Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fine-scale population genetic structure of the eastern Pacific bay pipefish, Syngnathus leptorhynchus de Graaf, Ramona Christine 2006

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

Item Metadata

Download

Media
831-ubc_2007-0072a.pdf [ 10.32MB ]
Metadata
JSON: 831-1.0093068.json
JSON-LD: 831-1.0093068-ld.json
RDF/XML (Pretty): 831-1.0093068-rdf.xml
RDF/JSON: 831-1.0093068-rdf.json
Turtle: 831-1.0093068-turtle.txt
N-Triples: 831-1.0093068-rdf-ntriples.txt
Original Record: 831-1.0093068-source.json
Full Text
831-1.0093068-fulltext.txt
Citation
831-1.0093068.ris

Full Text

Fine-scale population genetic structure of the eastern Pacific bay pipefish, Syngnathus leptorhynchus. by Ramona Christine de Graaf B.Sc. (Hons, Coop), The University of Victoria, 1999 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Zoology) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A December 2006 © Ramona Christine de Graaf, 2006 Abstract Seascapes are complex systems with features that both restrict and enhance dispersal and gene flow among individuals. Understanding factors shaping genetic diversity and genetic population substructure can assist in planning effective conservation areas. Syngnathus leptorhynchus is an eelgrass-dependent fish species ranging from Alaska to Mexico. Pipefish males brood eggs in a specialized brood pouch until their emergence as fully developed young. S. leptorhynchus has a body form that presumably reflects adaptations to mimic Zostera marina L . blades which may, in combination with the provision of prolonged parental care, limit its dispersal potential and promote population subdivision. Syngnathus leptorhynchus were collected from 17 localities i i i Barkley Sound, Vancouver Island, British Columbia. A total of 156 alleles were detected with 5 microsatellite loci with an average of 31 ( S T D V 13.9) alleles per population. Genetic diversity was high and the observed multi-locus heterozygosity was 0.91. Genetic diversity and allelic richness were lowest in North Barkley Sound encompassing the Broken Group Islands, Pacific R i m National Park Reserve (BGI). Pipefish genetic differentiation revealed patterns of local heterogeneity at small spatial scales within the B G I archipelago, but regional homogeneity at larger spatial scales. Genetic differentiation measured over the five loci was weak but statistically significant ( F $ T = 0.005, P = 0.0001) over a maximum distance between localities of 83 km. Individual pairwise F s T comparisons ranged from 0.00 to 0.017. There was evidence for closed population dispersal consistent with geography and two nascent genetic subpopulations encompassing East Barkley and South/North Barkley Sound. Pipefish at Gibralter Island (BGI) were significantly diverged from other localities. Fjordal environments appeared to restrict pipefish gene flow between the two genetic subpopulations. Archipelagos and deep-water channels restricted pipefish gene flow and increased genetic differentiation. Genetic > i i connectivity of pipefish in localities within the B G I archipelago was restricted and localities were significantly differentiated from each other. Pipefish in two B G I localities, however, revealed unrestricted connectivity and were genetically similar to neighbouring localities outside this oceanographically complex archipelago. Coastal habitats maintain connectivity among pipefish in eelgrass beds located in South/North Barkley Sound and were important in promoting an isolation-by-distance pattern among coastal eelgrass beds. Pipefish genetic neighbourhoods ranged in linear distance from 40-60 kilometres. i i i Table of Contents Abstract • i i Table of Contents iv List of Tables ' vi List of Figures • vi i Acknowledgements vi i i Dedication x Chapter 1: General Introduction , 1 1.1 Population Structure 1 Population structure in marine habitats 2 1.2 Conservation and Genetic Markers 5 Genetic Drift and Migration , 6 Natural Selection . 9 1.3 Coastal Marine Habitats and Connectivity. 9 Coastal Marine Habitats: Seagrasses 11 Syngnathus leptorhynchus Girard 13 1.4 Thesis Objectives , 21 Chapter 2: Fine-scale population genetic Structure of the.. 24 eastern Pacific bay pipefish Syngnathus leptorhynchus 2.1. Introduction 24 2.1.1 Marine Species and population structure 24 2.1.2 Seascape Genetics of the eastern Pacific bay pipefish 26 2.1.3 General hypotheses and goals of research 34 2.2 Materials and Methods 35 2.2.1 Site Selection '. 35 2.2.2 Eelgrass bed size and distance calculations 36 2.2.3 Sample Collection and D N A Extraction 39 2.2.4 Isolation of and cloning of microsatellites 40 2.2.5 Screening of existing microsatellite primers 45 2.2.6 P C R Amplification, electrophoresis and scoring of variation 46 2.2.7 Genetic Data Analysis 49 2.3 Results 58 2.3.1 Microsatellite polymorphi sm 58 2.3.2 Al le l ic diversity and richness 59 2.3.3 Linkage disequilibrium : 67 2.3.4 Hardy-WeinbergEquilibrium 68 2.3.5 Heterozygosity and gene diversity 71 2.3.6 Microsatellite allelic variation 73 iv 2.3.7 Interpopulation microsatellite differentiation 75 spatial genetic population substructure Exact Tests of allelic differentiation 75 F S T analysis and A M O V A of Regions 83 P C A - G E N analysis 9 0 B A R R I E R analysis 91 S T R U C T U R E Analysis 94 Individual-based assignment analyses 9 4 Isolation by Distance 103 2.4 Discussion 109 2.4.1 Microsatellite Variation within populations 109 Linkage Disequilibrium 109 Hardy Weinberg Equilibrium 109 Gene Diversities 109 2.4.2 Fine-scale population genetic structure of the eastern 110 Pacific bay pipefish Population Genetic Structure 110 Comparisons with other taxa 114 Isolation by Distance 121 Isolation by Distance and Fjord habitats 124 2.4.3 Effect of seascape features on genetic subdivision of the 126 eastern Pacific bay pipefish, Syngnathus leptorhynchus Archipelagos 126 Deep-water channels and genetic divergence 130 Spatial area size of breeding groups of the eastern 131 Pacific bay pipefish Local Currents and Inter-Patch Gene Flow 134 2.5 Conclusion 137 Chapter 3: General Conclusions and Conservation Implications 140 3.1 General Conclusions 140 3.2 Conservation Implications. 141 3.3 Marine reserves and eelgrass ecosystem integrity 144 Literature Cited 149 Appendices 163 Appendix I: Allele frequencies for each locus .163 Appendix U : F S T values of pairwise comparisons among localities by 167 each individual locus. v Lis t of Tables Table 2.1 Sample Locations and descriptions 39 Table 2.2 Primer sequences (5' - 3') and P C R annealing temperatures 45 and MgCl2 concentrations. Table 2.3 L o c i Arrays and Size Ranges 58 Table 2.4 Samples sizes, number of alleles, expected and observed. . 61 heterozygosities, proportion of available alleles, number of private and rare alleles. Table 2.5 Al le l ic richness per locus and locality based on a minimum 66 sample size of 13 diploid individuals (Assists). Table 2.6 Comparison of Al le l ic richness differences among geographic 67 regions. Table 2.7 Linkage Disequilibrium by locus pair over all localities : . 67 Table 2.8 Hardy-Weinberg Equilibrium tests for each locality and each locus . . . . . . 69 Table 2.9 Gene diversity (multi-locus expected heterozygosity) differences 72 among geographic regions. Table 2.10 Log-likelihood (G) based exact tests of population differentiation 76 of each locus among all localities. Table 2.11 Fisher's Exact Tests of population differentiation on all possible 78 pairwise comparisons by each locus. Table 2.12 Fisher's Exact Test P values of population differentiation of 83 allele frequencies conducted by combining loci and pairwise among localities. Table 2.13 F S T values for each locus and overall five microsatellite loci 84 Table 2.14: F S T values between pairwise localities of the eastern Pacific bay 86 pipefish of five combined microsatellite loci . Table 2.15: P values of F S T estimates between pairwise localities.for 86 combined loci . Table 2.16: Comparison of F S T values of groups of localities classified into 87 regions and coastal and island habitat types as well as distinct archipelagos. Table 2.17: Hierarchical analysis of the distribution of genetic diversity 89 in Barkley Sound pipefish populations under various hypotheses. Table 2.18: Mean likelihood scores and standard deviation of S T R U C T U R E . . 95 for hypothesized populations of eastern Pacific bay pipefish inferred from variation at five microsatellite loci . Table 2.19: Results of G E N E C L A S S 2 assignment of individual fish to their 96 location of sampling. Table 2.20: Results of G E N E C L A S S 2 assignment of fish to geographic 98 regions of sampling. Table 2.21: Mantel and partial Mantel tests of the correlation of genetic 103 distance and factors of geographic distance and habitat types. Table 2.22: Mantel and partial Mantel tests of the correlation of genetic 107 distance and factors of geographic distance and putative deep-water barrier types. Data were logio transformed and tests were conducted using I B D software and 10,000 randomizations. v i Lis t of Figures Figure 1.1 Photographs of female and male Syngnathus leptorhynchus 15 Figure 2.1 Map of study sites in Barkley Sound 37 Figure 2.2 3 2 P autoradiography gel of microsatellite locusTl6 44 Figure 2.3 Alle le Frequencies of loci Slep9 and C S L 9 by each locality 74 Figure 2.4 Principal component analysis of allele frequency variation in 92 bay pipefish, Syngnathus leptorhynchus, at five microsatellite loci. Figure 2.5 Analysis of areas of limited gene flow as identified 93 by B A R R I E R Figure 2.6 Relationship between geographic distance (km) and pairwise 99 proportion of misassigned eastern Pacific bay pipefish. Figure 2.7 Connectivity estimates and gene flow dynamics of the bay 100 pipefish among geographic regions of Barkley Sound. Figure 2.8 The number of pairwise localities in distance categories 102 less than 35 km and greater than 35 km within each inter-regional grouping. Figure 2.9 Isolation by Distance of eastern Pacific bay pipefish 104 sample localities Figure 2.10 Mantel tests of correlation between genetic distance and 106 geographic distance of coastal and island habitat types. Figure 2.11: Multi-locus F S T values between pairwise localities by categories 108 used in Mantel test for the effect of putative deep-water barriers on genetic distance. Acknowledgements It is difficult to sum up the many contributions of Dr. E B (Rick) Taylor, my supervisor, to this project. Without his knowledge of all things fishy, support, and fairness, there would be no thesis. Thank you, Rick, for your patience and persistence and for the unforgettable opportunity to learn new tools. I would also like to acknowledge my committee members Dr. Sally Aitken and Dr. Rob DeWreede for their contributions to this thesis. For assisting with my transformation from gum-boot marine ecologist to molecular ecologist and for great friendship and laugher, I would like to thank the magicians and my friends from the Taylor lab: Al lan Costello, Anna E lz , Jen Gow, Katriina lives, Jenn McLean, Emi ly Rubidge, M i k e Stamford, Matt Sniatynski, and the "man" Patrick Tamkee. Without your help, I would have destroyed more lab equipment than I did. To Anna and her love of sequencing, I owe a huge debt in acquiring the sequences for my precious CSL9 marker. The encouragement, focus, and support from this incredible lab group w i l l not be forgotten. For constant encouragement, lively rants, big (marine) fish stories, inspiration and epic tales the prize goes to Dr. J. D . (Don) McPhai l . I also thank Sharon Jeffrey for providing me with valuable GIS help. I also thank Brad Mason, Fisheries and Oceans Canada, for his assistance in management of the eelgrass mapping data on the Community Mapping Network. I also thank the many helpful and encouraging staff members of the Zoology Department and the Zoology Computing Unit. I would also like to acknowledge Dr. A G Jones for helpful suggestions in optimizing the existing east coast pipefish microsatellite primers. Also , many thanks to Dr. A B Wilson for providing me with two species-specific microsatellite loci . v i i i I 'd like to acknowledge the legions of brave folks who helped me in the field including the many volunteers and dock "vagrants" from the Bamfield Marine Sciences Centre ( B M S C ) , B M S C Public Education Department, Al lan Costello, Katriina lives, and Marika Younker. True to form, my family were hardworking and enthusiastic field assistants and included my brother Raymond Degraaf, niece Hilary Degraaf and nephew Michael Degraaf. These relatives embody the adage that blood is thicker than (salt) water. For inspiration and great talks about the crazy bay pipefish, I remember with kindness the late Dr. Kristina D . Louie. Kristina was a young, passionate marine biologist. Her PhD thesis provided many insights into the phylogenetics of the bay pipefish and other eelgrass taxa. I benefited from our many late-night conversations sharing stories and discoveries. Science and conservation have lost a brilliant mind for she was taken from this world too soon. A n d to prove that there is life outside of university, I 'd like to acknowledge the support, ease of friendship, and journeys into nature so essential to my spirit: M i k e Poser, Susan Shirley, Barbara Lucas, Anne Stewart, Simone Runyan, Sylvain LaForest, and M r . S. Chuck— thanks for keeping the fire lit! I would like to acknowledge and thank my funding sources: National Science and Engineering Resource Council P G S A , Bamfield Marine Sciences Centre Graduate Student Scholarships, Wor ld Wildlife Fund Research Grant, Dr. E B Taylor and the Department of Zoology, U B C . Animals were collected under licence number de Graaf-2001-1 Canadian Department of Fisheries and Oceans, Bamfield Marine Sciences Centre collection permit and a Parks Canada permit. ix Dedication To my family: my legion of nieces and nephews, my brothers Howard, John and Raymond for your support, laughter and friendship. To Sylvia Knoller, for just being and sharing so much of your life. To my parents, Arthur and Hi lda Degraaf, for their never-ending acceptance, patience and encouragement. Only by their examples of perseverance and focus have I made it this far on my path. For my mom, I particularly thank you for believing in me and for your patience and encouragement throughout this time. Although my Dad did not live to see this thesis written, his pride in my accomplishments, no matter how small, his steady hand and brilliance, his teaching, and his daily presence in my life while I laboured for every word of this thesis ensured its completion. Y o u are the wind in my sails. x Chapter 1 General Introduction 1.1 Population substructure A species can be subdivided into smaller units for many reasons including historical events, geographic distance, discontinuous habitats, and/or ecological and behavioural factors (Hedrick 2000) . Population subdivision ultimately results in genetic variation among the different units (Hedrick 2000) due to the increased probability of random drift in semi-isolated populations of finite size, and this is true for a wide range of taxa (Balloux and Lugon-Moulin 2002) . The degree of population subdivision reflects, among other things, the absolute amount of genetic exchange among population units and forces acting on loci (such as gene flow and genetic drift) within each individual breeding group (Balloux and Lugon-Moulin 2002; Hellberg et al. 2002) . Estimates of migration and connectivity among population units may be deduced indirectly through the detection of genetic differentiation between groups (Hellberg et al. 2002) . These estimates, together with knowledge of a species' biology and habitat utilization, can benefit resource managers and protected areas planners in conserving species and communities of organisms. Relative to freshwater fish and anadramous fish species, marine fish generally have lower degrees of population subdivision which is commonly measured as the proportion of total genetic variance attributable to differences among populations or " F S T " (Gyllensten 1985; Ward et al. 1994). A s well , marine fish have higher heterozygosities and almost triple the average number of alleles per locus relative to freshwater species (DeWoody and Avise 2000) . Higher heterozygosities and the extreme polymorphism (the number of alleles per locus) of some loci shown by marine fishes is perhaps advantageous for certain applications such as parentage analysis; but at the same time, such high intrapopulation variation sometimes makes detection of divergence between populations difficult (Waples 1998; DeWoody and Avise 1 2000). Interpopulation variation measured using F S T w i l l be reduced due to the sensitivity of F S T to high levels of within population heterozygosity (Hedrick 2000; Taylor et al. 2003). The allelic polymorphism differences between marine and freshwater fish may be due, in part, to the higher effective population sizes, larger population range sizes, and more continuous habitats of marine fish relative to freshwater fish (Waples 1998; Ward et al. 1994; DeWoody and Avise 2000). Population structure in marine habitats To early mariners, the oceans seemed a vast and unchanging space. The 1876 expedition of the HMS Challenger was the first scientific voyage to investigate the physical, chemical and biological properties of the deep sea in the great oceanic basins (Webber and Thurman 1991). After this oceanographic expedition, one of the prevailing conclusions was the constancy of the ocean's salinity and, therefore, further proof of the homogeneity of this vast space (Lalli and Parsons 1993). With the aid of modern technology, this view has changed dramatically; and we now know the oceans' hemispheres to be divided by pronounced currents resulting in distinct biogeographical regions, each with unique temperature, salinity and hydrographic regimes (Lall i and Parsons 1993). Habitat heterogeneity exists from the largest of oceanic scales to smaller coastal scales (Feral 2002). The view that the marine realm was a large featureless world, with few barriers for marine species dispersal was, however, still prevalent in the thinking of many whom investigated marine taxa for evidence of stock or population structure (Waples 1998). Marine environments, and hence marine populations themselves, were deemed to be "open" systems meaning that larvae and individuals were free to move without impediment (Cowen et al. 2000). Dispersal can be defined as the movement of an individual from its natal habitat or group to other breeding areas or groups (Coulon et al. 2004). Recently, numerous 2 studies looking at direct and indirect measures of dispersal have shown that many marine systems do exhibit barriers to movement such that distinct genetic structuring of populations is common. Marine invertebrates and fish display a wide range of larval dispersal types, adult swimming abilities, and habitat usage and specialization. Processes providing opportunities for genetic population subdivision of high dispersal species