Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Intraguild predation is a mechanism of divergent selection in the threespine stickleback Miller, Sara Elizabeth 2016

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

Item Metadata

Download

Media
24-ubc_2016_september_miller_sara.pdf [ 9.55MB ]
Metadata
JSON: 24-1.0308707.json
JSON-LD: 24-1.0308707-ld.json
RDF/XML (Pretty): 24-1.0308707-rdf.xml
RDF/JSON: 24-1.0308707-rdf.json
Turtle: 24-1.0308707-turtle.txt
N-Triples: 24-1.0308707-rdf-ntriples.txt
Original Record: 24-1.0308707-source.json
Full Text
24-1.0308707-fulltext.txt
Citation
24-1.0308707.ris

Full Text

INTRAGUILD PREDATION IS A MECHANISM OF DIVERGENT SELECTION IN THE THREESPINE STICKLEBACK  by  Sara Elizabeth Miller  B.A., University of Chicago, 2006 M.S. University of Missouri, St. Louis, 2010  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2016  © Sara Elizabeth Miller, 2016   ii Abstract   Biotic interactions among species are thought to be important for the generation of phenotypic diversity. Intraguild predation is a common ecological interaction that occurs when a species preys upon another species with which it competes. This interaction is potentially a mechanism of divergence between intraguild prey populations, but it is unknown if cases of character shifts in intraguild prey are phenotypically plastic or an evolutionary response. I collected threespine stickleback (Gasterosteus aculeatus) from lakes with and without prickly sculpin (Cottus asper) and identified trait differences in armour and behaviour among populations in the wild. Differences in behavioural and morphological traits among freshwater populations persisted in a common garden, suggesting that adaptation to intraguild predation has a genetic basis. To date, the evolutionary effect that biotic selection has upon an organisms’ genome remains largely unknown in natural populations. I used whole genome re-sequencing to investigate the extent of genetic differentiation between stickleback from populations with and without sculpin. The main axis of genetic variation in these populations is strongly associated with the presence or absence of sculpin. I identified the regions of the genome that have differentiated in parallel between lakes with and without sculpin, and measured the strength of this divergence. The presence or absence of sculpin corresponds to widespread differentiation that is unevenly distributed across the stickleback genome. Adaptation to intraguild predation may involve hundreds of genes with diverse functions. Observations of extensive phenotypic and genetic differentiation between stickleback from lakes with and without sculpin provide indirect evidence that sculpin are the cause of trait differences.   iii Pelvic morphology is one of the most conspicuously varying traits among freshwater stickleback populations. This variation has been hypothesized to be the result of predation by fish and/or insect predators. I conducted a selection experiment to test if sculpin were an agent of selection for pelvic spine length. The results were combined with other experimental selection studies and used in a meta-analysis. Fish predators are an agent of selection for longer pelvic spines, but the role of insect predators is still unclear. Intraguild predation is a mechanism of divergent selection in threespine stickleback.    iv Preface  A version of chapter 2 has been published as Miller SE, Metcalf D, Schluter D. (2015) Intraguild predation leads to genetically based character shifts in the threespine stickleback. Evolution, 69:3194-3203. I designed the experiment, collected field samples, gathered morphometric measurements, conducted behavioural tests, and wrote the paper. Daniel Metcalf assisted with field collections, animal care, and analysis of behavioural videos. Dolph Schluter helped with experimental design, statistical analysis, and writing of the paper. Work for chapter 3 was conducted in collaboration with Dolph Schluter. I collected samples with assistance from various members of the Schluter lab. Daniel Bolnick, Jeffrey McKinnon, Sean Rogers, and Monica Yau generously provided additional stickleback specimens. Kevin Brix contributed the Na/Ca/Conductivity measurements used in table 3.1. I designed the experiment, prepared genomic libraries, created bioinformatics pipelines, and wrote the paper. Dolph Schluter contributed to the experimental design, bioinformatics methods, data analysis, and writing of the paper.  For Chapter 4, I designed and conduced the mesocosm experiment and meta-analysis and wrote the paper. Daniel Metcalf provided assistance with the set-up and monitoring of the mesocosm experiment. These new data were used in the meta-analysis along with studies referred to in Mirjam Barrueto’s thesis (Barrueto 2009) and more recent publications. Tuomas Leinonen, Andrew MacColl, and Kenyon Mobley provided additional raw data that was incorporated into the meta-analysis. Dolph Schluter contributed to the   v experimental design and analysis of the mesocosm experiment and meta-analysis and helped with writing the paper. A version of chapter 5 has been published: Miller SE, Samuk KM, Rennison DJ. (2016) An experimental test of the effect of predation upon behaviour and trait correlations in the threespine stickleback. Biological Journal of the Linnean Society. The trout predation experiment was designed and run by Diana Rennison as part of her PhD thesis. I designed the behavioural experiment, conducted statistical analysis, and wrote the paper with input from the other authors. Behavioural data was collected and analysed by all of the authors. Protocols requiring the use of live animals were approved by the UBC animal care committee (A07-0293, A11-0402). Permits for scientific collections were obtained from the British Columbia Ministry of Environment (NA-SU10-60714, NA-SU10-68002, NA-SU12-76311, NA-SU13-85151, NA-SU14-93473).   vi Table of Contents  Abstract ........................................................................................................................................ ii!Preface ........................................................................................................................................ iv!Table of Contents .................................................................................................................... vi!List of Tables .............................................................................................................................. x!List of Figures ........................................................................................................................... xi!List of Abbreviations ............................................................................................................ xiii!Acknowledgements .............................................................................................................. xiv!Dedication ................................................................................................................................. xv!Chapter 1: Introduction ......................................................................................................... 1!1.1! General Introduction ....................................................................................................... 1!1.2! The Study System ........................................................................................................... 2!1.3! Intraguild Predation and Threespine Stickleback ....................................................... 3!1.4! Summary of Studies ....................................................................................................... 4!Chapter 2: Intraguild Predation Leads to Genetically Based Character Shifts in the Threespine Stickleback .............................................................................................. 7!2.1! Introduction ..................................................................................................................... 7!2.2! Materials and Methods ................................................................................................. 11!2.2.1! Study Populations and Sample Collection .......................................................... 11!2.2.2! Common Garden and Plasticity Experiment ...................................................... 12!2.2.3! Morphology ............................................................................................................. 14!  vii 2.2.4! Stickleback Behaviour ........................................................................................... 16!2.2.5! Statistical Analysis ................................................................................................. 18!2.3! Results ............................................................................................................................ 18!2.3.1! Character Shifts in Wild-Caught Stickleback ...................................................... 18!2.3.2! Character Shifts Persisted in a Common Garden .............................................. 19!2.3.3! Sculpin Exposure Induced Character Shifts in Marine Stickleback ................ 20!2.3.4! Maternal Effects ..................................................................................................... 21!2.4! Discussion ....................................................................................................................... 21!2.4.1! Trait Shifts in Response to Intraguild Predation ............................................... 21!2.4.2! Trait Inducibility has been Lost in Freshwater Populations ............................ 23!Chapter 3: Intraguild Predation Leads to a Multitude of Genomic Changes but is Constrained by Genomic Architecture ....................................................................... 39!3.1! Introduction ................................................................................................................... 39!3.2! Materials and Methods ................................................................................................. 42!3.2.1! Sample Collection and Library Preparation ........................................................ 42!3.2.2! Bioinformatics Pipeline .......................................................................................... 44!3.2.3! Divergence Among Populations ........................................................................... 45!3.2.4! Candidate Genes .................................................................................................... 48!3.3! Results ............................................................................................................................ 48!3.3.1! Sculpin Presence is not Correlated with Abiotic Environment ........................ 48!3.3.2! Genomic Divergence is Associated with Presence/Absence of Intraguild Predator ................................................................................................................................ 49!  viii 3.3.3! Genetic Differentiation is Extensive but Unevenly Distributed ....................... 49!3.3.4! Candidate Adaptive Genes ................................................................................... 51!3.4! Discussion ....................................................................................................................... 51!Chapter 4: A Comparative Analysis of Experimental Selection on the Stickleback Pelvis ................................................................................................................... 66!4.1! Introduction ................................................................................................................... 66!4.2! Methods .......................................................................................................................... 71!4.2.1! Mesocosm Experiment .......................................................................................... 71!4.2.2! Comparison with Other Selection Studies .......................................................... 73!4.3! Results ............................................................................................................................ 75!4.3.1! Mesocosm Experiment .......................................................................................... 75!4.3.2! Meta-Analysis of Selection Studies ...................................................................... 75!4.4! Discussion ....................................................................................................................... 76!Chapter 5: An Experimental Test of the Effect of Predation Upon Behaviour and Trait Correlations in Threespine Stickleback ...................................................... 92!5.1! Introduction ................................................................................................................... 92!5.2! Methods .......................................................................................................................... 94!5.2.1! Experimental Design ............................................................................................. 94!5.2.2! Behavioural Assays ................................................................................................ 96!5.2.3! Armour Traits ......................................................................................................... 98!5.2.4! Statistical Analysis ................................................................................................. 99!5.3! Results ............................................................................................................................ 99!  ix 5.4! Discussion ..................................................................................................................... 101!5.4.1! Stickleback Behaviour ......................................................................................... 101!5.4.2! Correlations Between Morphology and Behaviour .......................................... 102!Chapter 6: Conclusion ........................................................................................................ 112!6.1! Overview ....................................................................................................................... 112!6.2! Broader Implications ................................................................................................... 116!References .............................................................................................................................. 120!Appendices .............................................................................................................................. 135!Appendix A ............................................................................................................................. 135!Appendix B ............................................................................................................................. 136!   x List of Tables  Table 2.1: Wild caught stickleback trait measurements .................................................. 36 Table 2.2: Trait loadings for principal component analysis of armour of wild stickleback ... 37 Table 2.3: Experimental stickleback trait measurements ................................................. 38 Table 3.1: Abiotic traits for lakes .................................................................................. 63 Table 3.2: Comparison of environmental variables between lake types ............................ 64 Table 3.3: Outlier windows and Hidden Markov Model results ......................................... 65 Table 4.1: Experimental studies of piscivorous predators ................................................ 88 Table 4.2: Experimental studies of insect predators ....................................................... 90 Table 5.1: Trait loadings for principal component analysis of armour of experimental stickleback ................................................................................................................ 108 Table 5.2 Linear model of behavioural traits, armour principal component score, and treatment ................................................................................................................. 109 Table 5.3: Behaviour and morphology trait values for each experimental pond .............. 110 Table 5.4: Spearman’s rank correlations between traits ................................................ 111     xi List of Figures  Figure 2.1 : Map of sampling locations used in the chapter ............................................ 26 Figure 2.2 : Schematic of crosses used in the common garden experiment ...................... 27 Figure 2.3 : Landmarks coordinates used for morphometrics .......................................... 28 Figure 2.4 : The set-up for behavioral assays ................................................................ 29 Figure 2.5 : Armour PC1 and vertical position in the water column of wild populations of stickleback .................................................................................................................. 31 Figure 2.6 : Armour PC1 for experimental stickleback in common garden ........................ 32 Figure 2.7 : Body shape axis 1 for experimental stickleback ........................................... 33 Figure 2.8 : Position in the water column for experimental stickleback ............................ 34 Figure 2.9 : Time spent near the experimental shoal ...................................................... 35 Figure 3.1: Map of sampling locations of sequenced stickleback ..................................... 56 Figure 3.2: Principal component analysis of abotic traits of lakes .................................... 57 Figure 3.3 : Principal component 1 of all SNPs ............................................................... 58 Figure 3.4 : Principal components 2 and 3 of all SNPs .................................................... 59 Figure 3.5 : CS’ and FST for 10,000 bp windows throughout the genome ......................... 60 Figure 3.6 : Genome-wide distribution of CS’ score ........................................................ 61 Figure 3.7 : CS’ score and hidden markov model for chromosome twelve ........................ 62 Figure 4.1 : Surviving threespine stickleback with clipped and unclipped pelvic spines ...... 81 Figure 4.2 : Forest plot of the effect size for all fish predation experiments ..................... 82 Figure 4.3 : Forest plot of the effect size for all insect predation experiments .................. 83 Figure 4.4 : Forest plot of the effect size for all fish predation studies ............................. 84   xii Figure 4.5 : Forest plot of the effect size for all insect predation studies. ......................... 85 Figure 4.6 : Effect size and standard error of experiments used in the meta-analysis ....... 86 Figure 4.7 : Forest plot of the effect size of standard length for all predation experiments 87 Figure 5.1 : Reaction norms for behavioural traits between control and predation ponds 105 Figure 5.2 : Correlations between behaviour and armour traits. .................................... 106    xiii List of Abbreviations  IGP  Intraguild Predation PCA  Principal Component Analysis CS’   Cluster Separation Score Metric SNP   Single Nucleotide Polymorphism   xiv Acknowledgements  I would like to thank my supervisor Dolph Schluter for his constant support and help throughout my PhD. He was instrumental in my growth as a scientist, writer, and thinker. I am also grateful to the valuable suggestions and advice of the members of my committee: Darren Irwin, Loren Rieseberg, and Eric Taylor. The many members of the Schluter lab provided an amazing community in which to work. My dissertation would not have been possible without the pioneering work on the sculpin/no sculpin lake system by Travis Ingram and Richard Svanbäck. A special thanks to the legions of undergraduate (and graduate, and post doc) volunteers who helped me in the field and in the fish room. Seth Rudman answered all of my questions about the outside. Kieran Samuk answered all of my questions about computers. Kevin Brix gave me an appreciation for physiology. Thor Veen convinced me of the wonders of base R. Greg Owens, although only an honorary member of the lab taught me much about bioinformatics. Dan Metcalf was an invaluable research assistant and kept me from being eaten by bears. And a big thank you to Diana Rennison for all of the above and for your steadfast support, help, advice, and humour.  I am also appreciative to many members of the Biodiversity Building and the UBC community. In particular, Kira Delmore was my right hand for many years – I would not be here without you! And to the members of CDWMB, Erin Fenneman, Jessica Lu, Kaeli Johnson, Kate McGrath and Virginia Woloshen, thank you for helping me to keep my sanity. Last but not least, to my Mom, thank you for your never-ending love and encouragement.   xv Dedication        In memory of my father.        1 Chapter 1: Introduction  1.1 General Introduction  Species do not exist in isolation. Species interact with each other and these interactions can be an important cause of natural selection. Thompson (2013) has suggested that, “Evolutionary rates are just as much about the pace at which interactions among species evolve as they are about the rates of genome evolution of each species”.  Interactions among species have been previously shown to lead to the evolution of divergent phenotypes. For example, it has been well established that interspecific competition for resources can lead to phenotypic divergence (“Character Displacement”) as character shifts that allow species to use alternative resources will decrease competition and be favoured by natural selection (Schluter 2000b; Stuart and Losos 2013). Similarly, experimental studies have found that presence of predators can lead to the evolution of trait divergence (e.g. McPeek 1995; Vamosi 2002; Langerhans et al. 2007).  Intraguild predation is a widespread ecological interaction in which a predator is also a competitor with its prey species (Polis et al. 1989; Arim and Marquet 2004). Intraguild predation has also been predicted to lead to the evolution of trait divergence but this prediction has not been tested (Schluter 2000b; Nosil 2012).  My dissertation uses an interdisciplinary approach combining field measurements, modern genetic techniques, and selection experiments to explore the impact that biotic selection from a single species has on the evolution of another species. I focus primarily on   2 the evolution of trait divergence in the threespine stickleback (Gasterosteus aculeatus) in response to intraguild predation.  1.2 The Study System The threespine stickleback is a small fish that is common in marine and freshwater habitats throughout much of the Northern Hemisphere. A striking characteristic of stickleback is the frequent parallel evolution among isolated freshwater populations in similar environments (McKinnon and Rundle 2002).  Lakes in southwestern British Columbia were created following the end of the last ice age, approximately 10,000 years ago, and were colonized by marine stickleback at that time (Bell and Foster 1994). In six of these lakes, a benthic and limnetic species of stickleback occur in sympatry in the same lake (Schluter and McPhail 1992; Gow et al. 2008). The remaining lakes contain only a single species of stickleback (McPhail 2007). Chapters 2-4 of thesis will focus on freshwater populations from lakes with a single species of threespine stickleback and chapter 5 will address populations containing two stickleback species.  An advantageous feature of lake stickleback is that the modern marine population is probably roughly equivalent to the original colonizing population. Comparisons between marine and freshwater stickleback can help us understand the direction of trait evolution in these populations. Furthermore, the Pacific Ocean marine population has a large population size and serves as a reservoir of standing genetic variation (Schluter and Conte 2009; Hohenlohe et al. 2010; Jones et al. 2012a). The recent origin of the freshwater populations means that little time has passed to allow for the origin and spread of new mutations and   3 as a result, standing genetic variation from the marine stickleback that colonized these populations is the likely source of most of the adaptive variation.  Lakes containing freshwater stickleback can vary in fish community composition (McPhail, 2007). Therefore, we can isolate the effect of intraguild predation on threespine stickleback by comparing stickleback between similar lakes that differ by the presence or absence of prickly sculpin (Cottus asper), an intraguild predator of freshwater stickleback. Sculpin are bottom dwelling fish that live in the littoral zone and lack a swim bladder (McPhail 2007). Sculpin consume benthic invertebrates, but once sculpin reach 70mm, other species of fish, including threespine stickleback, become a component of their diet (McPhail 2007). Predation on stickleback therefore provides a dual reward for sculpin: the direct benefit of a meal, and the indirect benefit of reduced competition.   1.3 Intraguild Predation and Threespine Stickleback Intraguild predation occurs when a predator kills and eats a species that is a potential competitor for shared resources (Polis et al. 1989). Intraguild predation is predicted to have a more complex effect on population dynamics than predation or competition alone (Polis et al. 1989; Holt and Polis 1997). Continual coexistence of intraguild predators (IG-predators) and intraguild prey (IG-prey) relies on several conditions: an IG-prey species should be a superior competitor on shared resources or should shift its niche in the presence of the IG-predators, and anti-predator defences may be important for IG-prey survival (Polis et al. 1989; Holt and Polis 1997; Daugherty et al. 2007; Kratina et al. 2010). As a result, intraguild predation may lead to changes in IG-prey phenotype for traits affecting competition and defence.    4 We predict that intraguild predation by prickly sculpin will cause evolutionary shifts in stickleback traits related to anti-predatory defence and foraging. The role of anti-predator defences has not been investigated in these populations. However, a previous study provided preliminary evidence that stickleback increased the proportion of zooplankton in their diet when sculpin are present. In a mesocosm experiment, a population of stickleback sympatric with sculpin consumed more zooplankton than stickleback from a population without sculpin (Ingram et al. 2012). In the wild, stickleback from lakes with sculpin have shifted to limnetic-like body shape compared to stickleback from lakes without sculpin (Ingram et al. 2012). A limnetic body shape has been correlated with increased feeding upon zooplankton in the open water (Willacker et al. 2010). However, these trait differences may be induced by the presence of sculpin and caused by phenotypic plasticity, not evolution.   I carried out a series of studies to test if intraguild predation has lead to the evolution of trait divergence between stickleback populations that occur with and without sculpin.   1.4 Summary of Studies In chapter 2, I establish that stickleback from populations sympatric with sculpin have genetically based trait differences. I collected stickleback specimens from lakes with and without sculpin and measured defensive armour and behaviour in wild samples. To determine if differences among populations have a genetic basis or are induced by the presence of sculpin (phenotypic plasticity), I raised stickleback from lakes with and without sculpin, as well as marine stickleback, in a common garden in the lab. To examine if the   5 presence of sculpin induced trait changes, each stickleback family was split in half and raised in the presence or absence of sculpin. For stickleback reared in the common garden, I measured armour, body shape, and behaviour and compared these traits among half-families that were raised in the control or sculpin treatment. Selection acts on phenotype, but a shift in phenotype is mediated through the evolution of genes that underlie those phenotypes. The effect that biotic selection has upon an organism’s genome is largely unknown for wild populations. In Chapter 3, I used whole genome re-sequencing to investigate the extent of genetic differentiation between stickleback from populations with and without sculpin. I developed a genome scan metric (CS’) to identify the regions of the genome that have differentiated in parallel between lakes with and without sculpin, and to quantify the strength of this divergence. I investigated the genome architecture of divergence between lakes with and without sculpin by calculating the number of genes and potentially the number of selective sweeps identified in regions that are differentiated among population types. I looked for overrepresentation of gene ontology (GO) terms for genes in outlier regions. Chapter 4 tests if prickly sculpin are an agent of selection on the length of stickleback pelvic spines. In chapter 2, I discovered that stickleback populations sympatric with sculpin had longer pelvic spines than stickleback populations without sculpin. The stickleback pelvis has been hypothesized to be an anti-predator defence against piscivorous predators. Increased insect predation on stickleback with longer pelvic spines has been proposed as an alternative hypothesis for the variation in pelvic spine length. I carried out a mesocosm experiment to determine if sculpin preferentially consumed stickleback with shorter pelvic spines. I physically modified the length of the pelvic spines of stickleback   6 from two populations sympatric with sculpin and then compared the mortality rate of stickleback with clipped and unclipped pelvic spines experimentally in the presence of sculpin. To evaluate the predation hypotheses, I used a meta-analysis approach to combine the results of the mesocosm experiment with other experimental studies of selection on stickleback pelvic morphology from fish and insect predators.  Finally, in chapter 5 I extend my analysis to examine if biotic selection by another species is a mechanism of divergent selection on behaviour in benthic and limnetic stickleback. Limnetic stickleback primarily live in the open water and eat zooplankton, while benthic stickleback consume macroinvertebrates in the littoral zone (Schluter and McPhail, 1992). The two species have diverged in shoaling preference and preferred position in the water column (Larson, 1976; Vamosi and Schluter, 2002; Wark et al., 2011). This behavioural divergence has been hypothesized to be the result of divergent selection driven in part by differential predation from coastal cutthroat trout (Oncorhynchus clarkii clarkii) on limnetics in the open water. To experimentally test this hypothesis, benthic-limnetic hybrids families were split and raised in large experimental ponds in a predation treatment with trout, or in a control treatment without trout. I measured shoaling preference and preferred position of the stickleback hybrids in the water column using the behavioural tests developed in chapter 2.      