., Boyer, M. C., and Gresswell, R. E.(2009c). Local habitat, watershed, and biotic factors influencing thespread of hybridization between native westslope cutthroat trout and in-troduced rainbow trout. Transactions of the American Fisheries Society,138(5):1036?1051.Murphy, M. A., Evans, J. S., and Storfer, A. (2010). Quantifying Bufoboreas connectivity in yellowstone national park with landscape genetics.Ecology, 91(1):252?261.Narum, S. R. (2006). Beyond Bonferroni: less conservative analyses forconservation genetics. Conservation Genetics, 7(5):783?787.Nelson, J. S. and Paetz, M. J. (1992). The fishes of Alberta. University ofAlberta.71Chapter 4. ReferencesNelson, R. and Beacham, T. (1999). Isolation and cross species amplificationof microsatellite loci useful for study of Pacific salmon. Animal Genetics,30(3):228?229.Nielsen, E. E., Bach, L. A., and Kotlicki, P. (2006). Hybridlab (version1.0): a program for generating simulated hybrids from population samples.Molecular Ecology Notes, 6(4):971?973.Northrup, S. and Godin, T. I. (2009). A review of rainbow trout strainperformance: historical and contemporary data from British ColumbiaLakes. Freshwater Fisheries Society of British Columbia.Olsen, J., Wenburg, J., and Bentzen, P. (1996). Semiautomated multilocusgenotyping of Pacific salmon (Oncorhnychus spp.) using microsatellites.Molecular Marine Biology and Biotechnology, 5(4):259.O?Reilly, P. T., Hamilton, L. C., McConnell, S. K., and Wright, J. M. (1996).Rapid analysis of genetic variation in Atlantic salmon (Salmo salar)by PCR multiplexing of dinucleotide and tetranucleotide microsatellites.Canadian Journal of Fisheries and Aquatic Sciences, 53(10):2292?2298.Ostberg, C. and Rodriguez, R. (2002). Novel molecular markers differentiateOncorhynchus mykiss (rainbow trout and steelhead) and the O. clarki(cutthroat trout) subspecies. Molecular Ecology Notes, 2(3):197?202.Paul, A. J. and Post, J. R. (2001). Spatial distribution of native and non-native salmonids in streams of the eastern slopes of the Canadian RockyMountains. Transactions of the American Fisheries Society, 130(3):417?430.Pimentel, D., Zuniga, R., and Morrison, D. (2005). Update on the environ-mental and economic costs associated with alien-invasive species in theUnited States. Ecological Economics, 52(3):273 ? 288.Pritchard, J. K., Stephens, M., and Donnelly, P. (2000). Inference of popula-tion structure using multilocus genotype data. Genetics, 155(2):945?959.R Development Team (2012). R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna.Rahel, F. J. (1984). Factors structuring fish assemblages along a bog lakesuccessional gradient. Ecology, pages 1276?1289.72Chapter 4. ReferencesRahel, F. J. and Hubert, W. A. (1991). Fish assemblages and habitat gra-dients in a Rocky Mountain?Great Plains stream: biotic zonation andadditive patterns of community change. Transactions of the AmericanFisheries Society, 120(3):319?332.Rasmussen, J. B., Robinson, M. D., and Heath, D. D. (2010). Ecologicalconsequences of hybridization between native westslope cutthroat (On-corhynchus clarkii lewisi) and introduced rainbow (Oncorhynchus mykiss)trout: effects on life history and habitat use. Canadian Journal of Fish-eries and Aquatic Sciences, 67(2):357?370.Rasmussen, J. B., Robinson, M. D., Hontela, A., and Heath, D. D. (2012).Metabolic traits of westslope cutthroat trout, introduced rainbow troutand their hybrids in an ecotonal hybrid zone along an elevation gradient.Biological Journal of the Linnean Society, 105(1):56?72.Reese, C. D. and Harvey, B. C. (2002). Temperature-dependent interac-tions between juvenile steelhead and sacramento pikeminnow in labora-tory streams. Transactions of the American Fisheries Society, 131(4):599?606.Rhymer, J. M. and Simberloff, D. (1996). Extinction by hybridization andintrogression. Annual Review of Ecology and Systematics, pages 83?109.Rousset, F. (2008). GENEPOP007: a complete re-implementation of thegenepop software for Windows and Linux. Molecular Ecology Resources,8(1):103?106.Rubidge, E., Corbett, P., and Taylor, E. B. (2001). A molecular analysisof hybridization between native westslope cutthroat trout and introducedrainbow trout in southeastern British Columbia, Canada. Journal of FishBiology, 59(sA):42?54.Rubidge, E. and Taylor, E. B. (2005). An analysis of spatial and envi-ronmental factors influencing hybridization between native westslope cut-throat trout (Oncorhynchus clarki lewisi) and introduced rainbow trout(O. mykiss) in the upper Kootenay River drainage, British Columbia.Conservation Genetics, 6(3):369?384.Ruzzante, D. E., Hansen, M. M., Meldrup, D., and Ebert, K. M. (2004).Stocking impact and migration pattern in an anadromous brown trout(Salmo trutta) complex: where have all the stocked spawning sea troutgone? Molecular Ecology, 13(6):1433?1445.73Chapter 4. ReferencesSchmetterling, D. A. (2001). Seasonal movements of fluvial westslope cut-throat trout in the Blackfoot River drainage, Montana. North AmericanJournal of Fisheries Management, 21(3):507?520.Selong, J. H., McMahon, T. E., Zale, A. V., and Barrows, F. T. (2001).Effect of temperature on growth and survival of bull trout, with applica-tion of an improved method for determining thermal tolerance in fishes.Transactions of the American Fisheries Society, 130(6):1026?1037.Shepard, B. B., May, B. E., and Urie, W. (2005). Status and conservationof westslope cutthroat trout within the western United States. NorthAmerican Journal of Fisheries Management, 25(4):1426?1440.Sih, A. (1987). Predators and prey lifestyles: an evolutionary and ecologicaloverview. Predation: direct and indirect impacts on aquatic communities,pages 203?224.Simberloff, D. (2005). Non-native species do threaten the natural environ-ment! Journal of Agricultural and Environmental Ethics, 18(6):595?607.Slettan, A., Olsaker, I., and Lie, ?. (1995). Atlantic salmon, Salmo salar,microsatellites at the SSOSL25, SSOSL85, SSOSL311, SSOSL417 