are defined by Palumbi (1994) as: barriers to gene flow (biogeographic barriers, ocean currents, temperature and salinity barriers, coastal environments), isolation-by-distance (spatial scales and arrangement of suitable habitats), behavioural limits (homing by adults and habitat selectivity by larvae), selection (adaptation to local environments) and recent history (historical vicariance and contemporary gene flow). The interplay between physical factors and biological constraints has made predicting the level of self-recruitment (Sponaugle et al. 2002) and population divergence based on larval dispersal potential complex (Shulman and Bermingham 1995). After numerous genetic studies using diverse marine taxa, what is striking is not that species with larvae with low dispersal potential show population substructure, but that many species with larvae capable of long pelagic dispersal times do (Palumbi 1994; Shulman and Bermingham 1995; Cowen et al. 2000). Large, pelagic species with exceptional swimming abilities such as migratory yellow fin tuna (Thunnus albacares) exhibit genetic population substructure at a spatial scale of entire oceans (Appleyard et al. 2001) while smaller, coastal (non-pelagic) species such as the brown rockfish (Sebastes auriculatus) exhibit population substructure at the scale of hundreds of kilometres along a coastline (Buonaccorsi et al. 2005). Marine coastal species confined to habitat types due to habitat specialization and physical limitations may demonstrate population genetic substructure. High levels of genetic divergence have been found in many coastal marine species of fishes and invertebrates (Burton and Feldman 1981; Maltagliati 1999; Mariani et al. 2002; Buonaccorsi et al. 2005). The 3 processes that are important in structuring coastal populations by generating barriers to gene flow include local currents (Shulman and Bermingham 1995; G i l g and Hilbish 2003; Knutsen et al. 2003; Sotka et al. 2004; Kenchington et al. 2006); salinity and temperature gradients (Jorgensen et al. 2005) as well as coastal habitats and seascape features (Burton and Feldman 1981; Planes et al. 1996; Johnson and Black 1998; Ruzzante et al. 2000; Johnson et al. 2001; Riginos and Nachman 2001; Congiu et al. 2002; Hoarau et al. 2002; Watts and Johnson 2004). Other isolating features important for population subdivision are isolation-by-distance processes (Buonaccorsi et al. 2005; Congiu et al. 2002); behavioural processes such as natal homing (Congiu et al. 2002; Gold and Turner 2002; Knutsen et al. 2003) and larval behaviour (Warner et al. 2000; Kingsford et al. 2002; Sponaugle et al. 2002). Estimates of genetic variation and gene flow among populations provide indirect measures of dispersal distances vital to our understanding of the degree of population connectivity in the marine realm. Using data from genetic studies where marine species showed a signal of isolation-by-distance, Kinlan and Gaines (2003) modeled dispersal distances and found that dispersal distance/generation ranged from < 1 to 100 kilometres for species with low dispersal potential, such as algae, and from 20 to 200 kilometres for species with higher dispersal potential such as marine fish. On average, dispersal distances for marine invertebrates were 10 kilometres and for marine fish 200 kilometres (Palumbi 2004). Evidence is mounting, however, that home range sizes for non-pelagic marine fishes are much smaller than predicted (Kramer and Chapman 1999; Palumbi and Warner 2003; Palumbi 2006). In addition, dispersal opportunities are more limited and dispersal distances shorter than the thousands of kilometers long claimed to correlate with a featureless, homogenous ocean (Jones et al. 1999; Swearer et al. 1999; Cowen et al. 2000). As more researchers show that populations do diverge at finer-4 geographical scales, it may be time to assert that for many marine species, it is, indeed, a small world after all . 1.2 Conservation and Genetic M a r k e r s Using genetic information to provide estimates of dispersal is an attractive alternative to collecting data of direct dispersal. Anthropogenically-induced degradation of our planet's physical and biological systems is occurring at temporal scales so brief and spatial scales so vast that species may lack the evolutionary variation to persist in such a rapidly depleted world (Davis and Shaw 2001; Will iams et al. 2003). Increasingly, conservation biologists seek methods to address urgent issues critical to species survival such as inbreeding effects in small populations, recolonization and dispersal abilities of species, and the efficacy of protected area strategies. While direct measures, such as mark-recapture methods, of parameters such as connectivity and dispersal are desirable, genetic estimates of critical population parameters for some species can be obtained where resources, logistics or time do not permit direct estimates. Just such an approach has been used for the North American Brown bear (Paetkau et al. 1998), endangered species of cetaceans (Hoelzel 1998) and desert fishes (Meffe and Vrijenhoek 1988). Direct measures of demographic parameters provide estimates only for a specific time and place. Indirect genetic measures allow extrapolation of demographic parameters over large spatial and temporal scales and also provide measures beyond movement (migration) of individuals such as genetic exchange through breeding (gene flow) (Neigel 1997). Population genetic studies employ a variety of molecular markers and techniques. Markers that are under the influence of selection, such as some allozymes (alternative alleles of proteins), allow one to test the influence of environmental gradients (eg. temperature or salinity) on a species but are not best suited to the pursuit of population or stock delineation 5 (Hellberg et al. 2002). The availability of a large number of allozyme loci that are not under the influence of selection and the low costs of this technique, however, still make the. use of some allozymes important and relevant in studies of genetic population substructure. The study of evolutionary relationships among mitochondrial D N A haplotypes is useful for phylogeographic and matrilineal studies (Avise 1998). Microsatellite markers are increasing in use for studies directed to detecting population genetic substructure and dispersal. Microsatellite loci are found throughout the nuclear genome and are composed of repeats of 2 - 10 base pair nucleotide segments (Hellberg et al. 2002). Microsatellites are highly abundant with high mutation rates considered to range from 10"2 and 10"5 (Weber and Wong 1993; Jarne and Lagoda 1996). The majority of microsatellite loci are neutral to the forces of selection due to their position in non-coding regions of the genome; however, due to the random distribution of these loci , occasionally they are found in coding regions (Jarne and Lagoda 1996). Microsatellite loci , which are typically assayed as allelic length variants, can present problems due to allele size homoplasy (two alleles differing in nucleotide sequence composition, but not length) (Hellberg et al. 2002). Overall, however, due to their high abundance in the nuclear genome, high variability in alleles (due to variable repeat lengths), and high mutation rates, microsatellites are considered to be more sensitive than other nuclear markers to population genetic processes such as genetic drift and migration (gene flow) (Jarne and Lagoda 1996). Genetic drift and migration Genetic drift influences changes in allele frequencies due to finite population sizes and is an important force in structuring genetic differentiation among groups of animals (Neigel 1997; Hedrick 2000). Discontinuous or naturally patchy habitats can isolate groups of animals allowing allele frequencies to drift randomly relative to processes acting on other populations 6 (Hedrick 2000; Hellberg et al. 2002). Opposing the action of genetic drift in promoting genetic divergence among populations is gene flow—the effective movement (migration followed by genetic exchange) of genes—which acts to homogenize populations as genetic exchange promotes similarities among populations (Hedrick 2000; Hellberg et al. 2002). The influence of genetic drift on genetic variation among populations varies inversely with effective population size (Hellberg et al. 2002). Genetic drift can also be influenced by historic subdivisions among contemporary populations as well as founder and bottleneck effects (Hedrick 2000). Marine fish species generally have large census population sizes; but due to the high rates of mortality of adults, effective population sizes are often only a fraction of census sizes (Hellberg et al. 2002). A s well, isolation of populations by various mechanisms ranging from behaviour to geography (Palumbi 1994) facilitates the action of genetic drift to increase genetic differentiation among populations. Non-pelagic species inhabiting coastal areas may be subject to the forces of genetic drift due to various factors including non-continuous habitats (Maltagliati 1999; Congiu et al. 2002; Watts and Johnson 2004) and coastline complexity (Withler et al. 2001). Migration rates can vary dramatically among marine species due to the myriad of different life-history strategies of marine taxa. Migration followed by successful reproduction provides gene flow and genetic connectivity between groups of animals. Theoretical models predict that the movement of just one individual per generation is sufficient to prevent substantial genetic differentiation among populations (Hedrick 2000). In marine species, rates of gene flow, even among distant populations are generally higher than a single individual/generation yet significant population genetic structure is often detected (Paulmbi 2004). High mutation rates (p) associated with microsatellites, high population sizes and high gene diversity of marine fishes (DeWoody and Avise 2000; Ward et al. 1994) can combine to 7 play a large role in determining levels of genetic differentiation and gene flow. High within population mutation rates, when u. approaches the migration rate (m), may lead to saturation (Hellberg et al. 2002), and coupled with the bias due to allelic homoplasy, mutation-based differences between individuals in separate populations can appear less different than they are (Hellberg et al. 2002). The result can be that gene flow and connectivity estimates between populations are overestimated, population substructure is underestimated and dispersal is overestimated (Hellberg et al. 2002). Migration-drift equilibrium refers to a balance between the gain of alleles (by gene flow) and the loss of alleles (by genetic drift) by populations (Neigel 1997; Hutchinson and Templeton 1999) assuming an otherwise Hardy-Weinberg population model (Hartl and Clark 1997). Populations that are not in equilibrium relative to the forces of genetic drift and gene flow (effective migration) include those affected by historical forces resulting in recent recolonization of habitats. Population models using F S T based measures assume equilibrium between the two forces of genetic drift and gene flow (Hellberg et al. 2002). A recolonized population wi l l look genetically similar to that of its founding individuals until sufficient time has passed to allow the action of genetic forces to reflect contemporary subdivision and gene flow (Neigel 1997; Hellberg et al. 2002). Species with low dispersal abilities w i l l require more time to reach equilibrium than species with high dispersal abilities due to the inverse relationship between the time to reach equilibrium and migration rates (Hellberg et al. 2002). Interpretation of genetic data can be challenging when populations are in disequilibrium. For instance, nonequilibrium conditions and can lead to an overestimation of gene flow between recolonized and source populations (Hellberg et al. 2002) and underestimates of population genetic subdivision relative to estimates made with direct measures of dispersal.and knowledge of a species' life-history and morphology. 8 Natural selection Loc i under the influence of selection can also greatly influence interpretation of genetic data (Neigel 1997). The influence of selective forces acting on genetic loci within populations may be an alternative hypothesis to demographic independence among populations deduced by the hypothesis of genetic divergence (Neigel 1997; Hellberg et al. 2002). Selection for different alleles in different geographic areas, the effects of stabilizing selection (where an intermediate genotype is favoured) or a selective sweep (where an allele is favoured throughout the region of study) can create signatures of genetic heterogeneity or genetic homogeneity, respectively, confounding estimates of genetic connectivity (Pogson and Fevolden (2003); Hedrick 2000; Hellberg et al. 2002) particularly i f morphological characters or environmental conditions are homogenous. A s previously described, microsatellite loci located in non-coding regions of the nuclear genome are deemed to be neutral to selective forces (Hedrick 2000), but unless loci have been shown to be unlinked to fitness-related loci (Feral 2002), selective forces are still important to consider when interpreting genetic data (Neigel 1997). Consideration of the consequences of genetic drift, gene flow, and selection are vital to interpreting genetic data. Also, these genetic forces do not act in isolation but are influenced by environmental heterogeneity and the spatial configuration, quality and health of local environments. 1.3 Coastal Marine Habitats and Connectivity Understanding processes promoting and restricting genetic connectivity at local seascape scales is important for management and conservation of coastal marine habitats. Populations often require a network of interconnected habitats to maintain important biological processes such as reproduction and recruitment to allow long-term persistence (Hanski 1999). 9 Understanding population connectivity within habitat networks is vital to designing marine protected areas that are effective in maintaining these basic biological processes (Palumbi 2002). Population connectivity is measurable by quantifying a species' genetic population substructure and patterns of gene flow at varying spatial scales (Palumbi 2002; Hellberg et al. 2002; Buonaccorsi et al. 2005). Marine reserves are often designed as a network of habitats where spatially discrete groups of individuals are assumed to be interconnected by "open" population dynamics where extensive larvae or adult dispersal promotes mixing and genetic similarity over large geographic scales (Cowen et al. 2000). A s this assumption of large-scale connectivity due to extensive dispersal is tested, in fact, many marine species display more "closed" population dynamics with genetically distinct populations and high self-recruitment rates over smaller geographic scales (Jones et al. 1999; Swearer et al. 1999; Cowen et al. 2000; Palumbi 2004; Hoffman et al. 2005). We can be less certain of the ailing convention that larval dispersal rates are high and that adult migration wi l l maintain vital demographic connections among geographically distant population units (Cowen et al. 2000; Palumbi 2004; Buonaccorsi et al. 2005). The degree of ecological connectivity and the ability of individual reserves and reserve networks to maintain population dynamics and genetic diversity are, generally, uncertain. Measuring population connectivity at local geographic scales where marine reserves are located is importance to assess reserve effectiveness in maintaining genetically fit populations (Cowen et al. 2000; Palumbi 2004). The spatial configuration of habitats and seascape features can influence gene flow patterns within and among populations. Little is known of the importance of seascape features and the spatial arrangement of eelgrass habitats in maintaining genetically fit populations of the eastern Pacific Bay pipefish. M y study focuses on describing the genetic connectivity of the eastern Pacific bay pipefish, Syngnathus leptorhynchus, an eelgrass-dependent fish species, within a complex seascape. 10 Coastal marine habitats: seagrasses Seagrasses are not seaweeds but are marine angiosperms confined to photic environments even more limited than seaweeds in both depth distribution and photic quality (Bortone 1998). Zostera marina L , commonly known as eelgrass, is the most widely distributed seagrass throughout the temperate waters of the Pacific and Atlantic oceans in the northern hemisphere (Olsen et al. 2004). In British Columbia, seagrasses ecosystems are found along a narrow coastal fringe ranging from the intertidal to a maximum depth of 10-20 metres. Eelgrass systems have been described as one of the richest in biodiversity in the sea (Phillips 1969). Eelgrass ecosystems play diverse roles in nearshore systems providing spawning, nursery and year-round habitat (Bortone 1998). Eelgrass also plays an important role in offshore food webs by providing primary production and globally by providing oxygen and acting as a carbon sink (Short and Wyllie-Echeverria 1996). Over the last decade, seagrass areas have been noted to be in decline worldwide, decreasing by an alarming 15 percent over that time due to both natural and anthropogenic causes (Short and Wyllie-Echeverria 1996; Green and Short 2003). Consequently, the loss of seagrass habitats is a major conservation concern. Eelgrass habitats are naturally patchy, or fragmented, in distribution but have become increasingly fragmented due to human perturbation (Short and Wyllie-Escheverria 1996). At the local scale (<1 to 100 metres) eelgrass beds are structurally variable in many components including shoot density and biomass due to various factors such as hydrodynamics (Hovel et al. 2002), photic environments, substrate composition, water quality and human perturbation (such as anchoring, dredging and pollution). A t the seascape scale (1 to 10s of kilometers), eelgrass habitat structure is varied and is found in narrow fringing beds along gentle slopes, 11 fragmented patches (eelgrass plants separated by sand) and large continuous meadows over extensive tidal flats (Precision Identification 2002). Distributed as a series of disjunct or fragmented habitat areas, the structural and habitat complexity of eelgrass distribution can affect species diversity, abundance and dispersal (Sogard 1989; Turner et al. 1999; Hovel et al. 2002; Skilleter et al. 2005) in ways perhaps analogous to that documented for terrestrial habitats such as forests. The area size and connectivity of habitat patches wi l l affect rates of movement and dispersal of animals and resources (Skilleter et al. 2005). For example, Sogard (1989) found that colonization rates and species diversity of some fishes and crustaceans were higher in seagrass patches that were in close proximity relative to seagrass patches spaced at larger distances from each other. Due to a wide variety of factors, however, the findings of Sogard (1989) do not hold across all taxa and species abundance and diversity can be lower