7 Chapter 2: Intraguild Predation Leads to Genetically Based Character Shifts in the Threespine Stickleback1  2.1  Introduction Interspecific resource competition can lead to increased phenotypic diversity as natural selection favours character shifts that decrease competition and promote the use of alternative resources (Schluter 2000a,b; Pfennig and Pfennig 2010; Stuart and Losos 2013). Other trophic interactions may also lead to divergence between closely related species (Schluter 2000b; 2003; MacColl 2011; Nosil 2012). Experimental studies have verified that divergence in traits in response to predation (Endler 1980; Vamosi 2002; Langerhans et al. 2007; Marchinko 2009) and parasitism (Hudson and Greenman 1998; Laine 2009) have a genetic basis. Intraguild predation has been predicted to increase phenotypic diversity between lineages (Schluter 2000b), but the evolution of character shifts in response to intraguild predation has not been tested. Intraguild predation occurs when a predator is also a competitor of its prey species (Polis et al. 1989; Holt and Polis 1997; Hart 2002; Arim and Marquet 2004). Competition from intraguild predators (IG-predators) can shift the diet of intraguild prey (IG-prey) to include alternative food sources (Polis et al. 1989; Vance-Chalcraft et al. 2007; Ingram et al. 2012). Simultaneously, predation from IG-predators can result in increased anti-predator behaviour and defensive structures of IG-prey (Polis et al. 1989; Kratina et al. 2010; 1 A version of this chapter has been published: Miller SE, Metcalf D, Schluter D. (2015) Intraguild predation leads to genetically based character shifts in the threespine stickleback. Evolution, 69:3194-3203.   8 Walzer and Schausberger 2013; Vanak et al. 2013), as well as behavioural shifts to alternative habitats to reduce predation on these same prey (Donadio and Buskirk 2006).  We investigated the evolution of character shifts in freshwater threespine stickleback (Gasterosteus aculeatus) in response to an intraguild predator. Freshwater populations formed when marine or anadromous (hereafter, “marine”) stickleback became isolated in numerous lakes at the end of the last ice age, approximately 12,000 years ago. These populations adapted rapidly to freshwater in isolation from each other and from the marine environment (Bell and Foster 1994). A subset of these lakes was also colonized by prickly sculpin (Cottus asper) (Dennenmoser et al. 2015), a freshwater teleost fish and intraguild predator of the threespine stickleback (McPhail 2007). Sculpin grow to larger size than stickleback and consume juvenile and adult stickleback up to 60% of their body length (Reimchen 1994; McPhail 2007). Prickly sculpin are cryptic ambush predators of stickleback and they also eat benthic invertebrates (McPhail 2007).  Preliminary evidence indicates that intraguild predation has led to phenotypic changes in stickleback that decrease competition and/or predation from sculpin. In the wild, stickleback from lakes with sculpin show a shift to a limnetic-like body shape. In contrast, stickleback from lakes without sculpin are more benthic-like with a deeper body, a wider caudal-peduncle, and a posterior shift in the first dorsal spine (Ingram et al. 2012). Differences in stickleback body shape correlate with diet (Willacker et al. 2010). In a mesocosm experiment, stickleback from a population sympatric with sculpin had more zooplankton in their diet than stickleback from a population without sculpin, whose diet consisted of more benthic prey. When sculpin were experimentally added to mesocosms, stickleback from both populations increased the proportion of zooplankton consumed   9 (Ingram et al. 2012). The addition of sculpin also increased stickleback mortality and reduced growth rate, but to a lesser extent in the stickleback population sympatric with sculpin than stickleback from the sculpin-absent lake (Ingram et al. 2012), suggesting that they are less susceptible to predation. Differences between populations in other traits such as armour and behaviour are likely, but have not been measured. The presence of predators is often associated with greater defensive armour in stickleback (Reimchen 1994; Vamosi and Schluter 2004; Willacker et al. 2010; Leinonen et al. 2011; Lescak and von Hippel 2011; Lacasse and Aubin-Horth 2012), as well as differences in sociality and shoaling (Vamosi 2002; Bell and Sih 2007; Dingemanse et al. 2007; 2009; Lacasse and Aubin-Horth 2012). Longer spines increase the body diameter of the stickleback, making them more difficult for gape-limited predators to ingest (Hoogland et al. 1956) and lateral plates provide structural support for spines (Reimchen 1983). Increased zooplankton in the diet suggests greater use of the water column by stickleback from lakes with sculpin, which may decrease the rate of encounter (Lima and Dill 1990). One approach to testing evolutionary character shifts in IG-prey is to ask whether putative cases fulfill criteria analogous to those routinely used to test for ecological character displacement (modified from Schluter and McPhail 1992): (1) Phenotypic differences have a genetic basis; (2) Differences are not due to chance; (3) Divergence should be the outcome of evolution rather than species sorting; (4) Shifts in phenotype reflect differences in resource use and/or predation risk; (5) Shifts are not the result of other environmental differences between sites with and without IG-predators; and (6) There is independent evidence that pre-shift IG-prey phenotypes compete with and suffer predation from the IG-predator.   10 Here we evaluate the first criterion. Character shifts in response to intraguild predation might be the result of either phenotypic plasticity or genetic change (West-Eberhard 2003). Plasticity can lead to rapid character shifts because the match of phenotype to environment occurs without waiting for the spread of adaptive alleles (West-Eberhard 2003; Schlichting and Pigliucci 1998). Adaptive phenotypic plasticity in IG-prey behaviour (Heithaus 2001; Janssen et al. 2007; Amarasekare 2008) or inducible anti-predator defences (Urbani and Ramos-Jiliberto 2010; Kratina et al. 2010; Nakazawa et al. 2010) have been shown to increase survival of IG-prey in theoretical models. Alternatively, genetic mapping studies based on crosses between marine and freshwater stickleback populations have found different QTL between populations inhabiting lakes with and without sculpin associated with body shape differences and armour components (Rogers et al. 2012), suggesting that many trait differences between the population types have at least a partial genetic basis. Distinguishing between phenotypic plasticity and genetic evolution is also important for predicting community dynamics (Cortez 2011; Yamamichi et al. 2011). However, experimental studies are required to test whether character shifts have a genetic basis (Scheiner 1993). In this study, we describe character shifts in body armour, body shape, and behaviour among natural populations of stickleback that occur with and without prickly sculpin. Stickleback were raised in a common garden to determine the relative role of genetics and phenotypic plasticity in these shifts. We assessed the inducibility of these traits by rearing split families in the presence and absence of sculpin. We included marine stickleback in the experiment to determine if phenotypic plasticity was present in the form representing the ancestral state. If prickly sculpin have led to the evolution of character   11 shifts in stickleback, individuals raised in a common garden will replicate the phenotypes of the parental populations and the presence of sculpin will not induced trait shifts.  2.2 Materials and Methods 2.2.1 Study Populations and Sample Collection Lake characteristics and information on fish community composition were obtained from Habitat Wizard (www.env.gov.bc.ca/habwiz) maintained by the British Columbia Ministry of Environment. We identified eight lakes (8.0-58.7 ha) in southwestern British Columbia with a simple fish community of threespine stickleback, coastal cutthroat trout (Oncorhynchus clarkii clarkii), and prickly sculpin (Cottus asper) and contrasted these populations with eight lakes (3.7-44.6 ha) containing only threespine stickleback and trout (Figure 2.1). Cutthroat trout are ubiquitous in lakes in this region. All lakes are in separate watersheds, ensuring no gene flow between populations. Lakes with and without sculpin did not differ in mean area (Mann-Whitney test, U=18, P=0.16), perimeter (U=16, P=0.10), mean depth (U=25, P=0.77), elevation (U=34, P=0.88), or distance to the ocean (U=28, P=0.72). The study populations also included “marine” stickleback from two geographically distinct populations. Modern marine (including anadromous) stickleback are thought to be phenotypically similar to the ancestral populations that initially colonized the freshwater lakes following the last ice age (Bell and Foster 1994). Marine stickleback have a diverse and largely uncharacterized predator community including several species of marine sculpin (McPhail 2007).  Adult stickleback were collected in May-June 2011 and 2012 using 10-15 baited minnow traps placed overnight along the shoreline of each lake and at the marine sites.   12 Specimens collected for morphological analysis from all populations (Table S1; n=7-26/population) were euthanized using buffered MS-222 (Argent Chemical Laboratories, Redmond, WA) and preserved in 95% ethanol. Some sites were sampled in subsequent years to increase sample size. Additional adult stickleback were collected for behavioural experiments in 2011, but a sufficient number of specimens was only available for seven lakes without sculpin, three lakes with sculpin and one marine population (n=12-27/population). Stickleback were transported to the aquatics facility at the University of British Columbia and allowed to acclimatize for one week prior to behavioural trials. In 2012, adult stickleback in reproductive condition were collected from three lakes with sculpin (Ambrose, Paq, and Rosseau), three lakes without sculpin (Trout, Cranby and Kirk) and from two marine populations (Oyster Bay and Little Campbell) for a common garden and plasticity experiment (Figure 2.2). Sculpin were collected from Paq Lake at this time and were transferred to the aquatics facility.  2.2.2 Common Garden and Plasticity Experiment We raised stickleback in a common garden laboratory environment in the presence and absence of sculpin. We created four families from each population by artificially crossing pairs of wild-caught fish at the lakeside. Eggs were obtained by gently pressing on the sides of females and placed into lake water. Males were euthanized with an overdose of MS-222 and testes were dissected, minced, and added to the eggs. We made reciprocal F1 crosses between stickleback from a lake with sculpin (Paq) and a lake without sculpin (Trout) to test for maternal effects on phenotypes. Paq and Trout lake populations have divergent body shape (Ingram et al. 2012) but are less differentiated in armour (Table 2.1).   13 Four crosses used females from Paq Lake and males from Trout Lake and four crosses used females from Trout Lake and males from Paq Lake.  Fertilized eggs were transferred to the University of British Columbia within 24 hours. At that time, each clutch was split. Half the eggs were assigned to a sculpin treatment and half to a control treatment. Each 100L experimental tank was divided in the center with window screen and contained three kilograms of coarse limestone gravel and 1ppt sodium chloride. Each half clutch was added to one side while the other side was left empty. A low concentration of methylene blue was added to inhibit fungal growth. Tanks were kept at 16L:8D photoperiod. One Ambrose clutch, one Rosseau, and two F1 clutches did not hatch (Figure 2.2).  The development of induced defences may depend upon the timing of exposure to the stimulus (Harvell 1990). Limited evidence suggests that even stickleback embryos can change behaviour in response to cues from trout predation (Golub 2013). Because we were uncertain of the stage at which exposure to sculpin might lead to induced defences, we provided sculpin cues for the duration of the experiment, from fertilized eggs until nine months of age. To provide possible olfactory cues, daily we added a 50ml aliquot of water from a tank containing four adult sculpin to unhatched eggs in each sculpin treatment tank. This continued until stickleback hatched and fry were four weeks of age. Dechlorinated water was added to the control treatment during this time. At four weeks post-hatching, stickleback fry were too large to pass through the window screen dividing each tank and were gathered and moved to a random side of the tank. At that time, we reduced the number of fry to 20 per tank. In the case of half clutches with fewer than 20 fry we reduced the number of fry to an equal density in the control and sculpin treatment tanks.   14 In the sculpin treatment, a single adult sculpin was added to the other side of the tank. In the control treatment, an equal biomass (four fish) of adult stickleback was added. Adult stickleback were F2 hybrids between Paxton Lake benthic and limnetic stickleback that had been raised in the laboratory for an unrelated study. The window screen dividing each tank allowed experimental stickleback to receive constant visual and chemical cues from the sculpin or the adult stickleback. Stickleback were fed hatched brine shrimp nauplii for the first four months, and then a mixture of brine shrimp and bloodworms for the remainder of the experiment. Adult stickleback in the control treatment were fed a 3:1 mixture of bloodworms and Mysis shrimp to satiation daily. Sculpin do not eat Mysis shrimp and were fed only bloodworms.  The experiment was stopped at 36 weeks post-hatching. Several adult control stickleback died during the experiment and were immediately replaced upon discovery. There were no sculpin mortalities. A Rosseau family was excluded from analysis after a sculpin jumped to the other side of the tank and consumed the experimental stickleback. The final sample size was 35 families in 70 tanks.   2.2.3 Morphology Samples stored in 95% ethanol were rehydrated, fixed in 10% formalin, and stained with alizarin red to highlight bony characteristics following standard procedures (Peichel et al. 2001). We measured standard length, gape width, first and second dorsal spine length, pelvic spine length, pelvic girdle length, and lateral plate number on both wild-caught and experimental stickleback (Figure 2.3). Spine measurements were made on the left side of the fish using digital calipers. Lateral plates were counted under a dissecting microscope. All   15 armour traits were log(x+1) transformed to homogenize variance. Experimental stickleback smaller than 28mm were excluded from analysis because the development of lateral plates may be incomplete in smaller stickleback (Hagen 1973; Bell 2001; Rennison et al. 2015). All wild-caught stickleback were >28mm. To compare traits among stickleback of different sizes, all traits except lateral plates were size-adjusted to the mean standard length of the wild-caught samples (46.3 mm). For each trait, we fit a linear model with standard length as a covariate and population as a factor. All measurements were adjusted using the residuals from each regression (Vamosi 2002). The wild-caught samples were size-corrected separately from the common garden stickleback.  To minimize trait redundancy, stickleback armour variation was summarized with the first principal component (PC1) based on the correlation matrix between size corrected spine traits and lateral plates, separately for wild-caught and experimental stickleback. All armour traits had significant positive loadings on PC1, which accounted for 74.6% and 79.7% of the variance in wild stickleback and lab-raised stickleback (Table 2.2). Principal Component 1 was the only principal component with an eigenvalue greater than one therefore PC2-5 were not examined further.  We examined body shape in the experimental stickleback. The left side of each stickleback was photographed using a Nikon D300 camera. We placed 20 landmarks outlining the shape of the fish as well as the insertion points of spines and fins (Figure 2.3; Walker 1997; Ingram et al. 2012). Landmarks were digitized using tpsDig 2.16 software (Rohlf 2008) and were centered, scaled, and rotated using the shapes package (Dryden 2012) in the R 3.0 environment (R Core Team, 2014). We performed a linear discriminant   16 analysis (LDA) with the MASS package (Venables and Ripley 2010) to visualize shape differences among lakes. We used the tank (half-family) as our classification variable, and thus the LDA did not a priori differentiate between treatment or population type. An LDA was preferable to other types of multivariate methods such as a principal components analysis because it ignores trait combinations that vary only within populations (Tabachnick and Fidell 2012) and those resulting from measurement error or specimen bending. The first and second linear discriminant axes (LD1 and LD2) accounted for 34.3% and 15.9% of the observed variation in shape among half-families.  2.2.4 Stickleback Behaviour We used a behavioural assay to measure position in the water column preferred by stickleback. Vertical position in the water column is a proxy for habitat use in guppies (Torres-Dowdall et al. 2012) and a lower position in the water column correlates with increased anxiety behaviour in zebrafish (Danio rerio, Egan et al. 2009; Cachat et al. 2010; Stewart et al. 2012). Limnetic stickleback from Paxton Lake prefer to be higher in the water column than benthic stickleback (Larson 1976).  Wild-caught stickleback in non-reproductive condition were transferred from their home tank to a holding basket next to the assay. Although Cachat et al. (2010) recommends a 1-hour recovery period, preliminary trials showed that a 15-minute acclimation period was sufficient. At the start of each trial, a focal fish was gently introduced to the top of an unfamiliar tank and was allowed to move freely for 330 seconds (Figure 2.4). The first 30 seconds of each trial were not analyzed, because the introduction of the stickleback into the tank often resulted in erratic movement. Trials were recorded   17 and videos were subsampled to 0.5 frames per second using VirtualDub (www.virtualdub.org). The x and y coordinate position of the focal fish was measured every two seconds using MtrackJ (Meijering et al. 2012) in ImageJ (Schneider et al. 2012). For each trial we calculated the mean vertical position and the total movement of the stickleback in pixels (distance traveled).  The water column height preference of the experimental stickleback was assayed at 28-31 weeks of age using ten stickleback chosen at random from each experimental tank. Tanks containing the same family were tested in the sculpin and control treatments sequentially in random order. We further characterized the behaviour of experimental stickleback using a shoaling assay (Vamosi 2002; Kozak and Boughman 2008; Wark et al. 2011). A 100L tank was divided into two end compartments and one center arena using window screen (Figure 2.4). An experimental shoal of 10 unfamiliar stickleback was added to one end and two stickleback were added to the other end (Wark et al. 2011). A focal fish was introduced into the center arena and its distance to the stimulus shoal arena was used as a measure of shoaling preference. Shoaling assays were conducted two days after the water column preference assay using ten randomly chosen stickleback from each experimental tank. Stimulus stickleback were chosen from a stock of laboratory reared Priest Lake benthic stickleback. The stimulus population was selected because stickleback were similar in age and size to the experimental fish and were unrelated to all of the experimental populations. Experimental stickleback were moved to holding baskets near the shoaling assay for a fifteen-minute acclimation period. At the start of each trial, the focal stickleback was introduced into the center arena. Trials were recorded for 630 seconds and the first 30   18 seconds of each trial were not analyzed. The x and y coordinate position of the focal fish was calculated as described above. For each trial, we calculated the time spent within one body length of the stimulus shoal as well as the distance traveled. 463 shoaling videos were scored (Table 2.3).  2.2.5 Statistical Analysis  We tested for differences in mean trait values between wild-caught stickleback from lakes with sculpin and lakes without sculpin using linear models. Tests involving freshwater fish from the common garden used the tank mean as the unit of replication because each half-family was raised in the same tank. The experiment was analyzed using a linear mixed effects model with treatment (sculpin or control), population type (from a lake with or without sculpin) and their interaction as fixed factors and lake and family as random factors. Inducibility in the marine population was assessed in a separate analysis using a linear mixed effects model with treatment (sculpin or control) as a fixed factor and lake and family as random factors. Maternal effects were tested by comparing F1 crosses raised without sculpin using direction of cross as a fixed factor and family as a random factor.  2.3 Results 2.3.1 Character Shifts in Wild-Caught Stickleback The presence of sculpin was associated with character shifts in armour and behaviour in wild populations of stickleback. Stickleback from lakes with sculpin had higher mean armour PC1 scores than stickleback from lakes without sculpin (Figure 2.5A; F1,14= 33.9, P<0.001). All individual armour traits were greatest in stickleback from lakes with   19 sculpin (Table 2.3). There was no difference in gape width (F1,14= 3.01, P=0.11) or standard length (F1,14= 1.86, P=0.19) between lakes with and without sculpin.  Stickleback from lakes with sculpin also preferred a higher mean vertical position in the water column than stickleback from lakes without sculpin (Figure 2.5B; F1,8= 8.0, P=0.02). Distance traveled was not different between stickleback from the two types of lakes (F1,8= 0.12, P=0.73).  2.3.2 Character Shifts Persisted in a Common Garden Common garden stickleback from lakes with sculpin had a higher mean armour PC1 score than populations from lakes without sculpin (Figure 2.6, filled circles; F1,4= 12.5, P=0.047). Individual armour traits were similar between stickleback raised in the control treatment of the common garden and values of wild caught stickleback from the same lake (Table 2.4). Exposure to sculpin did not induce a detectable change in PC1 score (Figure 2.6, open circles; Treatment: F1,19= 0.17, P=0.38; Treatment x Population Type: F1,19= 0.41, P=0.53). Stickleback from lakes with and without sculpin differed in mean body shape in the common garden (Figure 2.7, filled circles). Lakes with sculpin were significantly differentiated in both LD1 (F1,4= 13.2, P=0.022) and LD2 (F1,4= 31.1, P=0.005). Stickleback from lakes with sculpin had an anterior shift in first dorsal spine, decreased body depth, a narrower caudal-peduncle, larger eye diameter and a longer jaw. Exposure to sculpin did not induce a detectable difference in mean shape (LD1: F1,19= 0.0, P=0.995; LD2: F1,19= 0.26, P=0.62; all treatment x population type interactions were non-significant, P>0.1).   20 Common garden stickleback from lakes with and without sculpin also differed in behaviour. As we saw in wild-caught stickleback, lab-raised stickleback from lakes with sculpin preferred a higher mean position in the water column (Figure 2.8, filled circles; type: F1,4= 16.1, P=0.016). Stickleback from lakes with and without sculpin traveled a similar distance during the trials (type: F1,4= 0.8, P=0.41). In the shoaling assay, stickleback from lakes with sculpin spent less time near the stimulus shoal (decreased shoaling preference) than stickleback from lakes without sculpin (Figure 2.9, filled circles; F1,4= 18.1, P=0.013). Population types traveled a similar distance during the shoaling assay (F1,4= 0.9, P=0.39). Exposure to sculpin did not detectably alter any behaviour (water column position: F1,17= 0.1, P=0.76; water column distance: F1,17= 0.4, P=0.54; shoaling preference: F1,16= 2.6, P=0.13; shoaling distance: F1,16= 1.4, P=0.25; all treatment x population type interactions were non-significant, all P>0.1).   2.3.3 Sculpin Exposure Induced Character Shifts in Marine Stickleback Phenotypic plasticity was observed in several traits in marine stickleback. Marine stickleback raised in the sculpin treatment had higher armour PC1 scores than those raised in the control treatment (Figure 2.6; F1,7= 6.7, P=0.016). Adding the category “marine” as a population type to our previous analysis of experimental populations from lakes resulted in a significant treatment x population type interaction (PC1: F1,27= 5.65, P=0.025), hinting that the marines are more plastic than the freshwater populations. Body shape did not differ between treatments (LD1: F1,7= 0.1, P=0.81; LD2: F1,7= 0.1, P=0.76). In the water column preference assay, marine stickleback from the sculpin treatment showed a marginal but non-significant increase in mean water column position (Figure 2.8; F1,7= 4.5, P=0.07;   21 treatment x population type: F1,25= 3.39, P=0.08), and in the shoaling assay, marines in the sculpin treatment had a decrease in shoaling tendency (Figure 2.9; F1,5= 8.0, P=0.037) and a significant treatment x population type interaction (F1,22= 6.77, P=0.016).  2.3.4 Maternal Effects Armour traits in F1 hybrids between Trout Lake and Paq Lake stickleback were intermediate between the parental populations (Figure 2.6) and direction of cross did not affect trait value (PC1: F1,4= 8.0, P=0.11). Similarly, overall body shape was intermediate between the parental populations, but F1 families with Trout Lake mothers (without sculpin) had a larger mean LD1 score (LD1: F1,4= 9.5, P=0.037), than F1 families with Paq Lake mothers (with sculpin), indicating that maternal effects may impact body shape in these populations. There was no difference in LD2 (F1,4= 1.4, P=0.3).  2.4 Discussion 2.4.1 Trait Shifts in Response to Intraguild Predation  The presence of an IG-predator, prickly sculpin, is associated with character shifts in multiple traits in the threespine stickleback, and the results herein indicate that these trait shifts have a genetic basis. Wild populations of stickleback sympatric with sculpin show parallel increases in armour morphology, prefer to be higher in the water column, and have been previously shown to differ in body shape (Ingram et al. 2012). These differences in armour, shape, and behaviour persisted in a common garden. To our knowledge, this system is the first confirmed case of genetically based character divergence associated with intraguild predation.    22 Competition, predation, or both might produce character shifts in response to intraguild predation and disentangling these interactions will be challenging. Piscivorous predators have previously been associated with longer spines and an increased number of lateral plates in stickleback (Hagen and Gilbertson 1972; Moodie 1972; Bell et al. 1993; Reimchen 1994; Reimchen and Nosil 2002; Baker et al. 2010; Leinonen et al. 2011, Lescak and von Hippel 2011). Increased armour in lakes with sculpin might be a response to increased predation, though number of lateral plates might also affect buoyancy (Myhre and Klepaker 2009) and drag (Walker 1997). Alternatively, it is possible that shifts in armour are the indirect outcome of a habitat shift between sculpin and stickleback. Prickly sculpin prefer the littoral zone of lakes where there is easy access to cover and benthic invertebrates (McPhail 2007). Sculpin may displace stickleback into the pelagic environment either by decreasing benthic resources, increasing the threat of predation, or both. Because coastal cutthroat trout are more prevalent in the open water (Reimchen 1994), longer spines might be an adaptation to increased predation from trout, rather than a direct response to predation by sculpin. A third, less plausible, hypothesis is that sculpin predation on benthic invertebrates indirectly relaxes selection for reduced spines. Juvenile stickleback are eaten by large aquatic insects and studies suggest that some insects capture stickleback by grabbing the spines (Reist 1980; Reimchen 1980; Marchinko 2009; although see Lescak et al. 2012 and Mobley et al. 2013). Spine length might represent a balance between selection for longer spines by gape-limited predators and selection for shorter spines via predation by aquatic insects upon juveniles stickleback (Reimchen 1980).  Similarly, trait shifts in behaviour could also be attributed to either competition or predation. We found that in the wild and in the lab, stickleback from lakes with sculpin   23 preferred to be higher in the water column. A position higher in the water column might lessen risk of predation from sculpin. We also observed a decreased shoaling preference in stickleback from lakes with sculpin and in marine stickleback reared in the sculpin treatment. Sculpin are ambush predators, therefore shoaling may not be an effective method for escaping sculpin predation. Alternatively, differences in water column and shoaling preference may be a response to changes in foraging behaviour caused by resource depletion by sculpin. The presence of sculpin has been demonstrated to induce a higher proportion of zooplankton in the stickleback diet (Ingram et al. 2012), and zooplankton is most abundant in the open water. Trait shifts in behaviour could also interact with shifts in morphological traits. For example, diet preference and body shape vary with lateral plate number (Bjaerke et al. 2010). Intraguild predation may independently select for trait shifts in behaviour and morphology, or changes in behaviour may have led to selection for changes in morphology (or vice versa). These alternatives underscore the challenge of elucidating the relative impacts of competition, predation, and their interactions in character shifts via intraguild predation.  2.4.2 Trait Inducibility has been Lost in Freshwater Populations Phenotypic plasticity has been proposed as a possible explanation for trait shifts in IG-prey (Urbani and Ramos-Jiliberto 2010; Kratina et al. 2010; Nakazawa et al. 2010). Although adaptive plasticity has been reported in stickleback feeding morphology (Day et al. 1994; Day and McPhail 1996; Wund et al. 2008; Svanbäck and Schluter 2012) and body shape (Garduño-Paz et al. 2010; Svanbäck and Schluter 2012) we found no evidence for sculpin-induced plasticity in freshwater populations. However, marine stickleback reared in   24 the presence of sculpin exhibited slightly increased armour, an increase in preferred water column height, and a decrease in shoaling behaviour compared to the controls. To our knowledge, the increased armour in marine stickleback in the presence of sculpin is the first observation of induced structural defences in stickleback. Importantly, induced trait changes in the presence of sculpin were in the same direction as the trait shifts among freshwater stickleback populations with and without sculpin. Phenotypic plasticity in the ancestral colonizing population may have aided in the initial divergence between freshwater populations (Wund et al. 2008).  It should be noted that while stickleback in the sculpin treatment received lifelong visual and olfactory cues from sculpin, they were not exposed to predation. Stickleback in this treatment might not have recognized sculpin as a threat or constant exposure to sculpin may have resulted in habituation (Kelley and Magurran 2003). All behavioural assays were conducted without sculpin, and including sculpin cues during these assays might induce a change in behaviour.   