loci.Animal genetics, 26(4):281?282.Sloat, M. R., Shepard, B. B., White, R. G., and Carson, S. (2005). Influenceof stream temperature on the spatial distribution of westslope cutthroattrout growth potential within the Madison River basin, Montana. NorthAmerican Journal of Fisheries Management, 25(1):225?237.Smith, G., Dowling, T., Gobalet, K., Lugaski, T., Shiozawa, D., and Evans,R. (2002). Biogeography and timing of evolutionary events among GreatBasin fishes. Great Basin aquatic systems history, pages 175?234.Stonecypher, R. W., Hubert, W. A., and Gern, W. A. (1994). Effect ofreduced incubation temperatures on survival of trout embryos. The Pro-gressive Fish-Culturist, 56(3):180?184.Strahler, A. N. (1952). Hypsometric (area-altitude) analysis of erosionaltopology. Geological Society of America Bulletin, 63:1117?1142.Tamkee, P., Parkinson, E., and Taylor, E. B. (2010). The influence of Wis-consinan glaciation and contemporary stream hydrology on microsatelliteDNA variation in rainbow trout (Oncorhynchus mykiss). Canadian Jour-nal of Fisheries and Aquatic Sciences, 67(6):919?935.74Chapter 4. ReferencesTaniguchi, Y., Rahel, F. J., Novinger, D. C., and Gerow, K. G. (1998). Tem-perature mediation of competitive interactions among three fish speciesthat replace each other along longitudinal stream gradients. CanadianJournal of Fisheries and Aquatic Sciences, 55(8):1894?1901.Taylor, E. B. (2004). Evolution in mixed company: evolutionary inferencesfrom studies of natural hybridization in salmonidae. Evolution Illumi-nated: salmon and their relatives, pages 232?263.Taylor, E. B., Stamford, M., and Baxter, J. (2003). Population subdivisionin westslope cutthroat trout (Oncorhynchus clarki lewisi) at the northernperiphery of its range: evolutionary inferences and conservation implica-tions. Molecular ecology, 12(10):2609?2622.Taylor, E. B., Tamkee, P., Sterling, G., and Hughson, W. (2007). Mi-crosatellite DNA analysis of rainbow trout (Oncorhynchus mykiss) fromwestern Alberta, Canada: native status and evolutionary distinctivenessof Athabasca rainbow trout. Conservation Genetics, 8(1):1?15.The Alberta Westslope Cutthroat Trout Recovery Team (2013). AlbertaWestslope Cutthroat Trout Recovery Plan: 2012-2017. Alberta Environment and Sustainable Resource Development, Alberta Species at Risk Re-covery Plan No. 28. Edmonton, AB. 77 pp.Trotter, P. (2008). Cutthroat: Native Trout of the West, 2nd edn. Universityof California Press.US Fish and Wildlife Service. (1999). Status review for westslope cutthroattrout in the United States. U.S. Fish and Wildlife Service, Regions 1 and6 Oregon, and Denver: Portland.Van Houdt, J., Pinceel, J., Flamand, M.-C., Briquet, M., Dupont, E., Vol-ckaert, F., and Baret, P. (2005). Migration barriers protect indigenousbrown trout (Salmo trutta) populations from introgression with stockedhatchery fish. Conservation Genetics, 6(2):175?191.van Oosterhout, C., Hutchinson, W. F., Wills, D. P., and Shipley, P. (2004).Micro-checker: software for identifying and correcting genotyping errorsin microsatellite data. Molecular Ecology Notes, 4(3):535?538.Vannote, R. L., Minshall, G. W., Cummins, K. W., Sedell, J. R., and Cush-ing, C. E. (1980). The river continuum concept. Canadian Journal ofFisheries and Aquatic Sciences, 37(1):130?137.75Chapter 4. ReferencesVenter, O., Brodeur, N. N., Nemiroff, L., Belland, B., Dolinsek, I. J., andGrant, J. W. (2006). Threats to endangered species in Canada. Bioscience,56(11):903?910.Verspoor, E. and Hammart, J. (1991). Introgressive hybridization in fishes:the biochemical evidence. Journal of Fish Biology, 39(sA):309?334.Vonlanthen, P., Bittner, D., Hudson, A., Young, K., Mu?ller, R., Lundsgaard-Hansen, B., Roy, D., Di Piazza, S., Largiader, C., and Seehausen, O.(2012). Eutrophication causes speciation reversal in whitefish adaptiveradiations. Nature, 482(7385):357?362.Warner, S. C., Travis, J., and Dunson, W. A. (1993). Effect of ph variation ofinterspecific competition between two species of Hylid tadpoles. Ecology,pages 183?194.Webb, B. W., Hannah, D. M., Moore, R. D., Brown, L. E., and Nobilis,F. (2008). Recent advances in stream and river temperature research.Hydrological Processes, 22(7):902?918.Weigel, D. E., Peterson, J. T., and Spruell, P. (2003). Introgressive hy-bridization between native cutthroat trout and introduced rainbow trout.Ecological Applications, 13(1):38?50.Weir, B. S. and Cockerham, C. C. (1984). Estimating F-statistics for theanalysis of population structure. Evolution, pages 1358?1370.Welcomme, R. (1992). A history of international introductions of inlandaquatic species. In ICES Marine Science Symposia. Copenhagen[ICESMAR. SCI. SYMP.]. 1992.Widmer, A. and Baltisberger, M. (1999). Molecular evidence for allopoly-ploid speciation and a single origin of the narrow endemic Draba ladina(brassicaceae). American Journal of Botany, 86(9):1282?1289.Wilcove, D. S., Rothstein, D., Dubow, J., Phillips, A., and Losos, E. (1998).Quantifying threats to imperiled species in the United States. BioScience,48(8):607?615.Willi, Y., Van Kleunen, M., Dietrich, S., and Fischer, M. (2007). Geneticrescue persists beyond first-generation outbreeding in small populationsof a rare plant. Proceedings of the Royal Society B: Biological Sciences,274(1623):2357?2364.76Chapter 4. ReferencesXiao, J., Li, J., Yuan, L., and Tanksley, S. D. (1995). Dominance is themajor genetic basis of heterosis in rice as revealed by QTL analysis usingmolecular markers. Genetics, 140(2):745?754.77Appendix AChapter 2Table A.1: Summary of genetic data for sampling locationsin southwestern Alberta and southeastern British Columbia,Canada. Abbreviations: N, number of individuals sampled;Mean qwsct: average admixture value; FIS : inbreeding coeffi-cient; 95% CI: 95% confidence interval of FIS ; HO: observedheterozygosity; HE : expected heterozygosity; NA: number ofalleles.MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Altrude