among closely situated eelgrass patches relative to more distal ones. Hydrodynamics and larval retention dynamics (Hovel et al. 2002), microscale structural complexity of seagrasses (Orth et al. 1984), linkages with other coastal habitats (such as mangroves, Skilleter et al. 2005) all play different roles in species and community dynamics. These microhabitat variations of eelgrass habitats also play a role in the ecology of the eastern Pacific bay pipefish, Syngnathus leptorhynchus, as its presence is affected by different hydrodynamic regimes and sediment substrates (de Graaf, unpublished data). Investigations of dispersal or, indirectly gene flow, at landscape scales are not common in marine systems (Skilleter et al. 2005; Kenchington et al. 2006) and are even rarer for eelgrass ecosystems. The population genetic structure of Zostera marina L . has been studied at patch level spatial scales to trans-oceanic scales (Ruckelshaus 1998; Reusch et al. 2000; Reush et al. 2002; Olsen et al. 2004). Information on gene flow dynamics of eelgrass itself is highly valuable but a single-species approach has limited utility for conservation planning. The 12 population genetic structure of fish and invertebrates inhabiting eelgrass habitats on the eastern Pacific Ocean coast has only been investigated at wide geographic scales (Louie 2003; Wilson 2006) and not at small, local spatial scales. Understanding home range sizes and habitat utilization of dependent taxa, such as coral reef species, has been an invaluable tool in reef management (Kramer and Chapman 1999; Jones et al. 1999; Swearer et al. 1999). A multi-species approach can provide valuable information for monitoring eelgrass habitat function. A large number of fishes and invertebrates utilize eelgrass beds during some part of their life cycle, but few species are dependent on eelgrass beds throughout their entire development and life span. (Adams 1976). The eastern Pacific bay pipefish, Syngnathus leptorhynchus, is an eelgrass resident species and together with the eelgrass nudibranch, Phyllaplysia taylori, and the eelgrass limpet, Tectura depicta, are a few eastern Pacific species exclusively associated with Zostera marina L habitats throughout their life spans. The eastern Pacific bay pipefish, Syngnathus leptorhynchus The eastern Pacific bay pipefish, Syngnathus leptorhynchus (Girard 1854), is distributed in the eastern Pacific Ocean from Alaska to Baja, California (Fritzsche 1980; Dawson 1985) and lives in eelgrass habitats over its entire life cycle (Bayer 1980; Fritzsche 1980). S. leptorhynchus, like many other pipefish, is an abundant (Bayer 1980) and important member of Zostera marina L . fish assemblages. It reaches a maximum total length of 33 centimetres (Dawson 1985) and females reach greater lengths than males (personal observation). Within the order Gasterosteiformes, the family Syngnathidae is composed of pipefishes, pipehorses, sea moths, seahorses and sea dragons (Foster and Vincent 2004). While the taxonomy of syngnathids is uncertain, 320 species of pipefish in 55 genera have been described (Kuiter 2000) and 32-33 species of seahorses occur in one genus (Lourie et al. 13 1999). The fish species within the order Gasterosteiformes have concentric rings of bony plates (Clemens and Wi lby 1949). Pipefish are generally nearshore, coastal fish species restricted to shallow waters due to habitat specialization (Howard and Koehn 1985). Pipefish share with seahorses a unique reproductive biology with females transferring eggs to males where males fertilize the eggs. Pipefish in the genus Syngnathus brood eggs in a fully enclosed, specialized brood-pouch providing osmotic regulation, oxygen and nourishment to embryos until their emergence as fully developed young (Berglund et al. 1986; Carcupino et al. 2002). Pipefish are elongate, slender fishes without pelvic fins (Figure 1.1) (Clemens and Wi lby 1949). Many species are described as sit-and-wait predators (Dawson 1985) relying on crypsis and camouflage for prey capture and predator avoidance. L ike several other pipefish species, S. leptorhynchus has behaviours, colouration and body form that presumably reflect adaptations to mimic Zostera marina L . blades resulting in reduced swimming ability (Dawson 1985). Such extreme changes due to habitat specialization have resulted in some habitat-specialist fishes having higher population genetic structure than non-habitat specialists (Smith and Fujio 1982). Syngnathus leptorhynchus is affected by variable environmental conditions along its range. S. leptorhynchus shows reproductive and high morphological variation along its distributional range in response to environmental variation among habitats from Alaska to Baja, California. Its breeding season varies along a latitudinal cline with temperature (Herald 1941 and Bayer 1980) ranging from year-round mating in California and decreasing in duration northward with cooler water temperatures. Morphological variation in vertebrate count, tail rings and dorsal fin rays are highly variable with latitudinal gradient and correlate well with temperature and local environmental conditions (Hubbs 1922; Fritzsche 1980; Fritzsche 1984). Based on 14 A . Female eastern Pacific bay pipefish, Syngnathus leptorhynchus B. Male eastern Pactfic b ay pipefish, Syngnathus leptorhynchus. The black co lour of the brood pouch indicates late stage of development of young. This male gave birth a few days after this photo was taken. Figure 1.1: Photographs of female and male Syngnathus leptorhynchus taken during sampling in Barkley Sound. Photo credit: R.C.de Graaf 15 morphological variation, S. leptorhynchus was previously described as consisting of northern .and southern subspecies (Herald 1941) until a recent'revision of the subfamily by Fritzsche (1980). High morphological variation along its range resulted in Fritzsche (1980) to hypothesize that discrete populations of S. leptorhynchus may occur at the level of individual bays. A s part of an investigation into the phylogenetics of Syngnathus leptorhynchus, Louie (2003) detected deep phylogenetic breaks along the eastern Pacific Ocean coast from Alaska to California, but these did not occur where taxonomic work had predicted (Herald 1941; Fritzsche 1980). Instead, major breaks were defined by a glacial recolonization hypothesis. Genetic population structure revealed a strong geographic break between northern and southern localities at Puget Sound, Washington State, and localities in the southern part of the range showed evidence of range expansion likely due to recolonization (Louie 2003; Wilson 2006). Louie (2003) also reported differing levels of gene flow among proximal localities due to environmental heterogeneity along the coast and did not find evidence of unique populations within bays (Fritzsche 1980, Louie 2003). Range expansion has been documented in the northern part of its range, where, in Alaska, Orsi et al. (1991) documented an 850 kilometre range expansion. These data provide important information for predicting finer-scale population genetic structure of S. leptorhynchus. Evidence of contemporary range expansion by 5. leptorhynchus may influence allele frequency-based tests such as FST due to lack of equilibrium conditions (Hellberg et al. 2002). While such variation among S. leptorhynchus populations along its ecological range illustrates the richness of this species, it does make predictions of its population dynamics challenging. Palumbi (1994) and Waples (1998) reviewed characteristics of marine species that influence the extent of population substructuring of marine fishes including level of vagility, 16 larval dispersal, population sizes, population stability, effects of habitat and oceanographic processes. Relative to other marine fishes common to coastal British Columbia, Syngnathus leptorhynchus displays many traits that could result in high genetic differentiation and isolation of populations for a marine teleost. While much about the biology of the eastern Pacific bay pipefish has not been studied, information from other syngnathid species may provide useful analogues. Syngnathid species are described as some of the least mobile of marine teleosts (Foster and Vincent 2004) with extreme levels of morphological adaptations for crypsis and camouflage (Dawson 1985; Foster and Vincent 2004). While some pipefishes possess prehensile tails (without caudal fins) to grasp eelgrass blades further reducing swimming abilities (Howard and Koehn 1985), Syngnathus leptorhynchus possesses a caudal fin and swims horizontally as well as vertically in association with upright eelgrass blades (personal observation). S. leptorhynchus is capable of good horizontal swimming ability as it moves along the seabed among eelgrass shoots and across sand patches (personal observation). If adult fishes are constrained in their dispersal abilities, larval stages can provide mechanisms for larger range dispersal i f larvae have suitable planktonic strategies. Pipefish juveniles emerge directly from the male brood pouch without yolk sacs and as free-swimming individuals (Dawson 1985). Syngnathus leptorhynchus newborn juveniles are darkly pigmented, approximately 1.2 cm in length, have high growth rates, and are active swimmers being observed, in captivity, to swim directly to food sources and orientate themselves with eelgrass blades, and grasping eelgrass blades with their tails. These newborn juvenile characteristics would suggest that S. leptorhynchus does not undergo a planktonic dispersal phase or undergoes a shortened dispersal phase relative to other fishes. Kendrick and Hyndes (2003) reported similar attributes of newborn Syngnathus argus and suggested that 17 pelagic dispersal, i f at all, is brief. In plankton tows near eelgrass beds, newborn juveniles have not been observed (personal observation). A l l age classes of the eastern Pacific bay pipefish (inferred by total length data) from newly emerged juveniles to reproductive adults have been found together in eelgrass beds over several months (de Graaf, unpublished data). Bayer (1980) also reported the co-occurrence of S. leptorhynchus juveniles and adults inferring a strong association of juveniles and adults for eelgrass habitats; and similar affinity for habitat among age classes has also been reported. for the seagrass pipefishes Syngnathus argus (Kendrick and Hyndes 2003) and Syngnathus typhle (Ahnesjo 1992). Larval dispersal has been reported for several seahorse species (Foster and Vincent 2004) and inferred for some pipefishes (Kendrick and Hyndes 2003). Truncated dispersal times, however, have been found even in coral reef fish species with fully planktonic larvae due to the influence of local currents in returning larvae to their natal reefs (Jones et al. 1999; Swearer et al. 1999). High retention to local areas as well as behavioural characteristics promoting site fidelity can result in high levels of population genetic subdivision. Pipefish, like seahorses, can be grouped into species with high site fidelity and species known to range widely. Gronnell (1984) demonstrated that the pipefish Corythoichthys intestinalis exhibited high site fidelity. The pipefish, Nerophis lumbriciformis, had a high rate of recapture over 19 months of study (Monteiro et al. 2005). In contrast, five pipefish species studied in a Swedish eelgrass meadow did not demonstrate site fidelity to either a home range or an eelgrass meadow (Vincent et al. 1995). Site fidelity among seahorses, particularly monogamous species, is high (Foster and Vincent 2004). Home ranges of individual adult seahorses are small ranging from several metres to 1,650 m 2 (Perante et al. 2002, Moreau and Vincent 2004, Vincent et al. 2005). Sea dragons display variable home ranges ranging from 50 - 150 m for the weedy sea dragon, Phyllopteryx taeniolatus, (Sanchez-Camara and Booth 18 2004), and for individual leafy sea dragons, Phycodurus eques, ranging from less than one hectare to 88 hectares (Connolly et al. 2002). While the degree to which Syngnathus leptorhynchus exhibits site fidelity is unknown, at two different eelgrass beds in Barkley Sound, southwest coast of Vancouver Island, British Columbia, tagged pipefish were recaptured after two-three weeks (Bamfield Inlet site, Hornbeck 2004) and after two months (Fleming Island site, de Graaf, unpublished data). While many aspects of the biology of S. leptorhynchus would favour the hypothesis of limited realized dispersal correlating with high levels of population genetic subdivision, there are aspects of its biology that may promote moderate levels of dispersal. Marine fish species generally display high population sizes (Waples 1998) and high polymorphism of genetic markers (Ward et al. 1994; DeWoody and Avise 2000). Syngnathus leptorhynchus were numerically abundant in eelgrass beds sampled throughout Barkley Sound; and this may promote high genetic diversity and high polymorphism at microsatellite loci. FST values are influenced by levels of heterozygosity (Waples 1998; Hellberg et al. 2002; Taylor et al. 2003) and high genetic diversity, promoting high levels of variation within populations diminishing measures of interpopulation variance (Waples 1998; Hellberg et al. 2002). As previously mentioned, evidence exists of contemporary northern and southern range expansion and this may influence whether or not Barkley Sound populations of S. leptorhynchus, although located far from both southern and northern range extremes, have reached genetic equilibrium, affecting interpretation of genetic measures of population subdivision and connectivity. While large-scale demographic processes such as population expansion can affect the ability of genetic population models and markers to resolve population substructure, interpatch dispersal abilities wi l l also influence the depth of substructure at smaller landscape scales. 19 While habitat specialization can lead to increased levels of population genetic substructure (Smith and Fujio 1982), survival in a naturally fragmented habitat may promote dispersal abilities and gene flow among population units. For the eastern Pacific bay pipefish, tradeoffs may occur in terms of the advantages gained by remaining in a single eelgrass bed and possible risks, such as predation, while dispersing for acquisition of food resources as well as mating opportunities (Roelke and Sogard 1993). The northern pipefish, Syngnathus fuscus, was observed to move across areas of sand to eelgrass patches. Sogard (1989) and Lazzari and Able (1990) reported dispersal of this species in 10-20 metres of water 20 kilometres from shore during seasonal estuarine variations. Syngnathus leptorhynchus has good swimming abilities, and a high percentage of individuals are bi-coloured which may afford them camouflage within eelgrass shoots and against a dark sediment background (de Graaf, unpublished data). Many individuals have a dark brown dorsal stripe as well as variable lateral stripe colouration and ornamentation (de Graaf, unpublished data). Pigmentation patterns, ornamentation and colours ranging from browns, greens and reds provide at least 14 colour variants (de Graaf, unpublished data). When orientated vertically within eelgrass shoots, the brown and green lateral stripes assist the pipefish to blend in with the vertical habitat, and while swimming the brown dorsal stripe resembles the sand/mud sediment interface assisting in camouflage (personal observation). The eastern Pacific bay pipefish has been noted to raft amongst drifting vegetation at the surface (Dawson 1985, Louie 2003) and other pipefish have been associated with drift algae, or marine tumbleweeds, at the sediment surface (Kulczycki et al. 1981) both providing mobile corridors between disjunct habitats (Ffolmquist 1994). Pipefish have complex social behaviours (Gronell 1984; Vincent et al. 1995) and their mating systems may require them to disperse to find suitable mates (Berglund et al. 1986). Pipefish males are capable of brooding multiple egg batches throughout their lives, and 20 Syngnathus leptorhynchus males assayed from Barkley Sound, like several other pipefishes, are polygamous mating with two - three females (de Graaf, unpublished data). Recently, in samples from Alaska, Washington, Oregon and California, multiple polygamous mating has been detected (Wilson, A B , Northwest Fisheries Centre, Seattle, personal communication). Both males and females can be limited in suitable mates as there is a high level of choosiness by each sex (Berglund et al. 1986; Ahnesjo 1992) perhaps resulting in the need for both sexes to migrate. Movement among eelgrass beds wi l l result in greater gene flow and decrease the degree of population genetic substructure. A successful dispersal strategy may lead to home range sizes being larger than that of an individual eelgrass bed. Isolation of populations can ultimately influence the degree of genetic population substructure and connectivity. Aspects of the biology of Syngnathus leptorhynchus and the seascape features of Barkley Sound (Chapter 2) provide an interesting system in which to examine fine-scale genetic population structure and gene flow. 1.4 Thesis Objectives Microsatellite markers provide useful characters for assessing historical and contemporary forces on population structure. Traditional genetic distance measures (such as Wright's FST) together with more recent maximum likelihood and Bayesian techniques, spatial autocorrelation and approaches to detect the geographic boundaries of restricted gene flow (Monmonier's algorithm) provide useful tools for testing correlations between environmental factors and genetic differentiation (Manel et al. 2003). The utility of these contemporary analyses are greatly advancing the ability to resolve the spatial scale of dispersal and self-recruitment in marine systems which has long been a formidable challenge (Cowen et al. 2000; Co wen et al. 2006). 21 Little is known about the relative influence of contemporary seascape features on structuring genetic population structure, gene flow and self-recruitment of coastal, temperate marine fishes at local, fine-spatial scales (being geographic scales less than 100 kilometres). Even less is known about the importance of the spatial arrangement and availability of eelgrass habitats to the persistence of local, genetically fit populations of eelgrass dependent species such as the eastern Pacific bay pipefish, Syngnathus leptorhynchus. I used microsatellite markers to generate estimates of genetic population substructure, self-recruitment rates and gene flow patterns maintaining connectivity among Syngnathus leptorhynchus localities in a complex seascape. I also tested the influence of seascape features such as local currents, variability in eelgrass habitats, fjords, archipelagos, and deep-water barriers on genetic differentiation and population genetic structure of S. leptorhynchus. I chose to address these issues for two primary reasons. First, the availability of diverse habitats within Barkley Sound, southwest coast of Vancouver Island, British Columbia, the specialized morphology of the eastern Pacific bay pipefish, and the fragmented nature of eelgrass beds provided a natural laboratory to investigate the effects of seascape features on genetic divergence and genetic diversity of a marine fish species at microgeographic scales. Second, the Broken Group Islands, a component of the Pacific Rim National Park Reserve, is a marine park located in Barkley Sound. Fine-scale genetic studies are necessary to elucidate processes operating at local geographic scales to assess the ability of marine reserves to maintain genetic diversity. Understanding local processes promoting and restricting genetic connectivity throughout Barkley Sound is valuable for management and protection of eelgrass dependent fauna and eelgrass habitats both inside and outside of marine reserve boundaries. Syngnathus leptorhynchus is an excellent candidate for use as an indicator species of eelgrass ecological integrity (Chapter 3). The investigation of the population genetic structure and gene 22 flow patterns of S. leptorhynchus in Barkley Sound may provide a framework for understanding the ecological connectivity of eelgrass-dependent fauna and aid conservation programs in marine reserve design. 