This study provides evidence that intraguild predation leads to evolutionary divergence among stickleback populations (Schluter and McPhail 1992). Phenotypic differences between lakes with and without sculpin have a clear genetic basis. Character shifts have likely occurred in parallel across replicated populations, therefore these differences are not due to chance. Preliminary comparisons found no evidence of consistent environmental differences among lakes. However, the biotic and abiotic environment can influence species interactions and affect the structure of piscivorous communities (Jackson et al. 2001). To fully rule out the role of the environment in generating these evolutionary shifts will require further investigation of abiotic characteristics (e.g. pH, vegetation), and   25 the biotic community (e.g. aquatic insects, avian predators). Phenotypic differences between lakes with and without sculpin suggest that stickleback have evolved in response to competition and/or predation with sculpin.      26 Figure 2.1 : Map of sampling locations used in the chapter. Lakes 1-8 contain only stickleback, A-H indicates lakes that contain stickleback and sculpin. M1 and M2 are marine populations. The lakes are (1) Kirk, (2) Cranby, (3) Klein, (4) Trout, (5) Hoggan, (6) Bullocks, (7) Blackburn, (8) Stowell, (A) Cedar, (B) Ormond, (C) Pachena, (D) Rosseau, (E) Paq, (F) Ambrose, (G) North, (H) Brown, (M1) Little Campbell, (M2) Oyster Bay    484950510 50 100 kmVancouverIslandM1M2F HAGBCED 76251438N  27 Figure 2.2 : Schematic of crosses used in the common garden experiment. Four crosses were created for each lake. One Ambrose clutch, one Rosseau, and two F1 clutches did not hatch. An additional family from Rosseau Lake was excluded when a sculpin consumed the experimental stickleback.   Lakes&with&Sculpin&Lakes&without&Sculpin&Marine&Popula5ons&Paq& Trout& Kirk& Cranby& Oyster& Li=le&Campbell&Ambrose& Rosseau&F1&Cross&TYPE%LAKE%FAMILY%Paq&male&Trout&female&&Trout&male&Paq&female&&2&families&3&families& 4&families& 4&families& 4&families&4&families&4&families&4&families&4&families&2&families&  28 Figure 2.3 : Landmarks coordinates used for morphometrics Individual landmarks are indicated with numbers. Armour traits are abbreviated as follows: First dorsal spine (DS1), second dorsal spine (DS2), pelvic spine (PS), pelvic girdle (PG), and lateral plates (LP).        29 Figure 2.4 : The set-up for behavioral assays. (A) Water column preference assay tank. A focal stickleback is introduced into an unfamiliar 15L tank. Vertical position and distance traveled were measured. (B) Shoaling Assay Tank. A standard 100L aquarium tank was separated into a centre arena and two end compartments using window mesh (dotted outline). The tank was filled with 32cm of water. The back and sides of the assay tanks were covered with white paper to reduce external visual cues, and tanks were backlit to increase the contrast between the focal fish and the background. An experimental shoal with 10 stickleback was introduced into one end compartment and two stickleback were added to the other end compartment. A focal fish was introduced to the center arena and horizontal position and distance traveled were measured. All trials were recorded using wireless cameras (D-link DSC-932L).    30         31 Figure 2.5 : Character shifts in wild populations of stickleback. (A) Mean and standard error of armour PC1. Each point represents a single population. The solid horizontal lines give the means for each type of population. (B) Mean and standard error of vertical position in the water column (y-axis position) of wild caught stickleback measured in the lab in an unfamiliar tank. Each point is a single population. Horizontal lines give the means of each population type.       -3-2-10123Armour PC1 (74.5%)Lakes Lakeswithout withSculpin SculpinMarineA0.00.51.01.52.02.53.03.5Vertical Tank PositionLakes Lakeswithout withSculpin SculpinMarineB  32 Figure 2.6 : Mean armour PC1 for experimental stickleback from each family raised in the control common garden (filled) and the sculpin treatment (open). The F1 is a cross between fish from Trout (sculpin absent) and Paq (sculpin present) Lakes. The father is first and the mother is second for F1 crosses. The mean and standard error of each lake and treatment is given on the left.               -3-2-10123PC1 (79.7%)Cranby Kirk Trout Paq Ambrose Rosseau L Camp Oyster Paq x Trout Trout x PaqLakes without Sculpin Lakes with Sculpin Marine F1  33 Figure 2.7 : Mean value of shape axis 1 from stickleback families reared in a common garden in a control treatment (filled) and a sculpin treatment (open). The F1 is a cross between fish from Trout (sculpin absent) and Paq (sculpin present) Lakes. For F1 crosses, the father’s population is first and the mother’s population is second. The mean and standard error of each lake and treatment is given on the left.            -4-20246Shape Axis 1 (34.3%)Cranby Kirk Trout Paq Ambrose Rosseau L Camp Oyster Paq x Trout Trout x PaqLakes without Sculpin Lakes with Sculpin Marine F1  34 Figure 2.8 : Mean vertical position in the water column in an unfamiliar tank (y-axis position) of stickleback raised in a common garden. The control treatment is represented by closed symbols and the sculpin treatment is represented by open symbols. The mean and standard error of each lake and treatment is given on the left.             678910Vertical Tank PositionLakes without Sculpin Lakes with Sculpin MarineCranby Kirk Trout Paq Ambrose Rosseau L Camp Oyster  35 Figure 2.9 : Mean time spent near the shoal. The control treatment is represented by closed symbols and the sculpin treatment is represented by open symbols. The mean and standard error of each lake and treatment is given on the left.    020406080100120Time Shoaling (s)Cranby Kirk Trout Paq Ambrose Rosseau L Camp OysterLakes without Sculpin Lakes with Sculpin Marine  36 Table 2.1 Mean and standard error of traits measured in wild caught stickleback. First dorsal spine (DS1), second dorsal spine (DS2), pelvic spine (PS), pelvic girdle (PG), and lateral plates (LP). All spine traits have been size corrected.  Lake% Type% Year%Sample%Size%Armour% DS%1%(mm)% DS%2%(mm)% PS%(mm)% PG%(mm)% LP% Armor%PC1%Sample%Size%Behaviour%Vertical%position%Blackburn* No*Sculpin* 2011* 20* 2.2*±*0.1* 3.0*±*0.1* 3.4*±*0.1* 6.3*±*0.2* 3.7*±*0.2* 91.84*±*0.14* 10* 0.87*±*0.38*Bullocks* No*Sculpin* 2011* 26* 0.8*±*0.2* 2.6*±*0.1* 2.8*±*0.1* 5.7*±*0.1* 3.3*±*0.2* 93.23*±*0.17* 0**Cranby* No*Sculpin* 2011* 19* 2.3*±*0.1* 3.1*±*0.1* 4.4*±*0.1* 1*±*0.1* 5.4*±*0.2* 90.97*±*0.1* 12* 1.72*±*0.43*Hoggan* No*Sculpin* 2011* 16* 2.3*±*0.1* 3.0*±*0.1* 3.4*±*0.1* 6.3*±*0.2* 4.7*±*0.15* 91.58*±*0.1* 11* 0.51*±*0.37*Kirk* No*Sculpin* 2011* 10* 2.8*±*0.1* 3.3*±*0.1* 3.7*±*0.1* 7.1*±*0.2* 3.1*±*0.5* 91.24*±*0.13* 15* 1.43*±*0.11*Klein* No*Sculpin* 2011* 20* 2.4*±*0.2* 3.2*±*0.1* 4.0*±*0.1* 7.3*±*0.1* 5.8*±*0.2* 90.85*±*0.13* 25* 2.95*±*0.33*Stowell* No*Sculpin* 2011* 22* 1.9*±*0.1* 2.6*±*0.1* 2.8*±*0.1* 6.7*±*0.2* 4.9*±*0.2* 92.06*±*0.1* 19* 0.88*±*0.25*Trout* No*Sculpin* 2011* 19* 3.0*±*0.1* 3.8*±*0.1* 4.4*±*0.1* 7.7*±*0.1* 4.4*±*0.2* 90.28*±*0.08* 22* 2.37*±*0.21*Ambrose* Sculpin* 2011* 19* 3.3*±*0.1* 4.0*±*0.1* 5.1*±*0.1* 8.4*±*0.1* 6.1*±*0.1* 0.40*±*0.07* 24* 2.11*±*0.30*Brown* Sculpin* 2011* 7* 3.2*±*0.2* 3.9*±*0.05* 4.7*±*0.1* 8.1*±*0.1* 6.4*±*0.2* 0.19*±*0.09* 0**Cedar* Sculpin* 2011* 17* 3.3*±*0.2* 4.2*±*0.1* 5.8*±*0.1* 8.7*±*0.1* 6.8*±*0.3* 0.77*±*0.1* 0**North* Sculpin* 2011* 7* 3.3*±*0.4* 4.2*±*0.2* 5.5*±*0.2* 8.9*±*0.4* 33.9*±*0.3* 1.69*±*0.18* 11* 2.68*±*0.49*Ormond* Sculpin* 2012* 25* 4.8*±*0.1* 5.4*±*0.1* 7.1*±*0.1* 9.7*±*0.1* 6.6*±*0.2* 2.08*±*0.07* 0**Pachena* Sculpin* 2012* 11* 3.0*±*0.1* 3.9*±*0.1* 5.3*±*0.1* 9.4*±*0.2* 6.5*±*0.2* 0.63*±*0.11* 0**Paq* Sculpin* 2011* 20* 3.3*±*0.1* 4.1*±*0.1* 5.6*±*0.1* 9.1*±*0.2* 6.3*±*0.2* 0.79*±*0.08* 15* 1.87*±*0.42*Rosseau* Sculpin* 2012* 19* 5.4*±*0.2* 6.0*±*0.2* 8.3*±*0.2* 11.1*±*0.1* 6.9*±*0.2* 2.92*±*0.08* 0**L*Camp* Marine* 2012* 11* 4.8*±*0.2* 4.8*±*0.1* 8.0*±*0.1* 11.0*±*0.1* 33.6*±*0.2* 3.26*±*0.08* 0**Oyster* Marine* 2011* 19* 3.9*±*0.1* 4.9*±*0.1* 6.5*±*0.1* 10.1*±*0.1* 27.1*±*0.5* 2.50*±*0.08* 14* 0.77*±*0.21*    37 Table 2.2: Principal component loadings for armour traits  * *WildECaught%Stickleback%*Experimental%Stickleback%* * * * * * * * * * * * ** *PC1* PC2* PC3* PC4* PC5**PC1* PC2* PC3* PC4* PC5*Trait%* * * * * * * * * * **Dorsal*Spine*1* 0.4342* 0.3085* 90.8432* 90.0731* 90.0062**0.4483* 0.2983* 90.5389* 90.6259* 0.1669**Dorsal*Spine*2* 0.4702* 0.2709* 0.2974* 0.5534* 90.5575**0.4446* 0.4918* 90.0651* 0.6865* 0.2914**Pelvic*Spine* 0.4860* 0.1598* 0.2896* 0.1534* 0.7943**0.4830* 0.0677* 0.1522* 0.0180* 90.8595**Pelvic*Girdle* 0.4706* 90.0256* 0.3034* 90.7925* 90.2403**0.4514* 90.1303* 0.7490* 90.2851* 0.3701**Lateral*Plates* 0.3644* 90.8973* 90.1572* 0.1919* 90.0221**0.4054* 90.8048* 90.3480* 0.2352* 0.1077*%    38 Table 2.3: Mean and standard error of traits measured in experimental stickleback in ‘control’ treatment. Sample size is number of families measured. First dorsal spine (DS1), second dorsal spine (DS2), pelvic spine (PS), pelvic girdle (PG), and lateral plates (LP). All spine traits have been size corrected.  Lake% Type% N% DS%1%(mm)% DS%2%(mm)% PS%(mm)% PG%(mm)% LP% Armor%PC1%Cranby* No*Sculpin* 4* 3.4*±*0.1* 4.1*±*0.1* 5.0*±*0.1* 8.6*±*0.1* 5.9*±*0.3* 14.1*±*0.5*Kirk* No*Sculpin* 4* 4.0*±*0.02* 4.6*±*0.1* 5.0*±*0.1* 7.9*±*0.2* 4.4*±*0.3* 15.3*±*0.1*Trout* No*Sculpin* 4* 4.5*±*0.03* 5.0*±*0.1* 6.2*±*0.2* 8.6*±*0.1* 4.9*±*0.1* 19.3*±*0.5*Ambrose* Sculpin* 3* 4.5*±*0.3* 5.4*±*0.3* 6.3*±*0.1* 9.0*±*0.1* 6.7*±*0.1* 19.6*±*0.8*Paq* Sculpin* 4* 4.1*±*0.1* 5.3*±*0.1* 7.3*±*0.1* 10.7*±*0.1* 7.1*±*0.2* 21.9*±*0.3*Rosseau* Sculpin* 2* 6.1*±*0.1* 6.4*±*0.1* 8.8*±*0.2* 11.7*±*0.2* 7.9*±*0.3* 27.8*±*0.7*L*Camp* Marine* 4* 5.8*±*0.1* 6.0*±*0.1* 8.9*±*0.2* 11.7*±*0.1* 33.6*±*0.2* 27.4*±*0.04*Oyster* Marine* 4* 5.6*±*0.1* 5.8*±*0.1* 8.5*±*0.2* 10.7*±*0.1* 31.8*±*1.6* 26.2*±*0.4*Paq*Male* F1* 4* 4.7*±*0.1* 5.2*±*0.1* 7.0*±*0.04* 9.8*±*0.1* 6.2*±*0.1* 21.3*±*0.4*Trout*Male* F1* 2* 4.5*±*0.1* 5.1*±*0.1* 6.7*±*0.2* 9.6*±*0.02* 6.4*±*0.02* 20.4*±*0.2*  39 Chapter 3: Intraguild Predation Leads to a Multitude of Genomic Changes but is Constrained by Genomic Architecture  3.1  Introduction The evolution of a species is governed both by the abiotic environment and by biotic interactions with other species in the environment (Thompson 2013). Biotic natural selection has been shown to be an important mechanism for the generation of phenotypic diversity (Kingsolver et al. 2001; Rieseberg et al. 2002). Although selection acts on phenotypes, ultimately changes in phenotype are mediated through the evolution of genes. A full comprehension of how organisms adapt to each other therefore requires an understanding of the number, identity, distribution, effect size, and source of genes under selection. To date, the evolutionary effect of biotic selection upon an organism’s genome remains largely unknown in natural populations.   This impact of biotic selection is especially interesting in the case of rapid adaptation. We identified multiple populations of threespine stickleback (Gasterosteus aculeatus) from similar lakes in southwestern British Columbia that differed mainly by the presence or absence of prickly sculpin (Cottus asper), an intraguild predator of stickleback. These populations originated approximately 10,000 years ago when marine stickleback from the Strait of Georgia colonized newly formed lakes following the melting of the glaciers at the end of the last ice age (McPhail 2007). As a result, stickleback in these lakes have independently adapted to new lakes, either with or without the same biotic agent of selection, over a short period of time. Comparing the genomes of stickleback from lakes   40 with sculpin and without sculpin would give insight into how the genes and genome of one species change in response to the presence of a single other species. It is difficult to predict the number of genes that are under selection from a single biotic agent. Genetic studies of single traits under selection from other species often identify at least one gene with a large effect on fitness. For example, selection for cryptic coat colour by predators has caused the fixation of adaptive mutations affecting expression of the Agouti gene in deer mice living in soils of different colour (Linnen et al. 2009). In human populations, the Duffy blood group locus conferring resistance to malaria occurs at a high frequency in sub-Saharan Africa but is rare in regions without malaria (Hamblin and Di Rienzo 2000). However, these traits might not be representative of all those affected by biotic selection. Methods such as QTL mapping, used to identify allelic variants between populations or species with differing phenotypes, are biased towards detecting genes with large phenotypic effects and may underestimate the number of genes under selection (Rockman 2012).  We are only aware of two studies that have attempted to quantify genome-wide adaptation in one species due to another species. Bonhomme et al. (2015) used whole genome re-sequencing of inbred lines of a legume species to examine adaptation to root associated microorganisms. The authors identified 190 genes in 58 regions that had putatively undergone selective sweeps. Similarly, comparison of sequence divergence among four populations pairs of stick insects (Timema cristinae) living on different host plant species revealed 1000 SNPs that were FST outliers in all four population pairs (Soria-Carrasco et al. 2014). Together these results suggest that many genes may be responding   41 to biotic interactions. However, it is unclear if these results apply to other animal species or are typical of biotic selection in the wild. Recent advancements in next generation sequencing are now making it possible to gain insight into the genomic architecture of adaptation by estimating the number and location of genes that have become differentiated in association with a selective agent, and potentially even the number of selective sweeps (Stapley et al. 2010). Studies of wild populations have primarily utilized reduced representation genome scans (e.g. Genotyping by Sequencing (GBS) or Restriction-site Associated DNA Sequencing (RADseq)). These methods produce greatly increased marker coverage compared to previous technologies. However, they only provide data for a limited portion of the genome and can introduce bias from loss of data at restriction cut sites or from PCR amplification (Andrews et al. 2016). Also problematic is that genetic differences not in linkage disequilibrium with markers will go undetected. A comprehensive understanding of the genes under selection requires whole genome re-sequencing to provide the increased precision needed to estimate the number and distribution of genes involved in biotic selection. The small genome size and high quality reference genome (Jones et al. 2012a) makes the threespine stickleback an ideal organism with which to answer these questions Here we report the results of a genome-wide analysis examining the genetic basis of stickleback adaptation to the presence or absence of one other species, prickly sculpin, an intraguild predator. The presence of sculpin in lakes has been shown to be strongly associated with genetically based character differences in many stickleback traits including defensive armour, body shape, and behaviour (Miller et al. 2015). A major challenge to studying the genome-wide response to biotic selection in natural populations lies in   42 isolating the effect of a single agent of selection. Furthermore, demographic processes such as genetic drift or population bottlenecks change the frequency of non-adaptive neutral alleles and can produce false signatures of selection (Klopfstein 2005; Excoffier and Ray 2008). The unique natural history of these lakes allows us to overcome these challenges. By comparing multiple threespine stickleback populations of a similar age that have independently adapted to the presence/absence of sculpin, we can isolate the effect of an agent of selection in the wild, provided that the shifts are not caused by correlated factors. This project is the first of its kind to use whole genome re-sequencing to examine the evolutionary response of a single agent of selection on a vertebrate species in the wild.   3.2 Materials and Methods 3.2.1 Sample Collection and Library Preparation  Up to 25 adult threespine stickleback were collected during the breeding season in May-June 2012-2014 from each of eight freshwater lakes containing a fish community of threespine stickleback and coastal cutthroat trout (Oncorhyncus clarkii clarkii) and nine lakes containing threespine stickleback, cutthroat trout, and prickly sculpin. Cutthroat trout are found in virtually all lakes throughout southwest British Columbia. In some cases, lakes are connected via small streams to other lakes within the same watershed. However, all study lakes were in separate watersheds, ensuring that there is no contemporary gene flow between populations. Marine stickleback were collected from six localities (23 populations total, Figure 3.1). The Pacific Ocean marine population is thought to be largely undifferentiated with high gene flow (Jones et al. 2012a,b).   43 We tested if sculpin presence was correlated with environmental differences among lakes. We gathered information on the area, perimeter, maximum depth, mean depth, and pH of each lake from Habitat Wizard (www.env.gov.bc.ca/habwiz). We used Google maps (www.maps.google.com) to determine the elevation and shortest straight-line distance from the lake to the ocean. Water samples were collected from some lakes and sodium concentration (Na), calcium concentration (Ca) and conductivity were determined using a flame photometer (Table 3.1). Abiotic variables were log transformed. We performed a principal components analysis (PCA) of the correlation matrix for abiotic traits using the ‘nipals’ option in the pcaMethods package because this algorithm is capable of handing a small amount of missing data using a non-linear iterative partial least squares method (Stacklies et al. 2007).  Stickleback were euthanized with an overdose of MS-222 anaesthetic (Argent Chemical Laboratories, Redmond, WA) and stored in 95% ethanol. Samples were stained with alizarin red (Peichel et al. 2001) and the left side of each stickleback was photographed using a Nikon D300 camera. We placed 20 landmarks outlining the shape of the fish and the insertion points of spines and fins (Walker 1997; Ingram et al. 2012). Landmarks were digitized using tpsDig 2.16 software (Rohlf 2008) and were centered, scaled, and rotated using the shapes package in R (Dryden 2012). For each population, we did a PCA of morphological landmarks, and chose as a single representative fish from each population, that female fish closest to the centroid of PC1 and PC2. Due to sample limitations, a male stickleback was used for Paq (sculpin), Cedar (sculpin) and Black Lakes (non-sculpin).    44 This strategy of sequencing a single individual per lake was chosen to maximize the number of populations sampled rather than the number of individuals. When lakes were originally colonized, rapid population growth likely occurred coincident with adaptation. This can lead to false signatures of selection if neutral rare alleles in the founding population increase in frequency as a result of genetic drift (Klopfstein 2005).  Genomic DNA was extracted from a fin clip from the single fish from each population using a standard phenol/chloroform method. The DNA samples were standardized to 20ng/ul with a QuBit 2.0 fluorometer. Paired-end whole genome libraries were prepared for each fish using the Illumina TruSeq sample kit (Illumina, San Diego CA) and quantified using High-Sensitivity Bioanalyzer chips (Agilent Technologies, Inc.). Libraries were sequenced using Illumina HiSeq 2000 at the University of British Columbia and at Genome Quebec.  3.2.2 Bioinformatics Pipeline Reads were aligned to the stickleback reference genome (gasAcu1 2006 assembly; Jones et al. 2012a) using the BWA aligner (version 0.7.6) (Li and Durbin 2009). Single Nucleotide Polymorphisms (SNPs) were identified using the UnifiedGenotyper tool in GATK (version 3.2.2) following the best practices recommendations for version 3.2.2 (DePristo et al. 2011; Van der Auwera et al. 2013). Picard (http://broadinstitute.github.io/picard) was used in conjunction with GATK to manipulate sequencing reads. Details of the bioinformatics pipeline used to generate SNPs are given in Appendix A. A BED file of the location of repeat regions was created using the RepeatMasker track in the USCS stickleback genome table browser (sticklebrowser.stanford.edu). Those SNPs with a   45 mapping quality score less than 100 (--minGQ), a mean read depth of less than 6 (--min-meanDP), or SNPs mapping to previously identified repeat regions were filtered with vcftools (version 0.1.11) (Danecek et al. 2011).   3.2.3 Divergence Among Populations We used principal component analysis (PCA) on genotype values at SNPs to visualize the overall pattern of divergence in our populations. In each population, SNPs were given a numerical value relative to the reference sequence (e.g. REF/REF = 0, ALT/ALT = 1; REF/ALT = 0.5). Missing values were filled in using the average value of that SNP across all populations. The PCA of the covariance matrix was calculated for all SNPs using the ‘svd’ option in the pcaMethods package (Stacklies et al. 2007).  To identify the regions of the genome that have differentiated in parallel between lakes with and without sculpin, and to measure the strength of this divergence, fish were grouped into lakes with sculpin and lakes without sculpin. We calculated FST between these groups in 10,000 bp sliding widows with a step size of 5,000 bp. FST was calculated using the Weir and Cockerham formula (Weir and Cockerham 1984). However, because there is no gene flow between lakes with sculpin (or lakes without sculpin), these groups are not true subpopulations. Therefore FST may not be the appropriate measure of genetic differentiation for these populations. We generated a modified cluster separation score (CSS) between fish from different lake types (groups) (Jones et al. 2012a) in windows across the genome. The CSS metric distinguishes between highly divergent regions of the genome for isolated populations adapting to the same ecological conditions. We modify this method by using principal   46 components analysis rather than multi-dimensional scaling (MDS) and weight the score by sequence coverage. This method preserves covariance of the data and is less computationally taxing than CSS. To do this, the genome was analyzed within 10,000 bp sliding windows with step size of 5,000 bp. A PCA was conducted for each window. We retained the first two principal components in each window and then measured the amount of divergence by calculating the distance between the scores for all pairs of individual fish from different lake types, adjusting for the average distance between pairs of fish within groups. The formula used is:   CS' = ! !", !!!!!!!!! (!") − 1! + ! !", ! + 1!!!!!!(! − 1)2 + ! !", ! + 1!!!!!!(! − 1)2  D is the Euclidean distance in the first two principal component axes between a pair of fish, i and j are individual fish from different groups, and s and n are the sample sizes of each group. To control for variation in sequence coverage, we divided CS’ by the number of sequenced bases within a window (both variant and invariant sites). Windows containing less than 500 bp or containing fewer SNPs than the total number of fish were dropped. Higher CS’ values indicate greater divergence between groups. A negative value is possible and signifies that the average pairwise distance between fish in different lake types is less than the pairwise distance between fish of the same lake type. We assessed the statistical significance of CS’ values using permutation tests. Within each window, we randomly redistributed the individual fish to the two groups, keeping the number of fish in each group the same, 10,000 times and calculated a CS’ score each time. We generated a P-value by calculating the proportion of times in which the value obtained from the permutated data exceeded the CS’ score calculated from the real data. Windows   47 were considered outliers based on a P-value threshold defined by false a discovery rate (FDR) of 0.05 (P<0.001). The FDR threshold was determined using the ‘fdrtool’ package. A Χ2 goodness of fit test was performed to test if outlier windows were evenly distributed among chromosomes, adjusting for chromosome size. The boundaries of divergent genomic regions between lakes with and without sculpin may be larger than 10,000 bp. Matching this prediction, we often found that neighbouring windows were identified as outliers. To define the boundaries of divergent regions we used a two state Hidden Markov Model (HMM) of log CS’ scores using default parameter values in the R package depmixS4 (Visser and Speekenbrink 2010).  The CS’ metric gives a conservative estimate of the regions under selection because only the regions that differentiate sculpin from non-sculpin stickleback repeatedly across multiple independent populations will be identified. Because CS’ measures differences in DNA sequence, our approach also requires that virtually the same alleles are involved in adaptation to sculpin presence/absence across lakes. Standing genetic variation is common in natural populations (Barrett and Schluter 2008). Reuse of standing genetic variation has been shown to be important in the repeated evolution of freshwater stickleback (Jones et al. 2012a). For example, reduction in lateral plates in freshwater stickleback occurs from the reuse of the Ecotodysplasin (EDA) ‘low’ allele present as standing genetic variation in the colonizing marine stickleback populations (Colosimo et al. 2005, Jones et al. 2012a).      48 3.2.4 Candidate Genes  To identify the genes that are divergent between lakes with and without sculpin we looked at the number and identity of genes within outlier windows. We used the biomaRt package (Durinck et al. 2009) to identify genes that occur within the window. We then looked for enrichment of gene ontology (GO) terms within outlier windows using GOwinda (Kofler and Schlotterer 2012). Portions of the genome are not sequenced with Illumina whole genome re-sequencing because of limitations of this technology. GOwinda controls for sequence coverage by comparing outlier SNPs to the total SNPs from which sequence is available. GOwinda also uses a permutation approach to control for bias in gene length. We generated a GO annotation file by obtaining zebrafish GO annotations from FuncAssociate 2 (Berriz et al. 2009) and then matched stickleback genes to the zebrafish orthologs using biomaRt. We ran GOwinda twice, first using SNPs only within outlier windows and then including SNPs 2000-bp upstream or downstream of outlier windows to account for nearby regulatory SNPs.  3.3 Results 3.3.1 Sculpin Presence is not Correlated with Abiotic Environment  There were no consistent environmental differences between lakes with and without sculpin for individual variables and lakes overlapped broadly in their abiotic traits along PC1 (U=28, P=0.48) and PC2 (U=36, P=1.0) (Table 3.2, Figure 3.2).    49 3.3.2 Genomic Divergence is Associated with Presence/Absence of Intraguild Predator We generated a total dataset of 5.7 million filtered SNPs from 23 populations and performed a PCA using all SNPs. The first principal component (PC1) explained 14.7% of the variation and separated individuals from lakes with sculpin from individuals from lakes without sculpin (F2,20=52.9, P<1e-8) (Figure 3.3A). Stickleback sympatric with sculpin had genome PC1 values that were more similar to the marine stickleback than stickleback from sculpin-absent lakes. This finding suggests that stickleback from lakes with sculpin share more marine genotypes at SNPs having high trait loadings on PC1. We performed a second PCA using only the individuals from lake populations (reduced dataset of 4.6 million SNPs). The first PC of this second PCA similarly distinguished lakes with and without sculpin (F1,15=26.3, P=0.0001) and explained 11.6% of the variation (Figure 3.3B). The main axis of genetic variation in these populations is strongly associated with the presence/absence of sculpin. For PC2 to PC22 populations are somewhat differentiated by geography but this signal is not strong (Figure 3.4). For example, PC2 separates some but not all of the lakes from Vancouver Island from all other lakes, while PC3 distinguishes some of the populations from the Sunshine Coast.  3.3.3 Genetic Differentiation is Extensive but Unevenly Distributed To identify genomic regions that have strongly differentiated in parallel between stickleback from lakes with and without sculpin, we calculated a cluster separation score CS’ between stickleback from the two lakes types. Across all sliding windows, CS’ is strongly positively correlated with FST (r=0.9) (Figure 3.5). Sufficient data were available for 88,711   50 sliding windows genome wide. Overall, 1473 windows were identified as outliers, accounting for 1.7% of sampled windows. Each outlier window had an average of 0.83 genes (sd=0.8), and individual genes often spanned multiple outlier windows. Combined, outlier windows contained more than 500 genes (Appendix B).  We compared the distribution of outlier windows with the expected distribution based upon the number of windows per chromosome for which sufficient sequence was available (at least 500 informative bases). Outlier windows were not randomly distributed throughout the genome (Χ2 =2392, df=21, P< 2.2e-16). For example, large portions of chromosomes 4, 7, and 12 showed elevated divergence between lakes with and without sculpin, while other regions such as chromosomes 14 and 15 were undifferentiated between lake types (Figure 3.6).  The number of genes or windows differentiating populations does not count the number of selective sweeps because a selective sweep will likely encompass multiple outlier windows. Although lakes are isolated from each other, the same marine population presumably originally colonized the lakes, potentially bringing similar standing variation each time. If these populations experienced selection upon the same standing genetic variation present in the colonizing marine fish, then we should be able to detect the regions that have undergone selective sweeps repeatedly in independent lakes as one or more adjacent outlier windows. A hidden Markov model (HMM) was implemented to define the boundaries of these outlier regions. The HMM estimates the location of state shifts between divergent regions and regions having little or no divergence (Visser and Speekenbrink 2010). This model collapsed the 1,473 outlier windows into 164 distinct outlier regions   51 across the genome (Figure 3.7, Table 3.2). The median width of regions in the ‘selected’ state containing an outlier window was 130,000 bp.  3.3.4 Candidate Adaptive Genes Within the outlier windows are many putative candidate genes for phenotypic differences between stickleback from lakes with and without sculpin. Several of the genes found within outlier windows have known roles in zebrafish development and might correspond to the phenotypic differences observed among populations of stickleback from lakes with and without sculpin. Such genes include SATB2, which is involved in the development of the vertebrate jaw (Fish et al. 2011), PDLIM7, a gene necessary for pectoral fin development (Camarata et al. 2010), GIGYF1, a modifier of IGF-I signalling (Giovannone et al. 2003), and KCTD12, a gene that may influence zebrafish thigmotaxis behaviour (Lee et al. 2014). Nevertheless, there was no significant enrichment of any GO terms associated with SNPs within outlier regions or with SNPs within 2000 bp up or downstream of outlier windows. We have no evidence that intraguild predation preferentially leads to selection on genes associated with a particular cellular component, molecular function, or biological process.  