Altrude(middle)AM 32 28 51.2335 -116.0421 2007,20080.988 0.003 (-0.178-0.142)0.22 0.22 2.0Bow Altrude Arnica Lake AR 33 28 51.2219 -115.9984 2008 0.993 -0.025 (-0.209-0.123)0.20 0.20 2.1Bow Altrude Boom Cr,lowerPC-LBC 31 30 51.2494 -116.0213 2006 0.929 0.253 ( 0.043-0.393)0.14 0.18 2.6Bow Altrude Boom Cr,upperPC-UBC 29 21 51.2597 -116.0718 2006 0.909 0.094 (-0.109-0.224)0.20 0.22 3.2Bow Altrude Lower TwinLakeLTwin 35 31 51.2018 -115.9813 2008 0.991 0.094 (-0.062-0.221)0.19 0.21 3.1Bow Altrude Smith Lake SML 30 52 51.2493 -115.9253 2007 0.991 0.133 ( 0.032-0.217)0.24 0.28 2.678AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Altrude Upper TwinLakeUT 34 30 51.2125 -115.9843 2008 0.985 -0.062 (-0.206-0.065)0.20 0.19 2.2Bow Bow mainstem Bow RiverupperBR 4 30 51.6509 -116.3839 2006 0.988 0.361 ( 0.210-0.477)0.22 0.30 3.3Bow Cascade CutheadCreekCC 13 30 51.4404 -115.6958 2008 0.993 0.026 (-0.112-0.114)0.23 0.24 2.4Bow Cascade Elk Lake PC-ElL 25 24 51.2883 -115.6509 2006 0.989 -0.020 (-0.155-0.078)0.37 0.36 2.6Bow Cascade SawbackLakeSBL 18 31 51.3498 -115.7675 2008 0.993 0.960 ( 0.842-1.000)0.01 0.08 1.3Bow Cascade River Block Lake BLK 17 29 51.3724 -115.8339 2008 0.995 -0.003 (-0.168-0.151)0.21 0.21 2.2Bow Corral Corral Creek CC 11 56 51.4487 -116.1162 2009 0.977 0.101 ( 0.032-0.162)0.23 0.24 2.4Bow Corral Corral Creek PC-CC 12 25 51.4487 -116.1162 2006 0.946 0.135 (-0.029-0.265)0.36 0.42 3.3Bow Elbow CanyonCreekAFW-CC57 30 50.8952 -114.7772 2006 0.974 0.004 (-0.132-0.097)0.38 0.39 3.7Bow Elbow CanyonCreekJ-E15 58 33 50.8955 -114.7753 2008 0.941 -0.052 (-0.159-0.023)0.40 0.38 4.2Bow Elbow Prairie Creek AFW-PC60 27 50.8799 -114.8794 2006 0.985 -0.068 (-0.385-0.044)0.25 0.23 1.5Bow Elbow Quirk Creek AFW-QC71 28 50.8073 -114.7544 2006 0.892 0.269 (-0.003-0.402)0.24 0.32 2.8Bow Elbow RangerCreek(higher up)J-E10a 55 26 50.9106 -114.7070 2007 0.650 -0.072 (-0.196-0.002)0.68 0.64 4.7Bow Elbow SilvesterCreekAFW-SiC62 25 50.8648 -114.7238 2006 0.992 0.216 ( 0.082-0.303)0.24 0.30 2.1Bow Elbow Tributary toCanyonCreekJ-E15b 56 37 50.9139 -114.9164 2008 0.975 -0.010 (-0.133-0.085)0.36 0.36 3.4Bow Fish Fish Creek AFW-FC64 21 50.8509 -114.6192 2006 0.912 0.070 (-0.100-0.166)0.43 0.46 4.579AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Forty MileCreekMystic Lake OCMY 28 29 51.2783 -115.7498 1999 0.992 -0.136 (-0.318--0.001)0.32 0.28 2.2Bow Ghost Johnson Cr AFW-JC15 17 51.3904 -115.0910 2006 0.992 0.194 (-0.189-0.338)0.27 0.32 2.6Bow Ghost MargaretCreekAFW-MaC14 30 51.4026 -115.0711 2006 0.970 0.151 (-0.117-0.290)0.26 0.30 2.8Bow Ghost MeadowCreekAFW-MeC19 27 51.3388 -115.0836 2006 0.970 0.173 (-0.074-0.310)0.22 0.26 3.0Bow Ghost WaiparousCreekAFW-WC23 29 51.2869 -114.8395 2006 0.985 0.069 (-0.053-0.147)0.29 0.31 3.1Bow Healy Healy Lake#1HL1 44 21 51.0829 -115.8555 2008 0.981 0.002 (-0.194-0.107)0.31 0.31 2.4Bow Healy Healy Lake#2HL2 45 30 51.0853 -115.8592 2008 0.992 -0.092 (-0.251-0.035)0.20 0.19 2.2Bow Healy Healy Lake#3HL3 42 25 51.0911 -115.8627 2008 0.982 -0.028 (-0.208-0.093)0.28 0.28 2.5Bow Helen KatherineLakePC-KL 1 26 51.6857 -116.3914 2006 0.075 0.252 ( 0.082-0.341)0.28 0.37 3.2Bow Highwood Bear Creek J-P16 104 30 50.3483 -114.4595 2007 0.117 0.033 (-0.039-0.065)0.74 0.77 8.5Bow Highwood Bear Creek J-P16a 106 27 50.3393 -114.4926 2008 0.787 0.149 (-0.031-0.247)0.39 0.46 5.2Bow Highwood CutthroatCreekAFW-CuC92 26 50.4763 -114.4912 2006 0.990 0.081 (-0.117-0.215)0.34 0.36 2.9Bow Highwood Deep Creek J-H11 120 29 50.4275 -114.4829 2007 0.993 0.017 (-0.131-0.121)0.31 0.38 3.3Bow Highwood EtheringtonCreek abovebarrierJ-H24b 107 30 50.3364 -114.6246 2007 0.980 -0.105 (-0.254-0.044)0.22 0.20 2.0Bow Highwood EtheringtonCreek belowbarrierJ-H24a 108 30 50.3393 -114.6247 2007 0.988 -0.116 (-0.313-0.073)0.22 0.19 1.9Bow Highwood Flat Creek AFW-FlC93 30 50.4654 -114.5587 2006 0.975 0.214 ( 0.065-0.320)0.30 0.38 3.380AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Highwood GreenfeedCreek (southfork)J-P17a 109 29 50.3229 -114.4718 2007 0.952 0.058 (-0.085-0.175)0.25 0.27 2.7Bow Highwood Hay Creek J-St14f 112 30 50.3048 -114.3374 2007 0.454 0.004 (-0.086-0.058)0.70 0.71 6.3Bow Highwood J-H11a DeepCreektributaryJ-H11a 96 30 50.4362 -114.5061 2007 0.992 -0.002 (-0.123-0.080)0.25 0.25 2.0Bow Highwood J-H4hSullivanCreekJ-H4h 86 29 50.5504 -114.5708 2007 0.582 -0.007 (-0.097-0.048)0.70 0.69 7.9Bow Highwood J-H4iSullivanCreekJ-H4i 87 30 50.5375 -114.5913 2007 0.720 0.061 (-0.054-0.145)0.58 0.61 7.0Bow Highwood J-H7aUnnamedtrib to FlatCreekJ-H7a 95 30 50.4483 -114.5012 2007 0.994 -0.168 (-0.305--0.071)0.32 0.28 1.6Bow Highwood LoomisCreek(lower)J-H33 94 22 50.4632 -114.8128 2008 0.968 0.040 (-0.076-0.107)0.40 0.42 3.9Bow Highwood MarstonCreekJ-H12 100 30 50.4124 -114.5202 2007 0.677 0.083 (-0.048-0.182)0.52 0.57 4.2Bow Highwood McPhailCreek(lower)J-H30 101 30 50.4162 -114.7490 2007 0.429 0.011 (-0.096-0.085)0.73 0.74 6.6Bow Highwood McPhailCreek(upper)J-H30a 98 30 50.4298 -114.8044 2007 0.706 0.041 (-0.070-0.106)0.57 0.61 6.1Bow Highwood Muir Creek J-H30b 99 25 50.4235 -114.7891 2007 0.897 -0.131 (-0.240--0.05)0.55 0.60 3.1Bow Highwood NorthSullivanCreek ( 1.5km up)J-H4c 88 28 50.5266 -114.4345 2007 0.041 0.156 ( 0.061-0.207)0.62 0.73 7.481AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Highwood PekiskoCreek(headwaters)J-P21 114 30 50.2856 -114.4657 2007 0.962 0.041 (-0.123-0.167)0.24 0.25 2.6Bow Highwood PekiskoCreek abovefallsAFW-PeC105 27 50.3440 -114.4551 2006 0.947 -0.072 (-0.204-0.052)0.29 0.28 2.0Bow Highwood Picklejar #4 AFW-PjL#490 29 50.5188 -114.7831 2006 0.991 0.101 (-0.138-0.264)0.20 0.23 1.6Bow Highwood Picklejar L#2AFW-PjL#290 26 50.5182 -114.7759 2006 0.990 0.144 (-0.170-0.403)0.11 0.12 1.4Bow Highwood Salt Creek J-P20 113 30 50.3041 -114.4611 2007 0.959 0.005 (-0.160-0.140)0.22 0.22 2.6Bow Highwood SheppardCreek (southfork)J-St14d 110 30 50.3206 -114.3967 2008 0.932 -0.071 (-0.183-0.004)0.39 0.37 3.3Bow Highwood