23 Chapter 2 Fine-scale population genetic structure of the eastern Pacific bay pipefish Syngnathus leptorhynchus 2.1 Introduction 2.1.1 Marine species and population structure Marine species have been investigated over several decades at spatial scales appropriate for phylogenetic studies. Phylogeography has elucidated the influence of historical forces such as the glacial history of coastal western North America on gene flow and speciation and postglacial dispersal by marine and estuarine species e.g., sea cucumbers, Cucumaria miniata and C. pseudocurata (Arndt and Smith 1998), eulachon, Thaleichthys pacificus (McLean 1999), northern clingfish, Gobbiesox maeandricus (Hickerson and Ross 2001), gobies, Clevelandia ios and Eucyclogobius newberryi (Dawson et al. 2002), eastern Pacific bay pipefish, Syngnathus leptorhynchus (Louie 2003), eelgrass seahare, Phyllaplysia taylori Dal l (Louie 2003), eelgrass limpet, Tectura depicta Hinds (Louie 2003), and the staghorn sculpin, Leptocottus armatus (Louie 2003)). The merits of phylogeographic studies at large spatial scales to understand the influence of historical events on genetic diversity of species are invaluable; however, local, small spatial scale studies are necessary to assess and perhaps predict factors influential in maintaining genetically diverse populations. Landscape-scale genetic studies have mainly been conducted dn terrestrial and freshwater species perhaps due to difficulties inherent with working with marine species (Chapter 1) (Waples 1998, Hellberg et al. 2002). In the marine realm, however, local studies in different regions are also important. At local geographic scales, negative impacts of pollution, oi l spills, and coastal development threaten local populations making smaller geographic scale studies important to undertake. As well, marine reserves and marine conservation areas generally cover small geographic ranges 24 of a species distribution and likely capture minimal numbers of genetic populations. Understanding the effects of seascape features on gene flow and population genetic structure can assist in designing and spacing marine reserve networks to ensure populations are well connected by dispersal (Cowen et al. 2000, 2006; Palumbi 2004). Species' ranges often cover vast spatial distances and encompass large numbers of genetic populations and variable habitats. Knowing the degree to which these genetic populations are ecologically connected through migration and interbreeding of individuals is important for conservation planning. Environmental factors influencing the effective movement of individuals (successful migration and reproduction) determine levels of gene flow and patterns of connectivity among these distinct populations (Coulon et al. 2005). Contemporary features of the environment can retain or enhance the mixing of individuals (Hellberg et al. 2002). Correlations between habitat structure (contiguous or fragmented) and genetic structure (homogenous or heterogeneous) of animals has been shown in numerous species including the European roe deer, Capreolus capreolus (Coulon et al. 2005); the damsel fly, Coenagrion mercuriale (Watts et al. 2004); the Eurasian red squirrel, Sciurus vulgaris (Trizio et al. 2005); bull trout, Salvelinus confluentus (Costello et al. 2003); brook char, Salvelinus fontinalis (Castric et al. 2001); Atlantic cod (Ruzzante et al. 2000, Knutsen et al. 2003) and Atlantic herring, Clupea harengus (Bekkevold et al. 2005; Jorgensen et al. 2005). Landscape genetics, a merging of landscape ecology (relationship between landscape structure and species biology) and molecular population genetics, provides a powerful tool for understanding processes and spatial patterns of gene flow operating at small geographic scales (Manel et al. 2003). Tools of landscape genetics, or fine-scale population genetics, can be applied to marine systems to investigate the complex nature of marine "seascape genetics". 25 Like landscapes, seascapes are also complex systems with features that can both restrict and enhance dispersal and gene flow of individuals influencing local patterns of genetic connectivity. Gene flow and dispersal of plants and animals are directly linked and hence influenced by the degree of connectivity of the habitat ( K i m et al. 1998; Sork et al. 1999). Contiguous habitats are more likely to enhance gene flow and maintain genetic homogeneity. Fragmented habitats or impassible barriers are more likely to restrict effective movement or isolate individuals enhancing forces such as genetic drift resulting in genetic heterogeneity among groups and, perhaps, loss of genetic variation (Taylor et al. 2003). Thus, landscape features play an important role in structuring a species' population genetic structure (Costello et al. 2003; Bekkevold et al. 2005; Coulon et al. 2005; Jorgensen et al. 2005), and knowing how local seascape features influence gene flow at small spatial scales can assist in efforts to maintain and protect genetic diversity. 2.1.2 Seascape genetics of the eastern Pacific bay pipefish M y study was designed to examine the seascape genetics of the eastern Pacific bay pipefish, Syngnathus leptorhynchus, in Barkley Sound, southwestern Vancouver Island, British Columbia for several key reasons. First, S. leptorhynchus is a good species to use to investigate local seascape effects on genetic population structure due to its purported limited dispersal abilities and high specialization to its eelgrass habitat (Chapter 1). S. leptorhynchus, like other pipefish species, is an important member of the eelgrass faunal community in terms of both abundance, distribution and trophic function (Bayer 1980; Vincent et al. 1995; Kendrick and Hyndes 2003; Louie 2003). Understanding processes promoting and restricting genetic connectivity throughout Barkley Sound has value for management and protection of populations of the eastern Pacific bay pipefish. 26 Second, the biology of both the eastern Pacific bay pipefish, Syngnathus leptorhynchus, the heterogeneity of eelgrass, Zostera marina, habitats (see Chapter 1) as well as the diversity of oceanographic features of Barkley Sound, provide an interesting system to test seascape effects on the genetic connectivity of pipefish populations. Eelgrass beds in Barkley Sound are found along sheltered and exposed coastal mainland and island zones and within steep-sided fjords. Island eelgrass habitats are separated by deep-water channels (waters greater than 3 0 - 35 metres depth) from other island habitats as well as from mainland, coastal habitats. Deep-water channels may restrict pipefish gene flow among eelgrass habitats and geographic regions. Mainland coastal eelgrass habitats within Barkley Sound are located in an almost linear manner along the Vancouver Island mainland shore and may promote greater connectivity of pipefish. Increased abundance of some fish species is often correlated with proximity of eelgrass patches (Sogard 1989). Physical isolation of groups of pipefish due to geographically restricted dispersal may also lead to genetic differentiation among populations (Wright 1943). Individuals located in geographically proximal, neighbouring habitats are more likely to exchange genetic material through effective gene flow maintaining local homogeneity (Hutchison and Templeton 1999). As effective movement decreases due to increasing geographic distance, individuals become more genetically divergent due to the forces of genetic drift (Hedrick 2000 , Hellberg et al. 2002) . With time and stabilization of population sizes and environmental forces, populations may achieve equilibrium between the forces of gene flow and genetic drift resulting in a positive and monotonic relationship between genetic distance measures such as FST and geographic distance (Slatkin 1 9 8 7 , . Hutchison and Templeton 1999). Linearity of habitats, such as mainland, coastal eelgrass habitats can promote the process of isolation-by-distance. Coastal marine rockfish species have been found 27 to show significant patterns of isolation-by-distance of genetic structure along linear, coastal habitat sites (Buonaccorsi et al. 2004, 2005). Third, local currents play an important role in the transport and retention of larvae and adult animals (Warner et al. 2000) and can affect the "open" or "closed" nature of population dynamics (Cowen et al. 2000). Local currents have been shown to promote open population dynamics of seahorses (Casey 1999) while in other examples, current patterns have been correlated to closed systems resulting in retention of larval fish (Jones et al. 1999), and extreme discontinuities and isolation of gene flow of Caribbean reef fishes (Taylor and Hellberg 2005; Cowen et al. 2006). In the northeastern Pacific Ocean, local coastal current patterns have been demonstrated to influence genetic population structure of coastal rockfish species (Withler et al. 2001; Buonaccorsi et al. 2004; Buonaccorsi et al. 2005). In Barkley Sound, northward flood and southward ebb tidal cycles move predictably throughout the Sound and Alberni Inlet, but are noted to move erratically when encountering narrow passages among islands and shallower water bodies (Stronach et al. 1993). Net northward tidal currents flow through Trevor Channel then westward via Junction Passage to Imperial Eagle Channel (Stronach et al. 1993). Current speeds are low averaging 0.35 km/hr (Stronach et al. 1993). The physical oceanographic attributes of the Alberni Inlet fjord are different from other areas of Barkley Sound. Estuary circulation is the outflow of surface freshwater toward the sea and inward flow of more saline seawater at depth (Thomson 1981). The Alberni Inlet fjord is influenced by estuarine flows with a major freshwater source at its head (Somas River) and smaller sources throughout its course (Doe 1952). Seasonally, the Somas River discharge is significant and brackish water moves southward to mix with the saline waters of Trevor Channel, moves westward via Junction Passage to Imperial Eagle Channel and through the passages throughout the Broken Group Islands to Loudoun Channel 28 (Doe 1952). If Syngnathus leptorhynchus utilize local currents for swimming or perhaps for rafting among floating vegetation, local currents should bias migration to favour movement from the eastern area of Alberni Inlet to southern and northern Barkley Sound. Predominant swell and winds from the west and the northwest should bias drift vegetation or drifting pipefish moving toward southern Barkley Sound. In their investigation of the winged kelp Alaria marginata, Kusomo and Druehl (2000) revealed population substructure on both a regional and local scale and demonstrated the direction of gene flow within the Barkley Sound population was concordant with the prevailing counterclockwise currents in the region. Counteracting local tidal currents or estuarine flow, however, is the ability of eelgrass beds to dampen advection (horizontal movements) and mixing due to wind and tidal currents (Worchester 1995) reducing water velocity such that sediment and plankton are entrained within the eelgrass bed. Due to the effect that eelgrass beds have on hydrodynamics (Worchester 1995), as well as by clinging to blades, pipefish adults and juveniles may be able to resist currents that would otherwise set them adrift. Syngnathids have variable retention periods in their home ranges (Chapter 1). Individually tagged Syngnathus leptorhynchus have been recaptured over periods of two to three weeks (Hornbeck 2004) to 2.3 months (de Graaf, unpublished data) in two different eelgrass environments (sheltered inlet versus a wave exposed beach) in Barkley Sound. During storm events, however, whole eelgrass plants are often dislodged from subtidal eelgrass margins and pipefish could be set adrift. While the benefits of remaining in a patch are many, an effective dispersal strategy is important to limit intraspecific competition for resources as well as maintaining high levels of genetic diversity. Syngnathid social and mating systems may influence their behavioural decisions to remain in or near natal habitats. Genetic analysis of Syngnathus leptorhynchus broods revealed 29 that males carry eggs from two to three females (de Graaf, unpublished data). Numerous pipefish species are polygamous with both males and females seeking multiple pairings (Jones and Avise 1997; Jones et al. 1999) a behaviour that may promote the need for inter-patch dispersal (Jones and Avise 1997; Jones et al. 1999). Vincent et al. (1995) found that resightings of individual pipefish in a single eelgrass meadow were low, which is in contrast with high resightings and site fidelity of monogamous seahorses (Jones et al. 1998; Jones and Avise 2001; Foster and Vincent 2004; Vincent et al. 2005). A dispersal mechanism such as rafting amongst floating eelgrass blades or other algae as well as utilization of net tidal currents and prevailing wind patterns of surface waters may be important mechanisms providing safe corridors between patchy habitats (Highsmith 1985). Pipefish have been noted to move with drifting seaweeds and eelgrass clumps at the surface (Dawson 1985; Fritzsche 1980; Louie 2003). Pipefish females have been observed drifting with detached eelgrass blades as these blades hang vertically in the water column and the pipefish is effectively camouflaged (personal observation). Dispersal by association with marine drift vegetation has been noted for other organisms (Highsmith 1985) and is a major form of dispersal in areas such as the Sargasso Sea. Fourth, the Alberni Inlet fjord is heavily influenced by freshwater sources; and therefore, salinity and temperature differences associated with brackish waters may influence pipefish migration and survival (Power and Attr i l l 2003). Recently, genetic substructure of Atlantic herring was shown to be affected by environmental gradients due to salinity differences in the Baltic Sea (Bekkevold et al. 2005; Jorgenson et al. 2005). Although detailed eelgrass maps for this region are lacking, the presence of eelgrass in the Alberni Inlet fjord would be limited due to fjord's steep sides that limits the accumulation of favourable substrates and depths suitable available for eelgrass growth. Eelgrass beds were only found in areas of 30 river outlets, small pocket coves, and Snug Basin (personal observation). A s well, much eelgrass habitat has been lost due to extensive historical development (canneries and wharves) and contemporary development (aquaculture, logging booms and resort marinas). The Alberni Inlet lacks the feature of an archipelago and eelgrass beds are arranged in a linear manner along both shorelines. Fifth, complex seascapes such as reef systems and archipelagos can serve to retain and even isolate larval and adult stages potentially disrupting gene flow (Johnson 2001). At local spatial scales, the influence of such seascape complexity on genetic subdivision can be strong (Johnson et al 2001). Isolated, finite groups are subject to forces of genetic drift (Hedrick 2000) which can affect their evolutionary potential through fixation of deleterious mutations; or isolation of groups can be advantageous by promoting adaptations to local environments. Isolation can create "sink" habitats reliant on immigrants from other areas to maintain their persistence in habitat patches. Within archipelagos, populations have been found to exhibit greater levels of genetic subdivision at smaller spatial scales relative to comparisons involving mainland coastal populations at larger spatial scales. Evidence of the effect of archipelagos on patterns of local heterogeneity within archipelagos, and regional homogeneity at larger spatial scales has been documented in tropical coral reef species including the Hawaiian milkfish, Chanos chanos, (Winans 1980); the Hawaiian limpet, Cellana exarata, (Shaklee and Samollow 1980); the Indonesian Banggai archipelago cardinal fish, Pterapogon kauderni (Hoffman et al. 2005); the surgeonfish, Acanthurus triostegus, of French Polynesia (Planes et al. 1996 and Planes et al. 1998) and several invertebrate species and a fish from the Western Australian Houtman Abrolhos archipelago (Johnson et al. 1994, Johnson and Black 1996, Johnson et al. 2001; Watts and Johnson 2004). 