3.4 Discussion The presence/absence of prickly sculpin, an intraguild predator of stickleback has resulted in widespread but unevenly distributed divergence across the threespine stickleback genome. Although the freshwater populations have been isolated from each   52 other for only 10,000 years, we observed extensive parallel differentiation in more than 1.7% of the measured genome between populations of stickleback from lakes with and without sculpin. Our methodology only identifies outlier regions that have diverged in parallel between lakes with and without sculpin. Genes that have diverged in only one lake will go undetected therefore these results are a conservative estimate of the proportion of the genome implicated in adaptation to the presence or absence of sculpin. The presence/absence of a single biotic agent had a rapid and profound effect on genomic divergence in stickleback. More than 500 genes were identified in outlier windows, suggesting that the presence/absence of the intraguild predator has resulted in selection on a large number of genes. Outlier windows in the present study were identified by parallel evolution and are unlikely to be caused by neutral evolutionary processes. It is generally considered that lake populations evolved independently of each other after colonization by the common marine ancestor therefore changes in genetic variation caused by neutral evolutionary processes (e.g. population bottlenecks) would be unique to each lake. Genome scans do not provide information on the effect sizes that candidate genes have on the phenotype or fitness of an organism. Thus we cannot quantify the number of genes of large and small effect. However, the large number of candidate genes identified in outlier windows is consistent with polygenic adaptation of many alleles of small effect on fitness.  Why is the presence/absence of a single biotic agent correlated with differentiation of so many genes? There are several possible explanations. First, intraguild predation is associated with character shifts involving many traits including body shape, defensive   53 armour, diet, and behaviour (Ingram et al. 2012; Miller et al. 2015). These traits are likely to have a polygenic basis and QTL studies of some of these same traits in stickleback often find them mapping to multiple genomic regions (e.g. Rogers et al. 2012). Selection upon many traits may necessitate change in many genes. Second, although prickly sculpin are a single agent of selection, and represent the only consistently observed difference between the two types of lakes, intraguild predation may lead indirectly to multifarious selection upon stickleback. Sculpin directly select on stickleback phenotypes and they may also result in downstream effects by changing how stickleback interact with other components of the lake ecosystem. For example, the absence of sculpin may allow stickleback to colonize the shallow benthic environment, and because coastal cutthroat trout live primarily in the open water (Reimchen 1994), this habitat shift would change the stickleback diet and indirectly decrease predation from trout. Finally, outlier regions occur more often on chromosomes with large regions of low recombination (Roesti et al. 2013). Linkage disequilibrium is increased in regions of low recombination (Hartl and Clark, 1997). As a result, neutral or even deleterious alleles can hitchhike along with linked genes under selection, which could lead to an overestimate of the number of ‘selected’ genes (Excoffier and Ray, 2008), although not of the number of selected sweeps. Within the list of outlier genes, we identified candidate genes for phenotypic traits previously found to differ among populations with and without sculpin (see chapter 2), such as fin position, mouth shape, and behaviour. Our genome scan was also able to identify candidate genes for many other phenotypes that may also be under selection, including immune function, brain development, and muscle structure (Appendix B). Interestingly, our analysis found no significant enrichment of GO terms in outlier windows. Adding (or   54 subtracting) a single species from the environment leads to selection on many genes but does not preferentially cause selection on genes with a certain function. Outlier windows were clustered in the genome. Other genome scan studies have found that heterogeneous genomic divergence – variation in genetic differentiation across the genome – is common (e.g. Nosil et al. 2009; Lawniczak et al. 2010; Delmore et al. 2015). Clustered architectures are predicted when differentiation occurs with gene flow because nearby co-adapted loci are less likely to be broken up by recombination than neutral loci (Yeaman 2013). During the colonization phase, lakes would have probably experienced gene flow with stickleback in the marine environment. However, this period was most likely short, and subsequently, lakes would have diverged in allopatry. Heterogeneous genomic divergence may instead occur as a by-product of constraints caused by underlying features of the genome such as variation in recombination rate (Cruickshank and Hahn 2014). Further investigation is necessary to understand the mechanism causing heterogeneous genomic divergence among these populations. Other whole genome resequencing studies examining genomic divergence between contrasting environments have reported fewer genes under selection. Jones et al. (2012a) examined the genetic basis of adaptation of marine stickleback to freshwater environment and found that 0.18 – 0.26% of the genome was differentiated between replicate populations of marine and freshwater stickleback. A study looking at parallel adaptation to hypoxic conditions at high altitude in 6 breeds of dog reported genomic differentiation at 28 regions containing 141 candidate genes (Gou et al. 2014). Adaptation of Arabidopsis lyrata to serpentine soils found significant allele frequency differences in only 96 of the 8.4 million SNPs identified (Turner et al. 2010). Although quantitative analysis of the genomic   55 architecture of selection between biotic and abiotic agents will require more studies, preliminarily evidence suggests that biotic selection is associated with higher genomic divergence.  Our findings have several implications. First, biotic selection affects many genes. The presence or absence of sculpin, a single species, appears to have led to differentiation in 1.7% of the genome. Yet sculpin are not the only biotic agent of selection in lakes. If the amount of differentiation is typical, a large percentage of the genome may be under selection as a consequence of interactions with other species. Second, adaptation to other species is not necessarily slow. A change to a new optimum can occur quickly when the initial genetic variance in the population is large (Stephan 2016). Lastly, stickleback from populations sympatric with sculpin retained more marine genetic variants. This strongly suggests that it was release from selection by sculpin that was the cause of genetic divergence between population types. Consequently, biotic selection appears to have had a profound effect upon the stickleback genome.             56 Figure 3.1 : Locations of populations sampled. Lakes with sculpin are in red: (A) Cedar, (B) Ormond, (C) Ambrose, (D) North, (E) Brown, (F) Paq, (G) Rosseau, (H) Pachena. Lakes without sculpin in in green: (1) Tom, (2) Cranby, (3) Kirk, (4) Klein, (5) Trout, (6) Hoggan, (7) Bullock, (8) Stowell, (9) Black. Marine populations are in blue: (M1) Seyward Estuary, (M2) Oyster Lagoon, (M3) Little Campbell River, (M4) Salmon River, (M5) West Creek, (M6) Bamfield.   -126 -125 -124 -123 -1224849500 50 100 kmVancouverIslandM3M2M6M1M5M4CEADBHFG736248519  57 Figure 3.2 : Principal component analysis (PCA) of abiotic traits of lakes with sculpin (red) and lakes without sculpin (black). Abiotic trait values are given in Table 3.1.      -2 -1 0 1 2-1012PC1(38.3%)PC2(23.1%)AmbroseBlackBrownBullocks CedarCranbyHogganKirkKleinNorthOrmondPachenaPaqRosseauStowellTomTrout  58 Figure 3.3 : Principal component analysis of all SNPs from (A) all populations and from (B) only freshwater populations. Only the first principal component (PC1) is shown. Each point is a single individual from a population. Lakes with and without sculpin are separated from each other along PC1.      -600-400-2000200400600PC1 (14.8%)Without WithSculpin Sculpin MarineA-600-400-2000200400600PC1 (11.6%)Without WithSculpin SculpinB  59 Figure 3.4 : Plot of principal components 2 and 3 from the principal component analysis of all SNPs from all populations presented in figure 3.3A. Each point is a single individual from a population. Lakes with sculpin are in red, lakes without sculpin are in black, and marine populations are in blue.       -600 -400 -200 0 200 400-600-2000200400600Genome PC2 (6.6%)Genome PC3 (5.1%)AmbroseBamBlackBrownBullockCedarCranbyHogganKirkKlein LCMNorthOrmondOysterPachenaPaqRosseauSalmonSeywardStowellTomTrout WCM  60 Figure 3.5 : Plot of CS’ and FST for 10,000 bp windows throughout the genome. Points in red are outlier windows for CS’ score. FST was calculated by comparing allele frequency between lakes with sculpin and lakes without sculpin.        61 Figure 3.6 : Genome-wide distribution of CS’ score. Points have been averaged over ten windows of 10,000 bp. Stickleback populations from lakes with and without sculpin are highly differentiated at many sites across the genome. All chromosomes are plotted on the same scale.    62  Figure 3.7 : CS’ score between lakes with and without sculpin for chromosome twelve. Outlier windows are indicated in red. State changes in the hidden markov model are shown in blue. Adjacent outlier windows are frequently grouped next to each other and are likely part of the same selective sweep.       051015Millions of bpCS’0 10 20  63 Table 3.1 : Abiotic traits measured from lakes with and without sculpin. Area, perimeter, max depth, mean depth, and pH were obtained from HabitatWizard. Elevation and the distance from the lake to the nearest ocean (To Sea) were calculated using google maps.      64 Table 3.2 : Mann-Whitney test results from environmental variables between lakes with and without sculpin.          65 Table 3.3 : Genome-wide distribution of windows identified as outliers by permutation test (Outlier Windows) and regions that both contain outlier windows and were identified by the hidden Markov Model (HMM).  Chromosome(Outlier(Windows( HMM(1" 104" 17"2" 42" 11"3" 4" 3"4" 305" 29"5" 4" 2"6" 4" 3"7" 436" 22"8" 70" 9"9" 61" 12"10" 3" 3"11" 27" 11"12" 156" 11"13" 7" 4"14" 1" 1"15" 7" 3"16" 16" 6"17" 4" 2"18" 13" 2"19" 23" 3"20" 109" 7"21" 16" 3"Un" 59" NA"Total" 1471" 164"   66 Chapter 4: A Comparative Analysis of Experimental Selection on the Stickleback Pelvis  4.1 Introduction Natural selection leads to changes in trait distribution when agents of selection cause differential fitness among individuals with different phenotypes (Endler 1986). Measurements of selection in the wild have now become commonplace (Kingsolver et al. 2001). However, identifying the mechanisms of selection presents a greater challenge because to determine the cause of natural selection requires demonstrating a link between the agent of selection, differential fitness, and a change in trait distribution.  Observational studies of natural selection in the wild are correlational and provide indirect evidence that the putative agent of selection is the cause of changes in trait distribution (Wade and Kalisz 1990; Schluter 2009). Natural selection may instead be the result of variation in the environment or the presence of another agent of selection. Experimental studies of selection are a powerful tool for identifying agents of selection. Researchers can manipulate the trait of interest to measure selection in isolation. However, selection experiments are notoriously difficult to perform. Experimental selection studies often lack the sample size to accurately estimate the effect size, resulting in wide confidence intervals for the estimate of selection on the trait of interest. Measurements of selection in the wild find that selection on morphological traits is typically weak to moderate (Hoekstra et al. 2001; Kingsolver et al. 2001) and conducting experimental studies with sufficient replicates to detect weak selection can be prohibitive for logistical reasons.   67 Combining results from multiple experimental studies of the same agent of selection might offer a solution to these difficulties by producing an aggregate estimate of selection that is more precise than any individual study (Arnqvist and Wooster 1995; Hersch and Phillips 2004; MacColl 2011). Meta-analysis of multiple studies has been used to estimate an effect size for diverse traits including Daphnia response to predator kairomones (Riessen 1999), and local adaptation in plants (Leimu and Fischer 2008). Threespine stickleback (Gasterosteus aculeatus) from isolated populations display a wide range of phenotypic variation for many traits (Bell and Foster 1994). Importantly, stickleback inhabiting similar environments frequently show parallel changes in the same traits, suggesting that trait divergence is caused by ecological differences among environments (e.g. Kaeuffer et al. 2012; McKinnon and Rundle, 2002). This link between phenotypic variation and environmental variation has made the stickleback a model organism for investigating the mechanisms of selection driving these parallel trait changes. One of the most conspicuous differences among stickleback populations is variation in pelvic morphology (Bell and Foster 1994). The stickleback pelvis is a bony structure consisting of a pelvic girdle and two hinged pelvic spines (Bell 1988). When extended, the pelvic spines brace against the pelvic girdle making them lock open (Reimchen 1983). Complete loss of the pelvic structure has occurred independently in multiple populations of threespine stickleback, ninespine stickleback (Pungitius pungitius), and brook stickleback (Culaea inconstans) (Nelson 1969; Nelson and Atton 1971; Klepaker et al. 2013). Pelvic loss has a genetic basis and loss of pelvic morphology has evolved multiple times in Gasterosteus (Kingsolver et al. 2001; Shapiro et al. 2004; Chan et al. 2010) and in Pungitius (Bell and Foster 1994; Shapiro et al. 2009; Shikano et al. 2013). Variation in   68 pelvic morphology can occur in the same lake both within (Bell 1988; Lescak et al. 2013) and between stickleback species (Reimchen 1983; McPhail 1992) and can be stable over multiple generations (Lescak et al. 2013) or can vary among stickleback of different size classes (Reimchen and Nosil 2002). Predation has been hypothesized as the driver of variation in pelvic morphology. Pelvic spines are predicted to be an anti-predator defence against gape-limited piscivorous predators. It is theorized that spines help stickleback escape from predatory fish by piercing the mouth parts of predators and/or by increasing the effective diameter of the stickleback, thereby making it more difficult for the stickleback to be swallowed (Hoogland et al. 1956; Hagen and Gilbertson 1972). Several lines of observational evidence support the hypothesis that longer pelvic spines provide protection from fish predators. In laboratory feeding trials, pike (Esox lucius) preferentially consumed de-spined stickleback (Hoogland et al. 1956). In the wild, stickleback in the stomach contents of trout (Oncorhynchus clarkii) have shorter spines than stickleback collected using seine nets (Moodie 1972). Lastly, an increase in the abundance of predators or in the number of piscivorous predator species in the wild is repeatedly correlated with longer pelvic spines in stickleback and an increase in other armour traits (Moodie 1972; Vamosi 2003; Marchinko 2009; Miller et al. 2015). Combined, this evidence suggests that natural selection from fish predators increases pelvic armour. Predation on juvenile stickleback by large aquatic insects has been hypothesized as the agent of selection for the reduction or loss of pelvic armour in many freshwater populations (Hoogland et al. 1956; Hagen and Gilbertson 1972; Reimchen 1980). Dragonfly nymphs (Aeshna sp.) can eat 1-2 juvenile stickleback per day (Hoogland et al. 1956; Reimchen 1980). Aquatic insects do not occur in the marine environment, and stickleback   69 from freshwater populations typically have reduced pelvic armour compared with the marine form (Klepaker et al. 2013). Two mechanisms have been proposed to explain why a reduction in the number and size of pelvic spines may provide a selective advantage against insect predation. Spines may provide a convenient “handhold” for insect predators to capture and hold on to stickleback (Reimchen 1980). Consequently, stickleback with shorter or absent spines will be able to avoid capture more easily. An alternative mechanism hypothesizes that individual stickleback with more armour might grow more slowly because investment in armour traits requires resources that would otherwise be used for growth (Marchinko and Schluter, 2007). Increased armour might thus prolong the length of time during which juvenile stickleback are small in size and most vulnerable to insect predation.  Several experimental studies have tested the role of predators as agents of selection on pelvic spines, but these experiments have produced somewhat inconsistent results. In some experiments, predatory fish more readily consume stickleback with shorter pelvic spines (e.g. Reist 1980; Lescak and Hippel 2011) while other experiments show no significant differences in fitness between stickleback with different pelvic morphology (e.g. Reist 1980; MacColl and Chapman 2011). Similarly, insect predators preferentially consume stickleback with longer pelvic spines in some experiments (Reist 1979; Marchinko 2009), while other experiments report non-significant estimates of selection on pelvic morphology (e.g. Lescak et al. 2012; Zeller et al. 2012; Mobley et al. 2013). In all cases, experimental estimates of selection on pelvic morphology are based on small sample sizes and have wide confidence intervals, indicating that estimates of the effect size are highly uncertain. In this paper we address the causes of selection on stickleback pelvic spine length with an experiment and a meta-analysis. Our experiment focuses on stickleback from lakes   70 that contain prickly sculpin (Cottus asper), an intraguild predator that eats stickleback and competes with stickleback for benthic resources. Stickleback sympatric with sculpin consistently have longer pelvic spines than stickleback from lakes in which prickly sculpin are absent (Miller et al. 2015). Variation in pelvic spine length among populations has a genetic basis (Rogers et al. 2012; Miller et al. 2015). A previous mesocosm experiment found higher mortality from sculpin predation on stickleback from lakes without sculpin (Ingram et al. 2011). However, stickleback sympatric with sculpin also exhibit genetically based difference in behaviour which may be important for escaping sculpin predation (Ingram et al. 2012; Miller et al. 2015), therefore the decrease in mortality cannot be directly attributed to longer pelvic spines. Furthermore, the presence of sculpin could also be correlated with another agent of selection for longer spines. For example, sculpin may displace stickleback from the benthic habitat into the open water where predation by coastal cutthroat trout is the agent of selection for longer pelvic spines. Therefore, experimental manipulation of the putative agent of selection is necessary to test the mechanism causing the association between pelvic spine length and fitness.  We tested if prickly sculpin are an agent of selection on stickleback pelvic spines by isolating the effect of spines as an anti-predator defence against sculpin in a mesocosm experiment. We physically modified the length of the pelvic spines of stickleback from two populations sympatric with sculpin and then compared the mortality rate of stickleback with clipped and unclipped pelvic spines experimentally in the presence of sculpin. The results of our mesocosm experiment were combined with previously published experimental selection studies to address the problem of low power and wide confidence intervals of the effect of   71 pelvic armour. We used a meta-analysis approach to determine the magnitude and direction of selection on pelvic morphology by both fish and insect predators.  4.2 Methods  4.2.1 Mesocosm Experiment Experimental mesocosms were established in 20 plastic 1136L cattle tanks 1m deep by 2 m wide. Mesocosms were filled with water and seeded with benthic mud and zooplankton collected from nearby experimental ponds. To stimulate primary production, 0.05 g KH2PO4 and 1.0 g KNO3 was added to each mesocosm. A 25 cm diameter open-ended cylinder constructed from stiff black 7 mm plastic mesh was attached to the side of each cattle tank and suspended 0.5 m above the bottom to provide shade and a refuge from predation. Mesocosms were allowed to settle for two weeks prior to the addition of fish.  Adult stickleback were collected from Paq Lake and Ambrose Lake and sculpin were collected from Paq Lake using minnow traps and by dipnet. Fish were transported to 100L holding tanks in the aquatic facility at the University of British Columbia and allowed to recover for several days. Paq and Ambrose are in separate watersheds in the Sechelt Peninsula. Both lakes contain a simple fish community composed of threespine stickleback, prickly sculpin, and coastal cutthroat trout.  To create variation in the length of the pelvic spine, stickleback were briefly anesthetised in MS-222 (1g/L) and pelvic spines were clipped to 2.5mm (the average length of stickleback pelvic spines from lakes without sculpin (Miller et al. 2015)). Control stickleback were anaesthetized and handled in a similar manner but pelvic spines were not   72 modified. Stickleback were returned to the 100L tanks for 24 hours of observation. There was no mortality following spine clipping.  The standard length of each stickleback was measured prior to introduction (36.5-60.6 mm). Four size-matched clipped and unclipped stickleback were added to each mesocosm (eight total). Paq Lake stickleback were used for 10 mesocosms and Ambrose Lake stickleback were used for the remaining 10 mesocosms. Following the first set of trials, sufficient Paq lake stickleback were available for six additional trials (N=26 trials total). A single sculpin (95-105mm) was added to each mesocosm two days after the stickleback introduction. A visual survey of the number of stickleback in each mesocosm was conducted daily. Dead stickleback that did not show evidence of sculpin predation were replaced with a similar-sized individual having the same pelvic phenotype. A trial was considered to be complete when half of the stickleback were consumed. At that time, the sculpin was removed and remaining stickleback were collected. We carefully examined each stickleback for signs of injury and recorded standard length and pelvic phenotype. Over the course of the experiment, visibility in mesocosms decreased. As a result, several trials were stopped when greater or fewer than four stickleback remained. To ensure that all surviving stickleback were collected, each mesocosms was trapped with minnow traps for 48 hours.   Paired t-tests were performed separately for Paq and Ambrose Lake mesocosms to compare the frequency and size of surviving clipped and unclipped stickleback using trial as the replicate. The log odds ratio was calculated for each trial and then a summary log odds ratio was estimated using the Peto method (Borenstein et al, 2009). A positive log odds ratio indicates stickleback with unclipped spines were more likely to survive, and a negative log odds ratio indicates increased survival of stickleback with clipped spines.   73  4.2.2 Comparison with Other Selection Studies  We conducted a meta-analysis by searching the literature for experimental studies measuring selection on stickleback pelvic morphology from insectivorous or piscivorous predators. Variation in pelvic morphology could be naturally occurring variation, physical modifications, or F2 or backcross hybrids between populations having divergent phenotypes. Studies were only included in the meta-analysis if sufficient information was available to allow us to calculate the standard error of the effect size, which also required multiple independent trials. Using these criteria, we excluded Ziuganov and Zotin (1995) because the study had a single uncontrolled experimental replicate. We were also forced to leave out Reimchen (1980) because results from multiple replicates were pooled, which loses all information on the variance between trial outcomes, and the original data was no longer available. Although Reist (1979, 1980) presented pooled data across replicates, the results for most trials were available in Reist (1978). When data for individual trials was not available, we contacted the authors of the original study. Leinonen et al. (2011), McColl et al. (2011), and Mobley et al. (2013) generously provided raw data for individual trials.  We used standardized mean difference in trait values between treatments, d (predation – control), as our measure of effect size. For studies reporting a continuous measure of pelvic spine length or standard length, d was calculated using the formula for independent groups (Borenstein et al, 2009). This metric is similar to the standardized selection differential (i) (equation 6.1 in Endler, 1986) except that d uses the pooled standard deviation across groups, whereas i uses the standard deviation from only the control treatment. The values of the two measures were always similar. For studies   74 measuring selection on the presence/absence of the pelvic structure, either with experimental manipulation or using natural existing variation, a log odds ratio was calculated from the proportion of survivors with and without pelvic spines/girdles in the two treatments for each trial. Then, an overall summary log odds ratio was calculated for each experiment using the Peto method (Borenstein et al, 2009). Summary log odds ratios were converted to d to facilitate comparisons across studies (Hasselblad and Hedges 1995).   Experimental design, target population, and stickleback species varied among studies, therefore the summary effect for the meta-analysis was calculated using a random effect model separately for insect and piscivorous predators. Effect sizes were weighted using the inverse of the sampling variance of the experiment (Borenstein et al, 2009). The random effects model assumes that the true effect size may vary from study to study (Borenstein et al, 2009). The summary effect is therefore an estimate of the mean distribution of the true effect size of pelvic morphology on the probability of survival. To minimize bias from the inclusion of multiple experiments from a single study, we calculated a second summary effect for each predator type using a single estimate for each study. A fixed effects meta-analysis was used to estimate the summary effect for each study. As before, a random effect model was then used to calculate an overall summary effect across studies. Standard length was reported in fewer studies. A summary effect for length was similarly calculated with a random effect model separately for insect and piscivorous predators. All summary effects were calculated using the ‘meta’ package (Schwarzer 2015) in the R statistical environment (R Development Core Team, 2015).    75 4.3 Results 4.3.1 Mesocosm Experiment  Trials took 15-50 days to reach 50% stickleback mortality. None of the surviving stickleback showed evidence of wounds from unsuccessful predation attempts. We found no significant difference in survival of stickleback with clipped and unclipped pelvic spines (Figure 4.1; Paq: t=0.75, df=15, P=0.47; Ambrose: t=0, df=9, P=1). The summary log odds ratio for all trials was 0.118 (95% CI: -0.358, 0.594) representing an 11.1% increase in survival probability for stickleback with unclipped pelvic spines, but this result was not significant and the confidence intervals were wide. Results were similar when comparing each lake individually (Paq Lake: 0.189, 95% CI: -0.414, 0.793; Ambrose Lake: 0.00, 95% CI: -0.775, 0.775). Surviving clipped and unclipped stickleback did not differ in standard length (Paq: t=-0.09, df=15, P=0.93; Ambrose: t=-1.45, df=9, P=0.18). There was no difference in mean standard length at the start of the experiment compared to mean standard length of the survivors (Paq: t=-1.19, df=15, P=0.25; Ambrose: t=-0.48, df=9, P=0.65).  4.3.2 Meta-analysis of Selection Studies We identified 25 published and unpublished experiments that met our criteria. Combined, these experiments represented 213 independent trials measuring selection on pelvic morphology in the presence of fish or insect predators. Studies included three species of stickleback, four species of insect predators, and four species of fish predators. Most experiments were conducted by adding stickleback with variation in pelvic morphology to a mesocosm containing a predator. Discrete variation in pelvic morphology was generated by   76 experimentally manipulating pelvic spine length, or by using study populations with naturally occurring variation in pelvic spine presence/absence. Several studies used F2 or backcrosses to create continuous variation in pelvic morphology. Standard length data was available for three fish predation experiments and five insect predation experiments. Details and effect sizes for all studies are reported in Tables 4.1 and 4.2.  For fish predation experiments, longer pelvic spines increased survival (Figure 4.2), with a mean effect size of 0.13 (95% CI: 0.02, 0.23; P=0.02). This is equivalent to an increase in the mean pelvic spine length by 0.13 of a standard deviation in the presence of fish predators. Insect predation favoured slightly shorter pelvic spines, with a mean effect size of -0.05 (95% CI: -0.28, 0.17; P=0.65), but this result was not significant (Figure 4.3). Using a single estimate for each study, the summary mean effect was 0.14 (95% CI: 0.004, 0.28; P=0.04) for fish predation experiments (Figure 4.4) and 0.04 (95% CI: -0.19, 0.28; P=0.71) for insect predation experiments (Figure 4.5). Insect predation experiments had a larger variance in effect size than fish predation experiments (Figure 4.6). Fish predators had no effect on standard length (Figure 4.7). The summary mean effect fish predation was  -0.02 (95% CI: -0.83, 0.78; P=0.96). Insect predators preferentially consumed smaller fish (Figure 4.7), with a summary mean effect of 0.27 (95% CI: -0.14, 0.67; P=0.20), but this effect was not significant.   