SouthSullivanCreek (6 kmup)J-H4f 91 30 50.5004 -114.5535 2007 0.731 0.038 (-0.094-0.114)0.53 0.55 6.2Bow Highwood SouthSullivanCreek trib (5km up)J-H4g 89 30 50.5162 -114.5659 2007 0.769 -0.048 (-0.157-0.033)0.61 0.58 5.2Bow Highwood Unnamed toPekisko CrJ-P15 103 21 50.3570 -114.4592 2007 0.080 0.073 (-0.016-0.119)0.63 0.68 7.0Bow Highwood Unnamed toupperPekiskoJ-P19 111 28 50.3098 -114.4468 2007 0.949 0.021 (-0.119-0.121)0.23 0.23 2.4Bow Highwood ZephyrCreekJ-H18 102 30 50.3851 -114.5757 2007 0.986 0.050 (-0.072-0.154)0.33 0.35 2.4Bow Johnston JohnstonCreekJC 27 17 51.2827 -115.8212 2006 0.293 0.192 (-0.011-0.347)0.22 0.27 2.982AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Johnston Luellen LakeU LavalClipsOCLL 20 17 51.3402 -115.9169 1999 0.436 -0.010 (-0.245-0.066)0.46 0.46 3.8Bow Jumpingpound Coxhill Crbelow fallsJ-J11 50 35 51.0072 -114.8186 2007 0.900 -0.025 (-0.113-0.034)0.44 0.43 4.8Bow Jumpingpound JumpingpoundCreekAFW-JuC54 27 50.9666 -114.9578 2006 0.973 -0.101 (-0.217--0.029)0.38 0.34 2.9Bow Jumpingpound Pine Cr.middleJ-J7 46 17 51.0184 -114.7674 2007 0.962 0.058 (-0.114-0.171)0.28 0.30 3.6Bow Jumpingpound Unnamed toCoxhill CrJ-J11c 49 33 51.0085 -114.8257 2007 0.901 0.024 (-0.101-0.114)0.38 0.39 3.3Bow Jumpingpound Unnamed toUp Jumping-poundJ-J19 51 30 50.9890 -114.9547 2007 0.965 -0.086 (-0.212-0.020)0.32 0.29 3.1Bow Jumpingpound Unnamed toUp Jumping-poundJ-J20 52 21 50.9757 -114.9537 2007 0.941 -0.046 (-0.180-0.028)0.40 0.38 2.8Bow Kananaskis Evan-ThomasCreekJ-EVTH 61 34 50.8820 -115.1219 2009 0.993 -0.092 (-0.243-0.028)0.20 0.18 2.0Bow Kananaskis Spotted WolfCreekAFW-SWC81 25 50.8424 -115.3445 2006 0.889 0.188 ( 0.047-0.296)0.22 0.27 3.1Bow Kananaskis Spotted WolfCreekB-K1 83 15 50.6611 -115.0973 2008 0.939 -0.089 (-0.260-0.012)0.30 0.28 2.7Bow Moraine Babel Creek-ConsolationLakeBCC 21 27 51.3099 -116.1491 2008 0.989 0.043 (-0.143-0.172)0.31 0.33 2.8Bow Mosquito Mosquito Cr PC-MC 3 30 51.6592 -116.3176 2006 0.981 0.204 ( 0.006-0.318)0.25 0.31 3.0Bow Outlet Outlet Creek OC 16 29 51.4001 -116.1225 2008 0.994 0.001 (-0.240-0.239)0.07 0.07 1.3Bow Pipestone Big FishLakeBFL 5 62 51.6423 -116.1985 2009 0.993 0.487 ( 0.407-0.551)0.19 0.36 3.083AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Pipestone Deer Lake DL 6 47 51.6271 -116.1625 2009 0.994 0.043 (-0.131-0.187)0.11 0.12 1.6Bow Pipestone Little FishLakeLFL 9 47 51.6436 -116.1809 2009 0.994 0.068 (-0.094-0.225)0.12 0.13 1.7Bow Pipestone Moose Lake ML 2 49 51.6634 -116.2010 2009 0.994 0.097 (-0.043-0.181)0.11 0.13 1.6Bow Pipestone Pipestone R PC-PR 8 20 51.6103 -116.1321 2006 0.959 0.302 ( 0.115-0.411)0.23 0.31 2.3Bow Redearth Black RockLPC-BRL 36 27 51.1210 -115.9152 2006 0.959 0.260 ( 0.043-0.415)0.18 0.22 2.3Bow Redearth Egypt L PC-EgL 41 27 51.0989 -115.9166 2006 0.956 0.128 (-0.030-0.252)0.26 0.30 2.9Bow Redearth Haiduk L PC-HL 37 25 51.1201 -115.9431 2006 0.432 0.168 ( 0.039-0.232)0.47 0.56 4.4Bow Redearth Mummy L PC-ML 43 21 51.0892 -115.9139 2006 0.017 0.088 (-0.107-0.249)0.38 0.42 2.9Bow Redearth Pharaoh Cr PC-PC 38 30 51.1196 -115.8982 2006 0.969 -0.022 (-0.174-0.100)0.28 0.28 2.4Bow Redearth Pharaoh L PC-PL 39 27 51.1143 -115.9110 2006 0.985 -0.004 (-0.215-0.170)0.25 0.25 2.1Bow Redearth Scarab L PC-SL 40 28 51.1005 -115.9039 2006 0.590 0.137 ( 0.015-0.210)0.44 0.50 3.8Bow Sheep Coal Creek(upperreach)J-S12a 85 30 50.5682 -114.5825 2007 0.977 0.013 (-0.153-0.130)0.24 0.24 2.7Bow Sheep Death ValleyCreekJ-T11d 79 29 50.7111 -114.5236 2007 0.010 0.099 (-0.013-0.173)0.62 0.69 7.6Bow Sheep Fisher Creek AFW-FeC69 25 50.8081 -114.6199 2006 0.775 0.228 ( 0.099-0.316)0.40 0.51 4.8Bow Sheep Gorge Creek( 10 kmabove falls)J-S17b 82 29 50.6595 -114.7462 2007 0.991 -0.088 (-0.254-0.057)0.20 0.18 1.784AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NABow Sheep Gorge Creek(3 km abovefalls)J-S17a 80 30 50.6646 -114.6860 2007 0.954 0.009 (-0.165-0.184)0.19 0.19 2.2Bow Sheep MuskegCreek(upper)J-T20a 73 25 50.7762 -114.6859 2007 0.068 0.074 (-0.014-0.128)0.69 0.74 7.4Bow Sheep North CoalCreekJ-S12b 84 31 50.6125 -114.5835 2008 0.929 0.197 ( 0.081-0.274)0.37 0.45 4.8Bow Sheep Ware Creek(headwaters)J-T11g 78 30 50.7138 -114.6471 2007 0.715 0.032 (-0.080-0.109)0.60 0.62 6.4Bow Spray Gloria L PC-GL 66 28 50.8645 -115.6049 2006 0.989 0.638 ( 0.181-0.914)0.02 0.04 1.3Bow Spray Marvel LakeU LavalClipsMLULaval,OCM63 34 50.8777 -115.5569 1999 0.993 0.088 (-0.107-0.249)0.12 0.12 1.9Bow Spray Smuts Cr AFW-SC68 25 50.8424 -115.3445 2006 0.897 0.201 ( 0.027-0.308)0.23 0.29 2.9Bow Spray Terrapin L PC-TeL 65 30 50.8648 -115.5918 2006 0.983 0.505 (-0.129-0.778)0.03 0.07 1.5Bow Spray Upper SprayRiverPC-USP 75 31 50.7412 -115.3934 2006 0.992 -0.104 (-0.252-0.013)0.19 0.17 2.0Bow Spray WatridgeCreekAFW-WatC67 17 50.8486 -115.4201 2006 0.966 0.194 (-0.167-0.376)0.15 0.19 2.3Bow Taylor O?Brien L PC-O?B 26 28 51.2866 -116.0829 2006 0.937 0.204 ( 0.091-0.276)0.34 0.42 3.9Bow Taylor Taylor Creek TC 22 26 51.3076 -116.0250 2008 0.910 0.222 ( 0.080-0.312)0.31 0.39 4.2Bow Taylor Taylor L PC-TaL 24 20 51.2958 -116.0957 2006 0.018 0.032 (-0.130-0.103)0.38 0.39 3.5Columbia Kootenay 2ndHoneymoonCreek2HC 47 29 51.0315 -115.9903 2008 0.981 -0.079 (-0.216-0.025)0.34 0.31 2.7Columbia Kootenay Daer Creek DC 72 29 50.8129 -115.9603 2008 0.943 -0.054 (-0.159-0.009)0.40 0.38 3.785AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NAColumbia Kootenay Dolly VardenCreekDVC 70 30 50.8272 -116.0146 2008 0.916 0.044 (-0.098-0.138)0.40 0.42 4.2Columbia Kootenay HoneymoonCreekHC 48 29 51.0275 -115.9856 2008 0.981 -0.147 (-0.283--0.048)0.33 0.29 2.3Columbia Kootenay Lost Creek LC 59 29 50.8975 -116.0710 2008 0.964 -0.114 (-0.239--0.027)0.41 0.37 3.2Columbia Kootenay MeadowCreekMCA 74 35 50.7638 -115.9528 2008 0.963 -0.057 (-0.146-0.004)0.39 0.37 3.9Columbia Kootenay Nixon Creek NC 76 27 50.7423 -115.9211 2008 0.896 -0.037 (-0.165-0.046)0.47 0.45 4.0Columbia Kootenay Pitts Creek PC 77 37 50.7378 -115.9008 