31 Spatial effects on the gene flow of the planktonically dispersing intertidal limpet, Siphonaria kurracheensis, differed according to spatial scales (Johnson et al. 2001). Within an archipelago at a spatial scale of about 10 kilometres, local heterogeneity between island locality pairs was found (Johnson et al. 2001). This pattern, however, changes at larger spatial scale of 70 and 400 kilometres between island and mainland locality pairs where genetic homogeneity was found (Johnson et al. 2001). Certain attributes of the archipelago had the potential to disrupt the gene flow of this planktonic disperser; and the directionality of the prevailing currents permitted connectivity of gene flow between sites on the extreme edges of the archipelago and mainland sites (Johnson et al. 2001). The surgeonfish, Acanthurus triostegus, showed significant population genetic substructure within archipelagos yet maintained connectivity with populations existing in nearby archipelagos. Population genetic studies of S. kurracheensis and A. triostegus, two different organisms with different life-history traits, both demonstrated a similar effect of the seascape influencing local heterogeneity but regional processes promoting genetic homogeneity at large spatial scales (Planes et al. 1996, Johnson et al. 2001). Within Barkley Sound there are two large archipelagos. The Deer Group Islands and the Broken Group Islands archipelagos both contain eelgrass beds with abundant pipefish populations. Both archipelagos were formed when glacial melt water submerged former high-elevation ridges leaving island chains surrounded by deep-water channels (depths ranging from 100 - 200 metres deep). Several attributes are different between these two archipelagos. Unlike the Broken Group Islands, the Deer Group Islands exist in a narrow chain of predominately single, large islands separated by deep-water contours and surrounded mainly by two deep-water channels yet in close proximity to coastal eelgrass habitats to the north and to the east. Eelgrass beds exist in isolated bays and a few shallow passages between nearby 32 islands and are exposed to seasonal southeasterly and northwesterly prevailing winds. The physical setting of the Broken Group Islands is different. The Broken Group Islands is a complex system of numerous islands, islets and shallow passages surrounded by two deep-water channels, and many eelgrass habitats are sheltered in basins and protected from prevailing winds. Narrow, deeper water channels separate the outer island group (Pacific Ocean side) from the inner island groups (Vancouver Island side) creating a complex maze of eelgrass, sand, and reef habitats. Where discontinuities in habitat have been identified, they often appear to contribute to genetic divergence (Riginos and Nachman 2001) and separation of individuals by water barriers is another feature of seascapes. Eelgrass beds are naturally patchy in distribution both along mainland and island coastlines. Inherent in an archipelago is the feature of tracts of water barriers separating individual island localities as well as isolating the archipelago from mainland coastal populations. Eelgrass growth is limited due to reduced light availability. at water depths greater than 20 metres (Dennison W C 1997; Hemminga and Duarte 2000) resulting in discontinuous habitat for the eastern Pacific bay pipefish. In Barkley Sound, Trevor Channel, with depths ranging from 150 - 200 metres and approximately two - four kilometres wide (Stronach et al. 1993), separates the Deer Group Island archipelago from coastal sites. Imperial Eagle Channel includes areas with depths of up to 100 metres and is a nine kilometre wide deep-water channel separating southern and northern Barkley Sound. Water channel widths between island localities are variable. Deep-water tracts act as barriers to gene flow and are correlated with genetic differentiation among populations. This has been demonstrated in marine invertebrates (Mariani et al. 2002), fishes (Maltagliati 1999; Congiu et al. 2002), and bats, Pteropodidae (Roberts 2006). 33 The sand smelt, Atherina boyeri is an anadromous coastal species found in the Mediterranean Sea (Congiu et al. 2002). This fish has a strong natal homing behaviour to brackish lagoons for spawning and rearing of juveniles (Congiu et al. 2000. These lagoons are discontinuous along coastlines of mainland and island sites. In populations separated by wide, open sea tracts, genetic distances were higher than among discontinuous habitats situated along mainland coasts (Congiu et al. 2002). Similar results were found for the Mediterranean brackish water killifish, Aphanius fasciatus (Maltagliati 1999) and the benthic dwelling cockle, Cerastoderm glaucum (Mariani et al. 2002) sampled over a similar spatial area and localities as A. boyeri. A deep-water barrier was also invoked as being responsible for greater genetic divergence among populations of plaice, Pleuronectes platessa, on either side of the barrier relative to well connected populations along a continental shelf (Hoarau et al. 2002). At finer spatial scales, Riginos and Nachman (2001) reported greater genetic distance between populations of the blennioid, Axoclinus nigricaudus, separated by the open waters of the southern region of the Gulf of Califorina than those separated by discontinuous coastal habitats. Other studies have demonstrated restricted gene flow across water barriers in coastal marine species (Bell et al 1982; Stepien & Rosenblatt 1991; Doherty et al 1995). 2.1.3 General hypotheses and goals of research Due to the eastern Pacific bay pipefish's high degree of specialization to eelgrass habitats, and the numerous seascape features that may restrict gene flow, significant population genetic structuring is likely at small geographic scales and should be correlated with seascape features. I tested the hypotheses that: (i) population genetic substructure deviates from panmixia, (ii) pipefish populations are largely self-recruiting, and (iii) that seascape features 34 are important in structuring gene flow patterns, genetic differentiation and genetic diversity among geographic regions and subpopulations. I predicted that the morphological specialization of Syngnathus leptorhynchus and the natural patchiness and heterogeneity of eelgrass beds would promote the existence of distinct genetic subpopulations of pipefish. I also predicted that these populations would be separated by a combination of deep-water channels and archipelagos that act as barriers to gene flow and enhance population differentiation and self-recruitment (Johnson etal. 1994, 2001). I further predicted that both pipefish behaviour and certain seascape features w i l l promote some genetic connectivity. Coastal stepping-stone routes for migration, potential for rafting using local currents (Highsmith 1985) and the need to be "good dispersers" (Harrison and Hastings 1996), may promote moderate to high connectivity among subpopulations and geographic regions. Finally, I predicted that local currents would bias migration to favour exchange between South and North Barkley Sound while decreasing the possibility of exchange with East Barkley Sound. 2.2 Materials and Methods 2.2.1 Site Selection Individual eelgrass beds were chosen after an extensive survey of available eelgrass habitat. The survey consisted of using the British Columbia Provincial Coastal Resource Atlas (Booth et al. 2000) and boat surveys. Candidate sites were assessed for certain parameters such as distance between sites (using nautical chart measurements), location with respect to barriers (deep-water channels), protected and semi-protected locations, sites along shore-lines and within archipelagos, and accessibility (water and road). Seventeen sites were sampled in Barkley Sound (Figure 2.1; Table 2.1). 35 2.2.2 Eelgrass bed area and distance calculations The area of eelgrass beds was determined by walking and boating the perimeter of the bed using a GPS (Global Positioning System). Latitudinal and longitudinal coordinates were imported into ArcView GIS (ESRI 2005) software to calculate eelgrass bed area size. Geographic distance between eelgrass patches was calculated using Nobeltec 7 (Garmin 2002) navigational software. 36 49" 20" 40' 10' -128' 55* -1g7" 3Q -126 05 • 124' 40" -123' 15" TOP 51 00" r • m / - Southern British Columbia 4»* 35" 51" 00' 49" 35' Bark ley Sounc >" 20' -125" 10" -12 iO' -124"" 40* 49' 20' North Broken Gioup Islands I A l W r c i i Coital T 49* 10' 49* 00" Ueluetet \3 Junction Past age J ' ' ' r ff / D ' * » G * 0 U P Island* 48" 50' • X? 1 Btuafteld 49* 00' 4B* 50' 48" 40' 48* 40' Pacific Ocean South Scate=lS M km •125*30' -125*20' -125* 10" -125* 00" -124*50" -124*40' A . Figure 2.1: A . Place names in Barkley Sound, southwest coast of Vancouver Island, British Columbia. 37 •125* 30* -125* 20 ' - 1 2 5 * 10" - 1 2 5 * 0 0 ' - 1 2 4 * 5 0 ' - 1 2 4 * 4 0 ' 49° 20" 4 9 s 10' 12 ToquartBay 13 Stopper Isis 14 Pinkertons 15 Jaques/Jarvis Isis 16 Gibraltar Isi 17 Turrett Isi 14 4 9 ' 0 0 ' 48° 5 0 ' 1 ) C h i n a 2Nahmint 1/ Creek y j 3 Snug Basin } j«F p ^ A s s i t s gJ$->J y& 5NumukamisBay 8 ^ ° ^ £ f j Grappler Inlet 8 Dodger Channel 7Jamfield Inlet ^*V/ 9 Fleming Isi 49 " 20" 4 9 ' 10 ' 49* 00 ' 10 Port Alberni Yacht Club I I 4 8 ' 50 ' 4 8 " 4 0 ' 48" 4 0 ' - 1 2 5 * 3 0 - 1 2 5 * 2 0 * - 1 2 5 * 10' - 1 2 5 * 0 0 ' - 1 2 4 * 5 0 ' - 1 2 4 * 4 0 ' B . Figure 2.1: B. Study sites located in Barkley Sound, southwest coast of Vancouver Island, British Columbia. 38 Table 2.1. Sample locations, collection date, site codes and number of fish sampled. Number Site Name Collection Date Site Code Alphabetical Site Code Numerical of fish collected Latitude Longitude Barkley Sound China Creek July 28, 2001 CC 1 37. 49.15 124.80 Nahmint Bay July 26, 2001 NB 2 25 49.06 124.87 Snug Basin July 28, 2001 SB 3 37 49.03 125.03 Assits July 25 & Aug. 24, 2001 Al 4 17 48.94 125.02 Numukamis Bay August 27, 2001 NUM 5 39 48.92 125.01 Grappler Inlet June 12 and 17, 2001 Gl 6 30 48.84 125.11 Bamfield Inlet June 12 and 17, 2001 BI 7 30 48.82 125.14 Dodger Channel August 10, 20, 25, 2001 DC 8 49 48.84 125.19 Fleming Island August 8 and 9, 2001 FL 9 64 48.88 125.15 Port Alberni Yacht Club August 8 and 9, 2001 PC 10 39 48.89 125.12 Useless Inlet August 10, 2001 Ul 11 42 48.99 125.03 Toquart Bay . August 3 and 4, 2001 TB 12 50 49.03 125.36 Stopper Islands August 16, 2001 SI 13 43 49.00 125.34 Pinkertons July 14, 2001 PI 14 34 48.97 125.29 Jaques/Jarvis Islands August 3, 2001 JJ 15 61 48.92 125.28 Gibralter Island August 30, 2001 GIB 16 78 48.9 125.25 Turrett Island August 15, 2001 TI 17 38 48.9 125.34 2.2.3 Sample collection and DNA extraction Before sampling individuals, the abundance of pipefish was first assessed to ensure that beds could sustain the harvest of 30-50 animals per site. Pipefish were obtained from each site by a variety of seining techniques. A 1.6 m x 0.82 m pole seine net with a 4 mm mesh size was used at shallow sites where two people could walk through the eelgrass bed while seining together. A 11.12 m x 3.32 m net with a variable mesh size ranging from 4 mm to 6 mm was designed for deeper sites by attaching it to the stern or the bow of a 14 or 16 foot aluminum boat with 30 H P outboard engine. While driving with the net deployed, the net was dragged through the eelgrass bed for three to five minutes. Eelgrass plants were not damaged by this technique, and it allowed eelgrass beds to be accessed at tide heights higher than those that 39 / could be accessed by beach seining. This large net was also used for beach seining by two-three persons where the net was walked out into the deeper end of the bed, encircled and moved through an area of eelgrass. Pipefishes were collected whole or by taking a small area of the dorsal fin, after being anesthetized using MS-222. A l l whole animals and tissue samples were stored in 95% ethanol and maintained at 4°C. A l l samples were obtained under permits from Parks Canada, the Bamfield Marine Sciences Centre and the Canadian Department of Fisheries and Oceans (de Graaf 2001-1). Genomic D N A was obtained from fin tissues by using the Puregene K i t (Gentra Systems). For whole animals, the left pectoral fin was used for all fishes with a small portion of the dorsal fin used i f more tissue was needed. Liver tissue was tried for D N A extraction, but due to the high quantity of impurities, was found to be unreliable. D N A was suspended in 50 -95 ul of T E solution ( l O m M Tris, I m M E D T A in H29; p H 8.0) and stored at -20°C. Dilution plates were made by diluting the stock D N A with autoclaved distilled water to approximately 100 ng/ul. 2.2.4 Isolation and cloning of microsatellite loci D N A was isolated from liver tissue of two pipefish and mixed for digestion. The protocol for isolation and cloning was that found in Glenn (1995) with modifications as noted. Approximately 50 ug of whole genomic D N A was digested with a cocktail of restriction enzymes Alul, Haelll, Hindi, and Rsal. Twenty ug of Bluescript II plasmid vector (Stratagene) was restricted with Smal enzyme and dephosphorylated with alkaline phosphatase (New England Biolabs). A n aliquot of digested plasmid vector and digested pipefish D N A was run side by side on 1.5% low-melting point agarose (Ultra Pure Gibco B R L ) gel in a I X T A E 40 buffer. Restriction fragments between 100 and 800 base pairs in size were extracted from the gel and purified using "beta" Agarose I (New England Biolabs) digestion. Fragments were ligated into the Bluescript plasmid vector. Ligants were transformed into E. coli supercompetent cells (Stratagene) following procedures outlined by the manufacturer and Glenn (1995). The cells were incubated at 37°C degrees for four hours in a shaker. Aliquots of E. coli cells were plated on to L B agar plates containing the antibiotic ampicillin and X-gal and incubated at 37°C for 24 hours to allow for screening of positive transformants. Bluescript vector contains an ampicillin resistance gene whereas wi ld type E. coli cells do not contain this gene, and only cells containing the Bluescript vector wi l l survive. Positive transformants, those containing Bluescript plasmid/pipefish D N A restriction fragments, can be detected by the presence of white colonies whereas blue colonies indicate the presence of the Bluescript plasmid transformant only. White colonies were picked from the agar plate using sterile pipette tips and sterile toothpicks and transferred to Petri dishes containing L B agar and incubated overnight at 37°C. The white colonies were then transferred to Hybond nylon hybridization membranes (Amersham) and agar plates placed back in 37°C for incubation. Membranes were probed for microsatellite repeat sequences using a series of 3 2 P end-labeled synthetic oligonucleotides ( G T ) 1 5 and ( C T ) i 0 probes in Westneat buffer (7% SDS, 1 m M E D T A , 0.263 M Na2P04, 1% B S A , at 55 °C)-followed by stringency washes to remove excess radio-labeled probe (2XSSC/0 .1% SDS; 5 minutes at 20°C, 20 minutes at 55°C). The membranes were exposed to autoradiography f i lm (Kodak B ioMax) for 48 hours. Positive colony screen positives were picked with sterile pipette tips and stored in 25 uL of sterile distilled water in 0.2 m L P C R tubes. Ten positive colonies were screened for microsatellite repeats by sequencing. 41 M l 3 primers complementary to the plasmid flanking regions were used to P C R amplify the plasmid insert D N A . P C R reactions were performed in 40 ul volumes containing: 3 ul of the D N A solution picked from the positive colonies, 800 n M dNTPs, l x P C R reaction buffer (Invitrogen), 0.5 U of Taq D N A polymerase (Invitrogen), 3.0 m M MgC12, and 5 p M each of the M l 3 forward (5' T G T A A A A C G A C G G C C A G T 3 ' ) and reverse (5' C A G G A A A C A G C T A T G A C C 3 ' ) primers. P C R reaction conditions were as follows: 1 cycle of 120 second denaturation at 95°C, 60 second annealing at 52°C and 90 second extension at 72°C; 4 cycles of 60 second denaturation at 94°C, 60 second annealing at 52°C and 90 second extension at 72°C; 25 cycles of 60 second denaturation at 94°C, 60 second annealing at 50°C and 90 second extension at 72°C; and 5 minute extension at 72°C. After successful P C R of a candidate clone, reactions were increased to a total volume of 50 uL for sequencing. Prior to sequencing, P C R products were purified using Quiagen "Quiaquick" P C R purification columns and eluted in 50 uL of autoclaved, distilled water. Sequencing reactions were performed in 20 uL volumes containing Taq termination premix (8 ul containing Taq D N A polymerase, reaction buffer and ddNTPs) obtained from the U B C N A P S unit), 1 uL 0.5% D M S O , 3.2 p M M 1 3 forward primer, approximately 100 ng purified D N A and sterile distilled water to make up the total volume. Sequencing P C R conditions were as follows: 1 cycle at 95°C denaturation for 4 minutes, 25 cycles at 95°C for 30 seconds, 55°C annealing for 15 seconds, and 72°C elongation for 45 seconds, followed by 1 cycle at 60°C extension for 4 minutes. Excess primers and ddNTPS were removed from the P C R product with Centri-Sep (Applied Biosystems) columns. Sequencing of the product was then performed by the U B C N A P S Unit. For some of the colonies, reverse sequences were also obtained using the reverse M l 3 primer. 42 O f the 10 putative microsatellite sequences, two contained microsatellite sequences. Primer sets were developed for the flanking regions of the microsatellite sequences using the program Primer 3 (Rozen and Skaletsky 1998) and by eye. The microsatellite loci were termed CSL1 and C S L 9 with the designation " C " for clone, and " S L " taken from the genus name and species name. The primer sets for each locus were then put through a series of "cold" P C R (non-radioactively-labeled primers) optimization reactions to determine optimal annealing temperatures, annealing and elongation times, MgC12 concentrations, and the use of any P C R adjuncts ( D M S O or K C L ) . P C R products were run on 1.5% or 2% agarose gels in 0.5X T B E buffer to judge P C R conditions. Forward and reverse primers for each locus were then endlabeled with 3 2 P in a "hot" P C R reaction with the optimal P C R conditions and resolved on 6% polyacrylamide gels and autoradiography film. C S L 1 was difficult to resolve and primers were not redesigned largely due to the inability to obtain a readable sequence to allow for enough base pairs in the flanking region for optimal primer design. C S L 9 was polymorphic and worked consistently (Table 2.2). Figure 2.2 illustrates an autoradiograph of locus T16. 43 Figure 2.2: Autoradiograph of P labeled microsatellite locus T16. Each lane is an individual fish. Fish identifier number represents the position of an individual fish on the gel. Standard reference fish represents individuals run on all gels for individual loci to ensure consistency in scoring. Base pair sizes are presented for a few alleles. 44 Table 2.2: Primer sequences, annealing temperatures and M g C l 2 concentrations. * indicates the primer labeled with P or fluorescent label. Microsatellite Cloned repeat Primer Sequence Annealing [MgCL2] Citation Loci Temperature (mM) (°C) T16 G A T G 1 5 F CAG GAC ACG CTG GAA AGA C 60 1.5 Jones AG etal. (1999) *R GCA ACA CCT TGA AGA GGA AAG T T12 (CA) 1 2CT(CA) 5 • F GCG TCC CAT TCA CTG ACT TGA TTG 56 1.5 Jones AG etal. (1999) *R CCC CAT GCT TCA GGC TTT CAC TAT CSL9 (GT)9GC(GT) *F ATG CCA AAT GAA CAC CTT CC R GCT CTC TGG TTC CTG TGG 55/53 1.5 Taylor, unpublished Slep9 (GT) 4 0 *F AAG TGA GTC ATT TGC TGC TAT GG R CGA CAG ACA GGT CAA GAT TTG G 50 2.5 Wilson AB (2006) Slep3 (TG) 4 2 *F AAG GAT GCA TTG CTT CAT GC R AGT CAT TAC CTG GCC CAT TG 50 2.5 Wilson AB (2006) 3 2 P end-labeling was carried out in one microlitre reactions (per sample) containing: I X reaction buffer (New England Biolabs), 0.25 U Polyonucleotide kinase (New England Biolabs), 0.5. u M primer, 0.024 u L 3 2 P and 0.802 u L sterile distilled water. The reaction was incubated at 37°C for one hour and then at 65°C for five minutes to denature the enzyme. A l l reactions were carried out in a M J Research, Ltd. P T C 100 thermocycler P C R machine. 