4.4 Discussion Pelvic morphology is a highly variable trait among and often within stickleback populations. The mechanisms of selection that produce this trait variation are still uncertain. Stickleback in lakes with sculpin have longer pelvic spines than stickleback from lakes   77 without sculpin (Miller et al. 2015). To test if prickly sculpin are an agent of selection on stickleback pelvic spines, we experimentally modified the length of pelvic spines and measured differential mortality between stickleback with clipped and unmodified pelvic spines. We observed an 11% increase in the probability of survival for stickleback with unclipped pelvic spines. However, the confidence intervals for this estimate overlapped with zero and this effect was not statistically significant. From this data alone, we were unable to conclude that prickly sculpin preferentially consumed stickleback with shortened pelvic spines. There are four possible reasons why our experiment failed to detect selection on pelvic spine length from prickly sculpin predation. (1) Selection on stickleback pelvic spines is caused by other factors and sculpin are not an agent of selection on this trait. For example, lakes with sculpin also contain coastal cutthroat trout and a variety of avian predators, which may be an alternative agent of selection favouring increased pelvic spines (although lakes without sculpin also contain these predators) (Miller et al. 2015). (2) A challenge of selection experiments is choosing the correct size class of both the agent of selection and target of selection (Endler 1986). Natural selection may favour greater pelvic morphology but only when stickleback and/or sculpin are at a different size class than that used in the experiment. For example, none of the surviving stickleback in this experiment had defensive wounds, in contrast to a similar experiment by Lescak and von Hippel (2011) that identified wounds caused by the predator (trout) in 40% of trials. The current experiment used adult sculpin near the upper limit of the size range of sculpin in Paq Lake (personal observation), whereas smaller sculpin might be gape limited and thus select for greater armour. (3) Manipulations of spines were not effective because the spines were not   78 scaled to body size. (4) Natural selection indeed favoured longer pelvic spines but we were unable to detect an effect because our experiment was underpowered (type II error). We observed a trend towards increased survival of stickleback with unclipped pelvic spines, but as in most other experiments of this kind (Figure 4.2), this result was not statistically significant and confidence intervals for treatment effects were large.  Partly to overcome the lower power of individual studies, we compiled a meta-analysis of experimental studies of selection on pelvic morphology from insect and fish predators. We found that fish predators indeed selected for longer pelvic spines, with a summary effect size of 0.13 units of a standard deviation. The effect sizes were similar when experiments were combined into a single estimate for each study. If we assume that pelvic spine length has a heritability of 0.38 (Leinonen et al. 2011b), using the 95% confidence interval of our estimate of effect size, we predict that the mean pelvic spine length would increase by one standard deviation in 12 – 132 generations. This value represents a small to moderate effect on fitness and is comparable to 0.14, the mean absolute value for linear selection differentials from the Kingsolver selection dataset (Kingsolver et al. 2001). On the basis of our meta-analysis, and in agreement with observational studies, we conclude that fish predators are an agent of selection favouring increased pelvic spines. In contrast, it is still unclear if insect predators are an agent of direct selection on the length of stickleback pelvic spines. The summary effect size indicated a very small increase in survival for stickleback with shorter pelvic spines. However, the large confidence interval for this estimate ranges from -0.28 to 0.17 preventing us from ruling out either selection for increased or decreased pelvic morphology by insects. Although Reimchen’s   79 (1980) hypothesis has been frequently cited, there is as of yet no convincing evidence in support of insect predators selecting for smaller spines by the “handhold” mechanism.  Studies included in the meta-analysis measured selection at a range of body size in multiple stickleback species for several species of insect predator. This variation in methodology may obscure the effect of insect predation. For example, Lescak et al. (2013) observed that dragonfly naiads preferred to eat stickleback with pelvic armour when the fish were smaller than the dragonfly but preferred stickleback without pelvic armour when the fish was larger then the dragonfly. The summary effect of insect predation revealed that insects preferred to eat smaller stickleback, but this effect was not significant. This leaves open the question of what is the selective mechanism underlying loss or reduction of the pelvis in many stickleback populations. Armour reduction may occur as a by-product of selection on another trait. Reduced pelvic armour has been proposed to increase buoyancy (Myhre and Klepaker 2009) and manoeuvrability (Reimchen 2000). One hypothesis is that indirect selection against pelvic spines occurs because investment in armour is costly in energy and materials thereby limiting minerals such as calcium or phosphorus available for growth (Giles, 1983). Small juvenile stickleback are eaten by insects therefore a slower growth rate increases the length of time that juveniles are vulnerable to insect predation. Direct selection for increased growth rate could lead to indirect selection for decreased armour. However, support for the ion limitation hypothesis is mixed. Bell (1993) compared the frequency of pelvic reduction with calcium concentration for 179 Alaskan lakes. When native piscivorous predators were present, none of the lakes had a reduction in pelvic structure. When predatory fish were absent from the lakes, pelvic reduction was associated with low calcium concentrations. However, pelvic reduction has   80 been observed in Canadian lakes with high concentrations of calcium (Klepaker at al. 2013). Marchinko and Schluter (2007) raised stickleback with differing numbers of lateral plates, another type of bony armour, in freshwater and saltwater. Supporting the ion limitation hypothesis, in freshwater, stickleback with more lateral plates grew more slowly compared to stickleback with fewer lateral plates. However, a later study by Rollins et al. (2014) did not detect higher growth rate of stickleback with decreased pelvic armour in the lab or in the wild, suggesting that ion limitation may vary among traits or that the effect was too small to detect in that study. We find support for the hypothesis that fish predation selects for longer pelvic spines in stickleback. The magnitude of this selection was small to moderate. At this time we are unable to support or reject the hypothesis that insect predators favour shorter pelvic spines.     81 Figure 4.1 : Number of surviving threespine stickleback with the clipped and unclipped pelvic spine treatment at the end of the mesocosm experiment. Lines connect stickleback from a single mesocosm replicate. Each trial started with four fish of each phenotype.      Unclipped Clipped01234AmbroseNumber of SticklebackUnclipped Clipped01234PaqNumber of Stickleback  82 Figure 4.2 : Forest plot of the effect size for all fish predation experiments using multiple estimates from each study. The grey box indicates the mean of the effect size (d) for each experiment and the lines give the 95% confidence interval of the effect. The weighted mean was calculated using a random effects model. W is the weight of the study in the model. The weighted mean is indicated with the dotted line and the 95% confidence interval of this estimate is contained within the red triangle.        StudyWeighted MeanReist AReist BReist CReist DReist EReist FReist GReist HLeinonen - no refugeLeinonen - refugeLescak - FishMacColl & ChapmanThis Study - PaqThis Study - Ambrose-2 -1 0 1 2 0.13 0.14-0.18 0.26 0.16-0.03-0.10 0.43-0.07 0.43 0.12 0.36-0.04 0.10 0.0095%-CI[ 0.02; 0.23][-0.16; 0.44][-0.51; 0.16][-0.06; 0.59][-0.23; 0.56][-0.38; 0.32][-0.49; 0.30][-0.12; 0.98][-0.62; 0.48][-1.55; 2.41][-1.84; 2.08][ 0.13; 0.59][-1.02; 0.95][-0.23; 0.44][-0.43; 0.43]W100%12.0% 9.8%10.3% 7.0% 9.0% 6.9% 3.6% 3.6% 0.3% 0.3%20.1% 1.1% 9.9% 6.0%d  83 Figure 4.3 : Forest plot of the effect size for all insect predation experiments using multiple estimates from each study. The grey box indicates the mean of the effect size (d) for each experiment and the lines give the 95% confidence interval of the effect. The weighted mean was calculated using a random effects model. W is the weight of the study in the model. The weighted mean is indicated with the dotted line and the 95% confidence interval of this estimate is contained within the red triangle.         StudyWeighted MeanReist IReist JReist KReist LMarchinko - PaxtonMarchinko - McKayBarrueto - 1Barrueto - 2Lescak - InsectZeller et al.Mobley et al.-1 0 1-0.05 0.17-0.16-0.51-0.20-0.10-0.84-0.89 0.54 0.35 0.27-0.0195%-CI[-0.28;  0.17][-0.18;  0.51][-0.74;  0.42][-0.93; -0.08][-0.71;  0.30][-1.23;  1.03][-1.76;  0.07][-1.99;  0.21][-0.46;  1.54][ 0.06;  0.63][-0.42;  0.97][-0.27;  0.24]W100%14.1% 8.8%12.1%10.2% 3.3% 4.7% 3.5% 4.1%15.8% 6.9%16.7%d  84 Figure 4.4 : Forest plot of the effect size for all fish predation studies with a single estimate from each study. The grey box indicates the mean of the effect size (d) for each study and the lines give the 95% confidence interval of the effect. For studies containing multiple experiments, the mean effect size of each study was calculated using a fixed effects model. The weighted mean for all studies was calculated using the inverse variance method with a random effects mode. W is the weight of the study in the model. The weighted mean effect is indicated with the dotted line and the 95% confidence interval of this estimate is contained within the red triangle.      StudyWeighted MeanReistLescak et al.MacColl & ChapmanThis StudyLeinonen-1.5 -1 -0.5 0 0.5 1 1.5 0.14 0.07 0.36-0.04 0.06 0.2795%-CI[ 0.00; 0.28][-0.07; 0.20][ 0.13; 0.59][-1.02; 0.95][-0.20; 0.33][-1.11; 1.66]W100%50.7%25.3% 1.9%21.2% 1.0%d  85 Figure 4.5 : Forest plot of the effect size for all insect predation studies with a single estimate from each study. The grey box indicates the mean of the effect size (d) for each study and the lines give the 95% confidence interval of the effect. For studies containing multiple experiments, the mean effect size of each study was calculated using a fixed effects model. The weighted mean for all studies was calculated using the inverse variance method with a random effects mode. W is the weight of the study in the model. The weighted mean effect is indicated with the dotted line and the 95% confidence interval of this estimate is contained within the red triangle.       StudyWeighted MeanReistMarchinkoBarruetoLescak et al.Zeller et al.Mobley et al.-1.5 -1 -0.5 0 0.5 1 1.5 0.04-0.16-0.55-0.16 0.35 0.27-0.0195%-CI[-0.19; 0.28][-0.48; 0.16][-1.26; 0.16][-1.55; 1.24][ 0.06; 0.63][-0.12; 0.66][-0.27; 0.24]W100%21.7% 8.2% 2.6%23.7%18.0%25.8%d  86 Figure 4.6 : The effect size and standard error of (A) fish predation experiments and (B) insect predation experiments used in the meta-analysis.      -1.0 -0.5 0.0 0.5 1.01.00.80.60.40.20.0Effect sizeStandard error-1.0 -0.5 0.0 0.5 1.01.00.80.60.40.20.0Effect sizeStandard errorA B  87 Figure 4.7 : Forest plot of the effect size of standard length for all predation experiments. A random effects model was run for fish and insect predation studies. The grey box indicates the mean of the effect size (d) for each experiment and the lines give the 95% confidence interval of the effect. The weighted mean was calculated using a random effects model. W is the weight of the study in the model. The weighted mean is indicated with the dotted line and the 95% confidence interval of this estimate is contained within the red triangle.   Insect StudiesWeighted MeanMarchinko - PaxtonMarchinko - McKayBarrueto - 1Barrueto - 2Zeller et al.-1.5 -1 -0.5 0 0.5 1 1.5 0.27 0.69 0.21-0.06 0.64 0.0995%-CI[-0.14; 0.67][-0.48; 1.85][-0.67; 1.09][-1.11; 0.99][-0.31; 1.59][-0.60; 0.79]W100%12.0%21.1%14.8%18.2%33.9%Fish StudiesWeighted MeanMacColl & ChapmanLeionen - no refugeLeionen - refuge-1 0 1-0.02-0.03-0.02 0.0295%-CI[-0.83; 0.78][-1.02; 0.95][-1.98; 1.94][-1.94; 1.98]W100%66.3%16.8%16.8%dd  88 Table 4.1 : Experimental studies of selection on stickleback pelvic morphology by piscivorous predators. Variation in the length of pelvic spines (source of variation) was obtained from populations with naturally occurring variation, by creating crosses between populations with divergent pelvic spine lengths, or by experimentally modifying spines. All effect sizes were converted to the standardized mean difference (d). Significant effect sizes are given in bold. Trait abbreviations are as follows: PG: pelvic girdle, PS: pelvic spine, ST: spine triangle (triangle formed by pelvic and dorsal spines), SL: standard length. Unreplicated experiments were excluded from the meta-analysis (#).      89  Author' Year' Experiment' Species' Source' Population' Predator' Traits' Size'(mm)' Trials' Fish' logodds' 95%'CI' d' d'95%'CI'Reist& 1980& A&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Esox&lucius& PS& 20>29.9& 7& 205& 0.259& (>0.291,&0.808)& 0.142& (>0.160,&0.446)&& &B& Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Esox&lucius& PS& 30>39.9& 6& 170& >0.318& (>0.921,&0.286)& >0.175& (>0.508,&0.157)&& & C&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Esox&lucius& PS& 40>49.9& 6& 171& 0.473& (>0.117,&1.063)& 0.261& (>0.064,&0.586)&& & D&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Esox&lucius& PS& 20>29.9& 4& 115& 0.300& (>0.420,&1.020)& 0.165& (>0.231,&0.562)&& &E& Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Esox&lucius& PS& 30>39.9& 5& 150& >0.052& (>0.684,&0.581)& >0.029& (>0.377,&0.320)&& & F&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Esox&lucius& PS& 40>49.9& 4& 115& >0.172& (>0.893,&0.548)& >0.095& (>0.492,&0.302)&& & G&Culea&inconstans&modified&wild&caught&Wakomao&Lake,&AB& Esox&lucius& PS& 30>39.9& 2& 60& 0.777& (>0.221,&1.77)& 0.428& (>0.122,&0.978)&& &H& Culea&inconstans&modified&wild&caught&Wakomao&Lake,&AB&Esox&lucius& PS& 40>49.9& 2& 60& >0.129& (>1.123,&0.867)& >0.071& (>0.620,&0.478)&Ziuganov&&&Zotin& 1995& fish&Pungitius&pungitius& wild&caught&Levin&Navolok,&Russia&Perca&fluviatilis& PG& 51>62& 1&#& 200& 1.800& && 0.993& &&Lescak&et&al.& 2011& &&Gasterosteus&aculeatus& wild&caught&Wallace&Lake,&AK&Oncorhynchus&mykiss& PG& 37>45& 26& 260& 0.650& (0.228,'1.073)' 0.359& (0.126,'0.592)'Leionen&et&al.& 2011& &Gasterosteus&aculeatus& half>sib&crosses& Baltic&Sea& Esox&lucius& PS&6&months& 2& 325& & ' 0.175& (>1.789,&2.140)&&2011&&Gasterosteus&aculeatus& half>sib&crosses& Baltic&Sea& Esox&lucius& PS&6&months&2& 325&& ' 0.173& (>1.790,&2.137)&MacColl&&&Chapman& 2011& &&Gasterosteus&aculeatus& F2&Marine&x&Hoggan,&BC& Cottus&asper& PS,&SL&33.2>43.3& 16& 160& && && >0.037& (>1.025,&0.950)&This&study& & Paq&Gasterosteus&aculeatus&modified&wild&caught& Paq&Lake,&BC& Cottus&asper& PS,&SL& 36>51& 16& 128& 0.189& (>0.414,&0.793)& 0.104& (>0.228,0.437)&& &Ambrose& Gasterosteus&aculeatus&modified&wild&caught&Ambrose&Lake,&BC&Cottus&asper& PS,&SL& 41>63& 10& 80& 0.00&& (>0.775,&0.775)& 0.00& (>0.427,&0.427)&     90 Table 4.2 : Experimental studies of selection on stickleback pelvic morphology by insect predators. Variation in the length of pelvic spines (source of variation) was obtained from populations with naturally occurring variation, by creating crosses between populations with divergent pelvic spine lengths, or by experimentally modifying spines. All effect sizes were converted to the standardized mean difference (d). Significant effect sizes are given in bold. Trait abbreviations are as follows: PG: pelvic girdle, PS: pelvic spine, SL: standard length. One study estimated stickleback length from mean values (^). Experiments were excluded from the meta-analysis when results from multiple trials were presented as pooled data (*), or when experiments were unreplicated (#).  Author' Year' Experiment' Species' Source' Population' Predator' Traits' Size'(mm)' Trials' Fish' logodds' 95%'CI' d' d'95%'CI'Reist& 1979& I& Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Lethocerus&americanus& PS& 20>29.9& 11& 149& 0.306& (>0.320,&0.931)& 0.169& (>0.176,&0.513)&& & J&Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Lethocerus&americanus& PS& 30>39.9& 4& 53& >0.288& (>1.339,&0.764)& >0.159& (>0.738,&0.421)&& &K& Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Dysticus&spp.& PS& 20>29.9& 7& 98& >0.918& (K1.684,'K0.151)' >0.506& (K0.929,'K0.083)'& & L&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Dysticus&spp.& PS& 20>29.9& 5& 69& >0.371& (>1.289,&0.547)& >0.205& (>0.711,&0.301)&& & M&Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Dysticus&spp.& PS& 30>39.9& 1#& 14& 0.575& & 0.317& && &N& Culea&inconstans& wild&caught&Wakomao&Lake,&AB&Aeshna&spp.& PS& 20>29.9& 1#& 11& 2.99& ' 1.646& '&& && O& Culea&inconstans& wild&caught&Wakomao&Lake,&AB& Aeshna&spp.& PS& 20>29.9& 7*& 91& >0.308& && >0.170& &&Reimchen& 1980& &Gasterosteus&aculeatus& wild&caught&Boulton&Lake,&BC& Aeshna&spp.& PG& 15>25& 7*& 408& >0.159& & >0.088& &Ziuganov&&&Zotin& 1995& insect&Pungitius&pungitius& wild&caught&Levin&Navolok,&Russia&Odonata&spp.&Dysticus&spp.& PG& 51>62& 1&#& 200& >1.520& && >0.838& &&     91 Author' Year' Experiment' Species' Source' Population' Predator' Traits' Size'(mm)' Trials' Fish' logodds' 95%'CI' d' d'95%'CI'Marchinko& 2009& Paxton& Gasterosteus&aculeatus& F2&hybrids& Paxton&x&Marine&Aeshna&spp.&Notonecta&spp.&PS,&SL&& 10>18& 6& 477&& &>0.100& (>1.232,&1.032)&&& && McKay& Gasterosteus&aculeatus& F2&hybrids& McKay&x&Marine&Aeshna&spp.&Notonecta&spp.&PS,&SL&& 10>23& 10& 767& && && >0.842& (>1.757,&0.073)&Barrueto& 2009& 1& Gasterosteus&aculeatus&modified&lab&raised&Salmon&River& Notonecta&spp.& PS,&SL& 11>22& 7& 423& & & >0.888& (>1.986,&0.210)&&& && 2& Gasterosteus&aculeatus& backcrosses& Paxton&Lake,&BC&Notonecta&spp.& PG,&SL& 9>18& 8& 573& && && 0.538& (>0.460,&1.536)&Lescak&et&al.& 2012& &Gasterosteus&aculeatus& wild&caught&Wallace&Lake,&AK& Aeshna&spp.& PG& 23>57& 11& 220& 0.630& (0.112,'1.150)' 0.348& (0.062,'0.634)'Zeller&et&al.& 2012& && Gasterosteus&aculeatus& wild&caught&Bern,&Switzerland&Aeshna&spp.& PS,&SL& adults& 16& 960& && && 0.272& (>0.424,&0.968)&Mobley&et&al.& 2013& &Pungitius&pungitius&modified&wild&caught&Bothnian&Bay,&Sweden&Aeshna&spp.& PS& 29.9>35.1^&20& 200& >0.026& (>0.484,&0.432)& >0.014& (>0.267,&0.238)&  92 Chapter 5: An Experimental Test of the Effect of Predation Upon Behaviour and Trait Correlations in Threespine Stickleback1  5.1 Introduction Ecological speciation occurs when reproductive isolation evolves as a consequence of divergent natural selection between contrasting environments (Schluter 2009; Nosil 2012). While there are many examples of ecological speciation in nature, our understanding of the underlying mechanisms remains incomplete (Rundle and Nosil 2005; Nosil 2012). Divergent selection can occur in response to differences in resource availability and as a result of biotic interactions such as predation, competition, or intraguild predation (Schluter 2000; 2009; Miller et al. 2015). Experimental studies have shown that differential predation can lead to the evolution of divergent morphological traits (e.g. Jiggens et al. 2001; Vamosi and Schluter 2002; Rundle et al. 2003; Nosil and Crespi 2006; Langerhans et al. 2007; Diabaté et al. 2008; Marchinko 2009; Svanbäck and Eklöv 2011). However, less attention has been given to the role of divergent selection in the evolution of behavioural diversity.  Benthic and limnetic threespine stickleback (Gasterosteus aculeatus sp.) are a classic example of ecological speciation. The two species have evolved in sympatry in five lakes in coastal British Columbia (Schluter and McPhail, 1992). The species differ in many morphological and behavioural traits. Relative to benthics, limnetics have longer spines and more lateral plates (McPhail 1984; Vamosi 2002). Nesting males show habitat isolation 1 A version of this chapter has been published: Miller SE, Samuk KM, Rennison D. (2016). An experimental test of the effect of predation upon behaviour and trait correlations in threespine stickleback. Biological Journal of the Linnean Society   93  (Southcott et al. 2013). Limnetics have an increased shoaling preference (Vamosi and Schluter 2002; Wark et al. 2011), and are generally higher in the water column (Larson 1976). In comparison, benthics are more often solitary (Vamosi and Schluter 2002; Odling-Smee et al. 2008; Wark et al. 2011), and prefer to be lower in the water column (Larson 1976). Limnetics primarily eat zooplankton in the open water while benthics consume macroinvertebrates in the littoral zone (Schluter and McPhail 1992). In the open water, limnetics encounter coastal cutthroat trout (Oncorhynchus clarkii clarkii) more frequently (Reimchen 1994). Consequently, many of the phenotypic differences between the species are thought to be the result of differential predation on limnetics by trout (Vamosi and Schluter 2002).  Indirect evidence from observational or comparative studies is insufficient to determine if a trait is the target of divergent selection (Schluter 2009). The presence of aquatic predators can co-vary with environmental factors (e.g. abiotic conditions, food resources) (Jackson et al. 2001). Controlled experiments manipulating the presence/absence of predators are necessary to confirm that trait shifts are caused by divergent selection from predation. Comparing trait shifts between species is further problematic because species have fixed differences in many traits. As a result, it is difficult to separate the trait(s) that are the target of divergent selection from those traits that are genetically linked but not under direct selection. Predation may also lead to selection for correlations between advantageous combinations of behaviour and defence morphology (Sinervo and Svensson 2002; Murren 2012). Creating advanced generation crosses between species with divergent phenotypes can create trait combinations not normally seen in the   94 wild. When such crosses are combined with predator exposure, it is possible to test if predation is responsible for changes in traits and trait correlations.  We experimentally tested the hypothesis that differences in behaviour between benthic and limnetic stickleback are the result of divergent selection from coastal cutthroat trout predation. Benthic-limnetic hybrid families were introduced into large, naturalistic experimental ponds in the presence/absence of trout predation. Experimental stickleback reproduced annually in the ponds and underwent two generations of differential selection prior to measurement in behavioural assays. We measured two putative anti-predator behaviours, which have been previously shown to differ between the two species - preferred position in the water column and shoaling preference (Larson 1976; Vamosi 2002; Kozak and Boughman 2008; Wark et al. 2011). Behaviours that differ consistently between control and predation ponds can be interpreted to arise in response to trout predation. We then tested for correlations between behaviour and defensive armour, and compared the strength of these correlations between treatments. If trout predation selects for combinations of behaviour and defensive armour, trait correlation will be greater in the predation treatment.   5.2 Methods 5.2.1 Experimental Design In May 2011, four F1 crosses were made between wild-caught benthic females and limnetic males from Paxton Lake, Texada Island. The F1 crosses were reared in 300L tanks in the laboratory without predators for one year until adulthood. In May 2012, adult stickleback were collected from First Lake, an advanced generation hybrid population. First   95 Lake is a small shallow lake on Texada Island that was founded in 1981 with Paxton Lake benthic x limnetic F1 stickleback (McPhail 1993). We consider this population to be a single family of ~F29 benthic-limnetic hybrids at the time of sampling. The First Lake population was included in the study because the greater number of recombination events this population has undergone affords us the opportunity to investigate the effect of linkage on adaptation.  In May 2012, the five hybrid families (Four F1s and one First Lake) were introduced in a split plot design to pairs of semi-natural ponds (n=21-31 individuals/pond; 10 ponds total) at the University of British Columbia’s experimental pond facilities. Each paired pond contained a single family. Stickleback bred in all experimental ponds creating F2s or ~F30s (First Lake ponds) in the summer of 2012. In the summer of 2013, the F2/F30 stickleback bred to form a F3/F31 generation. All behavioural assays were conducted on adult stickleback from the 2013 (F3/F31) cohort. The experimental ponds are 25m x 15m with a shallow littoral area and a 6m deep open water region. These ponds contain a natural assemblage of food resources and contain invertebrate and avian predators. For each set of paired ponds, one pond was randomly assigned to a predation treatment and the other pond to a control treatment. Adult coastal cutthroat trout were collected from Placid Lake in the Malcolm Knapp Research Forest located 50 km east of Vancouver, BC. Two trout were added to each predation pond in September 2012. The trout died in the summer of 2013 and were replaced with three new trout in September 2013.     96 5.2.2 Behavioural Assays Behavioural assays were conducted from November 8-14, 2013, in tanks adjacent to the experimental ponds. Twelve randomly chosen stickleback were collected from each pond with unbaited minnow traps (n=120 total). Paired ponds were tested sequentially, alternating between treatments. Sticklebacks were transferred in a bucket from the pond to the behavioural assay area for a 15-minute acclimation period prior to the start of the behavioural trials. At that time, each stickleback was placed into an individual mesh basket inside a larger aquarium so that we could follow the behaviour of individuals across assays. Behavioural tests were conducted in the following order: stickleback were tested in the novel tank test, returned to the holding basket for 15 minutes, and then tested in the shoaling assay.   The novel tank diving test measures stickleback movement and position in a new tank. Vertical position in the water column of a tank has been used as a proxy for habitat usage in guppies (Poecilia) and stickleback (Larson 1976; Torres-Dowdall et al. 2012; Miller et al. 2015). In zebrafish, anxiety (e.g. following exposure to alarm pheromones) leads to a reduction in exploration and a lower position in a tank (Egan et al. 2009; Cachat et al. 2010; Stewart et al. 2012). During the trial, a focal fish was gently introduced to the top centre of an empty unfamiliar 35.5 cm x 22 cm x 20 cm tank and allowed to move freely for 630 seconds. All assays were recorded with wireless D-Link DCS-930L webcams (DLink Corporation, Taiwan). We excluded the first 30 seconds of each assay as the introduction of a stickleback often resulted in erratic movement (Miller et al. 2015). Videos were subsampled to 0.5 frames per second using VirtualDub software (www.virtualdub.org). The MtrackJ plugin (Meijering et al. 2012) in ImageJ (Schneider et al. 2012) was used to   97 measure the x and y coordinates of the focal fish every 2 seconds. We calculated the mean vertical position of the focal fish, the latency to enter the upper half of the tank, and the distance that the focal fish travelled during the assay.  The second assay assesses shoaling preference by measuring the time that the focal stickleback spends near a stimulus shoal (Vamosi 2002; Kozak and Boughman 2008; Wark et al. 2011). Assay tanks were 75 cm x 30 cm x 46 cm with two 10 cm end compartments on either side of the tank that were separated from a large centre arena with window screen (Figure 2.4). Ten stimulus stickleback (shoal) were added to one end compartment and two stimulus stickleback (distractor) were added to the other end compartment (Wark et al. 2011). The stimulus sticklebacks were limnetic stickleback from Priest Lake reared at the experimental pond facility. This population was unrelated and unfamiliar to the experimental stickleback and was chosen because individuals have a high shoaling tendency (Wark et al. 2011) and were similar in size to the experimental stickleback. At the start of the shoaling assay, the focal stickleback was gently introduced into the centre arena and was allowed to move for 630 seconds. We measured the x and y coordinates of the focal fish every 2 seconds following the method used in the novel tank test. We used two metrics to assess shoaling behaviour: the mean horizontal position in the tank (shoaling position), and the time that the focal fish spends within one body length of the experimental shoal (shoaling preference). As a result of camera error, two trials were not analysed. Following Wark et al. (2011), we excluded trials in which the focal fish did not move during the trial (novel tank n=10; shoaling n=12). In total, 110 novel tank trials and 108 shoaling trials were measured.    98  5.2.3 Armour Traits Immediately following the shoaling assay, stickleback were euthanized in MS-222 and fixed in 10% formalin. Specimens were later stained with alizarin red to highlight bony structures following established protocols (Peichel et al. 2001). On the left side of each stained specimen we measured the length of length of the first and second dorsal spines, pelvic spine, pelvic girdle, the number of lateral plates and standard length. Specimens lacking an armour component were assigned a value of zero. Lateral plate number and standard length were not significantly correlated. All other armour traits were positively correlated with standard length and were size corrected to the average length (43.82 mm) using the equation !! = !! − !!(!! − L). Where !! is the size-adjusted trait, !! is the original trait,!! is the regression coefficient of the original trait values on standard length, !! is the standard length of the individual and L is the average length (Vamosi 2002). For second dorsal spine, pond had a significant effect on β and thus this trait was size corrected independently for each pond (pond did not have a significant effect for other traits). A principal component analysis (PCA) of the correlation matrix of size-corrected armour traits was used to visualize the overall defensive armour of each stickleback. The first principal component (PC1) accounted for 40.9% of the variation in stickleback armour and primarily describes the pelvic spine and pelvic girdle (Table 5.1). The second principal component (PC2) accounted for 25.8% of the variation and describes the length of the first and second dorsal spine.    99 5.2.4 Statistical Analysis  A linear mixed effects model was used to test if performance in behavioural assays differed between treatments and if armour traits affected these behaviours. Principal component score, treatment, and population (Paxton Lake or First Lake) were fixed factors. Pond and family were random factors. Population was not a significant covariate and was dropped from the final model.   