2008 0.964 -0.049 (-0.145-0.021)0.41 0.39 3.7Columbia Kootenay VermillionCreekVR 53 33 50.9823 -115.9505 2008 0.965 0.129 ( 0.030-0.195)0.34 0.39 3.1Oldman Castle Carbondale(middle)ACA-152133 18 49.4521 -114.4091 2009 0.911 0.031 (-0.093-0.090)0.38 0.39 4.0Oldman Castle Carbondale(middle)ACA-59 136 21 49.4372 -114.4313 2009 0.957 -0.040 (-0.147-0.023)0.37 0.35 3.9Oldman Castle Carbondale(upper)ACA-61 139 16 49.4059 -114.4981 2009 0.993 0.035 (-0.136-0.133)0.31 0.32 2.6Oldman Castle CarbondaleRiverAFW-CaR138 22 49.4218 -114.4671 2006 0.988 -0.126 (-0.276--0.016)0.38 0.34 2.9Oldman Castle GardinerCreekD-C3 142 29 49.3983 -114.4592 2007 0.993 0.028 (-0.102-0.122)0.33 0.34 3.0Oldman Castle Gladstone(middle)ACA-52 144 15 49.3664 -114.2158 2009 0.945 -0.158 (-0.295--0.097)0.50 0.43 4.1Oldman Castle Goat Creek ACA-57 129 29 49.4954 -114.5397 2009 0.992 0.518 ( 0.262-0.735)0.48 0.99 1.5Oldman Castle Gorge Creek ACA-78 131 29 49.4702 -114.4274 2009 0.939 0.134 (-0.135-0.326)0.16 0.18 2.8Oldman Castle Lost Creek AFW-LoC134 27 49.4492 -114.4870 2006 0.982 0.160 ( 0.011-0.256)0.31 0.37 3.586AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NAOldman Castle Lost CreektribACA-62 135 26 49.4484 -114.4966 2009 0.989 0.027 (-0.097-0.119)0.31 0.32 2.7Oldman Castle Lynx Creek AFW-LyC130 30 49.4741 -114.4713 2006 0.990 0.215 (-0.112-0.453)0.1 0.13 1.7Oldman Castle Lynx Creek(lower)ACA-83 132 30 49.4630 -114.4468 2009 0.992 0.010 (-0.115-0.100)0.07 0.08 1.5Oldman Castle Lynx CreektribACA-121128 29 49.5235 -114.5229 2009 0.993 0.150 (-0.150-0.472)0.05 0.05 1.5Oldman Castle MacDonaldCreekD-C1 140 30 49.3997 -114.5227 2007 0.986 -0.018 (-0.156-0.105)0.38 0.37 3.2Oldman Castle OHagenCreekD-C4 137 30 49.4252 -114.3911 2007 0.992 0.068 (-0.061-0.167)0.17 0.18 2.0Oldman Castle Scarpe Creek D-C6 149 34 49.2339 -114.2560 2008 0.984 0.070 (-0.071-0.164)0.31 0.34 3.5Oldman Castle South CastleRiver (lower)ACA-65 141 19 49.3801 -114.3332 2009 0.874 -0.052 (-0.181-0.015)0.49 0.47 4.1Oldman Castle South CastleRiver(upper)ACA-71 150 18 49.2225 -114.2285 2009 0.985 -0.015 (-0.175-0.081)0.30 0.30 2.4Oldman Castle West CastleRiver (lower)ACA-159143 20 49.3748 -114.3720 2009 0.850 0.098 (-0.127-0.227)0.44 0.48 5.2Oldman Castle West CastleRiver (lower)ACA-84 147 28 49.3445 -114.4098 2009 0.947 0.078 (-0.079-0.208)0.35 0.38 4.2Oldman Castle West CastleRiver(upper)ACA-68 148 27 49.2375 -114.3496 2009 0.994 0.064 (-0.111-0.205)0.12 0.13 1.5Oldman Castle WhitneyCreekAFW-WhC146 25 49.3599 -114.9252 2006 0.939 0.129 (-0.011-0.237)0.38 0.43 3.9Oldman Castle WhitneyCreek(lower)ACA-60 145 29 49.3587 -114.1525 2009 0.925 -0.067 (-0.162--0.006)0.12 0.13 1.5Oldman Crowsnest BlairmoreCreekAFW-BC125 26 49.6682 -114.4377 2006 0.931 0.218 ( 0.026-0.329)0.31 0.39 4.087AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NAOldman Crowsnest CrowsnestD-Cr1 nrtourist officeD-Cr1 126 30 49.6300 -114.6090 2007 0.994 0.128 (-0.064-0.281)0.15 0.17 1.6Oldman Crowsnest Island Creek AFW-IC 127 23 49.6284 -114.7013 2006 0.968 0.253 ( 0.122-0.354)0.38 0.39 3.3Oldman Livingstone Deep Creek AFW-DC120 30 49.9761 -114.5565 2006 0.978 0.305 ( 0.107-0.458)0.19 0.27 2.9Oldman Livingstone LivingstoneRiver abovefallsAFW-LR117 24 50.1664 -114.4610 2006 0.989 0.149 (-0.069-0.296)0.29 0.34 3.1Oldman Livingstone LivingstoneRiver belowfallsD-O2 118 31 50.0881 -114.4271 2008 0.864 0.113 (-0.005-0.194)0.39 0.44 4.8Oldman OldmanmainstemBeaver Creek D-O1 123 30 49.8055 -113.9663 2008 0.969 -0.001 (-0.117-0.079)0.41 0.41 4.5Oldman OldmanmainstemHiddenCreekAFW-HC119 25 49.9762 -114.5566 2006 0.991 -0.020 (-0.155-0.084)0.34 0.33 2.4Oldman OldmanmainstemSharplesCreekD-O3 121 29 49.8809 -114.0692 2008 0.990 0.013 (-0.107-0.103)0.38 0.38 3.2Oldman Racehorse NorthRacehorse Crabove fallsAFW-NRC122 27 49.8434 -114.5745 2006 0.990 0.169 ( 0.025-0.247)0.32 0.38 3.2Oldman Racehorse Vicary Creek AFW-VC124 21 49.7538 -114.4885 2006 0.990 0.093 (-0.075-0.216)0.38 0.41 3.4Oldman Waterton CarthewLakeCA 152 30 49.0286 -113.9912 2008 0.904 0.045 (-0.087-0.140)0.34 0.35 3.4Oldman Waterton Crypt Lake CLR 154 27 49.0012 -113.8407 2008 0.764 0.035 (-0.091-0.130)0.37 0.39 3.3Oldman Waterton Goat Lake GO 10 27 51.4471 -115.8592 2008 0.982 -0.042 (-0.201-0.076)0.25 0.24 2.2Oldman Waterton LinehamLakeLIN 153 28 49.0258 -114.0675 2008 0.770 0.081 (-0.040-0.159)0.40 0.43 3.7Oldman Waterton Lone Lake LOL 151 35 49.0887 -114.1313 2008 0.843 0.066 (-0.053-0.146)0.33 0.35 3.188AppendixA.Chapter2MajorDrainageSub-basin Stream/Lake Samplesite# onFig. 2.1N Latitude Longitude SampleYearMeanqwsctFIS (95%CI)HO HE NAOldman Willow Corral Creek DW4 115 30 50.2548 -114.4225 2007 0.992 0.026 (-0.112-0.114)0.34 0.34 2.8Oldman Willow JohnsonCreekDW2 116 28 50.2158 -114.4047 2007 0.827 -0.031 (-0.160-0.006)0.36 0.44 4.589AppendixA.Chapter2Table A.2: Variable data of streams used in RANDOMFORESTto generate admixture model. Site code: sampling location (seeTable. A.1 for location details); qwsct: average admixture; meand:mean depth at sample site (m); maxd: maximum depth at samplesite (m); temp: water temperature at time of sampling ( ?C); bar-rier: distance to nearest barrier (m); elev: elevation (m), streamo:Strahler stream order; road: euclidean distance to nearest road(m); pipeline: euclidean distance to nearest gas/oil pipeline (m);powerline: euclidean distance to nearest power line (m); railline:euclideaen distance to nearest rail line (m); stockint: total numberof stocking events at the stocking site; snumber: number of fishstocked at the stocking site; year2: total years between 2010 andyear of last stocking at stocking site; diststock: fluvial distance tonearest stocking site (m).SiteCodeqwsct meand maxd temp barrier elev streamo road pipeline powerline railline stockint snumber year2 diststockACA-1210.993 0.48 0.80 11.1 8905 1605 3 1094 10388 11707 10819 10 87500 59 8929ACA-1520.911 0.80 1.50 10.7 833 1355 6 126 3305 13437 12037 3 26500 59 409ACA-1590.850 0.54 