2.2.5 Screening of existing microsatellite primers Other polymorphic microsatellite loci for the eastern Pacific bay pipefish were investigated by screening published.loci from other syngnathid species (seahorses and pipefish) as well as from the threespined stickleback, a distant relative to this group of fishes, using empirical P C R methods. In total 21 primer sets were screened. L o c i screened from the seahorse Hippocampus angustus were Han03, Han 05, Han06, and H a n l 5 (Jones et al. 1998). L o c i screened from the threespined stickleback, Gaterosteus aculeatus, were Gac4, Gac7, Gac9, G a c l 4 (Taylor 1998), and CICR5152 (Rico et al. 1993). While some success was 45 obtained with loci from G. aculeatus (Gac 4, Gac9, Gac l4 , CICR5152) these proved to be unreliable using 3 2 P endlabeled primers. These loci, however, may prove reliable with fluorescently labeled primers. Microsatellite loci from two syngnathid pipefish were screened using empirical P C R techniques. Syngnathus typhle is endemic to Sweden and Syngnathus scovelli is endemic to the Gulf of Mexico, and microsatellites were developed for these two species by Jones et al.. (1999) and Jones and Avise (1997) and cross amplified in the Florida endemic pipefish Syngnathus floridae (Jones and Avise 1997). Microsatellite primers for loci T04, T18, T12, T16, T33, and T44 from Syngnathus typhle, and u25.10, u25.22, u l l . l , and u22.3 loci from Syngnathus scovelli were screened. Three loci optimized for "hot" P C R (loci endlabeled with 3 2 P ) were T12, T16, and u25.10. Only microsatellite markers T12 and T16 from S. typhle proved to be reliable for scoring using 3 2 P labeled primers (Figure 2.2). Locus u25.10 was highly resolvable but the high degree of stutter in certain alleles made it too difficult to accurately score using 3 2 P but may be suitable for use with fluorescently-labeled primers. Two primers developed for Syngnathus leptorhynchus by A . B . Wilson (University of Konstanz, 78457 Konstanz, Germany) were provided to our lab and are described as Slep 3 and Slep 9. T12, T16 and C S L 9 were resolved using 3 2 P end-labeled primers and polyacrylamide gels. Slep3 and Slep 9 were resolved using fluorescently labeled primers and the Beckman-Coulter C E Q 8000 autoanalysis system. i 2.2.6 PCR Amplification, electrophoresis and scoring of variation P C R reactions for C S L 9 , T12, and T16 loci were performed in 10 ul volumes containing 50- 100 ng of D N A template and using 96-well plates. For loci C L S 9 , T12 and T16, all reaction conditions contained I X P C R buffer (Invitrogen), 0.5 U Taq D N A polymerase (Invitrogen), 2.5 m M M g C l 2 , 200 u M of each d N T P (dATP, dGTP, dCTP, and 46 dTTP), 6 p M unlabeled primer, 0.5 p M 3 2 P end-labeled primer and 2.5 p M unlabeled primer that was used for end-labeling. For the C S L 9 locus, the forward primer was end-labeled and for the T12 and T16 loci the reverse primers were end-labeled. The P C R program used for locus T16 was 1 cycle 94°C/3min, 5 cycles 92°C/ lmin , 60 °C/ lmin , 72°C/ lmin , 30 cycles 92°C/30secs, 60°C /30secs, 72°C/30secs, and 1 cycle 72°C/15min. C S L 9 was 1 cycle 95°C/3min, 5 cycles 92°C/ lmin , 55°C / l m i n , 72°C/ lmin , 30 cycles 92°C/30 sec, 53°C /30secs, 72°C/30secs, and 1 cycle 72°C/8min. The P C R program used for locus T12 was 1 cycle 94°C/3min, 5 cycles 92°C/ lmin , 56°C/lmin, 72°C/ lmin, 24 cycles 92°C/30 sees, 56°C /30 sees, 72°C/45 sees, and 1 cycle 72°C/15min. Following P C R , 5-7 ul of formamide based loading buffer (US Biochemicals sequencing stop buffer) was added to each completed P C R reaction. Some products were placed in a -20°C freezer until loading on gels. Prior to loading, samples were denatured at 95°C for 15-30 minutes and kept on ice until loaded; loading times varied between 3 - 5 minutes/gel. Between 5 - 7 ul of P C R product was loaded and run on a 6% Long Ranger (Mandel Scientific) polyacrylamide gels along with an M 1 3 ladder and standardized reference individuals for each locus. Gels were electrophoresed vertically at 55-60 watts. For C S L 9 and T12 gels were run for 2-2.5 hours and for T16 for 4.5 - 5 hrs. The polyacrylamide gel was transferred to 3 M M Whatman filter paper, dried and then exposed to Kodak autoradiography film (Kodak Biorad) and exposed for 24-48 hours for C S L 9 , 72-96 hours for T16 and 2.5 - 3 weeks for T12. Initial gels were run with the M l 3 reference ladder and alleles were scored and named by base-pair size. Approximately 6 - 1 0 individual fish were then used as allele size references and used to score all subsequent gels. These animals were chosen for allelic patterns showing strong signal strength as well as a wide range of possible allele sizes and consistent stutter patterns. In addition to using standard reference fish to ensure consistency in 47 scoring, several individuals would be rerun on subsequent gels run on the same day. As well, over the course of one day, PCRs and a series of gels were run with reference individuals and 8-16 randomly selected animals from each locality to retest PCR efficiency and scoring consistency. An allele library for each locus was compiled for genotyping. Alleles for loci Slep9 and Slep3 were resolved using the CEQ 8000 System. PCR reactions were cared out using 96-well plates and included 2-4 standard reference animals. PCR programs for the locus Slep3 was 1 cycle 95°C/3min, 5 cycles 94°C/30secs, 57°C/30secs, 72°C/lmin, 30 cycles(92°C/30secs, 55°C/30secs, 72°C/70secs, and 1 cycle 72°C/10min (see Table 2.2). PCR program for locus Slep9 was 1 cycle 94°C/3min, 5 cycles 94°C/30secs, 60°C/30secs, 72°C/lmin, 30 cycles (92°C/30secs, 58°C/30secs, 72°C/lmin), and 1 cycle 72°C/7min. PCR products were diluted to l/20 t h - 1/40* concentration and added to 40 uL of SLS solution and 0.2 uL of reference ladder for running on the CEQ analyzer. Some plates were successfully run by mixing individual PCR products of each fish for the two Slep loci. Weak signals with the Slep3 locus were common and these animals were rerun separately. Alleles were scored according to peak height and consistency of the PCR reaction conditions, dyes and the CEQ analyzer was checked using the reference "positive" animals. An allele library was compiled and used to score all animals. Common difficulties with scoring included strong stutter patterns separated by two base pairs and heterozygotes differing by only two base pairs. A consistent scoring system was applied across all samples. Slep9 proved to be a reliable marker with few difficulties with weak signal strengths. Slep3 often proved to be a difficult marker to use. Commonly, animals were scored with a missing allele due to weak and unresolved peaks even after re-extraction of DNA and multiple runs. Alleles that could not be resolved at a set threshold peak height were 48 scored as nulls even i f a weak signal was seen resulting in numerous heterozygotes being scored as "000, Al le le" . It is likely that this wi l l result in higher homozygosities than expected for Slep3. 2.2.7 Genetic data analysis Genetic polymorphism can be defined as the occurrence in any one population of two or more alleles each at substantial frequencies (Hedrick 2000) and is calculated as the percentage of loci that are not fixed at a single allele set at the criteria of 0.95 and 0.99. Microsatellite polymorphism was calculated using the software program Tools for Population Genetic Analysis ( T F P G A 1.3, Mi l l e r 1997). The G E N E P O P (3.4) program was used to calculate summary statistics, allele (genie) and genotypic frequencies (Raymond and Rousset 1995). The total number of alleles for each locus, averages over all loci , and allelic diversity (the total number of alleles detected over combined loci) were calculated for each locality. Tests for differences among localities and groups were performed using a non-parametric Wilcoxon/Kruskal-Wallis test and J M P - I N version 4.0.3. The proportion of available alleles is the number of alleles found at the locality divided by the number of alleles found at that locus. Percentages of private and rare alleles were calculated for each locality. A private allele is one that occurs in only one locality. A "rare allele" was described as one occurring in two to six localities. A "common" allele appears in more than seven localities. Geographic regions were defined by seascape features (such as deep-water channels and fiords) with localities located in the Alberni Inlet area (China Creek, Nahmint Bay and Snug Basin) defined as East Barkley Sound; localities in southern Barkley Sound (Assits, Numukamis Bay, Grappler Inlet, Bamfield Inlet, Dodger Channel, Fleming Island, Port Alberni Yacht Club, and Useless Inlet) defined as South Barkley Sound; and localities in 49 northern Barkley Sound (Toquart Bay, Stoppers Islands, Pinkertons, Jaques/Jarvis Islands, Gibralter Island, and Turrett Island) defined as North Barkley Sound. These geographic regions are similar to those referred to in Stronach et al. (1993). Tests of linkage disequilibrium are conducted to investigate whether allelic variation at each locus could be treated as independent measures of genetic diversity. Linkage disequilibrium was tested for all combinations of locus pairs and locus pairs within localities using a Markov chain method in G E N E P O P (3.4). Basic descriptive statistics of microsatellite variation including number of alleles (N a ) , expected heterozygosity (H e ) , observed heterozygosity (H 0 ) were calculated using G E N E P O P . . G E N E P O P was also used to test for a heterozygote deficit and deviations from Hardy-Weinberg equilibrium expectations for each locus-locality using the exact test method of Raymond and Rousset's (1995) at 10,000 iterations. Frs values calculated followed that of Weir and Cockerham (1984). The program F S T A T (2.9.3.2) was used to calculate allelic richness, A R , (corrected for sample size using the rarefaction method) statistics and to calculate differences among groups using a randomized permutation approach with 10,000 permutations (Goudet 2001). Gene diversity, multi-locus expected heterozyogosity, is a common measure of genetic variation providing a useful statistic to compare genetic variation among localities (Ffedrick 2000). Differences between observed and expected multilocus heterozygosities were tested by first transforming values (using an arcsine, square root transform) to conform to a chi-square distribution and then using a goodness of fit test. Gene diversities were calculated as multilocus expected heterozygosities for each locality. Gene diversities, FJS, and tests of differences among groups were calculated using G E N E P O P and F S T A T . 50 Exact tests of population differentiation test the hypothesis of allele frequency differences among localities using statistics based on Fisher's Exact procedures (Raymond and Rousset 1995). Tests assuming random mating are made on allele frequencies differences; and tests assuming non-random mating are made on genotype frequencies differences. Tests for differentiation of allele frequencies using a log-likelihood (G) based exact test for single locus and combined loci over pooled localities, were carried out in F S T A T with 10,000 permutations. To test allele frequency differences of each locus between pairwise localities, a log-likelihood (G) based exact test was carried out in G E N E P O P at an alpha level of 0.05 and over combined loci using T F P G A . Significance for all tests was measured using the sequential Bonferroni procedure of Rice (1989). G E N E P O P was also used to calculate Weir and Cockerham's (1984) estimators of F-statistics. These statistics divide total heterozygote deficiency as: compared to all populations under panmixia (F), into components due to deficiencies within populations (f) and subdivision among populations (0). Variances and robustness of F-statistics estimates were assessed by resampling procedures using jack-knife and bootstrap methods over loci to generate P values and 95% confidence intervals using F S T A T . Significance for all tests was measured using the sequential Bonferroni procedure of Rice (1989). F-statistics follow the infinite-allele model of microsatellite evolution and identity versus non-identity of alleles (Hedrick 2000). The program A R L E Q U I N 2.0 was used to test how genetic variance was partitioned among groups using the Analysis of Molecular Variance ( A M O V A ) option of Excoffier et al. (1992) (Schneider et al. 2000). A M O V A , analogous to an A N O V A procedure, partitions the percentage of genetic variance (based on allele frequency differences) among groups, among populations within groups, and within populations. F-statistics based estimators were chosen for these calculations as they are noted to perform better than R-statistics (i.e., those 51 incorporating differences in allele sizes as well as frequencies) when overall genetic differentiation is low as commonly found in.recently diverged populations (Balloux and Lugon-Moulin 2002). P C A - G E N 1.2 (Goudet 1999) is a program for conducting principal component analysis that clusters sampled localities by new components based on the level of genetic variance among these sites using allele frequency data. Graphical plots allow a method to resolve similarities among sampled localities. The significance of each new component was evaluated by permutation tests set at 15,000 runs. P C A - G E N is available at http://www.unil.ch/izea/softwares/pcagen.html. The program B A R R I E R 2.2 (Manni et al. 2004) detects regions of genetic discontinuity (restricted gene flow) due to putative barriers using matrices of pairwise genetic distance between localities and spatial coordinates of each locality. Localities on either side of a barrier are more similar than localities on different sides of the boundary. Spatial coordinates are used to generate a network by Vironoi tessellation of the geometric space and Delauney triangulation which connects geographically neighbouring localities by triangulation. The map was then edited to join geographically neighbouring localities that were missed by the Delauney triangulation. Spatial areas of discontinuities of gene flow are estimated by using Monmonier's maximum distance algorithm and the genetic distance matrix. In my study, pairwise Fgx values of combined loci were used. B A R R I E R prompts the user to select a number of putative barriers. The algorithm draws a line along the Vironoi tessellation network until the line encloses the localities or hits another barrier. The most significant barrier is indicated in ascending numerical order with the first barrier being the most important one. If other distance matrices are available (genetic or environmental), they can be added and a 52 bootstrap like procedure is implemented where consensus barriers among matrices can be assessed (Manni et al. 2004). A Bayesian model-based approach was used to estimate the number of likely contemporary subpopulations represented in the sites sampled using the software program S T R U C T U R E 2.0 (Pritchard et al. 2000). S T R U C T U R E uses a Markov chain Monte Carlo ( M C M C ) sample method to infer the number of subpopulations (K) that best describes a data set and is based on known genotypes of individuals. Groups are clustered based on similarities of genotype makeup of groups. The number of clusters wi l l reflect the level of intrapopulation differentiation based on these genotypes. A l l models were run with default parameters in S T R U C T U R E . The first ancestry model chosen was that of admixture of individuals largely due to the results from tests for population differentiation and the small spatial scale of this study. This model infers mixed ancestry of individuals from a single, mixed population. As well as being the recommended model for this program, for the eastern Pacific bay pipefish this seems to be a more conservative model relative to that of pure ancestry or discrete populations defined by the no admixture model which infers substructure in the data (Pritchard et al. 2000). Correlated allele frequencies rather than independent allele frequencies among localities option was selected as a parameter based on previous tests of population differentiation and the small spatial scale of this study. The independent allele option is used when allele frequencies are likely to be dissimilar and merges subpopulations with similar allele frequencies (Pritchard et al. 2000). The correlated alleles option is used when there is reason to believe that some of the putative subpopulations are likely to share allele frequencies (Pritchard et al. 2000). The correlated allele frequency option allows the S T R U C T U R E program to detect assignments even if populations are closely related (Pritchard et al. 2000). 53 Model A was run using a pre-run or "burnin" of 50,000, and 100,000 subsequent M C M C runs. Each such simulation was repeated five times. These parameters were adjusted to minimize the range of alpha values among runs. Alpha values appeared to level out with these higher run parameters. The ancestry model of "prior population information", was performed with individuals coded to their locality of sampling, adding an element of discreteness in contrast to the admixture model. Correlated allele frequencies option was also chosen for this model. Model B (i) had a burnin of 50,000; M C M C 100,000 and 5 runs. Model B (ii) had a burnin of 100,000; M C M C 100,000 and 10 runs. A s differentiated localities can bias K values, analyses were also run excluding the locality Gibralter Island (Pritchard et al. 2000). Central to the inference of population structure and assignment of individuals to clusters in S T R U C T U R E is "g" the proportion of an individual's genome that originates from any number of A'populations (Pritchard et al. 2000). The proportion of an individual's genome originating from a number of K is modeled by the Dirichlet distribution (Pritchard et al. 2000). The intricacies of the M C M C aside, alpha is important. Alpha values should stabilize among runs as the Markov chain converges (Pritchard et al. 2000). Failure to stabilize between runs likely means that there is no detectable population substructure or that the number of burnin replicates and M C M C runs is inadequate (Pritchard et al. 2000). Variability in alpha has serious implications for "q" and the assignment of individuals to K number of clusters (Pritchard et al. 2000). In short, for an alpha > 1, a model of admixture, few to no subpopulations, is most likely as the genome of the individual l ikely originates from all K populations (Pritchard et al. 2000). For alpha <1 or approaching 0, the model of no admixture is most probable and individuals likely originate from a number of discrete populations (Pritchard et al. 2000). 