All traits were not normally distributed. Therefore, Spearman’s rank correlations were used to evaluate the correlations between armour and behavioural measurements. Confidence intervals for trait correlations were calculated by bootstrapping (1000 replicates) with RVAideMemoire (Hervé 2014). For traits with significant correlations, we compared the magnitude of the correlations between treatments using the Wilcoxon signed-rank test on Spearman rank correlations calculated separately for each pond. All statistical analysis were conducted in R (version 3.1) (R Core Team, 2014)  5.3 Results The presence of trout did not have a measurable effect upon stickleback behaviour (Table 5.2; Figure 5.1). Predation and control ponds did not differ in vertical position in the water column, the latency to enter the upper half of the tank, or distance travelled during the novel tank assay. Fish from all ponds spent more time shoaling than the random expectation, regardless of treatment (one sample t-test: t=9.29, P<0.0001, df=10). In the shoaling assay, we observed a trend of increased time spent with the shoal (shoaling preference) in the control ponds for four of the five families (Treatment: F1,4 =3.24,   100 P=0.15), and focal fish from control ponds travelled more during the assay (Figure 5.2; Treatment: F1,4 =5.69, P=0.08), although these results were not significant.  We observed variation in armour traits among experimental families (Table 5.3). The first PC differentiated stickleback with robust pelvic armour (limnetic-like) and stickleback with reduced pelvic armour (benthic-like), while PC2 separated individuals with longer dorsal spines (limnetic-like) from those with reduced dorsal spines (benthic-like). Predation and control ponds did not differ in PC1 (Treatment: F1,4=0.43, P=0.55), PC2 (Treatment: F1,4=2.5, P=0.18), or standard length (Treatment: F1,4=0.19, P=0.69).  There was a positive correlation between PC1 score and mean vertical position during the novel tank test (Figure 5.2A; Spearman’s rank correlation coefficient, ρ= 0.261, P=0.006, 95% CI: 0.068-0.442). Individuals with increased pelvic armour preferred a higher vertical position in the water column (PC1: F4,97=4.10, P=0.045). There was a negative correlation between PC2 and distance travelled during the novel tank test (Figure 5.2C; ρ= -0.260, P=0.006, 95% CI: -0.428 , -0.071). Scores along PC2 and distance travelled during the shoaling assay were not correlated with each other (Table 5.4), but there was a significant Treatment x PC2 interaction (F1,95=4.52, P=0.04). One individual had an extreme value for PC2; however, the correlation between these traits remained significant when this point was removed (without point, ρ= -0.245, P=0.01). Behaviour was not correlated with standard length (Table 5.4). All other armour and behaviour correlations were non-significant (Table 5.2, Table 5.4).  Trout predation did not change the strength of the correlations between PC1 and water column position (Figure 5.2B; Wilcoxon signed-rank test, z=9, n=5, P=0.812), or PC2 and distance travelled during the water column assay (Figure 5.2D; z=5, n=5, P=0.625).   101  5.4 Discussion Divergent selection from trout predation has been hypothesized to be an important driver of behavioural differences between benthic and limnetic stickleback (Larson 1976; Vamosi 2002; Vamosi and Schluter 2004; Wark et al. 2011). To test this hypothesis, we reared families of benthic-limnetic hybrids in natural-like experimental ponds in the presence or absence of trout predation. Contrary to predictions, there was no significant difference in behaviour between predation and control ponds. Instead, armour morphology was a stronger predictor of behaviour than trout predation.   5.4.1 Stickleback Behaviour Preferred position in the water column did not differ between predation and control ponds. Stickleback in predation ponds had a decreased shoaling preference, but this result was non-significant. If differences in benthic and limnetic behaviour are not caused by divergent selection from trout predation, then behavioural differences may be the result of selection from other factors that differ between the benthic and limnetic habitats. For example, benthics forage for invertebrates in the littoral zone, while limnetics eat zooplankton near the surface of the water (Larson 1976; Odling-Smee et al. 2008). Therefore differences in water column preference may be caused by divergence in diet and/or foraging behaviour between the two species. Similarly, limnetics are frequently observed in large aggregations (Larson 1976) and have a stronger shoaling preference than benthics (Vamosi 2002; Kozak and Boughman 2008; Wark et al. 2011). The differences in shoaling behaviour in the lakes may be due to differences in the structural complexity and   102 amount of open space between the two environments (Odling-Smee et al. 2008) rather than a consequence of increased trout predation. A shift in resource or habitat use could also have driven changes in shoaling preference. Compared to control ponds, predation ponds had a decrease in population density and a shift in diet towards benthic resources (Rudman et al. 2016). Selection for benthic-like trophic characteristics may have led to a decrease in shoaling preference. Trout predation may have also led to non-consumptive changes in behaviour by reducing competition and increasing intimidation in the open water environment (Preisser et al. 2005). Our findings suggest that differential predation alone is unlikely to explain the differences in shoaling behaviour and water column preference observed in the wild.  The experimental ponds provide an improvement over behavioural studies conducted in mesocosms or in the laboratory because experimental subjects can be manipulated in a natural environment. However, the paired design limited the statistical power of this experiment to detect small differences in behaviour between treatments. Additionally, behaviours were assayed at a single end point; therefore, if paired ponds did not start at the same trait value this would decrease our ability to detect a treatment effect.  5.4.2 Correlations Between Morphology and Behaviour The likelihood that an individual escapes a predation event may be determined by an interaction between behavioural and morphological traits (e.g. Brodie 1992; Dewitt 1999; Buskirk and McCollum 2000; Relyea 2001). We found a correlation between behavioural traits and bony armour. Armour PC1 (increased pelvic armour) was associated with a higher position in the water column and armour PC2 (longer dorsal spines) was   103 associated with increased movement during the water column assay. These correlations may be underestimated because behavioural traits have high variance and any measurement error can decrease the correlation between traits (Whitlock and Schluter 2014). As a result, correlations between these traits in the wild are likely greater than reported in this study. Functionally these associations match the greater pelvic armour and preference for a higher water column position found in limnetics (Larson 1976). A previous study by Grand (2000) found that within benthic stickleback that those individuals with reduced pelvic armour were less bold than individuals with increased pelvic armour.  The observed correlations between armour morphology and behaviour could result from genetic linkage or pleiotropy (Schlosser and Wagner 2004). Several inferences can be made regarding the possible genetic basis of the correlations. Recombination events in advanced generation hybrids should uncouple many traits that were genetically linked in limnetics and benthics. Yet three generations of recombination were insufficient to break up the association between armour and behaviour in the F3 families and >30 generations of recombination in First Lake ponds did not decrease the correlation. The maintenance of these correlations in spite of genome-wide recombination indicates that genetic linkage or pleiotropy underlies these associations.  Prior studies in stickleback support a role for linkage or pleiotropy between behaviour and morphology. Lateral plate number and body orientation during schooling have been genetically mapped to the same chromosomal segment (Greenwood et al. 2013). A single gene (Ectodysplasin) in this low recombination region has been previously shown to have pleiotropic effects upon lateral plate development, neuromast position, schooling behaviour, and salinity preference (Barrett et al. 2009; Wark and Peichel 2009;   104 Wark et al. 2012; Mills et al. 2014). A recent study has also uncovered a correlation between anti-predator behaviour and pigmentation in juvenile stickleback (Kim and Velando 2015), suggesting that these correlations may be more widespread then previously appreciated.  When certain trait combinations are preferentially favoured, natural selection may directly or indirectly lead to an increase in the correlation between these traits (Sinervo and Svensson 2002; Murren 2012). While we describe a correlation between multiple armour and behavioural traits, the strength of these correlations did not differ between treatments. Therefore we were unable to support the hypothesis that trout predation is the causal mechanism for the associations. However, the lack of change in correlation between treatments could be a consequence of the limited power of our experiment, or insufficient variation in correlation for selection to act upon. Trout may have also played an important role during the historical divergence between benthic and limnetic stickleback. Therefore, while trout predation may not be the proximate cause for the correlation between defence morphology and behaviour, it cannot be ruled out as the ultimate cause for this association. Future work examining the genetic basis of these traits will be required to elucidate the role of pleiotropy and linkage in behaviour and armour morphology in stickleback.      105  Figure 5.1 : Mean value for behavioural traits between control and predation ponds presented as reaction norms. The standard error is given for each pond. Each family is represented with a separate colour with the First Lake family given in red.     control predation260300340380Water Column Position(a)control predation2000600010000Water Column Distance (pixels) (b)control predation50100150Shoaling Preference (seconds) (c)control predation56789Shoaling Position(d)control predation300050007000Shoaling Distance (pixels)(e)control predation35404550Standard Length (mm)(f)control predation-4-2012Armour PC1(g)control predation-2.0-1.00.01.0Armour PC2(h)  106 Figure 5.2 : (A) Association between the mean position in the water column and armour PC1 with linear regression line. Trait variation in PC1 (lateral plates and pelvic spines) is shown in red along the x-axis. Each point is an individual from either a predation (filled symbols) or control (open symbols) pond. (B) Spearman’s correlation coefficient between armour PC1 and mean vertical position in the water column for each pond. Paired ponds are connected with a line. The F3 families are circles and the family from First Lake is a square. (C) Association between armour PC2 and distance travelled during the water column assay with linear regression line. Trait variation in PC2 (first and second dorsal spines) is shown in red along the x-axis. Individuals from predation ponds are indicated with filled symbols and individuals from control ponds are shown with open symbols. (D) Spearman’s correlation coefficient between armour PC2 and distance in the water column assay. Paired ponds are connected with a line. The F3 families are circles and the First Lake family is a square.   107      108 Table 5.1 Trait Loadings from the principal component analysis (PCA).  ! !PC1$ PC2$Trait$! !!Dorsal!Spine!1! 0.111! 0.716!!Dorsal!Spine!2! 0.195! 0.659!!Pelvic!Spine! 0.641! 90.114!!Pelvic!Girdle! 0.632! 90.196!!Lateral!Plates! 0.372! 90.031!  109 Table 5.2 Results of the linear mixed effects model of behavioural traits, armour principal component score, and treatment. The 95% confidence intervals are given for each effect. Significant associations are in bold.  !! !! Treatment( !! !! !! PC1( !! !! !! !! Treatment(x(PC1( !!Novel(Tank(Test( df! F! P! 95%!CI!!df! F! P! 95%!CI!!df! F! P! 95%!CI!!Mean%Vertical%Position% 1,4! 1.42! 0.30! 229.4,!4.7!!1,97! 4.10! 0.05( 20.8,!17.3!!1,97! 0.14! 0.71! 215.0,!10.2!!Latency%to%upper%tank% 1,4! 1.12! 0.35! 237.3,!77.5!!1,97! 0.17! 0.68! 26.5,!35.6!!1,97! 2.69! 0.10! 255.1,!5.2!!Distance%traveled% 1,4! 0.98! 0.38! 24026,!1808!!1,97! 0.78! 0.38! 2886,!1027!!1,97! 0.61! 0.44! 2820,!1888!! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !Shoaling(Assay(! ! ! ! ! ! ! ! ! ! ! ! ! !!Mean%horizontal%position% 1,4! 0.71! 0.45! 21.5,!0.77!!1,95! 0.04! 0.83! 20.5,!0.4!!1,95! 0.03! 0.86! 20.7,!0.6!!Shoaling%preference% 1,4! 3.24! 0.15! 284.0,!11.9!!1,95! 0.24! 0.62! 223.7,!16!!1,95! 0.00! 0.98! 228.9,!29.7!!! Distance%traveled% 1,4! 5.69! 0.08! 23266,!252! !! 1,95! 0.00! 0.94! 2784,!511! !! 1,95! 0.52! 0.47! 2571,!1226!! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !!! !! Treatment( !! !! !! PC2( !! !! !! !! Treatment(x(PC2( !!Novel(Tank(Test( df! F! P! 95%!CI!!df! F! P! 95%!CI!!df! F! P! 95%!CI!!Mean%Vertical%Position% 1,4! 1.33! 0.31! 234.2,!15.7!!1,97! 0.49! 0.49! 216.4,!6.0!!1,97! 0.38! 0.54! 211.0,!20.8!!Latency%to%upper%tank% 1,4! 0.85! 0.41! 244.8,!86.6!!1,97! 0.02! 0.88! 239.6,!12.5!!1,97! 3.6! 0.06! 21.9,!75.6!!Distance%traveled% 1,4! 0.90! 0.40! 21838,!1585!!1,97! 1.69! 0.20! 21696,!609!!1,97! 0.02! 0.88! 21838,!1585!! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !Shoaling(Assay(! ! ! ! ! ! ! ! ! ! ! ! ! !!Mean%horizontal%position% 1,4! 0.75! 0.44! 21.2,!0.6!!1,95! 1.78! 0.19! 20.6,!0.5!!1,95! 2.3! 0.13! 21.4,!0.2!!Shoaling%preference% 1,4! 3.27! 0.15! 282.7,!12.8!!1,95! 0.81! 0.37! 238.0,!11.1!!1,95! 0.39! 0.54! 226.1,!49.9!!! Distance%traveled% 1,4! 5.88! 0.07! 23269,!189! !! 1,95! 0.09! 0.77! 2176,!1378! !! 1,95! 4.52! 0.04( 22380,!282!     110 Table 5.3 Measured trait values for each pond. First dorsal spine (DS1), second dorsal spine (DS2), pelvic spine (PS), and pelvic girdle (PG) are size corrected. Stickleback are hybrids between benthics and limentics from Paxton Lake. The population from Paxton Lake are from the F3 generation and the population from First Lake are from the F31 generation. The mean and standard error are given for each trait.  Population( Family( Treatment( N( DS(1((mm)( DS(2((mm)( PS((mm)( PG((mm)( LP( Length((mm)( Armour(PC1( Armour(PC2(Paxton! D1! control! 12! 1.8!±!0.3! 2.2!±!0.3! 2.3!±!0.4! 5.3!±!0.8! 3.9!±!0.3! 46.8!±!0.6! 0.20!±!0.4! 21.13!±!0.4!Paxton! D1! predation! 11! 1.7!±!0.3! 3.0!±!0.1! 2.0!±!0.5! 4.6!±!0.9! 3.5!±!0.3! 47.4!±!1.0! 0.05!±!0.44! 20.18!±!0.16!First!Lake! D221! control! 12! 1.9!±!0.3! 3.1!±!0.1! 0.2!±!0.2! 0.7!±!0.3! 1.8!±!0.4! 43.9!±!1.5! 21.95!±!0.15! 0.39!±!0.23!First!Lake! D221! predation! 12! 1.6!±!0.4! 3.2!±!0.1! 1.0!±!0.5! 2.5!±!0.8! 2.3!±!0.4! 41.5!±!1.6! 21.03!±!0.34! 0.30!±!0.30!Paxton! D222! control! 12! 0.6!±!0.3! 3.0!±!0.1! 2.2!±!0.4! 4.9!±!0.8! 3.8!±!0.2! 39.1!±!0.9! 0.20!±!0.31! 21.00!±!0.23!Paxton! D222! predation! 12! 0.5!±!0.3! 2.9!±!0.3! 3.0!±!0.3! 5.6!±!0.5! 3.8!±!0.2! 35.7!±!1.0! 0.74!±!0.29! 20.83!±!0.20!Paxton! D321! control! 11! 2.1!±!0.4! 3.4!±!0.1! 2.1!±!0.5! 4.6!±!0.7! 3.3!±!0.4! 43.0!±!1.1! 0.04!±!0.41! 0.23!±!0.33!Paxton! D321! predation! 12! 2.7!±!0.3! 3.4!±!0.1! 2.9!±!0.5! 4.4!±!0.8! 4.6!±!0.2! 45.2!±!1.3! 0.92!±!0.37! 0.75!±!0.27!Paxton! D322! control! 12! 2.4!±!0.3! 3.2!±!0.3! 2.8!±!0.4! 5.4!±!0.6! 3.8!±!0.3! 47.9!±!1.5! 0.92!±!0.31! 0.74!±!0.18!Paxton! D322! predation! 12! 2.2!±!0.3! 3.5!±!0.1! 2.0!±!0.4! 4.3!±!0.8! 2.7!±!0.4! 48.4!±!1.3! 20.08!±!0.38! 0.73!±!0.28!    111 Table 5.4 Spearman’s rank correlations between behavioural traits and morphological traits. Correlations were calculated between behavioural traits, and armour PC1, armour PC2, and standard length. The 95% confidence intervals (95% CI) were calculated for each correlation by bootstrapping. All significant correlations are in bold.  PC1$ !! !! !!!Rho$ P$ 95%$CI$Water!column!position! 0.261! 0.006$ 0.068,!0.442!Water!column!latency! 0.213! 0.086! :0.032,!0.423!Distance!water!column!assay! 0.169! 0.078! :0.028,!0.353!Shoaling!preference! :0.013! 0.890! :0.202,!0.179!Shoaling!position! :0.110! 0.261! :0.295,!0.085!Distance!shoaling!! 0.011! 0.909! :0.167,!0.183!! ! ! !! ! ! !PC2$ !! !! !!!Rho$ P$ 95%$CI$Water!column!position! :0.104! 0.283! :0.090,!0.283!Water!column!latency! :0.108! 0.387! :0.103,!0.334!Distance!water!column!assay! 0.260! 0.006$ :0.428,!:0.071!Shoaling!preference! 0.088! 0.367! :0.283,!0.127!Shoaling!position! 0.094! 0.333! :0.280,!0.112!Distance!shoaling!! :0.003! 0.978! :0.202,!0.205!! ! ! !! ! ! !Standard$Length$ !! !! !!!Rho$ P$ 95%$CI$Water!column!position! 0.147! 0.128! :0.051,!0.342!Water!column!latency! 0.131! 0.294! :0.090,!0.359!Distance!water!column!assay! 0.029! 0.764! :0.167,!0.208!Shoaling!preference! :0.065! 0.506! :0.242,!0.124!Shoaling!position! :0.094! 0.337! :0.265,!0.090!Distance!shoaling!! :0.026! 0.787! :0.225,!0.148!  112 Chapter 6: Conclusion  6.1 Overview Organisms experience selection from both the abiotic and the biotic environments. As a result, interactions among species can be a mechanism of evolution by natural selection within a species (Thompson 2013). My dissertation has attempted to quantify the impact that biotic selection from a single species has had on the evolution of another species. I focus primarily on the evolution of trait divergence in the threespine stickleback in response to intraguild predation. In chapter 2, I established that the presence of an intraguild predator, prickly sculpin, is associated with character shifts in multiple traits in the threespine stickleback. Wild populations of stickleback sympatric with sculpin were previously shown to differ in body shape (Ingram et al. 2012). I demonstrated that compared to stickleback from lakes without sculpin, populations of stickleback from lakes with sculpin show parallel increases in armour morphology and prefer to be higher in the water column in laboratory tests. These differences in armour, shape, and behaviour persisted when stickleback crosses from these populations were raised in a common garden, suggesting that trait differences have a genetic basis. I then examined if experimental exposure to sculpin induced trait changes. I found limited phenotypic plasticity in marine stickleback in response to sculpin exposure, but I did not observe an induced response in the freshwater stickleback. Behavioural and morphological trait differences between freshwater populations with and without sculpin thus have a genetic basis and suggest an evolutionary response to intraguild predation.   113 These findings are the first confirmed case of genetically based character divergence associated with intraguild predation.  The evolution of phenotypes is tied to the genes that underlie the phenotypes. In chapter 3, I investigated the genomic architecture of divergence between stickleback from lakes with and without sculpin. The effect of biotic selection upon the entire genome of an organism is largely unknown in natural populations. Therefore this study provides the first description of the genome-wide response of a vertebrate species to a single biotic agent in the wild. I used whole genome re-sequencing of eight populations from lakes without sculpin, nine populations from lakes with sculpin, and six marine populations. I found that genetic variation in these populations is strongly associated with the presence or absence of sculpin. Using a genome scan metric modified from Jones et al. (2012a), I identified regions of the genome that have differentiated in parallel between lakes with and without sculpin. I find that intraguild predation is associated with extensive genomic differentiation. Outlier windows between lakes with and without sculpin contain more than 500 genes with diverse functions. Genes identified in outlier windows contain many candidate genes for further study. However, I did not find overrepresentation of any GO terms for genes in outlier regions.  Outlier windows were distributed unevenly across the genome. Some chromosomes had a large number of outlier windows while other chromosomes were underrepresented. The number of genes or windows differentiating populations does not count the number of selective sweeps because a selective sweep will likely encompass multiple outlier windows. I used a hidden Markov model to estimate the location of state shifts between divergent regions and regions having little or no divergence to estimate the number of selective   114 sweeps. This model collapsed the 1473 outlier windows into 164 distinct outlier regions across the genome. This smaller number of outlier regions still represents a large portion of the genome that is differentiated.  Observing trait differences between populations with and without sculpin is insufficient evidence to conclude that sculpin are the agent of selection leading to the evolution of those traits (Wade and Kalisz 1990). The presence of sculpin could be merely correlated with another mechanism of selection (e.g. decreased time in benthic habitat, see Chapter 4). For example, stickleback from lakes with sculpin have pelvic spines that are on average 2.3mm longer than stickleback from lakes without sculpin (Chapter 2). Longer pelvic spine have been hypothesized to be a defence against fish predators because spines are sharp and increase the effective diameter of the stickleback making it more difficult to ingest (Hoogland et al. 1956; Hagen and Gilbertson 1972). In contrast, shorter pelvic spines may provide a selective advantage against insect predators (Reimchen 1980). In chapter 4, I used a mesocosm experiment to test if prickly sculpin are an agent of selection on stickleback pelvic spine length. Using clippers, I physically shortened stickleback pelvic spines of individuals from two populations of stickleback sympatric with sculpin and compared the mortality rates of stickleback with clipped and unclipped pelvic spines in the presence of sculpin. I observed an increase in the probability of survival of stickleback with unclipped pelvic spines, but this effect had wide confidence intervals and was not statistically significant. From the mesocosm experiment I was unable to conclude whether prickly sculpin preferentially consumed stickleback with shortened pelvic spines.  I combined the results of my mesocosm experiment with other experimental studies of selection on stickleback pelvic morphology from fish or insect predators. I used a meta-  115 analysis approach to evaluate if fish predators or insect predators were agents of selection on pelvic spine length. From this analysis I concluded that fish predators indeed favoured longer pelvic spines by preferentially eating shorter-spined fish. The summary effect size of fish predators on the length of pelvic spines, 0.13 units of standard deviation, was comparable to other measures of selection in the wild (Kingsolver et al. 2001). While this effect size is considered small to moderate, if consistent over multiple generations, and assuming a heritability of 0.3, it is sufficient to change the mean length of pelvic spines by one standard deviation in less than 100 generations. The summary effect of insect predators on stickleback pelvic spine length had wide confidence intervals and overlapped with zero. Therefore I was unable to rule out selection for increased or decreased pelvic morphology by insect predators. Many experimental measures of selection have small sample sizes with a low number of replicates, resulting in wide confidence intervals for the estimate of selection (Hersch and Phillips 2004). As this study demonstrates, combing results from multiple experimental studies can produced an aggregate estimation of selection that may be more precise than any single study (Hersch and Phillips 2004; MacColl 2011).  In my final chapter, I examined if biotic selection from trout is a mechanism of divergent selection on stickleback behaviour. Benthic and limnetic stickleback occur in sympatry in multiple lakes (Schluter and McPhail, 1992) and have diverged in behavioural traits and quantity of armour (Larson, 1976; Vamosi and Schluter, 2002; Wark et al., 2011). Behavioural divergence between these stickleback species has been hypothesized to be the result of divergent selection driven in part by increased predation from coastal cutthroat trout on limnetics in the open water. However, differences in behavioural traits between   116 benthic and limnetic stickleback could be the consequence of another mechanism of selection. For example, divergence in diet between the species may lead to different foraging strategies and consequently result in differences in preference in the water column. In a selection experiment, I tested if the presence or absence of coastal cutthroat trout predation was correlated with differences in behaviour. Split families of benthic-limnetic F1 hybrids were reared for three generations in large experimental ponds in the presence or absence of trout predation. I compared shoaling preference and preferred position in the water column of stickleback taken from the control and predation treatment ponds. Stickleback behaviour did not differ appreciably among treatments and estimates of the effect of trout on behavioural traits had wide confidence intervals. Therefore it is possible that either trout are not an agent of selection on benthic and limnetic behavioural differences, or that the study lacked the power to detect differences in behaviour between treatments.  6.2 Broader Implications The work described in my thesis has several implications. First, intraguild predation is an important mechanism of divergence. Intraguild predation is a taxonomically widespread species interaction (Polis et al. 1989; Arim and Marquet, 2004). Furthermore, intraguild predation may be more common then previously reported, as many examples of character shifts currently attributed to competition may instead be the result of intraguild predation. For example, a decrease in range overlap between large mammalian carnivores with similar carnassial tooth length is frequently cited as an example of character displacement (Davies et al. 2007). However, interspecific killings occur more often between   117 carnivore species with similar diets (Donadio and Buskirk 2006) and in regions with the most resources (Vanak et al. 2013). As a consequence, shifting to an alternative habitat decreases competition and predation (Donadio and Buskirk 2006) and may be an adaptive response to either (although see Pfennig and Pfennig 2010). Phenotypic plasticity in the ancestral marine colonizing population may have aided in the initial adaptation to freshwater populations (Wund et al. 2008), but subsequent trait divergence among lakes with and without sculpin is due to genetic differences. Phenotypic differences between populations in different environments can be caused by evolution or by phenotypic plasticity – the ability of a organism to modify its phenotype in response to environmental change (West-Eberhard 2003). In a common garden, I found that the presence of sculpin induced slightly increased armour, an increase in preferred water column height, and a decrease in shoaling behaviour in marine stickleback. Importantly, induced trait changes in the presence of sculpin were in the same direction as the trait shifts among freshwater stickleback populations with and without sculpin. In contrast, I found no evidence for sculpin-induced plasticity in stickleback from freshwater populations reared in the presence of sculpin. Instead, trait differences between stickleback populations from lakes with and without sculpin had a genetic basis. The role of phenotypic plasticity in driving diversification has been the subject of recent debate (Pfennig et al. 2010). The advantage of phenotypic plasticity, particularly in the case of rapid evolution, is that upon exposure to a novel environment, phenotypes can arise quickly without the need for the mutation and spread of adaptive alleles (West-Eberhard 2003). However, phenotypic plasticity can also be costly to maintain (Dewitt et al. 1998). For trait divergence among   118 stickleback populations with and without sculpin, the role of phenotypic plasticity was likely limited. My study suggests that the addition or removal of a single biotic agent can have a profound effect on genomic divergence in stickleback. More than 500 genes were differentiated between lakes with and without sculpin. Furthermore, the number of differentiated genes is likely an underestimate as my methodology only identifies genes in outlier regions that have diverged in parallel between lakes with and without sculpin. Intraguild predation was predicted to cause evolution of the genes for foraging and anti-predatory defence traits. I identified genes in outlier windows that are potential candidates for these traits. In addition, outlier windows contain genes for other phenotypes that may differ between stickleback from lakes with and without sculpin. These phenotypes, such as differences in brain development or immunity, would not have been easily detected by traditional morphological studies. Together, the presence or absence of sculpin has an effect on numerous genes and phenotypes. It will be necessary to examine the genome-wide response to biotic selection in other organisms to determine if these findings are typical. However, there are methodological constraints that researchers should consider when designing these studies. Previous studies of the genetic basis of selection have relied on a candidate gene approach in which the researcher selects genes to study a priori (Nadeau and Jiggins 2010) or reduced representation genome scans (e.g. Hohenlohe et al. 2010). These approaches have severe limitations. Candidate gene studies are constrained by both the number of divergent phenotypes identified as well as the researcher’s ability to identify putative candidate genes for those phenotypes. These approaches do not adequately address   119 quantitative or polygenic traits (Akey 2009). Additionally, reduced representation genome scans are unable to detect selection on genes that are not linked to markers. Therefore, by not examining the entire genome, these methods will fail to detect many genes thereby underestimating the number of genes under selection. An understanding of the genomic consequences of biotic selection will require more whole genome sequencing studies.  These findings have ramifications for the role of biotic interactions in the evolution of species. While comparisons are preliminary, more genes may be under selection from biotic interactions than from abiotic agents of selection. When a species adapts to another species, they are co-evolving with each other (Thompson 2013). Therefore adaptation to an abiotic agent is likely to be a stationary adaptive peak, but adaptation to another species may involve a constantly moving target. The multifarious selection hypothesis of speciation predicts that speciation is more likely the more traits that are under selection (Nosil et al. 2009). Consequently, biotic interactions may be an important mechanism for speciation.      120 References Akey, J. M. 2009. Constructing genomic maps of positive selection in humans: Where do we go from here? Genome Research 19:711–722. Amarasekare, P. 2008. Coexistence of intraguild predators and prey in resource-rich environments. Ecology 89:2786–2797.  Andrews, K. R., J. M. Good, M. R. Miller, G. Luikart, and P. A. Hohenlohe. 2016. Harnessing the power of RADseq for ecological and evolutionary genomics. Nature Reviews Genetics 17:81–92.  Arim, M., and P. A. Marquet. 2004. Intraguild predation: a widespread interaction related to species biology. Ecology Letters 7:557–564.  Baker, J. A., M. A. Wund, R. Y. Chock, L. Ackein, R. Elsemore, and S. A. Foster. 