1.00 9.1 8538 1372 5 330 8082 20709 19776 3 26500 59 5506ACA-570.992 0.20 0.30 4.0 7914 1581 4 139 11970 14956 14164 6 23400 71 8027ACA-590.957 0.81 1.30 10.3 2472 1393 6 71 5371 15602 14204 4 21710 66 1228ACA-600.925 0.27 0.35 10.0 15555 1436 3 193 347 10993 12596 1 6000 74 4084ACA-610.993 0.32 0.40 9.0 7503 1519 4 138 11291 21021 19641 3 12700 41 2474ACA-620.989 0.18 0.25 11.0 4224 1493 4 134 9620 17182 15826 6 23400 71 193390AppendixA.Chapter2SiteCodeqwsct meand maxd temp barrier elev streamo road pipeline powerline railline stockint snumber year2 diststockACA-650.874 1.06 2.00 9.2 7186 1356 5 211 6632 18648 19080 64 3101151 0 2797ACA-680.994 0.54 1.00 9.1 5124 1563 3 1193 16592 25806 26392 1 6300 45 1880ACA-710.985 0.34 0.60 11.6 12670 1521 5 153 10442 18419 18809 2 5960 41 5564ACA-830.992 0.61 1.40 14.2 296 1419 4 99 5924 13672 12296 3 26500 59 3084ACA-840.947 0.90 1.60 9.3 7594 1397 4 431 12305 24939 23537 1 5000 69 7032D-C3 0.993 0.20 0.50 8.0 8958 1529 4 2525 10931 22067 20668 3 12700 41 1726D-C4 0.992 0.40 1.00 10.0 3980 1435 3 12 4099 15885 14483 3 26500 59 3498D-C6 0.984 0.50 1.40 8.0 10301 1512 3 1312 11810 19596 20082 2 5960 41 3504D-Cr1 0.994 0.20 0.40 6.0 12656 1377 4 70 984 1211 483 15 121720 45 2028D-O3 0.990 0.20 1.50 11.0 19783 1386 3 9 2961 7586 32936 22 20935 18 3976D-W2 0.992 0.10 0.30 8.0 7972 1590 4 372 230 12258 47344 3 33300 59 6417DW4 0.992 0.20 0.40 6.0 4557 1586 3 3290 3259 13303 46204 2 8400 45 776GO 0.988 0.35 1.10 10.0 6477 1619 4 38 14444 19654 56283 7 117880 41 4926J-E10a 0.650 0.20 0.50 15.0 5241 1425 3 1052 3390 20423 26774 51 112568 0 1653J-H11 0.993 0.20 0.70 13.0 1000 1437 3 843 16479 23247 43767 1 5000 66 6074J-H11a 0.992 0.10 0.50 10.0 929 1514 2 2690 17979 25152 45174 1 5000 66 4817J-H12 0.677 0.15 0.30 12.0 1413 1482 3 89 19418 24481 46741 1 5000 66 3338J-H18 0.986 0.20 0.75 8.0 332 1485 4 682 17614 23505 51359 22 237767 25 575J-H24a 0.988 0.35 1.10 10.0 6450 1619 4 38 14444 19654 56283 7 117880 41 4926J-H30 0.429 0.30 1.00 9.0 2999 1599 4 1423 26643 11966 62449 1 2340 31 10J-H30a 0.706 0.20 0.80 8.0 4100 1712 4 3944 30374 8857 64836 1 2340 31 4173J-H30b0.897 0.15 0.75 8.0 4882 1703 3 3617 29118 9636 64302 1 2340 31 2916J-H33 0.968 0.40 1.00 7.0 858 1760 4 2058 33536 9167 63442 42 478769 0 4985J-H4c 0.041 0.20 1.00 15.0 6607 1350 4 1457 10624 17120 37196 15 312340 31 3639J-H4f 0.731 0.15 1.00 6.0 5172 1563 3 2840 19502 25567 45668 15 312340 31 529491AppendixA.Chapter2SiteCodeqwsct meand maxd temp barrier elev streamo road pipeline powerline railline stockint snumber year2 diststockJ-H4g 0.769 0.15 0.50 7.0 6568 1593 3 2886 18982 25371 45396 15 312340 31 5853J-H4h 0.582 0.30 0.50 6.0 2753 1577 4 210 17152 24019 43730 15 312340 31 7245J-H4i 0.720 0.20 0.40 9.0 4338 1634 4 60 19121 25914 45700 15 312340 31 8001J-H7a 0.994 0.10 0.40 9.0 1980 1482 3 3453 17525 25643 44582 1 5000 66 5828J-J11 0.900 0.30 2.00 8.5 312 1469 4 161 184 15355 16340 8 14980 24 3289J-J11c 0.901 0.20 0.50 9.0 630 1469 3 686 707 15271 16030 8 14980 24 3258J-J19 0.965 0.30 1.00 9.0 7747 1644 3 1 4714 7970 13752 3 1150 11 5623J-J20 0.941 0.15 0.30 8.5 8799 1675 3 1 4468 9046 15063 3 1150 11 6841J-P15 0.080 0.10 0.30 12.0 1355 1463 3 1084 14090 17765 44553 6 31000 40 2024J-P16 0.117 0.30 0.50 14.0 463 1477 3 1962 13353 17431 44813 3 30500 66 2380J-P16a 0.787 0.40 1.10 12.0 1860 1593 3 3438 11945 19390 47352 3 30500 66 941J-P17a 0.952 0.20 0.50 9.0 2229 1580 3 4900 10396 17518 46456 3 30500 66 1423J-P19 0.949 0.15 0.40 13.0 3936 1551 2 6007 9461 15449 45255 1 10000 71 2820J-P20 0.959 0.20 0.40 8.0 4316 1591 3 6724 8635 16300 46432 1 10000 71 1647J-P21 0.962 0.15 0.30 10.0 2991 1631 3 5371 6620 16324 47499 1 10000 71 2040J-S12a 0.977 0.20 1.00 5.0 948 1590 4 1861 17070 24148 43543 2 9200 60 7329J-S12b 0.929 0.20 1.00 12.5 1484 1483 4 4013 13176 23285 41662 2 9200 60 2605J-S12e 0.182 0.60 1.80 12.0 1375 1395 5 1532 10782 20873 38979 2 9200 60 1443J-S17 0.698 1.00 3.00 15.0 0 1460 5 320 13791 28087 45716 3 30000 60 3395J-S17a 0.954 0.40 1.20 16.0 163 1565 4 359 13312 27592 47119 2 50000 69 3717J-S17b 0.990 0.30 0.50 10.0 4147 1687 4 4652 16341 23884 51406 2 50000 68 1400J-St14d0.932 0.20 0.40 11.0 4929 1458 3 5538 8127 12253 41488 3 30500 66 6333J-St14f0.454 0.30 0.60 15.0 2684 1359 3 2080 4408 7739 38279 1 2125 42 3468J-T11d0.010 0.30 0.70 18.0 8071 1320 4 104 2549 19270 35001 3 32000 70 2275J-T11g 0.715 0.30 0.70 13.0 5324 1497 4 3 7316 27931 43660 1 10000 73 5400J-T20a 0.068 0.20 0.50 14.0 6493 1566 3 2439 4022 31901 40911 1 10000 73 247692Appendix BChapter 3Table B.1: Critical thermal minima, length and mass datafor each trout tested. Rainbow trout (RT); Blackwater Riverrainbow trout (BW); Blackwater River backcross (BWB);westslope cutthroat trout (WSCT); Blackwater River x Tzen-zaicut Lake F1 hybrid x Blackwater River x Tzenzaicut LakeF1 hybrid (F2); Tzenzaicut Lake rainbow trout (TZ); Tzen-zaicut Lake backcross (TZB).FishNumberTestGroupSpecies Acc.TempCTMin, ForkLengthMass?C ?C (cm) (g)1 BW RT 15 2.5 7.5 4.732 BW RT 15 1.1 8.1 5.273 BW RT 15 1.4 8.6 6.584 BW RT 15 1.4 7.5 4.435 BW RT 15 2.5 9.9 10.206 BW RT 15 1.8 7.7 4.837 BW RT 15 1.1 10.1 10.538 BW RT 15 3.3 8.5 6.439 BW RT 15 2.6 7.1 3.8110 BW RT 15 1.8 7.4 3.6811 BW RT 15 2.3 6.0 2.5612 BW RT 15 1.7 7.0 3.7713 BW RT 15 1.1 7.6 4.6214 BW RT 15 0.5 9.5 9.2915 BW RT 15 2.0 7.4 4.5116 BW RT 15 1.5 8.9 7.8917 BW RT 15 1.2 9.1 7.8818 BW RT 15 2.1 7.7 5.0919 BW RT 15 2.8 7.6 5.1120 BW RT 15 1.8 9.5 8.9693Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)21 BW RT 15 2.4 7.2 4.4822 BW RT 15 3.6 6.9 3.8223 BW RT 15 2.7 8.8 7.7624 BW RT 15 1.8 7.2 4.3125 TZ RT 15 1.2 10.3 10.2026 TZ RT 15 0.7 10.3 10.6727 TZ RT 15 0.6 8.3 5.7328 TZ RT 15 0.5 12.0 17.7529 TZ RT 15 0.9 9.7 8.9130 TZ RT 15 0.1 11.6 14.9031 TZ RT 15 0.8 10.7 11.3832 TZ RT 15 2.5 10.8 11.6433 TZ RT 15 1.0 9.1 6.7234 TZ RT 15 2.1 8.7 6.3035 TZ RT 15 0.3 9.8 8.6736 TZ RT 15 0.3 12.4 18.4237 F2 RT 15 1.1 9.8 9.3338 F2 