54 When evaluating results on K, (-) log likelihood values (posterior probability of "K") are important but can be misleading i f attention is not paid to the behaviour of alpha and assignment probabilities. Alpha values should not range more than 0.2 among runs (Pritchard et al. 2000). Assignment probabilities should be high to support a valid finding of population substructure, K>1 (Pritchard et al. 2000). In this study, with 17 sites, a random assignment probability is 1/17 or 0.0588, a finding of no substructure. Alpha values and assignment probabilities, as well as (-) log likelihood probabilities were monitored when assessing the number of K subpopulations chosen by the different ancestry models. A log likelihood ratio test was used to test for model differences when values for K were similar (Taylor et al. 2005). The test statistic is 2 ( l n L l - lnL2), where I n L l is the natural logarithm of the model with the lowest - l og likelihood score (the simplest model and lower K) and lnL2 the more general model with a higher - l o g likelihood score (and higher K) (Taylor et al. 2005). This test follows a chi-square distribution and the degrees of freedom are equal to the number of additional parameters (populations) of the more complex model. In the case of two populations, the degree of freedom would be one. Assignment tests were conducted in the program G E N E C L A S S 2 (Piry et al. 2004; version 2.0). The Bayesian method (Rannala and Mountain 1997) was used to test for assignment probabilities of individuals "belonging" to sample localities. This test relies on setting threshold values for the distribution of an individual's assignment probability (likelihood) compared to a probability obtained by sampling 100,000 simulated individuals from each potential source locality. Individuals were assigned to localities with the highest assignment probability and to which it met or exceeded a threshold of P > 0.05. The exclusion probability, or the probability that an individual is actually a migrant from another locality, is 55 then P < 0.05. Individual fish that were not assigned to the location from where they were sampled were classified as "misassigned". Isolation by distance models test the degree to which geographic distance can explain the genetic differentiation between a pair of study sites. If genetic differentiation is a result of increasing distances among localities, gene flow is said to follow a stepping-stone model where isolation by distance is important in shaping genetic differences among localities. Mantel and i partial Mantel tests were used to test for significance of the correlation between genetic differentiation and geographic distance. If the Mantel test is not significant and there is no correlation between genetic differentiation and geographic distance, it can be inferred that gene flow follows an island-model and that other factors other than distance are important in understanding gene flow among localities. The software program I B D was used with 10,000 permutations of matrices (Bohonak 2002). Partial Mantel tests of the influence of habitat types were performed by coding coastal localities (1) and island localities (2). Mantel and partial Mantel tests were conducted to test the correlation between genetic distance, geographic distance and putative deep-water barriers, and six categories were tested: (1) coast by coast within a region (where there are no deep-water barriers between localities as eelgrass beds occur along a contour line favourable to eelgrass growth); (2) coast by coast among three geographical regions where there are no deep-water barriers between localities); (3) island by island within a region where deep-water barriers range in width from 3.9 - 14.5 km in the Deer Group and 3.8 - 9.5 km in the Broken Group; (4) coast by island within regions where deep-water barriers range in width 8.4 - 21.9 kilometres (Trevor Channel (South Barkley Sound) and Louden Channel (North Barkley Sound) crossings; (5) island by island among regions where deep-water barriers range in width across Imperial Eagle Channel and Imperial Eagle Channel/Louden Channel crossings of 9.8 - 28 kilometres; and (6) coastal by 56 island among regions across Trevor Channel, Imperial Eagle Channel and Louden Channel ranging in combined widths of 39.8 - 76.3 kilometres. 57 2.3 Results 2.3.1 Microsatellite polymorphism At any locus and any locality, no allele was fixed. A l l microsatellite loci used in this study were polymorphic both at the 0.95 and 0.99 criteria. Max imum sizes of loci ranged from 20 to 116 nucleotide base-pairs. Each had a distinct repeat array; and two loci were di-nucleotide, two were di-nucleotide complexes, and one was a tetra-nucleotide repeat (Table 2.3). Table 2.3: Eastern Pacific bay pipefish, Syngnathus leptorhynchus, microsatellite loci . Locus name, repeat array, range of allele size and number of alleles resolved in this study. SIZE LOCUS RANGE NUMBER OF NAME REPEAT ARRAY TYPE OF REPEAT (bp) ALLELES *T16 (GATG) 1 5 tetranucleotide 282- 398 30 CSL9 (GT)9(GC)(GT) dinucleotide complex 100- 120 10 *T12 (CA) 1 2 CT(CA) 5 dinucleotide complex 254 - 324 29 Slep9 (GT) 4 0 dinucleotide 267 - 367 41 Slep3 (TG) 4 2 dinucleotide 111 - 219 46 * Jones et al 1999; Syngnathus typhle (broad-nosed pipefish) 58 2.3.2 Allelic diversity and richness The number of fish assayed for T16 ranged from 16 (Assits) to 79 (Gibralter Island); C S L 9 ranged from 17 (Assits) to 74 (Gibralter Island); T12 ranged from 15 (Assits) to 76 (Gibralter Island); Slep9 ranged from 17 (Assits) to 74 (Jaques/Jarvis Islands); and Slep3 ranged from 15 (Assits) to 51 (Jaques/Jarvis) (Table 2.4). The number of alleles at each locus ranged from 10 (CSL9), 29 (T12), 30 (T16), 41 (Slep9) and 46 (Slep3) with a mean of 31.2 ± 13.9 (STDV) alleles (Table 2.4). Over all loci and all localities, 156 alleles were detected. The highest allelic diversity was found at the Port Alberni Yacht Club site with a total of 114 alleles, average 22.8 ( S T D V 9) and 64 fish sampled (Table 2.4). The lowest diversity was at the Assits site with a total of 69 alleles, average 13.8 ( S T D V 5.1) and only 17 fish sampled. A non-parametric Wilcoxon/Kruskal-Wallis test of mean allele number over all loci revealed a significant difference of allelic diversity among localities (X 2 ( 0 .o5 , i6) 32.65, P = 0.008). Testing for allelic diversity differences among the three geographic regions using nonparametric Wilcoxon-Kruskal-Wallis tests revealed no differences among regions based on geography (excluding Useless Inlet) (X 2 ( 0 . 0 5 ,2) 0.82, P = 0.66). The proportion of alleles, the number of alleles detected (allelic diversity) divided by the total number of alleles, is a useful way to view allelic diversity. The proportion of the total available alleles that were present at any locality for T16 ranged from 0.53 (Assits/Turret Island) to 0.87 (Port Alberni Yacht Club); C S L 9 ranged from 0.5 (Assits/Numukamis Bay) to 1.0 (Snug Basin); T12 ranged from 0.48 (Useless Inlet) to 0.83 (Toquart Bay); Slep9 ranged from 0.39 (Bamfield Inlet) to 0.73 (Grappler Inlet); and for Slep 3 ranged from 0.33 (Assits) to 0.67 (Port Alberni Yacht Club) (Table 2.4). For combined' loci , the proportion of available alleles ranged from 0.44 (Assists) to 0.73 (Port Alberni Yacht Club) and the average was 0.61 (Table 2.4). 59 Private alleles were found in nine localities: China Creek (2), Nahmint Bay (1), Grappler Inlet (1), Dodger Channel (1), Fleming Island (3), Port Alberni Yacht Club (3), Useless Inlet (1), Stopper Islands (1) and Gibralter Island (1) (Table 2.4). Al le l ic richness was calculated over each locus over each locality based on a minimum sample of 13 individuals (Table 2.5). Al le l ic richness for T16 ranged from 11.7 (Turret Island) to 15.2 (Pinkerton Islands); for C S L 9 from 3.9 (Stopper Islands) to 6.6 (Bamfield Inlet); for T12 from 10.8 (Gibralter Island) to 15.4 (Dodger Channel); for Slep9 from 14.2 (Bamfield Inlet) to 16.5 (Fleming Island); and Slep3 from 14.1 (Assits) to 18.1 (China Creek). Combining alleles over loci , allelic richness differences were tested among groups of localities based on geographical criteria (East, South, and North Barkley Sound). Al le l ic richness for each geographic group was 13.43, 13.22, and 12.62 respectively. Significant differences in allelic richness were detected among groups with North Barkley Sound having the lowest allelic richness (P = 0.01) (Table 2.6). 60 Table 2.4: Sample size (N), numbers of observed alleles (N a ) , expected ( H E ) and observed (HQ) heterozygosities per locus and per locality sampled (Weir and Cockerham 1984), Proportion of available alleles (N p ) , number of private alleles ( N p r i v ) , percentage of private alleles (%priv), number of rare alleles (N r ) . Observed heterozygosities are direct counts and expected heterozygosities are corrected, unbiased estimates. Microsatellite Loci Average Combined Combined T16 CSL9 T12 Slep9 Slep3 Loci Loci Statistics over all localities Total Number of alleles 30 10 29 41 46 156 Average number of alleles over all localities (STDV) 20.7 (2.7) 6.9 (1.5) 19.6(2.6) 25.3 (4.2) 24.2 (4.4) 31.2(13.9) H e 0.94 0.74 0.94 0.97 0.97 0.91 H 0 0.96 0.7 0.95 0.86 0.85 0.87 Locality China Creek N 37 36 37 33 35 Na 18 6 20 24 30 98 19.6 (8.9) H e 0.92 0.75' 0.92 0.95 0.97 0.90 H 0 0.86 0.72 0.81 0.84 0.88 0.82 N P 0.6 0.6 0.69 0.59 0.65 0.63 N r 1 1 2 3 Npriv 2 %priv 14.30% F i s 0.088 Nahmint Bay N 22 25 25 25 25 Na 18 8 21 19 24 90 18 (6.04) H e 0.95 0.8 0.95 0.96 0.97 0.92 H 0 1 0.84 0.92 0.84 0.83 0.88 N p 0.6 0.80 0.72 0.46 0.52 0.58 N r 1 2 1 3 Npriv 1 %priv 7.10% F i s 0.044 Snug Basin N 40 43 42 43 42 Na 22 10 20 26 28 106 21.2 (7.0) H e 0.95 0.8 0.94 0.95 0.95 0.92 H 0 0.95 0.88 0.93 0.86 1 0.92 N p 0.73 1.00 0.69 0.63 0.61 0.68 N r 2 2 1 2 Npriv %priv 0% F i s -0.004 61 Table 2.4 continued: Average Combined Combined Locality T16 . CSL2 T12 Slep9 S!ep3 Loci Loci Assists N Na 16 . 16 17 5 15 15 17 18 15 15 69 13.8(5.1) H e 0.95 0.66 0.94 0.95 0.95 0.89 H 0 0.93 0.65 1 0.82 0.8 0.83 N p 0.53 0.50 0.52 0.44 0.33 0.44 N r 1 1 2 Npriv %priv F i s 0.00% 0.059 Numukamis Bay N Na 39 23 36 5 '31 20 37 25 38 28 101 20.2 (9) H e 0.95 0.68 0.94 0.96 0.97 0.9 H 0 0.97 0.66 1 0.87 0.76 0.85 N P 0.77 0.50 0.69 0.61 0.61 • 0.65 N r 1 1 Npriv %priv F i s 0% 0.057 Grappler Inlet N Na 28 19 30 6 26 16 29 30 28 24 95 19(9) H e 0.95 0.78 0.92 0.98 0.95 0.91 H 0 1 0.69 1 0.89 0.89 0.89 N P 0.63 0.60 0.55 0.73 0.52 0.61 N r 4 1 Npriv %priv F i s 1 7.14% 0.028 Bamfield Inlet N Na 25 17 28 8 26 16 20 16 23 23 80 16(5.3) H e 0.94 0.8 0.94 0.95 . 0.97 0.91 H 0 0.96 0.85 1 0.74 0.73 0.86 N p 0.57 0.80 0.55 0.39 0.50 0.51 N r 1 Npriv %priv F i s 0% 0.054 62 Table 2.4 continued: Average Combined Combined Locality T16 CSL2 T12 Slep9 Slep3 Loci Loci Dodger Channel N Na 45 20 45 6 34 21 27 22 16 18 87 17.4 (6.5) H e 0.93 0.63 0.95 0.96 0.97 0.86 H 0 0.91 0.55 1 0.92 0.92 0.82 N p 0.67 0.60 0.72 0.54 0.39 0.56 N r Npriv %priv F i s . 1 7.14% 0.037 Fleming Island N Na 41 21 39 6 23 18 37 27 35 26 98 19.6 (8.4) H e 0.94 0.73 0.93 0.96 0.95 0.9 H 0 1 0.73 0.96 0.86 0.79 0.86 N p 0.70 0.60 0.62 0.66 0.57 0.63 N r 1 2 1 Npriv %priv F i s 1 2 21.40% 0.040 Port Alberni Yacht Club N 64 63 56 62 59 Na 26 9 19 29 31 114 22.8 (9.0) H e 0.95 0.76 0.93 0.96 0.95 0.91 H 0 1 0.7 0.98 0.8 0.84 0.86 N P 0.87 0.90 0.66 0.71 0.67 0.73 N r 2 1 2 1 Npriv %priv F i s 1 2 21.40% 0.049 Useless Inlet N 43 43 23 41 32 Na 21 6 14 23 28 92 18.4 (8.6) H e 0.94 0.68 0.9 0.95 0.97 0.88 H 0 1 0.6 0.95 0.88 0.86. 0.84 N p 0.70 0.60 0.48 0.56 0.61 0.59 N r 1 1. Npriv %priv F i s 1 7.14% 0.04 63 Table 2.4 continued: Average Combined Combined Locality T16 CSL2 T12 Slep9 Slep3 Loci Loci Toquart Bay N N a He 45 23 0.94 44 8 0.74 48 24 0.93 37 27 0.95 28 19 0.95 101 0.89 20.2 (7.4) 0.91 0.68 0.89 0.86 0.88 0.84 Np 0.77 0.80 0.83 0.66 0.41 0.65 Nr Npriv %priv Rs 1 2 1 0% 0.057 Stopper Islands N N a r i 40 21 0.93 40 4 0.69 34 20 0.93 40 25 0.96 41 24 0.95 94 0.89 18.8 (8.5)' Ho 1 0.56 1 0.88 0.9 0.86 Np 0.70 0.40 0.69 0.61 0.52 0.6 Nr Npriv %priv R s 1 3 7.14% 0.034 Pinkertons N N a hi 34 22 0.95 34 6 0.74 . 31 18 0.93 31 22 0.96 22 20 0.96 88 0.9 17.6 (6.7) Ho 0.94 0.74 0.97 0.81 0.84 0.86 Np 0.73 0.60 0.62 0.54 0.43 0.56 N Npriv %priv R s 1 1 0% 0.048 64 Table 2.4 continued: Average Combined Combined Locality T16 CSL2 T12 Slep9 Slep3 Loci Loci Jaques/Jarvis Islands N 58 60 58 74 51 Na 21 7 21 29 26 104 20.8 (8.4) H e 0.95 0.73 0.94 0.96 0.95 0.91 H 0 0.98 0.64 0.95 0.88 0.88 0.86 Np 0.70 0.70 0.72 0.71 0.57 0.67 N ' 1 3 1 Npriv %priv F i s 0% 0.048 Gibralter Island N 79 74 76 54 45 Na 18 7 18 30 26 99 19.8 (8.8) H e 0.93 0.73 0.9 0.95 0.96 0.88 H 0 0.96 0.71 0.91 0.98 0.81 0.87 N P 0.60 0.70 0.62 0.73 0.57 0.63 N r 1 1 3 3 Npriv %priv F i s 1 7.14% 0.012 Turrett Island N 39 34 37 32 28 Na 16 7 22 22 22 89 17.8 (6.6) H e 0.92 0.75 0.94 0.95 0.95 0.9 H 0 1 0.88 0.94 0.81 0.85 0.9 N P 0.53 0.70 0.76 0.54 0.48 0.57 N r 2 Npriv %priv F i s 0% -0.005 65 Table 2.5 Al le l ic richness per locus and locality based on a minimum sample size of 13 diploid individuals. Locality Microsatellite Loci T16 CSL9 T12 Slep9 Slep3 Alberni Inlet China Creek 12.1 5.0 12.9 15.9 18.1 Nahmint Bay 14.9 6.3 15.6 14.9 17.7 Snug Basin 15.2 7.1 14.3 15.4 16.2 Trevor Channel/ Deer Group Islands -Assits 14.8 4.7 14.5 15.5 14.1 Numukarhis Bay 15.0 4.4 14.7 15.9 16.9 Grappler Inlet 14.5 5.4 12.6 19.2 16.1 Bamfield Inlet 13.9 6.6 13.5 14.2 17.1 Dodger Channel 13.6 4.8 15.4 16.0 18.0 Fleming Island 14.5 4.8 13.8 16.5 15.9 Port Alberni Yacht Club 14.8 6.4 12.8 15.9 15.9 Useless Inlet 14.1 4.6 11.8 15.2 17.8 Louden Channel/BGI Toquart Bay 14.0 5.2 13.5 16.0 14.5 Stopper Islands 13.4 3.9 13.9 16.0 15.4 Pinkerton Islands 15.2 4.7 13.3 15.6 16.1 Jacques/ Jarvis Isis 14.8 5.1 13.5 16.3 15.6 Gibralter Island 13.0 4.6 10.8 16.2 16.5 Turret Island 11.7 5.6 14.4 14.5 15.5 Average (AII_W) 14.5 5.3 13.8 16.1 16.8 66 Table 2.6: Al le l ic richness differences among geographic regions. Calculations performed with F S T A T 2.9.3.2 (Goudet 2001) with 10,000 permutations. Two sided P Geographic Region Allel ic R ichness value Locali t ies in each division East Barkley Sound 13.43 C C , N B , S B South Barkley Sound 13.15 A l , N U M , G l , B A M , D C , FL, P A Y C North Barkley Sound 12.62 T B , SI, P I N K S , J J , GIB, TI 0.01 2.3.3 Linkage disequilibrium Genotypic linkage disequilibrium was detected in two of ten linkage tests involving pairs of microsatellite loci after sequential Bonferroni correction of alpha levels both involving the locus Slep9 and one test involving the locus Slep3 (Table 2.7). Table 2.7: Probabilities for X 2 tests (Fisher's method) of linkage disequilibrium. Tests were performed for each locus pair across all localities using GENEPOP(3 .4 ) (Raymond and Rousset 1995) software. The table-wide adjusted significance value was 0.005. Bo ld indicates significant results. Locus Pair X 2 df P-Va lue Test Adjusted Va lue T16 and S lep9 Infinity 22 Highly sign 0.005 S lep 9 and S lep3 Infinity 22 Highly sign 0.006 T 1 6 a n d C S L 9 32.19 30 0.359 0.006 C S L 9 and S lep9 31.61 30 0.386 0.007 C S L 9 a n d T 1 2 26.33 30 0.658 0.008 T16 and S l e p 3 11.83 18 0.856 0.010 C S L 9 and S lep3 14.63 22 0.878 0.013 T12 and S lep9 13.21 26 0.982 0.017 T12 and S lep3 6.94 20 0.997 0.025 T 1 6 a n d T 1 2 5.64 26 1.00 0.050 67 2.3.4 Hardy-Weinberg Equilibrium (HWE) Tests for Hardy-Weinberg Equilibrium (HWE) were conducted for each locality by each locus resulting in 85 tests. After application of the Bonferroni correction for multiple comparisons, 15 tests (involving 10 localities) were significantly different from H W E expectations (Table 2.8). Seven of these tests involved the locus Slep9 and eight involved the locus Slep3. 68 Table. 2.8: Significance values of Hardy Weinberg Equil ibrium tests and F i s (Weir and Cockerham 1984) calculated for each locality at each locus (0.05 alpha level). The adjusted alpha is the Bonferroni sequential correction values. Significant values at a table-wide alpha level of 0.01 are noted in bold. F.s P Va lue A D J U S T E D L O C A L I T Y L O C U S (W&C) (TEST) S . E . A L P H A China Creek Slep3 0.098 0.011 0.003 0.01 T16 0.062 0.044 0.005 0.0125 Slep9 0.12 0.046 0.006 0.0167 T12 0.119 0.156 0.009 0.025 CSL9 0.037 0.394 0.005 0.05 Nahmint Bay Slep3 0.143 . 0.001 0.001 0.01 Slep9 0.12 0.017 0.003 0.0125 T12 0.034 0.130 . 0.009 0.0167 CSL9 -0.055 0.746 0.006 0.025 T16 -0.056 1.000 0.000 0.05 Snug Basin Slep9 0.095 0.009 0.002 , 0.01 T12 0.018 0.181 0.010 0.0125 T16 0.004 0.494 0.013 0.0167 CSL9 -0.107 0.914 0.004 0.025 Slep3 -0.049 1.000 0.000 0.05 Assits Island Slep9 0.137 0.023 0.004 0.01 Slep3 0.158 0.028 0.004 0.0125 CSL9 0.022 0.316 0.003 0.0167 T16 0.013 0.592 0.013 0.025 T12 -0.064 1.000 0.000 0.05 Numukamis Bay Slep3 0.216 0.000 0.000 0.01 Slep9 0.094 0.075 0.007 0.0125 CSL9 0.03 0.418 0.004 0.0167 T16 -0.026 0.760 0.012 0.025 T12 -0.06 1.000 0.000 0.05 Grappler Inlet CSL9 0.119 0.050 0.002 0.01 Slep3 0.068 0.069 0.007 0.0125 Slep9 0.085 0.034 0.005 0.0167 T12 -0.09 1.000 0.000 0.025 T16 -0.059 1.000 0.000 0.05 69 Table 2.8 continued: F, s P Va lue A D J U S T E D L O C A L I T Y L O C U S (W&C) (TEST) S . E . A L P H A Bamfie ld Inlet S lep3 S lep9 C S L 9 T16 T12 0.25 0.222 -0.067 -0.021 -0.067 0.000 0.012 0.813 0.817 1.000 0.000 0.002 0.009 0.005 0.000 0.01 0.0125 0.0167 0.025 0.05 Dodger C h a n n e C S L 9 S lep3 S lep9 T12 T16 0.129 0.046 0.035 -0.053 0.026 0.186 0.370 0.174 1.000 0.256 0.004 0.014 0.011 0.000 0.011 0.01 0.0125 0.0167 0.025 0.05 Fleming Island S lep9 S lep3 C S L 9 T12 T16 0.101 0.172 -0.003 -0.024 -0.061 0.003 0.009 0.614 0.789 0.001 0.003 0.005 0.011 0.000 0.01 0.0125 0.0167 0.025 0.05 Port Alberni Yacht C lub S lep9 S lep3 C S L 9 T12 T16 0.159 0.114 0.084 -0.06 -0.058 0.000 0.002 0.231. 