2010. Predation history and vulnerability: Conservation of the stickleback adaptive radiation. Biology Conservation 143:1184–1192.  Barrett, R. D. H., and D. Schluter. 2008. Adaptation from standing genetic variation. Trends in Ecology & Evolution 23:38–44.  Barrett, R. D. H., S. M. Rogers, and D. Schluter. 2008. Natural selection on a major armor gene in threespine stickleback. Science 322:255–257. Barrett, R.D.H., T.H. Vines, J.S. Bystriansky, and P.M. Schulte. 2009. Should I stay or should I go? The Ectodysplasin locus is associated with behavioural differences in threespine stickleback. Biology Letters 5:788–791. Barrueto, M. 2009. Adaptive significance of pelvic girdle loss in threespine stickleback. (M.S. Thesis). University of British Columbia, Vancouver, BC.   Bell, M. A. 1988. Stickleback fishes: Bridging the gap between population biology and paleobiology. Trends in Ecology & Evolution 3:320–324. Bell, A. M., and A. Sih. 2007. Exposure to predation generates personality in threespined sticklebacks (Gasterosteus aculeatus). Ecology Letters 10:828–834.  Bell, M. A. 2001. Lateral plate evolution in the threespine stickleback: getting nowhere fast. Genetica 112:445–461.  Bell, M. A., and S. A. Foster, eds. 1994. The Evolutionary biology of the threespine stickleback. Oxford University Press, Oxford, UK.    121 Bell, M. A., G. Orti, J. A. Walker, and J. P. Koenings. 1993. Evolution of pelvic reduction in threespine stickleback fish: a test of competing hypotheses. Evolution 47:906–914.  Bertram, B. C. R. 1978. Living in groups: predators and prey. Pp 64-96 in J.R. Krebs & N.B. Davies, eds. Behavioral ecology. Blackwell, Oxford.  Berriz, G. F., J. E. Beaver, C. Cenik, M. Tasan, and F. P. Roth. 2009. Next generation software for functional trend analysis. Bioinformatics 25:3043–3044.  Bjaerke, O., K. Ostbye, H. M. Lampe, and L. A. Vøllestad. 2010. Covariation in shape and foraging behaviour in lateral plate morphs in the three-spined stickleback. Ecology of Freshwater Fish 19:249–256.  Bonhomme, M., S. Boitard, H. San Clemente, B. Dumas, N. Young, and C. Jacquet. 2015. Genomic Signature of Selective Sweeps Illuminates Adaptation of Medicago truncatulato Root-Associated Microorganisms. Molecular Biology and Evolution 32:2097–2110. Brodie, E.D. III. 1992. Correlational selection for color pattern and antipredator behavior in the garter snake Thamnophis ordinoides. Evolution 1284–1298. Buskirk, V. 2000. Functional mechanisms of an inducible defence in tadpoles: morphology and behaviour influence mortality risk from predation. Journal of Evolutionary Biology 13:336–347. Cachat, J., A. Stewart, L. Grossman, S. Gaikwad, F. Kadri, K. M. Chung, N. Wu, K. Wong, S. Roy, C. Suciu, J. Goodspeed, M. Elegante, B. Bartels, S. Elkhayat, D. Tien, J. Tan, A. Denmark, T. Gilder, E. Kyzar, J. DiLeo, K. Frank, K. Chang, E. Utterback, P. Hart, and A. V. Kalueff. 2010. Measuring behavioral and endocrine responses to novelty stress in adult zebrafish. Nature Protocols 5:1786–1799.  Camarata, T., D. Snyder, T. Schwend, J. Klosowiak, B. Holtrup, and H.-G. Simon. 2010. Pdlim7 is required for maintenance of the mesenchymal/epidermal Fgf signaling feedback loop during zebrafish pectoral fin development. BMC Developmental Biology 2010 10:104. Chan, Y. F., M. E. Marks, F. C. Jones, G. Villarreal, M. D. Shapiro, S. D. Brady, A. M. Southwick, D. M. Absher, J. Grimwood, J. Schmutz, R. M. Myers, D. Petrov, B. Jónsson, D. Schluter, M. A. Bell, and D. M. Kingsley. 2010. Adaptive evolution of pelvic reduction in sticklebacks by recurrent deletion of a Pitx1 enhancer. Science 327:302–305. Colosimo, P. F., K. E. Hosemann, S. Balabhadra, G. Villarreal, M. Dickson, J. Grimwood, J. Schmutz, R. M. Myers, D. Schluter, and D. M. Kingsley. 2005. Widespread parallel evolution in sticklebacks by repeated fixation of Ectodysplasin alleles. Science 307:1928–1933. Cortez, M. H. 2011. Comparing the qualitatively different effects rapidly evolving and rapidly induced defences have on predator-prey interactions. Ecology Letters 14:202–209.    122 Cruickshank, T. E., and M. W. Hahn. 2014. Reanalysis suggests that genomic islands of speciation are due to reduced diversity, not reduced gene flow. Molecular Ecology 23:3133–3157. Danecek, P., A. Auton, G. R. Abecasis, C. A. Albers, E. Banks, M. A. DePristo, R. E. Handsaker, G. Lunter, G. T. Marth, S. T. Sherry, G. McVean, R. Durbin, 1000 Genomes Project Analysis Group. 2011. The variant call format and VCFtools. Bioinformatics 27:2156–2158. Daugherty, M. P., J. P. Harmon, and C. J. Briggs. 2007. Trophic supplements to intraguild predation. Oikos 116:662–677. Davies, T., S. Meiri, T. G. Barraclough, and J. L. Gittleman. 2007. Species co-existence and character divergence across carnivores. Ecology Letters 10:146–152. Day, T., and J. D. McPhail. 1996. The effect of behavioral and morphological plasticity on foraging efficiency in the threespine stickleback (Gasterosteus sp). Oecologia 108:380–388.  Day, T., J. Pritchard, and D. Schluter. 1994. Ecology and genetics of phenotypic plasticity: A comparison of two sticklebacks. Evolution 48:1723–1734.  Delmore, K. E., S. Hübner, N. C. Kane, R. Schuster, R. L. Andrew, F. Câmara, R. Guigó, and D. E. Irwin. 2015. Genomic analysis of a migratory divide reveals candidate genes for migration and implicates selective sweeps in generating islands of differentiation. Molecular Ecology 24:1873–1888. Dennenmoser, S., A. W. Nolte, S. Vamosi, and S. M. Rogers. 2015. Phylogeography of the prickly sculpin (Cottus asper) in north-western North America reveals parallel phenotypic evolution across multiple coastal-inland colonizations. Journal of Biogeograpy 42:1626-1638.  DePristo, M. A., E. Banks, R. Poplin, K. V. Garimella, J. R. Maguire, C. Hartl, A. A. Philippakis, G. del Angel, M. A. Rivas, M. Hanna, A. McKenna, T. J. Fennell, A. M. Kernytsky, A. Y. Sivachenko, K. Cibulskis, S. B. Gabriel, D. Altshuler, and M. J. Daly. 2011. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature Genetics 43:491–498.  Dewitt, T. J., A. Sih, and D. S. Wilson. 1998. Costs and limits of phenotypic plasticity. Trends in Ecology and Evolution 13:77–81. Dewitt, T., A. Sih, and J. Hucko. 1999. Trait compensation and cospecialization in a freshwater snail: size, shape and antipredator behaviour. Animal Behaviour 58: 397–407. Diabaté, A., R.K. Dabiré, K. Heidenberger, J. Crawford, W.O. Lamp, L.E. Culler, and T. Lehmann. 2008. Evidence for divergent selection between the molecular forms of Anopheles gambiae: role of predation. BMC Evolutionary Biology 8:5.   123 Dingemanse, N. J., F. Van der Plas, J. Wright, D. Réale, M. Schrama, D. A. Roff, E. Van der Zee, and I. Barber. 2009. Individual experience and evolutionary history of predation affect expression of heritable variation in fish personality and morphology. Proceedings of the Royal Society of London B 276:1285–1293.  Dingemanse, N. J., J. Wright, A. J. N. Kazem, D. K. Thomas, R. Hickling, and N. Dawnay. 2007. Behavioral syndromes differ predictably between 12 populations of three-spined stickleback. Journal of Animal Ecology 76:1128–1138.  Donadio, E., and S. W. Buskirk. 2006. Diet, Morphology, and Interspecific killing in carnivora. American Naturalist 167:524–536.  Dryden, I.L. 2012. Package 'shapes': Statistical Shape Analysis, v.1.1-3 Available at http://CRAN.R-project.org/package=shapes. (CRAN: Comprehensive R Archive Network, 2009).  Durinck, S., P. T. Spellman, E. Birney, and W. Huber. 2009. Mapping identifiers for the integration of genomic datasets with the R/Bioconductor package biomaRt. Nature Protocols 4:1184–1191. Egan, R. J., C. L. Bergner, P. C. Hart, J. M. Cachat, P. R. Canavello, M. F. Elegante, S. I. Elkhayat, B. K. Bartels, A. K. Tien, D. H. Tien, S. Mohnot, E. Beeson, E. Glasgow, H. Amri, Z. Zukowska, and A. V. Kalueff. 2009. Understanding behavioral and physiological phenotypes of stress and anxiety in zebrafish. Behavioural Brain Research 205:38–44.  Elmer, K. R., and A. Meyer. 2011. Adaptation in the age of ecological genomics: insights from parallelism and convergence. Trends in Ecology & Evolution 26:298–306. Endler, J. A. 1980. Natural selection on color patterns in Poecilia reticulata. Evolution 34:76–91.  Endler, J. A. 1986. Natural Selection in the Wild. Princeton University Press. Excoffier, L., and N. Ray. 2008. Surfing during population expansions promotes genetic revolutions and structuration. Trends in Ecology & Evolution 23:347–351. Fish, J. L., B. Villmoare, K. Köbernick, C. Compagnucci, O. Britanova, V. Tarabykin, and M. J. Depew. 2011. Satb2, modularity, and the evolvability of the vertebrate jaw. Evolution and Development 13:549–564. Garduño-Paz, M. V., S. Couderc, and C. E. Adams. 2010. Habitat complexity modulates phenotype expression through developmental plasticity in the threespine stickleback. Biological Journal of the Linnean Society 100:407–413.  Giovannone, B., E. Lee, L. Laviola, F. Giorgino, K. A. Cleveland, and R. J. Smith. 2003. Two Novel Proteins That Are Linked to Insulin-like Growth Factor (IGF-I) Receptors by the Grb10   124 Adapter and Modulate IGF-I Signaling. Journal of Biological Chemistry 278:31564–31573. Golub, J. L. 2013. Embryonic learning in threespine stickleback (Gasterosteus aculeatus) (Ph.D. Thesis). Clark University, Worcester, MA.   Gou, X., Wang, Z., Li, N., Qiu, F., Xu, Z., Yan, D., Yang, S., Jia, J., Kong, X., Wei, Z. and S. Lu 2014. Whole-genome sequencing of six dog breeds from continuous altitudes reveals adaptation to high-altitude hypoxia. Genome research 24:1308-1315.  Gow, J.L., S.M. Rogers, M. Jackson, and D. Schluter. 2008. Ecological predictions lead to the discovery of a benthic-limnetic sympatric species pair of threespine stickleback in Little Quarry Lake, British Columbia. Canadian Journal of Zoology 86:564-571.  Grand, T.C. 2000. Risk-Taking by Threespine Stickleback (Gasterosteus aculeatus) Pelvic Phenotypes: Does Morphology Predict Behaviour? Behaviour 137:889–906.  Greenwood, A.K., A.R. Wark, K. Yoshida, and C.L. Peichel. 2013. Genetic and Neural Modularity Underlie the Evolution of Schooling Behavior in Threespine Sticklebacks. Current Biology 23:1884–1888.  Hagen, D. W. 1973. Inheritance of numbers of lateral plates and gill rakers in Gasterosteus aculeatus. Heredity 30:303–312.  Hagen, D. W., and L. G. Gilbertson. 1972. Geographic variation and environmental selection in Gasterosteus aculeatus L. in the Pacific Northwest, America. Evolution 26:32–51.   Hamblin, M. T., and A. Di Rienzo. 2000. Detection of the Signature of Natural Selection in Humans: Evidence from the Duffy Blood Group Locus. The American Journal of Human Genetics 66:1669-1679. Hart, D. R. 2002. Intraguild predation, invertebrate predators, and trophic cascades in lake food webs. Journal of Theoretical Biology 218:111–128.  Hasselblad, V., and L. V. Hedges. 1995. Meta-analysis of screening and diagnostic tests. Psychological Bulletin 117:167–178. American Psychological Association. Hartl, Daniel L., Andrew G. Clark, and A. G. Clark. 1997. Principles of population genetics. Vol. 116. Sunderland: Sinauer associates.  Harvell, C. D. 1990. The ecology and evolution of inducible defenses. Quarterly Review of Biology 65:323–340.  Heithaus, M. R. 2001. Habitat selection by predators and prey in communities with asymmetrical intraguild predation. Oikos 92:542–554.    125 Hepher, B. 1988. Nutrition of Pond Fishes. Cambridge University Press.  Hersch, E. I., and P.C. Phillips. 2004. Power and potential bias in field studies of natural selection. Evolution 58:479–485.  Hervé, M. 2014. RVAideMemoire: Diverse Basic Statistical and Graphical Functions, Version 0.9-52. Available at https://cran.r-project.org/package=RVAideMemoire Hohenlohe, P. A., S. Bassham, P. D. Etter, N. Stiffler, E. A. Johnson, and W. A. Cresko. 2010. Population genomics of parallel adaptation in threespine stickleback using sequenced RAD tags. PLoS Genet 6:e1000862. Holt, R. D., and G. A. Polis. 1997. A theoretical framework for intraguild predation. American Naturalist 149:745–764.  Hoekstra, H. E., J. M. Hoekstra, D. Berrigan, S. N. Vignieri, A. Hoang, C. E. Hill, P. Beerli, and J. G. Kingsolver. 2001. Strength and tempo of directional selection in the wild. Proceedings of the National Academy of Sciences 98:9157–9160. Hoogland, R., D. Morris, and N. Tinbergen. 1956. The spines of sticklebacks (Gasterosteus and Pygosteus) as means of defence against predators (Perca and Esox). Behavior 10:205–236.  Hudson, P., and J. Greenman. 1998. Competition mediated by parasites: biological and theoretical progress. Trends in Ecology & Evolution 13:387–390.  Ingram, T., R. Svanbäck, N. J. B. Kraft, P. Kratina, L. Southcott, and D. Schluter. 2012. Intraguild predation drives evolutionary niche shift in threespine stickleback. Evolution 66:1819-1832.  Jackson, D. A., P. R. Peres-Neto, and J.D. Olden. 2001. What controls who is where in freshwater fish communities - the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences 58:157-170.  Janssen, A., M. W. Sabelis, S. Magalhães, M. Montserrat, and T. van der Hammen. 2007. Habitat structure affects intraguild predation. Ecology 88:2713-2719.  Jiggins, C.D., R.E. Naisbit, R.L. Coe, and J. Mallet. 2001. Reproductive isolation caused by colour pattern mimicry. Nature 411:302–305.  Jones, F. C., M. G. Grabherr, Y. F. Chan, P. Russell, E. Mauceli, J. Johnson, R. Swofford, M. Pirun, M. C. Zody, S. White, E. Birney, S. Searle, J. Schmutz, J. Grimwood, M. C. Dickson, R. M. Myers, C. T. Miller, B. R. Summers, A. K. Knecht, S. D. Brady, H. Zhang, A. A. Pollen, T. Howes, C. Amemiya, J. Baldwin, T. Bloom, D. B. Jaffe, R. Nicol, J. Wilkinson, E. S. Lander, F. Di Palma, K. Lindblad-Toh, and D. M. Kingsley. 2012a. The genomic basis of adaptive evolution in threespine sticklebacks. Nature 484:55–61.   126  Jones, F. C., Y. F. Chan, J. Schmutz, J. Grimwood, S. D. Brady, A. M. Southwick, D. M. Absher, R. M. Myers, T. E. Reimchen, B. E. Deagle, D. Schluter, and D. M. Kingsley. 2012b. A Genome-wide SNP Genotyping Array Reveals Patterns of Global and Repeated Species-Pair Divergence in Sticklebacks. Current Biology. 22:83–90. Kaeuffer, R., Peichel, C. L., Bolnick, D. I., and A.P. Hendry. 2012. Parallel and nonparallel aspects of ecological, phenotypic, and genetic divergence across replicate population pairs of lake and stream stickleback. Evolution 66:402-418.  Kelley, J. L., & Magurran, A. E. 2003. Learned predator recognition and antipredator responses in fishes. Fish & Fisheries 4:216-226.  Kim, S.Y., and A. Velando. 2015. Phenotypic integration between antipredator behavior and camouflage pattern in juvenile sticklebacks. Evolution 69:830–838. Kingsolver, J. G., H. E. Hoekstra, J. M. Hoekstra, D. Berrigan, S. N. Vignieri, C. E. Hill, A. Hoang, P. Gibert, and P. Beerli. 2001. The strength of phenotypic selection in natural populations. American Naturalist 157:245–261. Klepaker, T., K. Østbye, and M. A. Bell. 2013. Regressive evolution of the pelvic complex in stickleback fishes: a study of convergent evolution. Evolutionary Ecology Research 15:413–435. Klopfstein, S. 2005. The Fate of Mutations Surfing on the Wave of a Range Expansion. Molecular Biology and Evolution 23:482–490. Kofler, R., and C. Schlotterer. 2012. Gowinda: unbiased analysis of gene set enrichment for genome-wide association studies. Bioinformatics 28:2084–2085. Kozak, G. M., and J. W. Boughman. 2008. Experience influences shoal member preference in a species pair of sticklebacks. Behavioral Ecology 19:667–676.  Kozak, G.M., and J. W. Boughman. 2012. Plastic responses to parents and predators lead to divergent shoaling behaviour in sticklebacks. Journal of Evolutionary Biology 25:759–769. Kratina, P., E. Hammill, and B. R. Anholt. 2010. Stronger inducible defences enhance persistence of intraguild prey. Journal of Animal Ecology 79:993–999.  Lacasse, J., and N. Aubin-Horth. 2012. A test of the coupling of predator defense morphology and behavior variation in two threespine stickleback populations. Current Zoology 58:53–65.  Lahti, D. C., N. A. Johnson, B. C. Ajie, S. P. Otto, A. P. Hendry, D. T. Blumstein, R. G. Coss, K. Donohue, and S. A. Foster. 2009. Relaxed selection in the wild. Trends in Ecology & Evolution 24:487–496.   127 Laine, A. L. 2009. Role of coevolution in generating biological diversity: spatially divergent selection trajectories. Journal of Experimental Botany 60:2957–2970.  Langerhans, R. B., M. E. Gifford, and E. O. Joseph. 2007. Ecological speciation in Gambusia fishes. Evolution 61:2056-2074.  Larson, G. L. 1976. Social behavior and feeding ability of two phenotypes of Gasterosteus aculeatus in relation to their spatial and trophic segregation in a temperate lake. Canadian Journal of Zoology 54:107-121.  Lawniczak, M. K. N., S. J. Emrich, A. K. Holloway, A. P. Regier, M. Olson, B. White, S. Redmond, L. Fulton, E. Appelbaum, J. Godfrey, C. Farmer, A. Chinwalla, S.-P. Yang, P. Minx, J. Nelson, K. Kyung, B. P. Walenz, E. Garcia-Hernandez, M. Aguiar, L. D. Viswanathan, Y.-H. Rogers, R. L. Strausberg, C. A. Saski, D. Lawson, F. H. Collins, F. C. Kafatos, G. K. Christophides, S. W. Clifton, E. F. Kirkness, and N. J. Besansky. 2010. Widespread divergence between incipient Anopheles gambiae species revealed by whole genome sequences. Science 330:512–514. Leinonen, T., G. Herczeg, J. M. Cano, and J. Merilä. 2011. Predation-imposed selection on threespine stickleback (Gasterosteus aculeatus) morphology: a test of the refuge use hypothesis. Evolution 65:2916–2926. Lescak, E. A., and F. von Hippel. 2011. Selective predation of threespine stickleback by rainbow trout. Ecology of Freshwater Fish 20:308–314.  Lescak, E. A., F. von Hippel, B. K. Lohman, and M. L. Sherbick. 2012. Predation of threespine stickleback by dragonfly naiads. Ecology of Freshwater Fish 21:581–587.  Lescak, E. A., F. von Hippel, R. R. Bernhardt, and M. A. Bell. 2013. Pelvic girdle reduction and asymmetry in threespine stickleback from Wallace Lake, Alaska. Evolutionary Ecology Research 15:155–170. Li, H., and R. Durbin. 2009. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25:1754–1760. Linnen, C. R., E. P. Kingsley, J. D. Jensen, and H. E. Hoekstra. 2009. On the origin and spread of an adaptive allele in deer mice. Science 325:1095–1098. Lima, S. L., and L. M. Dill. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68:619–640.  MacColl, A. D. C. 2011. The ecological causes of evolution. Trends in Ecology & Evolution 26:514–522.  Maccoll, A., and S. M. Chapman. 2011. A benthic predatory fish does not cause selection on   128 armour traits in three‐spined stickleback Gasterosteus aculeatus (Gasterosteiformes: Gasterosteidae). Biological Journal of the Linnean Society 104:877-885. Marchinko, K. B. 2009. Predation's role in repeated phenotypic and genetic divergence of armor in threespine stickleback. Evolution 63:127–138.  Marchinko, K. B., and D. Schluter. 2007. Parallel evolution by correlated response: lateral plate reduction in threespine stickleback. Evolution 61:1084–1090. McPeek, M. A. 1995. Morphological evolution mediated by behavior in the damselflies of two communities. Evolution:749-769. McPhail, J.D. 1984. Ecology and evolution of sympatric sticklebacks (Gasterosteus): morphological and genetic evidence for a species pair in Enos Lake, British Columbia. Canadian Journal of Zoology 62:1402-1408. McPhail, J.D. 1993. Ecology and evolution of sympatric sticklebacks (Gasterosteus): origin of the species pairs. Canadian Journal of Zoology 71:515-523.  McPhail, J. D. 2007. The freshwater fishes of British Columbia. Univ of Alberta Press.  Meijering, E., O. Dzyubachyk, and I. Smal. 2012. Methods for cell and particle tracking. Methods in Enzymology 504:183-200.  Mills, M.G, A.K. Greenwood, and C.L. Peichel. 2014. Pleiotropic effects of a single gene on skeletal development and sensory system patterning in sticklebacks. EvoDevo 5: 1-10. Miller, S.E., D.M. Metcalf, and D. Schluter. Intraguild predation leads to genetically based character shifts in the threespine stickleback. Evolution 69:3194-3203. Mobley, K. B., Ruiz, R. C., Johansson, F., Englund, G., & Bokma, F. 2013. No evidence that stickleback spines directly increase risk of predation by an invertebrate predator. Evolutionary Ecology Research 15:189-198.  Moodie, G. E. E. 1972. Predation, natural selection and adaptation in an unusual threespine stickleback. Heredity 28:155-167.  Murren, C.J. 2012. The Integrated Phenotype. Integrative and Comparative Biology 52:64–76. Myhre, F., and T. Klepaker. 2009. Body armour and lateral-plate reduction in freshwater three-spined stickleback Gasterosteus aculeatus: adaptations to a different buoyancy regime? Journal of Fish Biology 75:2062-2074.  Nadeau, N. J., and C. D. Jiggins. 2010. A golden age for evolutionary genetics? Genomic studies of adaptation in natural populations. Trends in Genetics 26:484–492.    129 Nakazawa, T., T. Miki, and T. Namba. 2010. Influence of predator‐specific defense adaptation on intraguild predation. Oikos 119:418–427.   Nelson, J. S. 1969. Geographic Variation in the Brook Stickleback, Culaea inconstans, and Notes on Nomenclature and Distribution. Journal of the Fisheries Board of Canada. NRC Research Press Ottawa, Canada. Nelson, J. S., and F. M. Atton. 1971. Geographic and morphological variation in the presence and absence of the pelvic skeleton in the brook stickleback, Culaea inconstans (Kirtland), in Alberta and Saskatchewan. Canadian Journal of Zoology 49:343-352. Nosil, P. 2012. Ecological Speciation. Oxford University Press, Oxford, UK.  Nosil, P., and B.J. Crespi. 2006. Experimental evidence that predation promotes divergence in adaptive radiation. Proceedings of the National Academy of Sciences 103:9090–9095. Nosil, P., D. J. Funk, and D. Ortiz-Barrientos. 2009. Divergent selection and heterogeneous genomic divergence. Molecular Ecology 18:375–402. Odling-Smee, L.C., J.W. Boughman, and V.A. Braithwaite. 2008. Sympatric species of threespine stickleback differ in their performance in a spatial learning task. Behavioral Ecology and Sociobiology 62:1935–1945. Peichel, C. L., K. S. Nereng, K. A. Ohgi, B. L. E. Cole, P. F. Colosimo, C. A. Buerkle, D. Schluter, and D. M. Kingsley. 2001. The genetic architecture of divergence between threespine stickleback species. Nature 414:901–905.  Pfennig, D. W., and K. S. Pfennig. 2010. Character displacement and the origins of diversity. American Naturalist 176:S26–S44.  Pfennig, D. W., M. A. Wund, E. C. Snell-Rood, T. Cruickshank, C. D. Schlichting, and A. P. Moczek. 2010. Phenotypic plasticity's impacts on diversification and speciation. Trends in Ecology and Evolution 25:459–467. Polis, G. A., C. A. Myers, and R. D. Holt. 1989. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology & Systematics 20:297–330.   Preisser, E.L., D.I. Bolnick, and M.F. Benard. 2005. Scared to death? The effects of intimidation and consumption in predator-prey interactions. Ecology 86:501–509. Purcell, S., B. Neale, K. Todd-Brown, and L. Thomas. 2007. PLINK: A Tool Set for Whole-Genome Association and Population-Based Linkage Analyses. The American Journal of Human Genetics 81:559-575.   130 Reimchen, T. 1994. Predators and morphological evolution in threespine stickleback. Pp 240-276 in M.A. Bell and S.A. Foster, eds. The evolutionary biology of the threespine stickleback. Oxford Univ. Press, Oxford, U.K.  R Core Team 2015. R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria.  Reimchen, T. E. 1980. Spine deficiency and polymorphism in a population of Gasterosteus aculeatus: an adaptation to predators? Canadian Journal of Zoology 58:1232-1244.  Reimchen, T. E. 1983. Structural relationships between spines and lateral plates in threespine stickleback (Gasterosteus aculeatus). Evolution:931–946.   Reimchen, T. 1994. Predators and morphological evolution in threespine stickleback. Pp 240-273 in M.A. Bell and S.A. Foster, eds. The evolutionary biology of the threespine stickleback. Oxford Univ. Press, Oxford, U.K.  Reimchen, T. E. 2000. Predator handling failures of lateral plate morphs in Gasterosteus aculeatus: functional implications for the ancestral plate condition. Behaviour 137:1081-1096. Reimchen, T. E., and P. Nosil. 2002. Temporal variation in divergent selection on spine number in threespine stickleback. Evolution 56:2472–2483.  Reist, J. D. 1980. Selective Predation Upon Pelvic Phenotypes of Brook Stickleback, Culaea inconstans, by Northern Pike, Esox lucius. Canadian Journal of Zoology 58:1245–1252.  Relyea, R.A. 2001. Morphological and Behavioral Plasticity of Larval Anurans in Response to Different Predators. Ecology 82:523–540. Rennison, D. J., K. Heilbron, R. D. H. Barrett, and D. Schluter. 2015. Discriminating selection on lateral plate phenotype and its underlying gene, Ectodysplasin, in threespine stickleback. American Naturalist 185:150–156.  Rieseberg, L. H., A. Widmer, and A. M. Arntz. 2002. Directional selection is the primary cause of phenotypic diversification. Proceedings of the National Academy of Sciences 99:12242-12245. Roesti, M., D. Moser and D. Berner. 2013. Recombination in the threespine stickleback genome – patterns and consequences. Molecular Ecology 22:3014-3027. Rockman, M. V. 2012. The QTN program and the alleles that matter for evolution: all that's gold does not glitter. Evolution 66:1–17.   131 Rogers, S. M., P. Tamkee, B. Summers, S. Balabahadra, M. Marks, D. M. Kingsley, and D. Schluter. 2012. Genetic signature of adaptive peak shift in threespine stickleback. Evolution 66:2439–2450.  Rohlf, F. 2008. tpsDig2, digitize landmarks and outlines, version 2.12. Department of Ecology and Evolution, State Univer. of New York at Stony Brook.  Rudman S., J. Heavyside, D.J. Rennison, and D. Schluter. 2016. Piscivore addition causes a trophic cascade within and across ecosystem boundaries. Oikos. In press.  Rundle, H.D., and P. Nosil. 2005. Ecological speciation. Ecology Letters 8:336-352. Rundle H.D., S. Vamosi, and D. Schluter. 2003. Experimental test of predation's effect on divergent selection during character displacement in sticklebacks. Proceedings of the National Academy of Sciences 100:14943–14948. Scheiner, S. M. 1993. Genetics and evolution of phenotypic plasticity. Annual Review of Ecology & Systematics 24:35–68.  Schlichting, C. D., and M. Pigliucci. 1998. Phenotypic evolution: a reaction norm perspective. Sinauer Associates Incorporated.  Schluter, D. 2000a. Ecological Character Displacement in Adaptive Radiation. American Naturalist 156:S4–S16.  Schluter, D. 2000b. The ecology of adaptive radiation. Oxford University Press, Oxford, UK.  Schluter, D. 2003. Frequency dependent natural selection during character displacement in sticklebacks. Evolution 57:1142–1150.  Schluter, D. 2009. Evidence for ecological speciation and its alternative. Science 323:737–741. Schluter, D., and G. L. Conte. 2009. Genetics and ecological speciation. Proceedings of the National Academy of Sciences 106:9955–9962. Schluter, D., and J. D. McPhail. 1992. Ecological character displacement and speciation in sticklebacks. American Naturalist 140:85–108.  Schlosser, G, and G.P. Wagner. 2004. Modularity in Development and Evolution. Chicago, IL: University of Chicago Press. Schneider, C. A., W. S. Rasband, and K. W. Eliceiri. 2012. NIH Image to ImageJ: 25 years of image analysis. Nature methods 9:671-675.    132 Schwarzer, G. 2015. meta: General Package for Meta-Analysis. R package version 4.3-1. http://CRAN.R-project.org/package=meta  Shapiro, M. D., B. R. Summers, S. Balabhadra, J. T. Aldenhoven, A. L. Miller, C. B. Cunningham, M. A. Bell, and D. M. Kingsley. 2009. The Genetic Architecture of Skeletal Convergence and Sex Determination in Ninespine Sticklebacks. Current Biology 19:1140–1145.  Shapiro, M. D., M. E. Marks, C. L. Peichel, B. K. Blackman, K. S. Nereng, B. Jónsson, D. Schluter, and D. M. Kingsley. 2004. Genetic and developmental basis of evolutionary pelvic reduction in threespine sticklebacks. Nature 428:717–723. Shikano, T., V. N. Laine, G. Herczeg, J. Vilkki, and J. Merila. 2013. Genetic Architecture of Parallel Pelvic Reduction in Ninespine Sticklebacks. G3 3:1833–1842. Sinervo, B., and E. Svensson. 2002. Correlational selection and the evolution of genomic architecture. Heredity 89:329–338.  Soria-Carrasco, V., Z. Gompert, A. A. Comeault, T. E. Farkas, T. L. Parchman, J. S. Johnston, C. A. Buerkle, J. L. Feder, J. Bast, T. Schwander, S. P. Egan, B. J. Crespi, and P. Nosil. 2014. Stick Insect Genomes Reveal Natural Selection's Role in Parallel Speciation. Science 344:738–742. Southcott, L., L. Nagel, T. Hatfield, and D. Schluter. 2013. Weak habitat isolation in a threespine stickleback (Gasterosteus spp.) species pair. Biological Journal of the Linnean Society 110:466-476.  Stacklies, W., H. Redestig, M. Scholz, D. Walther, and J. Selbig. 2007. pcaMethods a bioconductor package providing PCA methods for incomplete data. Bioinformatics 23:1164–1167. Stapley, J., J. Reger, P. G. D. Feulner, C. Smadja, J. Galindo, R. Ekblom, C. Bennison, A. D. Ball, A. P. Beckerman, and J. Slate. 2010. Adaptation genomics: the next generation. Trends in Ecology & Evolution 25:705–712.  Stephan, W. 2016. Signatures of positive selection: from selective sweeps at individual loci to subtle allele frequency changes in polygenic adaptation. Molecular Ecology 25:79–88. Stewart, A., S. Gaikwad, E. Kyzar, J. Green, A. Roth, and A. V. Kalueff. 2012. Modeling anxiety using adult zebrafish: A conceptual review. Neuropharmacology 62:135–143.  Stuart, Y. E., and J. B. Losos. 2013. Ecological character displacement: glass half full or half empty? Trends in Ecology & Evolution 28:402–408.  Svanbäck, R., and P. Eklöv. 2011. Catch me if you can-predation affects divergence in a polyphenic species. Evolution 65:3515–3526.   133 Svanbäck, R., and D. Schluter. 2012. Niche specialization influences adaptive phenotypic plasticity in the threespine stickleback. American Naturalist 180:50–59.  Tabachnick, B. G., and L. S. Fidell. 2012. Using Multivariate Statistics. Pearson Higher Ed.  Thompson, J. N. 2013. Relentless Evolution. University of Chicago Press.  Torres-Dowdall, J., C. A. Handelsman, D. N. Reznick, and C. K. Ghalambor. 2012. Local adaptation and the evolution of phenotypic plasticity in Trinidadian guppies (Poecilia reticulata). Evolution 66:3432–3443.  Turner, T. L., E. C. Bourne, E. J. Von Wettberg, and T. T. Hu. 2010. Population resequencing reveals local adaptation of Arabidopsis lyrata to serpentine soils. Nature Genetics 42:260-263. Urbani, P., and R. Ramos-Jiliberto. 2010. Adaptive prey behavior and the dynamics of intraguild predation systems. Ecological Modeling. 221:2628–2633.  Vamosi, S. M. 2002. Predation sharpens the adaptive peaks: survival trade-offs in sympatric sticklebacks. Annales Zoologici Fennici. 39:237–248.  Vamosi, S. 2003. The presence of other fish species affects speciation in threespine sticklebacks. Evolutionary Ecology Research 5:717–730. Vamosi, S. M., and D. Schluter. 2002. Impacts of trout predation on fitness of sympatric sticklebacks and their hybrids. Proceedings of the Royal Society B: Biological Sciences 269:923–930. Vamosi, S. M., and D. Schluter. 2004. Character shifts in the defensive armor of sympatric sticklebacks. Evolution 58:376–385.  Van der Auwera, G. A., M. O. Carneiro, C. Hartl, R. Poplin, G. del Angel, A. Levy-Moonshine, T. Jordan, K. Shakir, D. Roazen, J. Thibault, E. Banks, K. V. Garimella, D. Altshuler, S. Gabriel, and M. A. DePristo. 2013. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Current Protocols in Bioinformatics 11-10.  Vanak, A. T., D. Fortin, M. Thaker, M. Ogden, C. Owen, S. Greatwood, and R. Slotow. 2013. Moving to stay in place: behavioral mechanisms for coexistence of African large carnivores. Ecology 94:2619–2631.  