RT 15 0.6 10.8 11.3539 F2 RT 15 0.2 11.6 15.5240 F2 RT 15 0.4 8.9 7.9441 F2 RT 15 1.0 8.4 6.5342 F2 RT 15 1.6 12.5 18.6343 F2 RT 15 2.0 9.5 8.0544 F2 RT 15 1.8 10.7 10.9545 F2 RT 15 1.6 9.1 7.9546 F2 RT 15 0.3 9.7 9.7647 F2 RT 15 0.3 9.8 9.6148 F2 RT 15 2 11.2 14.5149 TZ RT 18 1.1 12.2 16.6850 TZ RT 18 1.2 10.6 10.6651 TZ RT 18 1.7 11.0 12.6052 TZ RT 18 0.8 9.7 8.5153 TZ RT 18 0.5 10.4 10.6054 TZ RT 18 1.1 10.8 11.7255 TZ RT 18 3.1 9.5 7.9956 TZ RT 18 1.9 10.9 13.0057 TZ RT 18 1.5 10.3 10.4294Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)58 TZ RT 18 1.6 12.1 15.8559 TZ RT 18 0.8 12.0 15.2460 TZ RT 18 1.9 10.0 9.5861 TZ RT 15 1.0 9.3 7.2662 TZ RT 15 0.2 11.4 13.9663 TZ RT 15 2.0 10.3 9.7164 TZ RT 15 1.8 10.2 9.5665 TZ RT 15 0.5 10.0 9.2466 TZ RT 15 1.0 6.7 3.0367 TZ RT 15 3.3 9.9 8.9868 TZ RT 15 2.1 11.2 13.8069 TZ RT 15 1.2 9.7 8.7470 TZ RT 15 1.3 9.1 6.7071 TZ RT 15 0.8 10.2 9.7172 TZ RT 15 1.3 10.2 9.6673 F2 RT 18 3.7 10.1 11.1174 F2 RT 18 1.8 9.9 8.0875 F2 RT 18 2.1 11.2 13.0076 F2 RT 18 3.2 9.6 8.5277 F2 RT 18 2.2 10.6 11.9878 F2 RT 18 3.7 10.8 12.4979 F2 RT 18 5.1 10.3 9.6280 F2 RT 18 4.3 10.5 11.5481 F2 RT 18 1.6 10.5 10.8782 F2 RT 18 2.6 10.3 11.9083 F2 RT 18 1.7 11.2 14.4584 F2 RT 18 2.5 8.9 7.4185 F2 RT 15 0.8 10.9 12.1786 F2 RT 15 1.8 10.5 10.0287 F2 RT 15 0.7 9.1 6.7688 F2 RT 15 0.2 8.2 5.1589 F2 RT 15 1.6 9.1 5.6090 F2 RT 15 0.9 11.0 12.5391 F2 RT 15 2.0 7.6 4.0592 F2 RT 15 2.6 9.7 8.5293 F2 RT 15 0.3 8.2 5.4794 F2 RT 15 2.1 9.9 8.8595Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)95 F2 RT 15 0.3 10.5 11.3496 F2 RT 15 5.1 11.5 13.3697 F2 RT 18 0.5 10.2 10.6698 F2 RT 18 1.9 10.8 14.5199 F2 RT 18 0.9 10.3 10.26100 F2 RT 18 1.3 9.1 6.37101 F2 RT 18 1.9 9.1 7.69102 F2 RT 18 0.4 9.1 8.56103 F2 RT 18 2.5 11.3 14.38104 F2 RT 18 2.6 9.4 8.18105 F2 RT 18 0.7 7.6 4.90106 F2 RT 18 0.7 9.6 8.40107 F2 RT 18 2.4 9.1 7.86108 F2 RT 18 1.8 10.5 11.80109 BW RT 15 1.2 9.0 7.34110 BW RT 15 0.4 11.1 15.80111 BW RT 15 1.6 10.5 11.18112 BW RT 15 0.4 10.1 9.70113 BW RT 15 0.5 9.5 8.24114 BW RT 15 0.1 11.9 16.13115 BW RT 15 1.3 10.7 13.06116 BW RT 15 0.5 8.5 5.86117 BW RT 15 -0.1 8.3 5.43118 BW RT 15 1.1 9.1 7.44119 BW RT 15 0.3 8.1 5.32120 BW RT 15 1.5 9.8 8.47121 TZ RT 18 0.6 9.1 7.45122 TZ RT 18 0.6 10.6 11.52123 TZ RT 18 2.6 9.6 7.69124 TZ RT 18 1.7 9.2 7.26125 TZ RT 18 1.1 10.9 11.77126 TZ RT 18 0.0 8.8 6.13127 TZ RT 18 2.9 7.9 4.70128 TZ RT 18 3.0 9.8 8.03129 TZ RT 18 2.7 8.5 5.28130 TZ RT 18 0.8 10.9 11.92131 TZ RT 18 1.4 9.6 8.0696Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)132 TZ RT 18 1.2 10.4 9.74133 BW RT 18 0.0 9.7 9.86134 BW RT 18 1.5 8.7 5.88135 BW RT 18 1.9 8.0 5.22136 BW RT 18 0.7 10.1 10.64137 BW RT 18 1.7 9.2 8.24138 BW RT 18 2.7 11.4 15.11139 BW RT 18 1.6 7.4 4.72140 BW RT 18 0.1 8.0 3.65141 BW RT 18 0.7 9.4 8.35142 BW RT 18 1.3 8.9 7.45143 BW RT 18 2.1 7.9 5.21144 BW RT 18 0.6 10.8 13.27145 F2 RT 15 -0.1 9.8 10.42146 F2 RT 15 0.1 10.6 12.03147 F2 RT 15 0.9 9.8 9.51148 F2 RT 15 0.0 10.8 12.01149 F2 RT 15 3.2 6.8 3.61150 F2 RT 15 -0.2 9.9 10.03151 F2 RT 15 -0.2 11.7 14.88152 F2 RT 15 1.3 10.4 9.69153 F2 RT 15 1.1 9.8 10.28154 F2 RT 15 -1.1 10.1 10.90155 F2 RT 15 0.5 9.1 5.76156 F2 RT 15 0.2 8.3 6.84157 BWB RT 15 1.8 10.0 9.20158 BWB RT 15 0.1 9.0 7.56159 BWB RT 15 1.4 7.8 4.70160 BWB RT 15 0.1 8.7 6.64161 BWB RT 15 0.8 7.0 3.45162 BWB RT 15 1.2 7.3 3.84163 BWB RT 15 2.4 8.6 6.47164 BWB RT 15 2.3 8.3 5.40165 BWB RT 15 0.3 7.2 4.06166 BWB RT 15 0.6 10.1 10.68167 BWB RT 15 1.3 10.5 10.7168 F2 RT 18 1.7 9.1 7.3397Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)169 F2 RT 18 1.0 9.9 9.47170 F2 RT 18 2.4 8.5 6.07171 F2 RT 18 1.6 9.2 7.29172 F2 RT 18 3.0 8.2 5.53173 F2 RT 18 3.3 9.2 7.09174 F2 RT 18 1.4 10.1 9.63175 F2 RT 18 1.8 8.8 6.37176 F2 RT 18 1.4 9.9 9.23177 F2 RT 18 1 11.1 13.31178 F2 RT 18 1.4 8.1 5.29179 F2 RT 18 1.8 10.7 11.73180 BWB RT 15 3.7 10.4 9.85181 BWB RT 15 1.2 8.4 5.47182 BWB RT 15 1.6 11.7 15.83183 BWB RT 15 1.0 9.3 7.67184 BWB RT 15 2.4 11.9 15.39185 BWB RT 15 2.6 9.5 7.43186 BWB RT 15 5.1 8.5 5.97187 BWB RT 15 3.6 7.4 3.82188 BWB RT 15 1.3 9.0 7.25189 BWB RT 15 1.3 8.1 5.31190 BWB RT 15 1.7 9.3 7.82191 BWB RT 15 2.5 7.2 4.15192 BWB RT 18 1.8 10.5 9.47193 BWB RT 18 2.4 8.0 4.81194 BWB RT 18 2.2 8.8 6.08195 BWB RT 18 1.2 9.0 6.99196 BWB RT 18 0.9 9.2 6.70197 BWB RT 18 2.6 7.9 4.36198 BWB RT 18 3.4 9.1 6.57199 BWB RT 18 4.5 9.1 6.79200 BWB RT 18 3.0 9.3 6.20201 BWB RT 18 3.1 8.3 4.98202 BWB RT 18 1.9 7.8 4.60203 BWB RT 18 1.5 10.0 9.37204 TZB RT 15 1.0 9.7 8.03205 TZB RT 15 1.2 12.5 18.1098Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)206 TZB RT 15 1.3 10.3 10.67207 TZB RT 15 0.2 8.4 5.63208 TZB RT 15 0.8 9.2 7.10209 TZB RT 15 1.0 9.3 7.06210 TZB RT 15 2.1 7.9 4.50211 TZB RT 15 2.7 9.5 8.01212 TZB RT 15 2.3 7.9 4.40213 TZB RT 15 0.3 8.0 4.30214 TZB RT 15 1.0 9.9 8.71215 TZB RT 15 1.1 8.9 6.41216 BW RT 18 1.1 8.5 5.50217 BW RT 18 2.0 6.9 3.53218 BW RT 18 1.4 10.5 12.34219 BW RT 18 4.0 10.1 7.68220 BW RT 18 2.6 9.0 6.21221 BW RT 18 0.9 8.5 6.41222 BW RT 18 3.1 7.5 4.01223 BW RT 18 3.2 7.9 5.07224 BW RT 18 2.9 8.6 6.56225 BW RT 18 2.8 11.5 14.51226 BW RT 18 1.3 10.9 13.14227 BW RT 18 2.1 11.5 14.88228 BW RT 18 1.6 10.1 9.76229 BW RT 18 1.7 11.4 12.02230 BW RT 18 1.9 11.3 11.32231 BW RT 18 4.4 9.4 6.96232 BW RT 18 1.7 9.9 9.20233 BW RT 18 2.2 9.3 7.33234 BW RT 18 1.6 9.6 7.82235 BW RT 18 3.5 10.2 9.56236 BW RT 18 1.4 9.4 8.05237 BW RT 18 2.7 10.2 9.54238 BW RT 18 1.8 8.9 6.56239 BW RT 18 1.7 10.3 10.63240 CCT WSCT 15 0.6 7.8 4.35241 CCT WSCT 15 0.8 6.9 3.30242 CCT WSCT 15 1.4 6.6 2.6099Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)243 CCT WSCT 15 0.5 6.8 3.06244 CCT WSCT 15 1.7 7.0 3.37245 CCT WSCT 15 1.5 7.0 2.96246 CCT WSCT 15 1.2 6.6 3.12247 CCT WSCT 15 2.9 7.7 4.07248 CCT WSCT 15 0.5 6.8 2.96249 CCT WSCT 15 1.2 7.5 3.80250 CCT WSCT 15 2.2 6.1 2.31251 CCT WSCT 15 1.2 7.3 3.82252 CCT WSCT 15 2.1 6.6 3.02253 CCT WSCT 15 0.8 6.8 2.92254 CCT WSCT 15 0.3 7.6 4.40255 CCT WSCT 15 1.3 7.8 4.8 0256 CCT WSCT 15 1.6 6.5 2.32257 CCT WSCT 15 -0.1 8.0 4.99258 CCT WSCT 15 2.0 7.5 4.10259 CCT WSCT 15 2.3 10.0 10.69260 CCT WSCT 15 2.0 7.5 4.14261 CCT WSCT 15 0.0 7.8 4.49262 CCT WSCT 15 0.9 8.1 7.19263 CCT WSCT 15 0.4 8.3 5.85264 CCT WSCT 15 0.5 7.4 4.08265 CCT WSCT 15 0.1 9.0 7.66266 CCT WSCT 15 1.2 6.6 2.58267 CCT WSCT 15 1.1 7.2 3.96268 CCT WSCT 15 0.0 7.1 4.29269 CCT WSCT 15 0.2 6.8 3.33270 CCT WSCT 15 1.6 7.0 3.86271 CCT WSCT 15 2.0 6.4 2.40272 CCT WSCT 15 0.2 7.0 3.06273 CCT WSCT 15 -0.5 7.5 4.72274 CCT WSCT 15 1.9 7.0 3.57275 CCT WSCT 15 0.1 6.7 2.37276 TZ RT 18 3.1 9.4 6.36277 TZ RT 18 1.6 9.5 7.83278 TZ RT 18 1.4 8.7 6.48279 TZ RT 18 1.2 8.6 6.00100Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)280 TZ RT 18 1.4 11.0 11.19281 TZ RT 18 1.8 9.2 7.25282 TZ RT 18 4.8 10.2 9.02283 TZ RT 18 4.2 9.1 6.70284 TZ RT 18 2.0 9.4 7.24285 TZ RT 18 1.2 10.3 10.15286 TZ RT 18 2.8 9.6 8.55287 TZ RT 18 2.2 11.1 12.68288 TZB RT 18 2.2 10.3 10.70289 TZB RT 18 0.6 9.6 9.29290 TZB RT 18 0.9 11.3 13.58291 TZB RT 18 1.7 9.0 7.36292 TZB RT 18 1.3 10.0 10.10293 TZB RT 18 1.0 10.1 10.98294 TZB RT 18 2.6 9.1 6.91295 TZB RT 18 3.3 10.5 9.66296 TZB RT 18 2.2 10.5 11.92297 TZB RT 18 3.0 10.1 10.56298 TZB RT 18 3.0 8.7 6.17299 TZB RT 18 1.1 9.9 9.16300 CCT WSCT 18 2.5 7.7 4.53301 CCT WSCT 18 1.5 6.9 3.43302 CCT WSCT 18 1.7 7.1 3.68303 CCT WSCT 18 2.3 7.0 3.53304 CCT WSCT 18 1.9 8.9 7.13305 CCT WSCT 18 1.6 7.8 4.67306 CCT WSCT 18 1.8 7.0 3.66307 CCT WSCT 18 2.8 6.8 3.39308 CCT WSCT 18 1.7 8.3 5.92309 CCT WSCT 18 1.6 7.4 4.23310 CCT WSCT 18 3.6 6.8 3.37311 CCT WSCT 18 1.8 7.4 4.27312 TZ RT 15 1.5 11.2 11.66313 TZ RT 15 1.9 10.5 12.20314 TZ RT 15 1.0 10.9 13.37315 TZ RT 15 1.1 11.3 13.02316 TZ RT 15 2.1 10.1 9.45101Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)317 TZ RT 15 2.1 10.4 11.33318 TZ RT 15 1.7 9.3 8.01319 TZ RT 15 1.5 9.5 8.66320 TZ RT 15 2.1 10.1 10.31321 TZ RT 15 1.8 10.4 9.97322 TZ RT 15 2.2 11.3 14.30323 TZ RT 15 1.8 11.2 13.12324 CCT WSCT 18 1.8 6.2 2.18325 CCT WSCT 18 1.9 7.2 4.25326 CCT WSCT 18 1.6 8.0 5.63327 CCT WSCT 18 0.6 7.9 4.74328 CCT WSCT 18 1.1 7.4 4.47329 CCT WSCT 18 2.5 7.1 3.94330 CCT WSCT 18 3.0 7.4 4.94331 CCT WSCT 18 1.8 9.6 10.75332 CCT WSCT 18 0.8 7.2 4.18333 CCT WSCT 18 2.5 6.4 3.20334 CCT WSCT 18 2.8 6.9 4.15335 CCT WSCT 18 1.4 8.0 5.92336 CCT WSCT 18 2.1 7.6 4.01337 CCT WSCT 18 1.2 8.5 6.68338 CCT WSCT 18 1.5 9.1 10.03339 CCT WSCT 18 1.7 7.2 4.02340 CCT WSCT 18 2.3 6.5 3.01341 CCT WSCT 18 1.5 6.6 2.72342 CCT WSCT 18 2.5 7.9 5.16343 CCT WSCT 18 1.4 7.2 5.36344 CCT WSCT 18 2.4 6.3 2.41345 CCT WSCT 18 0.9 7.4 4.66346 CCT WSCT 18 1.8 8.9 7.84347 CCT WSCT 18 2.7 6.2 2.99348 BWB RT 15 0.1 11.3 16.87349 BWB RT 15 2.0 10.4 11.93350 BWB RT 15 1.6 10.3 12.36351 BWB RT 15 -0.1 10.0 11.45352 BWB RT 15 1.1 9.5 9.10353 BWB RT 15 -0.1 10.3 11.91102Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)354 BWB RT 15 0.1 9.7 10.64355 BWB RT 15 0.3 9.8 11.20356 BWB RT 15 1.8 11.5 16.14357 BWB RT 15 2.8 10.1 11.60358 BWB RT 15 2.5 9.5 9.17359 BWB RT 15 1.2 9.6 9.35360 BWB RT 18 2.2 9.3 8.43361 BWB RT 18 1.7 9.6 9.78362 BWB RT 18 1.4 9.1 8.63363 BWB RT 18 1.1 9.7 10.27364 BWB RT 18 1.9 10.0 10.42365 BWB RT 18 2.2 10.0 10.88366 BWB RT 18 3.3 9.4 9.19367 BWB RT 18 3.1 8.2 6.24368 BWB RT 18 2.0 10.8 12.57369 BWB RT 18 0.9 9.5 9.47370 BWB RT 18 1.8 8.2 5.82371 BWB RT 18 1.0 9.1 7.71372 BWB RT 18 2.9 10.8 13.33373 BWB RT 18 1.7 9.4 8.41374 BWB RT 18 2.1 9.8 10.37375 BWB RT 18 1.4 10.0 10.64376 BWB RT 18 1.4 9.9 10.90377 BWB RT 18 1.2 10.4 12.28378 BWB RT 18 2.4 9.3 9.32379 BWB RT 18 2.5 10.2 11.31380 BWB RT 18 2.0 8.8 8.10381 BWB RT 18 1.2 9.8 10.57382 BWB RT 18 1.6 8.4 7.07383 BWB RT 18 3.0 7.9 5.71384 TZB RT 18 2.5 9.0 8.22385 TZB RT 18 1.2 8.9 7.99386 TZB RT 18 1.5 9.0 7.69387 TZB RT 18 3.1 11.1 14.60388 TZB RT 18 1.2 9.1 8.05389 TZB RT 18 1.5 10.4 12.76390 TZB RT 18 1.6 9.0 7.76103Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)391 TZB RT 18 2.7 9.3 9.02392 TZB RT 18 2.5 9.6 9.88393 TZB RT 18 2.4 8.9 7.76394 TZB RT 18 2.1 9.8 10.46395 TZB RT 18 1.2 10.8 14.14396 TZB RT 18 0.6 9.9 10.86397 TZB RT 18 1.4 8.9 8.64398 TZB RT 18 0.5 8.4 7.71399 TZB RT 18 1.4 10.5 12.23400 TZB RT 18 1.7 7.7 5.49401 TZB RT 18 1.6 10.0 12.04402 TZB RT 18 3.0 9.1 7.74403 TZB RT 18 1.1 9.6 10.08404 TZB RT 18 0.7 9.3 9.30405 TZB RT 18 3.4 9.9 10.56406 TZB RT 18 1.3 8.4 6.47407 TZB RT 18 1.9 9.4 8.73408 TZB RT 15 1.6 9.9 10.70409 TZB RT 15 1.8 9.8 11.06410 TZB RT 15 1.6 10.0 13.25411 TZB RT 15 2.0 11.5 17.32412 TZB RT 15 1.8 9.8 10.80413 TZB RT 15 2.1 11.4 16.40414 TZB RT 15 2.8 9.6 10.48415 TZB RT 15 1.6 9.9 10.92416 TZB RT 15 1.6 11.4 16.43417 TZB RT 15 1.4 9.0 8.96418 TZB RT 15 1.9 10.4 13.05419 TZB RT 15 2.8 9.5 10.21420 TZB RT 15 1.3 8.7 8.81421 TZB RT 15 2.5 10.2 12.44422 TZB RT 15 1.6 10.0 10.53423 TZB RT 15 1.1 9.8 10.88424 TZB RT 15 1.6 9.5 9.70425 TZB RT 15 1.2 9.0 9.12426 TZB RT 15 2.9 9.4 10.40427 TZB RT 15 2.7 10.2 12.23104Appendix B. Chapter 3FishNumberTestGroupSpecies Acc.TempCtmin ForkLengthMass?C ?C (cm) (g)428 TZB RT 15 2.3 9.0 9.64429 TZB RT 15 1.6 8.7 8.18430 TZB RT 15 2.7 8.6 7.47431 TZB RT 15 2.6 9.8 11.48105Appendix B. Chapter 3Figure B.1: Spread of CTMin data of trout test groups acclimated at 15 ?C.Blackwater River rainbow trout (BW); Blackwater River backcross (BWB);westslope cutthroat trout (WSCT); Blackwater River x Tzenzaicut Lake F1hybrid x Blackwater River x Tzenzaicut Lake F1 hybrid (F2); TzenzaicutLake rainbow trout (TZ); Tzenzaicut Lake backcross (TZB).106Appendix B. Chapter 3Figure B.2: Spread of CTMin data of trout test groups acclimated at 18 ?C.Blackwater River rainbow trout (BW); Blackwater River backcross (BWB);westslope cutthroat trout (WSCT); Blackwater River x Tzenzaicut Lake F1hybrid x Blackwater River x Tzenzaicut Lake F1 hybrid (F2); TzenzaicutLake rainbow trout (TZ); Tzenzaicut Lake backcross (TZB).107

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0165667/manifest

Comment

Related Items