Vance-Chalcraft, H. D., J. A. Rosenheim, J. R. Vonesh, C. W. Osenberg, and A. Sih. 2007. The influence of intraguild predation on prey suppression and prey release: a meta-analysis. Ecology 88:2689–2696.  Venables, W. N., and B. D. Ripley. 2010. Modern Applied Statistics with S. Springer Verlag.   134  Visser, I., and M. Speekenbrink. 2010. depmixS4: an R package for hidden Markov models. Journal of Statistical Software 36:1-21.  Wade, M. J., and S. Kalisz. 1990. The Causes of Natural Selection. Evolution 44:1947. Walker, J. A. 1997. Ecological morphology of lacustrine threespine stickleback Gasterosteus aculeatus L. (Gasterosteidae) body shape. Biological Journal of the Linnean Society 61:3–50.  Walzer, A., and P. Schausberger. 2013. Phenotypic plasticity in anti-intraguild predator strategies: mite larvae adjust their behaviors according to vulnerability and predation risk. Experimental and Applied Acarology 60:95-115.  Wark, A.R., and C.L. Peichel. 2009. Lateral line diversity among ecologically divergent threespine stickleback populations. Journal of Experimental Biology 213: 108–117. Wark, A.R., M.G. Mills, L.H. Dang, Y.F. Chan, F.C. Jones, S.D. Brady, D.M. Absher, J. Grimwood, J. Schmutz, R.M. Myers, and D.M. Kingsley. 2012. Genetic architecture of variation in the lateral line sensory system of threespine sticklebacks. G3 2:1047–1056. Wark, A. R., B. J. Wark, T. J. Lageson, and C. L. Peichel. 2011. Novel methods for discriminating behavioral differences between stickleback individuals and populations in a laboratory shoaling assay. Behavioral Ecology and Sociobiology 65:1147–1157.  Weir, B. S., and C. C.Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution:1358-1370.  West-Eberhard, M. J. 2003. Developmental plasticity and evolution. Oxford University Press.  Willacker, J. J., F. A. von Hippel, P. R. Wilton, and K. M. Walton. 2010. Classification of threespine stickleback along the benthic-limnetic axis. Biological Journal of the Linnean Society 101:595–608.  Wund, M. A., J. A. Baker, B. Clancy, J. L. Golub, and S. A. Foster. 2008. A Test of the “Flexible Stem” Model of Evolution: Ancestral Plasticity, Genetic Accommodation, and Morphological Divergence in the Threespine Stickleback Radiation. American Naturalist 172:449–462.  Yamamichi, M., T. Yoshida, and A. Sasaki. 2011. Comparing the effects of rapid evolution and phenotypic plasticity on predator-prey dynamics. American Naturalist 178:287–304.    135 Appendices Appendix A    #!/bin/sh#  thesis_pipeline.sh#  ##lake='lake_name'project='sculpin'bwa='/filelocation/bwa-0.7.6a' #version 0.7.6picard='/filelocation/picard-tools-1.97' #version 1.97GATK='/filelocation/GATK3' #version 3.2.2ref='/filelocation/gasAcu1.fa' #reference genome Broad Instituted 2006 Assembly# find the suffix array (SA) coordinates of good hits of each individualbwa aln -q20 $ref $lake_R1.fastq > $lake_R1.saibwa aln -q20 $ref $lake_R2.fastq > $lake_R2.sai# read and generate alignments in the SAM formatbwa sampe $ref $lake_R1.sai $lake_R2.sai $lake_R1.fastq $lake_R2.fastq > $lake.q20.sam# add read groups, sort sam file, produce bam and bai files using Picard's AddOrReplaceReadGroupsjava -Djava.io.tmpdir=/tmp -jar $picard/AddOrReplaceReadGroups.jar RGID=$lake  RGLB=$lake RGSM=$lake RGPL=ILLUMINA RGPU=$project I=$lake.q20.sam O=$lake.bam SORT_ORDER=coordinate CREATE_INDEX=TRUE VALIDATION_STRINGENCY=LENIENT# create list of targets for realignment;# intervals file must end in ".intervals" or other approved extensionjava -Xmx2g -jar $GATK/GenomeAnalysisTK.jar -I $lake.bam -T RealignerTargetCreator -R $ref -o $lake.temp.intervals# run the realigner using list of realignment targets.# default LOD is 5.0 but lower recommended for low coveragejava -Xmx4g -Djava.io.tmpdir=/tmp -jar $GATK/GenomeAnalysisTK.jar -I $lake.bam -R $ref -T IndelRealigner-targetIntervals $lake.temp.intervals -o $lake.realigned.bam -LOD 0.4# mark pcr duplicatesjava -Djava.io.tmpdir=/tmp -jar $picard/MarkDuplicates.jar I=$lake.realigned.bam O=$lake.realigned.mkdup.bam M=$lake.realigned.markdup.metrics REMOVE_DUPLICATES=FALSE VALIDATION_STRINGENCY=LENIENT# make an index file of marked duplicate bam filejava -Djava.io.tmpdir=/tmp -jar $picard/BuildBamIndex.jar I=$lake.realigned.mkdup.bam O=$lake.realigned.mkdup.bai VALIDATION_STRINGENCY=LENIENT# call SNPs with GATK, includes invariant sitesjava -jar /Linux/GATK3/GenomeAnalysisTK.jar -R $ref -T UnifiedGenotyper -l INFO -I $lake.realigned.mkdup.bam -o $lake.invariant.vcf --output_mode EMIT_ALL_CONFIDENT_SITES --genotype_likelihoods_model BOTH  136 Appendix B   Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$GRM5!! ENSGACG00000007968! groupI! 6315506!ctsc! ENSGACG00000007985! groupI! 6368016!ZFP36! ENSGACG00000008351! groupI! 7086897!!ENSGACG00000008353! groupI! 7087527!cpd! ENSGACG00000008673! groupI! 7617542!gosr1!! ENSGACG00000008685! groupI! 7633993!abhd15! ENSGACG00000008738! groupI! 7711424!pde3a! ENSGACG00000008742! groupI! 7742778!slco1c1! ENSGACG00000008749! groupI! 7779850!dus4l! ENSGACG00000008772! groupI! 7799466!MNT! ENSGACG00000008790! groupI! 7804379!ELN! ENSGACG00000008825! groupI! 7895088!phc2b! ENSGACG00000009625! groupI! 9505875!ppp1r37! ENSGACG00000009629! groupI! 9508614!mrpl28! ENSGACG00000009641! groupI! 9544230!relb! ENSGACG00000009656! groupI! 9552408!snrpa! ENSGACG00000009838! groupI! 9799920!FLJ41131! ENSGACG00000009856! groupI! 9803940!itpkcb! ENSGACG00000009859! groupI! 9807301!!ENSGACG00000009863! groupI! 9816673!gfm1! ENSGACG00000009864! groupI! 9818367!neu4! ENSGACG00000009879! groupI! 9829619!wdr53! ENSGACG00000009884! groupI! 9833731!cux2! ENSGACG00000009915! groupI! 9866672!alcam! ENSGACG00000010444! groupI! 10881353!igsf9b! ENSGACG00000010671! groupI! 11124934!paf1! ENSGACG00000010818! groupI! 11597447!TAOK1!! ENSGACG00000010833! groupI! 11607936!tsr1! ENSGACG00000010843! groupI! 11619039!synrg! ENSGACG00000011112! groupI! 12185852!DUSP18! ENSGACG00000011118! groupI! 12219197!tada2! ENSGACG00000011119! groupI! 12229512!nbeab! ENSGACG00000011555! groupI! 13479008!TIAM1! ENSGACG00000011727! groupI! 13903005!!ENSGACG00000011739! groupI! 13926650!bach1a! ENSGACG00000011798! groupI! 13962476!usp16! ENSGACG00000011804! groupI! 13966627!rwdd2b! ENSGACG00000011812! groupI! 13971864!FOXO1!! ENSGACG00000011918! groupI! 14219642!cog6! ENSGACG00000011922! groupI! 14261590!lhfp! ENSGACG00000011929! groupI! 14281680!slc8a2b! ENSGACG00000012302! groupI! 15646519!srsf7a! ENSGACG00000012311! groupI! 15672536!   137 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$ccdc9! ENSGACG00000012335! groupI! 15743759!gdpd4a! ENSGACG00000013263! groupI! 17220794!myo7a! ENSGACG00000013269! groupI! 17244095!capn5a! ENSGACG00000013281! groupI! 17297365!ompa! ENSGACG00000013293! groupI! 17304763!ulk2! ENSGACG00000013524! groupI! 18392058!bcas3! ENSGACG00000013586! groupI! 18620444!myo18a! ENSGACG00000013682! groupI! 19641380!NCAM2! ENSGACG00000014231! groupI! 21388343!!ENSGACG00000014241! groupI! 21473205!gart! ENSGACG00000014244! groupI! 21479070!satb2! ENSGACG00000014437! groupI! 22087211!BNC1! ENSGACG00000014360! groupII! 3225917!KCTD12! ENSGACG00000015376! groupII! 8124609!epha6! ENSGACG00000015384! groupII! 8148242!LGR4! ENSGACG00000015514! groupII! 8790724!ZNF821! ENSGACG00000015798! groupII! 11172021!RFX7! ENSGACG00000015838! groupII! 11510431!NEDD4! ENSGACG00000015840! groupII! 11521136!PRTG! ENSGACG00000015845! groupII! 11545338!PBXIP1! ENSGACG00000015850! groupII! 11572026!pigb! ENSGACG00000015853! groupII! 11579796!rab27a! ENSGACG00000015855! groupII! 11586268!vrk3! ENSGACG00000015858! groupII! 11595660!FAM65A!! ENSGACG00000015863! groupII! 11605084!ctcf! ENSGACG00000015865! groupII! 11613372!ush2a! ENSGACG00000016262! groupII! 14069329!esrp2! ENSGACG00000016359! groupII! 14972131!NFATC3! ENSGACG00000016365! groupII! 14997944!!ENSGACG00000021256! groupII! 15394094!!ENSGACG00000016462! groupII! 15396267!tmed3! ENSGACG00000016467! groupII! 15396450!!ENSGACG00000017354! groupII! 21973529!!ENSGACG00000017356! groupII! 21977401!atp6v0d1! ENSGACG00000017358! groupII! 21978560!ppip5k1a! ENSGACG00000017465! groupII! 22253581!astn1! ENSGACG00000014303! groupIII! 3571387!atp1b3a! ENSGACG00000015750! groupIII! 9036495!gk5! ENSGACG00000015760! groupIII! 9042783!!ENSGACG00000017965! groupIII! 16675147!!ENSGACG00000017967! groupIII! 16679542!tmem255a! ENSGACG00000017424! groupIV! 6653772!zbtb33! ENSGACG00000017428! groupIV! 6662937!nkap! ENSGACG00000017430! groupIV! 6667356!    138 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$enox2! ENSGACG00000017504! groupIV! 7553451!BEND4! ENSGACG00000017530! groupIV! 7960082!ATP8A1! ENSGACG00000017532! groupIV! 7972945!scyl1! ENSGACG00000017569! groupIV! 8174384!ehd1b! ENSGACG00000017605! groupIV! 8233238!MAP1LC3C! ENSGACG00000017618! groupIV! 8247807!slc43a1a! ENSGACG00000017810! groupIV! 9284842!PPP2R2C! ENSGACG00000017871! groupIV! 9490428!wfs1b! ENSGACG00000017874! groupIV! 9499826!!ENSGACG00000022555! groupIV! 9503554!HRH2! ENSGACG00000017883! groupIV! 9586702!DOCK2! ENSGACG00000017885! groupIV! 9595096!TRPC7! ENSGACG00000017891! groupIV! 9701995!DIAPH1! ENSGACG00000017896! groupIV! 9760072!tmco6! ENSGACG00000017990! groupIV! 10696737!!ENSGACG00000017993! groupIV! 10703030!!ENSGACG00000017995! groupIV! 10714052!!ENSGACG00000018041! groupIV! 11075050!SIMC1! ENSGACG00000018042! groupIV! 11085492!cnot6! ENSGACG00000018046! groupIV! 11092615!kctd12! ENSGACG00000018268! groupIV! 12353365!!ENSGACG00000018286! groupIV! 12451969!znf185! ENSGACG00000018287! groupIV! 12459458!fut11! ENSGACG00000018289! groupIV! 12463399!nlgn3a! ENSGACG00000018296! groupIV! 12500111!unc5a! ENSGACG00000018337! groupIV! 13161092!PDLIM7! ENSGACG00000018344! groupIV! 13220801!cltb! ENSGACG00000018348! groupIV! 13247849!higd2a! ENSGACG00000018350! groupIV! 13252044!EBF1! ENSGACG00000018450! groupIV! 14093708!!ENSGACG00000018493! groupIV! 14882018!nxt2! ENSGACG00000018508! groupIV! 15058946!psmd10! ENSGACG00000018509! groupIV! 15062301!xiap! ENSGACG00000018511! groupIV! 15065812!stag2b! ENSGACG00000018514! groupIV! 15071091!tenm1! ENSGACG00000018534! groupIV! 15373076!POU3F4! ENSGACG00000018547! groupIV! 15610119!slc7a2! ENSGACG00000018601! groupIV! 15845227!mtmr7a! ENSGACG00000018602! groupIV! 15857017!cnot7! ENSGACG00000018605! groupIV! 15866651!zdhhc2! ENSGACG00000018608! groupIV! 15871647!her1! ENSGACG00000018614! groupIV! 15916641!her5! ENSGACG00000018615! groupIV! 15920212!pfdn6! ENSGACG00000018616! groupIV! 15923902!    139 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$mmgt1! ENSGACG00000018620! groupIV! 15928732!ddx26b! ENSGACG00000018623! groupIV! 15932627!!ENSGACG00000021515! groupIV! 15981048!!ENSGACG00000022231! groupIV! 15981371!gpc3! ENSGACG00000018655! groupIV! 15982395!gpc4! ENSGACG00000018656! groupIV! 16071855!brd8! ENSGACG00000018659! groupIV! 16098569!psd2! ENSGACG00000018679! groupIV! 16223245!nrg2! ENSGACG00000018682! groupIV! 16245967!!ENSGACG00000018696! groupIV! 16331382!GABRA2! ENSGACG00000018709! groupIV! 16385527!dcps! ENSGACG00000018734! groupIV! 16563045!ids! ENSGACG00000018736! groupIV! 16565926!aff2! ENSGACG00000018740! groupIV! 16586165!!ENSGACG00000022553! groupIV! 16690488!fmr1! ENSGACG00000018742! groupIV! 16711643!rab33a! ENSGACG00000018745! groupIV! 16726999!casp3b! ENSGACG00000018765! groupIV! 17650266!grm8a! ENSGACG00000018829! groupIV! 18215045!nampt! ENSGACG00000018838! groupIV! 18426809!gsap! ENSGACG00000018841! groupIV! 18438511!ccdc146! ENSGACG00000018843! groupIV! 18461196!fgl2a! ENSGACG00000018844! groupIV! 18488149!strip2! ENSGACG00000018938! groupIV! 19685754!ahcyl2! ENSGACG00000018942! groupIV! 19707744!tbc1d22a! ENSGACG00000019014! groupIV! 20565588!lmo3! ENSGACG00000019264! groupIV! 23585527!mgst1.1! ENSGACG00000019267! groupIV! 23687033!rint1! ENSGACG00000019309! groupIV! 23850745!dennd5b! ENSGACG00000019311! groupIV! 23864187!MPPED1! ENSGACG00000019323! groupIV! 24031409!SCUBE1! ENSGACG00000019325! groupIV! 24111028!samm50l! ENSGACG00000019330! groupIV! 24192525!aldh1l2! ENSGACG00000019333! groupIV! 24202091!!ENSGACG00000019334! groupIV! 24219970!slc41a2b! ENSGACG00000019336! groupIV! 24223728!kcnc2! ENSGACG00000019441! groupIV! 24763664!!ENSGACG00000019455! groupIV! 24852375!nav3! ENSGACG00000019472! groupIV! 25318371!arid2! ENSGACG00000019522! groupIV! 25849279!igbp1! ENSGACG00000019563! groupIV! 26204734!magt1! ENSGACG00000019568! groupIV! 26207852!fbxo38! ENSGACG00000019572! groupIV! 26215973!!ENSGACG00000021313! groupIV! 26242542!    140 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$!ENSGACG00000021446! groupIV! 26243268!csnk1a1! ENSGACG00000019578! groupIV! 26250725!stox2b! ENSGACG00000019586! groupIV! 26297130!enpp6! ENSGACG00000019588! groupIV! 26317520!fgf4! ENSGACG00000019608! groupIV! 26392351!SLC7A11! ENSGACG00000019609! groupIV! 26452010!il1rapl2! ENSGACG00000019644! groupIV! 27845019!!ENSGACG00000019647! groupIV! 28304039!cspp1b! ENSGACG00000019648! groupIV! 28308294!snx25! ENSGACG00000019651! groupIV! 28316948!!ENSGACG00000019652! groupIV! 28330947!fgl1! ENSGACG00000019653! groupIV! 28334582!TMEM60! ENSGACG00000019659! groupIV! 28429103!phtf2! ENSGACG00000019660! groupIV! 28438400!adipor2! ENSGACG00000019662! groupIV! 28457178!syt1a! ENSGACG00000020017! groupIV! 31925264!cyb5r3! ENSGACG00000020019! groupIV! 31945908!!ENSGACG00000020041! groupIV! 32090728!igf1! ENSGACG00000020042! groupIV! 32098461!KCNMA1! ENSGACG00000002248! groupV! 434056!xpo1b! ENSGACG00000002743! groupVI! 1376813!alox5a! ENSGACG00000007762! groupVI! 9151832!slc25a16! ENSGACG00000007784! groupVI! 9159664!spock2! ENSGACG00000008600! groupVI! 9823329!!ENSGACG00000018640! groupVII! 406357!CLEC19A! ENSGACG00000018643! groupVII! 411335!col4a5! ENSGACG00000018775! groupVII! 976428!!ENSGACG00000019337! groupVII! 4136442!lphn3! ENSGACG00000019498! groupVII! 5325617!PTPRD! ENSGACG00000019528! groupVII! 5878944!USP5! ENSGACG00000019533! groupVII! 6050601!myoz2! ENSGACG00000019535! groupVII! 6072903!SYNPO2! ENSGACG00000019536! groupVII! 6086475!sec24d! ENSGACG00000019537! groupVII! 6100518!dhx15! ENSGACG00000019549! groupVII! 6551731!!ENSGACG00000019590! groupVII! 6623513!rpl34! ENSGACG00000019593! groupVII! 6663776!ostc! ENSGACG00000019601! groupVII! 6666220!etnppl! ENSGACG00000019605! groupVII! 6668657!NFIB! ENSGACG00000019677! groupVII! 7005184!SNORD22! ENSGACG00000022786! groupVII! 7725103!SNORD29! ENSGACG00000022602! groupVII! 7727427!SNORD31! ENSGACG00000022446! groupVII! 7727697!SNORD22! ENSGACG00000022711! groupVII! 7728434!    141 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$!ENSGACG00000019819! groupVII! 7739783!!ENSGACG00000019888! groupVII! 8020685!!ENSGACG00000019893! groupVII! 8044555!!ENSGACG00000019895! groupVII! 8092741!rpgrip1! ENSGACG00000019896! groupVII! 8099373!fam113! ENSGACG00000019900! groupVII! 8107491!TDRD7! ENSGACG00000019913! groupVII! 8130248!wrap53! ENSGACG00000019924! groupVII! 8570726!gigyf1! ENSGACG00000019925! groupVII! 8583416!ENDOD1! ENSGACG00000019927! groupVII! 8599185!tmem256! ENSGACG00000019937! groupVII! 8805329!TMEM102! ENSGACG00000019941! groupVII! 8813564!fgf11! ENSGACG00000019942! groupVII! 8870275!chrnb1! ENSGACG00000019943! groupVII! 8932781!chrnb1l! ENSGACG00000019946! groupVII! 8950083!cldn7a! ENSGACG00000019947! groupVII! 8968580!!ENSGACG00000019950! groupVII! 8975875!ponzr5! ENSGACG00000019951! groupVII! 8986296!!ENSGACG00000019952! groupVII! 8987681!!ENSGACG00000019953! groupVII! 8995661!!ENSGACG00000019954! groupVII! 9000505!!ENSGACG00000019958! groupVII! 9005340!PTGDR2! ENSGACG00000019959! groupVII! 9007774!kirrela! ENSGACG00000020005! groupVII! 9351489!!ENSGACG00000020029! groupVII! 9728033!NRXN2!! ENSGACG00000020030! groupVII! 9805143!nrip1b! ENSGACG00000020154! groupVII! 12304843!wscd1b! ENSGACG00000020156! groupVII! 12410732!SIM2! ENSGACG00000020158! groupVII! 12494791!hlcs! ENSGACG00000020159! groupVII! 12512681!B3GAT1! ENSGACG00000020191! groupVII! 13248581!!ENSGACG00000020192! groupVII! 13266860!STT3A! ENSGACG00000020193! groupVII! 13275786!prkrir! ENSGACG00000020194! groupVII! 13295494!!ENSGACG00000020195! groupVII! 13300100!wnt11r! ENSGACG00000020196! groupVII! 13318639!uvrag! ENSGACG00000020197! groupVII! 13374257!dgat2! ENSGACG00000020198! groupVII! 13452421!mogat2! ENSGACG00000020199! groupVII! 13463877!map6! ENSGACG00000020200! groupVII! 13472213!umodl1! ENSGACG00000020201! groupVII! 13489019!zbtb21! ENSGACG00000020202! groupVII! 13505638!cxadr! ENSGACG00000020209! groupVII! 13593320!auts2a! ENSGACG00000020210! groupVII! 13627454!    142 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$cltca! ENSGACG00000020216! groupVII! 14183650!dhx33! ENSGACG00000020229! groupVII! 14498720!c1qbp! ENSGACG00000020230! groupVII! 14507117!MLXIPL! ENSGACG00000020231! groupVII! 14513150!SRRM3! ENSGACG00000020232! groupVII! 14555092!YWHAG!! ENSGACG00000020234! groupVII! 14603675!SSC4D! ENSGACG00000020235! groupVII! 14614771!BZRAP1! ENSGACG00000020236! groupVII! 14658991!!ENSGACG00000022230! groupVII! 14692593!!ENSGACG00000022453! groupVII! 14692704!supt4h1! ENSGACG00000020237! groupVII! 14706030!hpd! ENSGACG00000020238! groupVII! 14778775!MTMR4! ENSGACG00000020239! groupVII! 14793405!ca4! ENSGACG00000020240! groupVII! 14824723!gusb! ENSGACG00000020241! groupVII! 14852692!vkorc1! ENSGACG00000020242! groupVII! 14865123!NUPR1! ENSGACG00000020243! groupVII! 14871252!znhit3! ENSGACG00000020244! groupVII! 14873239!MYO19! ENSGACG00000020245! groupVII! 14876168!pigw! ENSGACG00000020246! groupVII! 14888367!ggnbp2! ENSGACG00000020247! groupVII! 14891145!dhrs11! ENSGACG00000020248! groupVII! 14896973!!ENSGACG00000020249! groupVII! 14910165!flot2! ENSGACG00000020250! groupVII! 14919931!!ENSGACG00000021516! groupVII! 14937030!!ENSGACG00000021526! groupVII! 14937156!eral1! ENSGACG00000020251! groupVII! 14938208!fam222b! ENSGACG00000020252! groupVII! 14964767!trpv1! ENSGACG00000020253! groupVII! 14974131!shpk! ENSGACG00000020254! groupVII! 14983177!emc6! ENSGACG00000020255! groupVII! 14987555!p2rx5! ENSGACG00000020256! groupVII! 14990159!ubc! ENSGACG00000020257! groupVII! 14996045!!ENSGACG00000020258! groupVII! 14996446!cenpv! ENSGACG00000020259! groupVII! 15002338!!ENSGACG00000020271! groupVII! 15260522!stip1! ENSGACG00000020272! groupVII! 15340712!rps6ka4! ENSGACG00000020273! groupVII! 15349635!FLRT1! ENSGACG00000020274! groupVII! 15405552!vps51! ENSGACG00000020279! groupVII! 15448825!tm7sf2! ENSGACG00000020280! groupVII! 15454238!adssl! ENSGACG00000020288! groupVII! 15567186!!ENSGACG00000022183! groupVII! 15576697!!ENSGACG00000021274! groupVII! 15576872!    143 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$ammecr1! ENSGACG00000020290! groupVII! 15598480!gabra3! ENSGACG00000020292! groupVII! 15640600!thoc2! ENSGACG00000020367! groupVII! 17555779!gria3! ENSGACG00000020368! groupVII! 17580031!atp6v1e1a! ENSGACG00000020388! groupVII! 17841023!NYAP1! ENSGACG00000020389! groupVII! 17857676!atp1b2b! ENSGACG00000020390! groupVII! 17868985!gltpd2! ENSGACG00000020391! groupVII! 17880340!chrne! ENSGACG00000020392! groupVII! 17887230!adamts15! ENSGACG00000020418! groupVII! 18219272!bco2a! ENSGACG00000020420! groupVII! 18286645!fhl1b! ENSGACG00000020432! groupVII! 18527504!slc9a6b! ENSGACG00000020433! groupVII! 18539190!il2rgb! ENSGACG00000020434! groupVII! 18548109!snx12! ENSGACG00000020435! groupVII! 18553876!abhd11! ENSGACG00000020497! groupVII! 19269612!cldnh! ENSGACG00000020498! groupVII! 19279730!rxfp2! ENSGACG00000020550! groupVII! 19898579!fry! ENSGACG00000020551! groupVII! 19937626!trpc4a! ENSGACG00000020568! groupVII! 20270318!synj1! ENSGACG00000020575! groupVII! 20539803!tiam1! ENSGACG00000020582! groupVII! 20634688!map3k7cl! ENSGACG00000020584! groupVII! 20695666!GRM5! ENSGACG00000020586! groupVII! 20721798!TYR! ENSGACG00000020587! groupVII! 20737985!chordc1a! ENSGACG00000020588! groupVII! 20741461!!ENSGACG00000020589! groupVII! 20749176!cdh2! ENSGACG00000002928! groupVIII! 95049!irf4b! ENSGACG00000004966! groupVIII! 3510444!exoc2! ENSGACG00000004977! groupVIII! 3518916!lrp8! ENSGACG00000006827! groupVIII! 7401543!kank4! ENSGACG00000006846! groupVIII! 7475544!FCHO1! ENSGACG00000007515! groupVIII! 8519939!rfx2! ENSGACG00000007541! groupVIII! 8542911!acsbg2! ENSGACG00000007559! groupVIII! 8571604!mllt1! ENSGACG00000007614! groupVIII! 8587594!acer1! ENSGACG00000007629! groupVIII! 8601701!myo1f! ENSGACG00000007641! groupVIII! 8608236!NR2F6! ENSGACG00000007766! groupVIII! 8767315!ptprsa! ENSGACG00000008773! groupVIII! 10168569!nr5a2! ENSGACG00000008896! groupVIII! 10503534!HYKK!! ENSGACG00000009062! groupVIII! 11269682!nmur1a! ENSGACG00000009069! groupVIII! 11278291!    144 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$!ENSGACG00000009076! groupVIII! 11289636!HTR3E! ENSGACG00000009077! groupVIII! 11296839!!ENSGACG00000009078! groupVIII! 11301927!KIAA0226! ENSGACG00000009086! groupVIII! 11312464!tsc22d2! ENSGACG00000009134! groupVIII! 11336696!lrrc15! ENSGACG00000010035! groupVIII! 12743498!CCDC50! ENSGACG00000010040! groupVIII! 12761712!p3h2! ENSGACG00000010066! groupVIII! 12787013!!ENSGACG00000010465! groupVIII! 13126242!arhgef18b! ENSGACG00000010468! groupVIII! 13135192!slc1a6! ENSGACG00000011239! groupVIII! 14314442!LTBP! ENSGACG00000011786! groupVIII! 15151890!CYP2J2! ENSGACG00000011790! groupVIII! 15158424!uqcrh! ENSGACG00000011853! groupVIII! 15173417!!ENSGACG00000011857! groupVIII! 15177443!clockb! ENSGACG00000015939! groupIX! 489361!polq! ENSGACG00000016100! groupIX! 1429086!ptpn9b! ENSGACG00000016107! groupIX! 1443105!usp53b! ENSGACG00000017796! groupIX! 8669923!myoz2b! ENSGACG00000017804! groupIX! 8678717!ctnna2! ENSGACG00000017985! groupIX! 9146194!DLC1! ENSGACG00000018034! groupIX! 9720025!kcnq5b! ENSGACG00000018097! groupIX! 9900166!scrn2! ENSGACG00000018155! groupIX! 10153416!!ENSGACG00000018157! groupIX! 10162047!CBX4!! ENSGACG00000018219! groupIX! 10496079!inab! ENSGACG00000018244! groupIX! 10749285!nt5c2! ENSGACG00000018247! groupIX! 10750399!jup! ENSGACG00000018328! groupIX! 10988528!kcnip2! ENSGACG00000018330! groupIX! 10996164!rgs12b! ENSGACG00000018356! groupIX! 11144114!dok7! ENSGACG00000018358! groupIX! 11180226!lrpap1! ENSGACG00000018360! groupIX! 11194695!!ENSGACG00000018365! groupIX! 11287767!cpz! ENSGACG00000018366! groupIX! 11321370!htra3a! ENSGACG00000018376! groupIX! 11334493!pi4k2b! ENSGACG00000018377! groupIX! 11339606!sclt1! ENSGACG00000018401! groupIX! 11484494!prkca! ENSGACG00000018519! groupIX! 12793088!cacng5! ENSGACG00000018521! groupIX! 12900795!cacng4a! ENSGACG00000018524! groupIX! 12915523!cacng1! ENSGACG00000018528! groupIX! 12926489!helz! ENSGACG00000018530! groupIX! 12938360!gne! ENSGACG00000018604! groupIX! 13630506!    145 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$clta! ENSGACG00000018609! groupIX! 13638390!evpla! ENSGACG00000019260! groupIX! 16159871!ten1! ENSGACG00000019266! groupIX! 16171112!acox1! ENSGACG00000019268! groupIX! 16175263!pik3r5! ENSGACG00000019372! groupIX! 16835067!ntn1a! ENSGACG00000019374! groupIX! 16854580!hn1a! ENSGACG00000019394! groupIX! 17081308!trim71! ENSGACG00000008309! groupX! 12801925!abat! ENSGACG00000004710! groupXI! 304385!prrg2! ENSGACG00000004841! groupXI! 428273!nags! ENSGACG00000005126! groupXI! 764494!b4galnt2.2! ENSGACG00000005131! groupXI! 798847!rnd2! ENSGACG00000009070! groupXI! 6291989!hs3st3b1a! ENSGACG00000010816! groupXI! 8484883!pmp22a! ENSGACG00000010819! groupXI! 8512176!!ENSGACG00000010914! groupXI! 8839051!GPRC5C! ENSGACG00000010915! groupXI! 8840666!btbd17b! ENSGACG00000010921! groupXI! 8847136!zgc:171489! ENSGACG00000011523! groupXI! 10379462!cdr2a! ENSGACG00000011526! groupXI! 10385184!sdr42e2! ENSGACG00000011530! groupXI! 10392237!!ENSGACG00000011652! groupXI! 10494542!mkl2b! ENSGACG00000013988! groupXI! 13726297!slc16a7! ENSGACG00000003483! groupXII! 2152217!!ENSGACG00000003497! groupXII! 2156242!PNPLA8! ENSGACG00000003501! groupXII! 2161697!kcnc4! ENSGACG00000007907! groupXII! 10128185!SLC6A17!! ENSGACG00000007913! groupXII! 10154388!adora1b! ENSGACG00000008072! groupXII! 10258103!def6! ENSGACG00000008311! groupXII! 10464460!!ENSGACG00000008347! groupXII! 10475511!DTX3! ENSGACG00000008546! groupXII! 10679024!rnd1l! ENSGACG00000008549! groupXII! 10697368!cacnb3! ENSGACG00000008568! groupXII! 10710520!ADCY6!! ENSGACG00000008575! groupXII! 10731068!ARHGEF25!! ENSGACG00000008996! groupXII! 11150617!ankrd33aa! ENSGACG00000009005! groupXII! 11188599!hoxc13b! ENSGACG00000009389! groupXII! 11575429!hoxc10! ENSGACG00000009394! groupXII! 11606252!csrp1! ENSGACG00000009706! groupXII! 12100734!phlda3! ENSGACG00000009727! groupXII! 12103970!tnni1a! ENSGACG00000009730! groupXII! 12109940!!ENSGACG00000009740! groupXII! 12113808!lad1! ENSGACG00000009743! groupXII! 12117810!    146 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$tnnt2a! ENSGACG00000009747! groupXII! 12120260!pkp1a! ENSGACG00000009752! groupXII! 12129960!tead3! ENSGACG00000009758! groupXII! 12143626!ube2t! ENSGACG00000009813! groupXII! 12215552!ETV7! ENSGACG00000009831! groupXII! 12218917!KCNA10! ENSGACG00000009836! groupXII! 12234412!KCNA2! ENSGACG00000009837! groupXII! 12251007!!ENSGACG00000009842! groupXII! 12262938!ngfa! ENSGACG00000009843! groupXII! 12310358!tspan2b! ENSGACG00000009844! groupXII! 12315252!!ENSGACG00000009866! groupXII! 12325491!TSHB! ENSGACG00000009897! groupXII! 12329010!CACNA2D3! ENSGACG00000010123! groupXII! 12488249!LRTM1! ENSGACG00000010150! groupXII! 12569996!wnt5a! ENSGACG00000010153! groupXII! 12609079!ERC2! ENSGACG00000010158! groupXII! 12648568!tlr9! ENSGACG00000010164! groupXII! 12738331!apeh! ENSGACG00000010172! groupXII! 12743186!capza1b! ENSGACG00000010181! groupXII! 12750847!cttnbp2nlb! ENSGACG00000010190! groupXII! 12758253!!ENSGACG00000010196! groupXII! 12771425!acss2! ENSGACG00000010216! groupXII! 12779521!mapre1b! ENSGACG00000010256! groupXII! 12798594!dnmt3! ENSGACG00000010262! groupXII! 12806093!dnmt4! ENSGACG00000010273! groupXII! 12823810!commd7! ENSGACG00000010283! groupXII! 12844822!lama5! ENSGACG00000010405! groupXII! 13049591!tuba1b! ENSGACG00000010436! groupXII! 13123145!RAB29! ENSGACG00000010473! groupXII! 13152549!NPBWR2! ENSGACG00000010477! groupXII! 13219071!oprl1! ENSGACG00000010479! groupXII! 13262291!FAM187B! ENSGACG00000010484! groupXII! 13309389!sox18! ENSGACG00000010505! groupXII! 13373226!xkr7! ENSGACG00000010506! groupXII! 13403771!ccm2l! ENSGACG00000010511! groupXII! 13448683!!ENSGACG00000010523! groupXII! 13483912!hsd17b10! ENSGACG00000010525! groupXII! 13489301!!ENSGACG00000022263! groupXII! 13493981!!ENSGACG00000022179! groupXII! 13494464!tfcp2! ENSGACG00000010929! groupXII! 13828331!csrnp2! ENSGACG00000010943! groupXII! 13840252!itga5! ENSGACG00000010945! groupXII! 13865837!PLXNA2!! ENSGACG00000011007! groupXII! 14055012!rgmb! ENSGACG00000009766! groupXIII! 9546174!    147 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$mef2cb! ENSGACG00000010270! groupXIII! 10417033!!ENSGACG00000021300! groupXIII! 19683692!ptrh1! ENSGACG00000014682! groupXIII! 19684305!slc27a6! ENSGACG00000018349! groupXIV! 14021776!lmnb1! ENSGACG00000018353! groupXIV! 14039862!!ENSGACG00000005133! groupXV! 1444644!fosaa! ENSGACG00000007617! groupXV! 4764418!mlh3! ENSGACG00000007623! groupXV! 4768572!crlf1a! ENSGACG00000009968! groupXV! 8210188!snap23.1! ENSGACG00000011055! groupXV! 10429950!sptb! ENSGACG00000011100! groupXV! 10442747!asmt! ENSGACG00000006621! groupXVI! 12969734!fer1l6! ENSGACG00000008352! groupXVI! 16508075!prkag3b! ENSGACG00000008373! groupXVI! 16525600!parp4! ENSGACG00000008521! groupXVI! 16908539!!ENSGACG00000008710! groupXVI! 17405163!!ENSGACG00000008714! groupXVI! 17408026!!ENSGACG00000008715! groupXVI! 17411595!ppcs! ENSGACG00000007405! groupXVII! 6273370!utp3! ENSGACG00000007417! groupXVII! 6274869!!ENSGACG00000007429! groupXVII! 6279497!fam83e! ENSGACG00000007430! groupXVII! 6282073!!ENSGACG00000007437! groupXVII! 6285207!SLC2A9! ENSGACG00000009129! groupXVII! 8676250!znf395! ENSGACG00000004727! groupXVIII! 1364627!ufl1! ENSGACG00000006287! groupXVIII! 4128680!HS3ST5! ENSGACG00000006332! groupXVIII! 4276771!mllt4a! ENSGACG00000006700! groupXVIII! 4873161!!ENSGACG00000006716! groupXVIII! 4892806!!ENSGACG00000006718! groupXVIII! 4904826!ANO1! ENSGACG00000002381! groupXIX! 959942!!ENSGACG00000011039! groupXIX! 13049496!kcnj11! ENSGACG00000011042! groupXIX! 13049738!mob2! ENSGACG00000011046! groupXIX! 13053541!osbpl5! ENSGACG00000011081! groupXIX! 13227769!ldha! ENSGACG00000011270! groupXIX! 13757826!tsg101a! ENSGACG00000011311! groupXIX! 13764439!hrasa! ENSGACG00000011340! groupXIX! 13775246!rag2! ENSGACG00000011461! groupXIX! 14489909!rag1! ENSGACG00000011465! groupXIX! 14493756!polr3b! ENSGACG00000011794! groupXIX! 15049715!rfx4! ENSGACG00000011830! groupXIX! 15070736!SPIRE1! ENSGACG00000004577! groupXX! 2462673!atp6v1c1a! ENSGACG00000004615! groupXX! 2501237!    148 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$TMEM74! ENSGACG00000005301! groupXX! 3788628!rbms3! ENSGACG00000006377! groupXX! 6102582!grb10! ENSGACG00000006438! groupXX! 6202336!nrsn1! ENSGACG00000006587! groupXX! 6912058!SLC6A3! ENSGACG00000006614! groupXX! 7033258!cdkal1! ENSGACG00000006622! groupXX! 7344820!tcea3! ENSGACG00000007286! groupXX! 8195753!kcnk5b! ENSGACG00000007871! groupXX! 9218218!EPB41! ENSGACG00000007879! groupXX! 9233627!!ENSGACG00000007893! groupXX! 9249276!CSNK2B! ENSGACG00000007897! groupXX! 9251806!cyp21a2! ENSGACG00000007916! groupXX! 9263386!!ENSGACG00000007933! groupXX! 9276661!!ENSGACG00000021296! groupXX! 9290682!!ENSGACG00000007971! groupXX! 9292404!rxrbb! ENSGACG00000007982! groupXX! 9297024!fhod3b! ENSGACG00000007994! groupXX! 9307622!sp8a! ENSGACG00000008062! groupXX! 9481109!macc1! ENSGACG00000008067! groupXX! 9490944!twist2! ENSGACG00000008070! groupXX! 9505813!!ENSGACG00000008075! groupXX! 9526242!hdac9b! ENSGACG00000008076! groupXX! 9526503!ankrd28b! ENSGACG00000008084! groupXX! 9563160!fam188a! ENSGACG00000008767! groupXX! 10155793!!ENSGACG00000008776! groupXX! 10173184!cdh17! ENSGACG00000009113! groupXX! 10687628!rad54b! ENSGACG00000009128! groupXX! 10704434!rnf41l! ENSGACG00000009133! groupXX! 10715828!!ENSGACG00000009138! groupXX! 10719553!esrp1! ENSGACG00000009169! groupXX! 10732757!epb41l4b! ENSGACG00000009175! groupXX! 10747464!ptpn3! ENSGACG00000009182! groupXX! 10758440!fam171a1! ENSGACG00000009230! groupXX! 10816365!nmt2! ENSGACG00000009238! groupXX! 10834107!styk1! ENSGACG00000010432! groupXX! 11900072!phc1! ENSGACG00000010438! groupXX! 11905700!m6pr! ENSGACG00000010442! groupXX! 11911904!NAT14! ENSGACG00000010519! groupXX! 12046215!aicda! ENSGACG00000010521! groupXX! 12050972!NECAP1! ENSGACG00000010530! groupXX! 12053199!epn1! ENSGACG00000010556! groupXX! 12077130!foxj2! ENSGACG00000010572! groupXX! 12085828!isoc2! ENSGACG00000010580! groupXX! 12090387!atg12! ENSGACG00000010597! groupXX! 12092527!    149 Gene$Name$ ENSEMBL$gene$id$ Chromosome$ Gene$Start$Position$rcv1! ENSGACG00000010601! groupXX! 12100423!!ENSGACG00000010614! groupXX! 12112453!ccdc106! ENSGACG00000010622! groupXX! 12120567!u2af2! ENSGACG00000010632! groupXX! 12125987!klf6a! ENSGACG00000001778! groupXXI! 1563702!smyhc2! ENSGACG00000002145! groupXXI! 2929787!C3orf19! ENSGACG00000000764! Un!!!ENSGACG00000000832! Un!!COL7A1! ENSGACG00000000833! Un!!GNAI2! ENSGACG00000000839! Un!!kbp! ENSGACG00000001219! Un!!!ENSGACG00000001222! Un!!!ENSGACG00000001226! Un!!kbp! ENSGACG00000001229! Un!!!ENSGACG00000001237! Un!!MON1B! ENSGACG00000001246! Un!!TCTA! ENSGACG00000001247! Un!!GLYCTK! ENSGACG00000001248! Un!!BSN! ENSGACG00000001561! Un!!!ENSGACG00000001670! Un!!!ENSGACG00000000873! Un!!KIAA0495! :! Un!!!ENSGACG00000001002! Un!!SLC12A7! ENSGACG00000001024! Un!!!ENSGACG00000001055! Un!!SASS6! :! Un!!TG! ENSGACG00000000896! Un!!CRYGN! ENSGACG00000018096! Un!!NYREN18! ENSGACG00000018101! Un!!!ENSGACG00000018093! Un!! 

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.24.1-0308707/manifest

Comment

Related Items