0.942 1.000 0.000 0.001 0.006 0.005 0.000 0.01 0.0125 0.0167 0.025 0.05 Use less Inlet C S L 9 S lep3 S lep9 T12 T16 0.128 0.108 0.081 -0.057 -0.068 0.069 0.070 0.093 0.679 1.000 0.002 0.008 0.008 0.011 0.000 0.01 0.0125 0.0167 0.025 0.05 Toquart Bay T12 S lep9 C S L 9 S lep3 0.034 0.095 0.072 0.074 0.028 0.146 0.155 0.227 0.004 0.010 0.006 0.011 0.01 0.0125 0.0167 0.025 T16 0.03 0.354 0.013 0.05 Table 2.8 continued: F, s P Va lue A D J U S T E D L O C A L I T Y L O C U S (W&C) (TEST) S . E . A L P H A Stopper Islands Slep9 0.086 0.030 0.005 0.01 CSL9 0.186 0.131 0.002 0.0125 Slep3 0.055 0.287 0.013 0.0167 T12 -0.075 1.000 0.000 0.025 T16 -0.076 1.000 0.000 0.05 Pinkerton Islands Slep9 0.159 0.000 0.000 0.01 Slep3 0.119 0.032 0.005 0.0125 T16 0.013 0.450 0.014 0.0167 CSL9 0.005 0.589 0.005 0.025 T12 -0.041 0.919 0.006 0.05 Jacques Jarves Slep9 0.089 0.002 0.001 0.01 Slep3 0.072 0.023 0.004 0.0125 CSL9 0.126 0.278 0.006 0.0167 T12 -0.011 0.280 0.011 0.025 T16 -0.032 0.839 0.009 0.05 Gibralter Island Slep3 0.158 0.000 0.000 0.01 CSL9 0.031 0.248 0.007 0.0125 T12 -0.011 0.394 0.011 0.0167 T16 -0.033 0.537 0.011 0.025 Slep9 -0.033 0.856 0.010 0.05 Turrett Island Slep 3 0.101 0.001 0.000 0.01 Slep 9 0.141 0.001 0.001 0.0125 T12 -0.004 0.466 0.014 0.0167 CSL9 -0.173 0.962 0.002 0.025 T16 -0.091 1.000 0.000 0.05 2.3.5 Heterozygosity and gene diversity Observed heterozygosities for each locus (over all localities) ranged from 0.70 (CSL9) to 0.96 (T16) with an average of 0.87 (Table 2.4). Expected heterozygosity ranged from 0.74 (CSL9) to 0.97 (Slep3) with an average of 0.91. Observed heterozygosities were lower than expected heterozygosities at three loci (CSL9, Slep9 and Slep3) and were higher than expected at two loci (T1.6 and T12), but these differences were not statistically significant after X 71 analysis (Table 2.4). Over the entire data set, the average multilocus expected heterozygosity was 0.91 Multilocus observed heterozygosities were high and ranged from 0.82 (China Creek and Dodger Channel) to 0.92 (Snug Basin) and expected heterozygosities (gene diversity) ranged from 0.85 (Dodger Channel) to 0.92 (Nahmint Bay) (Table 2.4). Observed multilocus heterozygosities were lower than expected at 15 of 17 localities and were higher at two localities (Snug Basin and Turret Island), but these differences were not statistically significant by X 2 analysis. Within localities, there appears to be a high number of heterozygotes indicating high gene diversity within individuals. F I S values were low and ranged from -0.005 (Turret Island) to 0.088 (China t Creek) suggesting that a high level of outbreeding occurs within localities (Table 2.4). Gene diversity (multilocus expected heterozygosity) was calculated among regions based on geography (Table 2.9). The region defined by localities from the Broken Group Islands/Loudoun Channel had the lowest calculated gene diversity of the three geographic regions (Table 2.9). N o statistically significant differences of gene diversity among regions was detected (two sided P = 0.074). Table 2.9: Gene diversity (multi-locus expected heterzygosity) differences among geographic regions. Calculations performed with F S T A T 2.9.3.2 (Goudet 2001) with 10,000 permutations. Site abbreviations are given in Table 2.1. Two sided P Geographica l Region G e n e Diversity value Local i t ies in each region East Barkley Sound 0.914 C C , N B , S B South Barkley Sound 0.902 A , N U M , G l , B A M , D C , FL , P A Y C North Barkley Sound 0.899 0.074 T B , SI, J J , GIB, TI 72 2.3.6 Microsatellite allelic variation In total, 156 alleles were, detected. Alleles were placed in three categories: private alleles (alleles found at only one locality) at generally low frequencies, rare alleles (alleles shared by no more than five localities) at varying frequencies, to common alleles found at more than six localities and often across all localities, albeit at varying frequencies (Figure 2.3). For locus Slep9, the allelic size distribution is characterized by a normal curve with private and rare, low-frequency alleles at the extreme ends and common alleles centered around the size distribution. Small base-pair and large base-pair alleles are generally found in low frequencies in the tails of the normal distribution and medium-sized alleles at higher frequencies (Figure 2.3). This trend is found in all loci used in this study with the exception of a private Slep 3 (allele size 145) found toward the middle of the distribution. In total, 14 private alleles were detected (2 at T12; 6 at Slep 9; 6 at Slep 3) and 24 rare alleles (4 at T16; 2 at C S L 9 ; 1 at T12; 8 at Slep9; and 9 at Slep3) (Table 2.4). Loc i T16 and C S L 9 did not have private alleles. One hundred and eighteen (76%) alleles were classified as common. Nine, or 64.3%, of the private alleles were found at island localities (Dodger Channel (1); Fleming Island (3); Port Alberni Yacht Club (3); Stopper Islands (1); Gibralter Island (1); and five, or 35.7%, at coastal sites (China Creek (2); Nahmint Bay (1); Grappler Inlet (1), Useless Inlet (1). Fifty percent of the private alleles were found in the Deer Group Archipelago. There was no significant difference between the total number of private alleles found among island sites and coastal sites (X (0.o5(i), 1.14) or between the total number of private alleles found among regions ( X o.o5(2) 3.71). Allele frequencies of each locus among localities differ but all localities share common alleles (Figure 2.3; Appendix I). L o c i Slep9 and C S L 9 represent loci with allele frequencies that are statistically different among localities. 73 0.2 i 0.18 j 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 Allele N u m b e r A. Locus Slep9 1 2 3 4 5 6 7 8 9 10 Allele Number B. Locus CSL9 Figure 2.3: Allele Frequencies of locus Slep9 (A) and CSL9 (B) by each locality. Data series represent allele frequencies at each allele for all 17 localities. Fisher's Exact Test and F § T values calculated using GENEPOP3.4 (Raymond and Rousset 1995) software. Allele frequencies are statistically different among localities for each locus. 74 2.3.7 Interpopulation microsatellite differentiation Spatial Genetic Population Substructure Summary L o w levels of differentiation were observed among localities with an overall value of F S T =0.005 (P = 0.0001, Table 2.13). Permutation tests revealed that only a small number of pairwise comparisons of F S T were statistically different from zero after sequential Bonferroni corrections for multiple comparisons (Table 2.14 and Table 2.15). A M O V A results revealed that over 99.7% of genetic variation was distributed within populations (Table 2.17). A M O V A results indicated a weak yet significant difference between East Barkley Sound and South/North Barkley Sound (Table 2.17), and principal component analysis revealed genetic differentiation between localities situated in East Barkley Sound and all other localities as well as a genetically divergent locality situated in the Broken Group Islands (Figure 2.4) Exact Tests of allelic differentiation Using log-likelihood (G) based exact tests, each individual locus showed statistically significant allele frequency differences among all localities (P = 0.001 - 0.022) whether the assumption of random mating (HWE) was made or not with the exception of Slep9, non-random mating (P = 0.379) (Table 2.10). 75 Table 2.10: Log-likelihood (G) based exact tests of population differentiation of each locus among all localities. Tests assuming random mating are calculated on allele frequencies. Tests assuming non-random mating are calculated on genotypic frequencies. Tests were performed using F S T A T 2.9.32. (Goudet 2001) software over 1000 randomizations. B o l d indicates significant results. P value Adjusted Va lue Assuming R a n d o m Mating within S a m p l e s T16 0.001 0.010 C S L 9 0.001 0.013 T12 0.001 0.017 S lep3 0.001 0.025 Slep9 0.035 0.050 Assuming non-random mating within samp les T16 0.001 0.010 T12 0.001 0 .013 C S L 9 0.022 0.017 Slep3 0.004 0.025 Slep9 0.379 0.050 Fisher's exact tests of allele frequency differences between pairwise localities revealed that over individual loci , 39 of 680 comparisons, or 6%, were significant at adjusted table-wide alpha levels (Table 2.11). The Gibralter Island locality had 25 significantly different pairwise comparisons, China Creek had 13, and Turret Island had 3. The number of significant comparisons for each locus at adjusted alpha levels was: 20 (T16), 1 (CSL9) , 17 (T12), 1 (Slep9), and 2 (Slep3). Fisher's exact tests with combined loci between pairwise localities revealed 39 of 136, or 29%, significant tests after Bonferroni correction at table-wide alpha levels (Table 2.12). Seven localities had from 6 - 1 6 significant pairwise comparisons ranging from 6 - 1 6 (China Creek 10; Snug Basin 6, Dodger Channel 6, Jaques/Jarvis Island 6; Gibralter Island (16); 76 Turret Island 6; Toquart Bay 6). A l l other localities had from 1 - 4 significant pairwise comparisons. Fisher's exact tests revealed that the localities within the Alberni Inlet (East Barkley Sound) are genetically homogenous and are genetically differentiated from localities within the Broken Group Islands/Loudoun Channel area (North Barkley Sound). The Nahmint Bay and Snug Basin localities reveal more genetic affinity with Trevor Channel localities (Assits, Numukamis Bay, Grappler Inlet and Bamfield Inlet). The Trevor Channel/Deer Group Islands localities (South Barkley Sound) were characterized by significant genetic homogeneity among localities within this region and revealed significant differentiation from sampled localities within the Broken Group Islands. The Useless Inlet locality, geographically central to both southern and northern regions of Barkley Sound, appears to be genetically homogenous with the majority of localities but genetically differentiated from both the Broken Group Islands and Alberni Inlet localities. The most genetically differentiated area is the Broken Group Islands archipelago. Intriguingly, within the Broken Group. Islands, all neighbouring sites are significantly differentiated from each other, and the Gibralter Island locality is significantly different from all other localities in this' study. There was no significant genetic differentiation among Toquart Bay, Stopper Islands and Pinkertons localities proximal to the Broken Group Islands localities (Table 2.11 and Table 2.12) 77 Table 2.11: Fisher's Exact Tests of population differentiation on all possible pairwise comparisons by each locus. Bold indicates significant at the Bonferroni corrected alpha level of 0.000368. Sample designations are as given in Table 2.1 Locus T16 Pairwise popns P value S.E. Adjusted Value Pairwise popns P value S.E. Adjusted P Value Pairwise popns P value S.E. Adjusted P Value CCxDC CCxPC CCxGB CCxTI DCxGB FLxGB UlxGB TBxGB CCxTB CCxGI CCxJJ CCxNM PCxGB SIxGB CCxBM JJxGB NBxGB CCxUI GBxTI NMxGB CCxSI SBxGB CCxPI CCxAl SBxTI CCxFL JJxTI SBxDC GlxTI SBxPC DCxJJ BMxGB GlxGB AlxGB SBxTB GlxTB GlxSI CCxSB UlxJJ FLxJJ SIxJJ PCxJJ NBxTI GlxDC NBxGI PIxGB TBxTI 0 0 0 0 0 0 0 0 0.00001 0.00004 0.00005 0.00010 0.00010 0.00012 0.00013 0.00013 0.00022 0.00025 0.00034 0.00063 0.0009 0.00098 0.0015 0.0020 0.0040 0.0041 0.0048 0.0051 0.0060 0.0066 0.0066 0.0075 0.0085 0.0101 0.0101 0.0102 0.0134 0.0273 0.0285 0.0287 0.0379 0.0384 0.0400 0.0477 0.0549 0.0561 0.0581 0 0 0 0 0 0 0 0 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.002 0.002 0.002 0.002 0.000368 0.00037 0.00037 0.00038 0.00038 0.00038 0.00038 0.00039 0.00039 0.00039 0.00040 0.00040 0.00040 . 0.00041 0.00041 0.00041 0.00042 0.00042 0.00042 0.00043 0.00043 0.00043 0.00044 0.00044 0.00045 0.00045 0.00045 0.00046 0.00046 0.00047 0.00047 0.00048 0.00048 0.00049 0.00049 0.00050 0.00050 0.00051 0!00051 0.00052 0.00052 0.00053 0.00053 0.00054 0.00054 0.00055 0.00056 BMxTI GlxUI TBxJJ SBxUI BMxDC NBxJJ BMxSI FLxTI NBxDC NMxTI SBxJJ SBxSI DCxTB NBxPC GlxPC DCxTI . UlxTI SBxNM BMxPC NMxDC PIxJJ SIxTI NBxTB GlxFL NBxBM BMxUI CCxNB GlxJJ PCxTI BMxTB NBxNM PIxTI DCxSI PCxUI NMxBM PCxTB SBxBM UlxTB BMxJJ FLxTB NMxPC SBxGI SBxFL NMxUI NMxGI AlxJJ DCxPC 0.06 0.06 0.07 0.07 0.08 0.09 0.09 0.10 0.10 0.11 0.13 0.14 0.16 0.16 0.16 0.17 0.18 0.18 0.21 0.21 0.22 0.22 0.22 0.23 0.23 0.23 0.23 0.25 0.25 0.26 0.29 0.29 0.29 0.30 0.30 0.31 0.34 0.37 0.37 0.37 0.38 0.39 0.40 0.40 0.42 0.42 0.43 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.003 0.002 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.004 0.004 0.003 0.003 0.003 0.003 0.004 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.004 0.005 0.004 0.005 0.005 0.004 0.004 0.005 0.004 0.004 0.005 0.000562 0!000568 0.000575 0.000581 0.000588 0.000595 0.000602 0.000610 0.000617 0.000625 0.000633 0.000641 0.000649 0.000658 0.000667 0.000676 0.000685 0.000694 0.000704 0.000714 0.000725 0.000735 0.000746 0.000758 0.000769 0.000781 0.000794 0.000806 0.00082 0.000833 0.000847 0.000862 0.000877 0.000893 0.000909 0.000926 0.000943 0.000962 0.00098 0.001 0.00102 0.001042 0.001064 0.001087 0.001111 0.001136 0.001163 AlxGI UlxSI FLxSI NMxFL NMxSI TBxPI AlxBM DCxUI PCxSI DCxPI NMxJJ DCxFL NBxSI BMxFL AlxDC NBxSB BMxPI SIxPI PCxPI GlxBM . GlxPI AlxPC NBxAl AlxUI AlxTI SBxAl NMxTB NMxPI F L x P C NBxPI TBxSI FLxUI NBxUI AlxFL UlxPI SBxPI NBxFL FLxPI AlxTB AlxNM AlxPI AlxSI 0.44 0.45 0.45 0.47 0.54 0.54 0.56 0.56 0.57 0.58 0.58 0.60 0.61 0.63 0.65 0.67 0.68 0.69 0.71 0.71 0.73 0.75 0.76 0.77 0.78 0.79 0.79 0.80 0.81 0.84 0.85 0.85 0.86 0.87 0.88 0.90 0.90 0.93 0.94 0.94 0.95 0.99 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.005 0.004 0.004 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.002 0.003 0.003 0.002 0.002 0.002 0.002 0.002 0.002 0.001 0.001 0.001 0.001 0.00119 0.00122 0.00125 0.001282 0.001316 0.001351 0.001389 0.001429 0.001471 0.001515 0.001563 0.001613 0.001667 0.001724 0.001786 0.001852 0.001923 0.002 0.002083 0.002174 0.002273 0.002381 0.0025 0.002632 0.002778 0.002941 0.003125 0.003333 0.003571 0.003846 0.004167 0.004545. 0.005 0.005556 0.00625 0.007143 0.008333 0.01 0.0125 0.016667 0.025 0.05 78 Table 2.11 continued: Locus CSL2 Pairwise popns P value S. Adjusted P| E. Value Pairwise popns P value S.E. Adjusted P Value Pairwise popns P value S.E. Adjusted P Value PCxGB SBxGB NBxDC DCxGB NBxNM GlxDC SBxSI NBxUI PCxSI NBxSI NBxAl SBxDC FLxPC PIxGB AlxFL BMxDC NBxPC NMxBM NMxFL AlxGB NBxGB SBxJJ GlxUI BMxSI NMxGB DCxPI SIxPI GlxPI BMxGB DCxPC AlxSI NMxGI GlxGB PCxJJ FLxPI PCxPI SBxFL NBxJJ NBxTB SBxNM PIxJJ GlxSI DCxTB BMxJJ DCxSI AlxBM AlxGI 0.00012 0.0007 0.0008 0.0024 0.0027 0.0037 6.0078 0.0113 0.0130 0.0132 0.0135 0.0140 0.0169 0.0174 0.0183 0.0193 0.0198 0.0225 0.0229 0.0240 0.0267 0.0289 0.0312 0.0313 0.0344 0.0350 0.0359 0.0363 0.0399 0.0409 0.0425 0.0438 0.0447' 0.0516 0.0518 0.0546 0.0552 0.0566 0.0580 0.0587 0.0619 0.0619 0.0632 0.0881 0.0888 0.0956 0.0960 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.002 0.002 0.002 0.002 0.002 0.002 0.001 0.002 0.002 0.002 0.002 0.002 0.000368 0.00037 0.00037 0.00038 0.00038 0.00038 0.00038 0.00039 0.00039 0.00039 0.00040 0.00040 0.00040 0.00041 0.00041 0.00041 0.00042 0.00042 0.00042 0.00043 0.00043 0.00043 0.00044 0.00044 0.00045 0.00045 0.00045 0.00046 0.00046 0.00047 0.00047 0.00048 0.00048 0.00049 0.00049 0.00050 0.00050 0.00051 0.00051 0.00052 0.00052 0.00053 0.00053 0.00054 0.00054 0.00055 0.00056 CCxDC UlxGB NMxPI FLxUI BMxUI AlxDC AlxJJ SBxUI GlxTI NMxDC GlxBM BMxFL NBxFL SBxTB AlxPI DCxFL NMxSI NMxJJ NBxTI . NBxPI-DCxTI TBxPI SBxGI BMxPC PCxUI CCxGB FLxTI GlxFL GBxTI SBxPC JJxGB SIxTI DCxJJ UlxPI CCxBM BMxTB GlxJJ CCxPC FLxTB CCxSI NMxPC FLxGB GlxPC CCxNB NMxTI NBxGI CCxFL 0.098 0.100 0.104 0.109 0.111 0.116 0.120 0.125 0.127 0.129 0.131 0.131 0.134 0.136 0.136 0.139 0.149 0.156 0.167 0.171 0.181 0.194 0.197 0.201' 0.206 0.211 0.217 0.237 0.245 0.251 0.259 0.265 0.280 0.285 0.287 0.299 0.304 0.316 0.320 0.325 0.330 0.331 0.333 0.337 0.343 0.355 0.355 0.002 0.003 0.002 0.002 0.002 0.002 0.002 0.003 0.002 0.002 0.002 0.002 0.003 0.003 0.002 0.002 0.002 0.003 0.003 0.003 0.002 0.003 0.003 0.003 0.003 0.003 0.003 0.002 0.003 0.003 0.004 0.003 0.004 0.003 0.003 0.003 0.003 0.003 0.004 0.003 0.003 0.004 0.003 0.003 0.003 0.003 0.003 0.000562 0.000568 0.000575 0.000581 0.000588 0.000595 0.000602 0.000610 0.000617 0.000625 0.000633 0.000641 0.000649 0.000658 0.000667 0.000676 0.000685 0.000694 0.000704 0.000714 0.000725 0.000735 0.000746 0.000758 0.000769 0.000781 0.000794 0.000806 0.00082 0.000833 0.000847 0.000862 0.000877 0.000893 0.000909 0.000926 0.000943 0.000962 0.00098 0.001 0.00102 0.001042 0.001064 0.001087 0.001111 0.001136 0.001163 UlxSI CCxAl PCxTB AlxTI DCxUI PIxTI SBxAl CCxGI GlxTB JJxTI BMxPI CCxNM UlxJJ PCxTI TBxGB AlxTB SBxTI AlxUI FLxSI CCxJJ NBxBM TBxJJ CCxTI CCxUI CCxSB CCxPI SBxPI NBxSB SIxJJ TBxSI NMxTB UlxTI UlxTB CCxTB SBxBM TBxTI AlxPC AlxNM BMxTI NMxUI SIxGB FLxJJ 0.36 0.36 0.37 0.38 0.40 0.41 0.41 0.41 0.44 0.45 0.45 0.45 0.47 0.48 0.53 0.53 0.54 0.54 0.55 0.59 0.60 0.62 0.64 0.65 0.65 0.68 0.69 0.72 0.75 0.76 0.77 0.79 0.79 0.81 0.83 0.85 0.85 0.86 0.87 0.91 0.93 0.96 0.003 0.002 0.004 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.004 0.003 0.003 0.002 0.003 0.004 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.003 0.002 0.003 0.003 0.002 0.002 0.002 0.002 0.002 0.002 0.001 0.001 0.001 0.001 0.001 0.00119 0.00122 0.00125 0.001282 0.001316 0.001351 0.001389 0.001429 0.001471 0.001515 0.001563 0.001613 0.001667 0.001724 0.001786 0.001852 0.001923 0.002 0.002083 0.002174 0.002273 0.002381 0.0025 0.002632 0.002778 0.002941 0.003125 0.003333 0.003571 0.003846 0.004167 0.004545 0.005 0.005556 0.00625 0.007143 0.008333 0.01 0.0125 0.016667 0.025 0.05 79 Table 2.11 continued: Locus T12 Pairwise popns P value S.E Adjusted P Value Pairwise popns P value S.E. Adjusted P Value Pairwise popns P value S.E. Adjusted P Value NBxGB 0 0 000 0.000368 NMxUI 0.06 0.002 0.00119 PCxTB 0.48 0.005 0.00119 SBxGB 0 0 000 0.00037 FLxTB 0.07 0.002 0.00122 NMxFL 0 48 0.004 0.00122 NMxGB 0 0 000 0.00037 NMxTB 0.07 0.002 0.00125 GlxTB 0 50 0.004 0.00125 DCxGB 0 0 000 0.00038 NMxPC 0.08 0.002 0.001282 CCxSB 0 52 0.004 0.001282 FLxGB 0 0 000 0.00038 DCxPI 0.09 0.002 0.001316 NMxGI 0 52 0.004 0.001316 PCxGB 0 0 000 0.00038 UlxPI 0.09 0.002 0.001351 BMxTB 0 53 0.004 0.001351 TBxGB 0 0 000 0.00038 NBxDC 0.09 0.002 0.001389 AlxUI 0 53 0.003 0.001389 SIxGB 0 0 000 0.00039 UlxJJ 0.10 0.002 0.001429 GlxBM 0 57 0.003 0.001429 JJxGB 0 0 000 0.00039 NMxSI 0.10 0.002 0.001471 NBxGI 0 57 0.004 0.001471 BMxGB 0 0 000 0.00039 CCxBM 0.11 0.002 0.001515 TBxTI 0 58 0.004 0.001515 PIxGB 0 0 000 0.00040 NBxFL 0.11 0.002 0.001563 GlxPC 0 59 0.004 0.001563 GBxTI 0 0 000 0.00040 CCxJJ 0.11 0.003 0.001613 PCxSI 0 59 0.004 0.001613 CCxFL 0 0 000 0.00040 SBxGI 0.14 0.003 0.001667 NBxTI 0 60 0.004 0.001667 UlxGB 0 0 000 0.00041 SBxBM 0.14 0.002 0.001724 PCxJJ 0 60 0.004 0.001724 CCxDC 0 0 000 0.00041 FLxSI 0.17 0.003 0.001786 AlxTB .0 62 0.004 0.001786 CCxSI 0 0 000 0.00041 GlxPI 0.1.9 0.003 0.001852 BMxUI 0 62 0.003 0.001852 CCxGB 0.001 0 000 0.00042 SBxAl 0.20 0.003 0.001923 AlxJJ 0 63 0.004 0.001923 AlxGB 0.001 0 000 0.00042 SBxJJ 0.21 0.004 0.002 AlxSI 0 64 0.003 0.002 GlxGB 0.002 0 000 0.00042 CCxAl 0.24 0.003 0.002083 CCxGI 0 66 0.004 0.002083 SBxSI 0.002 0 000 0.00043 GlxUI 0.25 0.003 0.002174 AlxGI 0 67 0.003 0.002174 SBxUI 0.003 0 000 0.00043 FLxTI 0.25 0.004 0.002273 NMxTI 0 69 0.003 0.002273 SBxFL 0.003 0 000 0.00043 NBxSB 0.26 0.003 0.002381 BMxSI 0 69 0.003 0.002381 SBxPC 0.004 0 000 0.00044 BMxTI 0.26 0.003 0.0025 PIxJJ 0 69 0.004 0.0025 CCxPI 0.005 0 000 0.00044 NBxBM 0.27 0.003 0.002632 NMxBM 0 71 0.003 0.002632 FLxJJ 0.005 0 000 0.00045 UlxTI 0.28 0.004 0.002778 BMxDC 0 71 0.003 0.002778 CCxTB 0.007 0 001 0.00045 NMxPI 0.29 0.003 0.002941 NBxNM 0 72 0.003 0.002941 CCxPC 0.011 0 001 0.00045 GlxSI 0.29 0.004 0.003125 NBxAl 0 73 0.003 0.003125 SBxTB 0.017 0 001 0.00046 SIxPI 0.29 0.004 0.003333 NMxDC 0 75 0.003 0.003333 SBxDC 0.018 0 001 0.00046 TBxJJ 0.30 0.004 0.003571 TBxSI 0 76 0.004 0.003571 FLxUI 0.022 0 001 0.00047 BMxJJ 0.30 0.004 0.003846 SBxNM 0 77 0.003 0.003846 NBxPC 0.023 0 001' 0.00047 CCxTI 0.31 0.004 0.004167 PIxTI 0 78 0.003 0.004167 DCxJJ 0.024 0 001 0.00048 JJxTI 0.31 0.004 0.004545 AlxBM 0 79 0.002 0.004545 FLxPC 0.026 0 001 0.00048 PCxPI 0.31 0.004 0.005 AlxFL 0 79 0.003 0.005 SBxPI 0.026 0 001 0:00049 SIxJJ 0.32 0.004 0.005556 AlxTI 0 85 0.002 0.005556 CCxUI 0.026 0 001 0.00049 BMxPI 0.33 0.004 0.00625 PCxUI 0 85 0.002 0.00625 GlxFL 0.028 0 001 0.00050 NMxJJ 0.34 0.004 0.007143 BMxPC 0 88 0.002 0.007143 DCxTB 0.031 0 001 0.00050 BMxFL 0.38 0.004 0.008333 AlxDC 0 89 0.002 0.008333 SIxTI 0.032 0 001 , 0.00051 DCxSI 0.40 0.004 0.01 AlxNM 0 90 0.002 0.01 DCxUI 0.033 0 001 0.00051 FLxPI 0.41 0.004 0.0125 UlxTB 0 90 0.002 0.0125 NBxUI 0.036 0 001 0.00052 PCxTI 0.42 0.004 0.016667 GlxTI 0 90 0.002 0.016667 NBxTB 0.041 0 001 0.00052 SBxTI 0.42 0.004 0.025 DCxFL 0 92 0.002 0.025 NBxSI 0.042 0 001 0.00053 CCxNB 0.42 0.004 0.05 AlxPI 0 95 0.001 0.05 DCxPC 0.045 0 002 0.00053 AlxPC 0.45 0.004 NBxPI 0.048 0 001 0.00054 TBxPI 0.47 0.004 DCxTI 0.049 0 002 0.00054 GlxJJ 0.47 0.004 NBxJJ 0.061 0 002 0.00055 CCxNM 0.48 .0.004 GlxDC 0.063 0 002 0.00056 UlxSI 0.48 0.004 80 Table 2.11 continued: Locus Slep9 Pairwise Adjusted P Pairwise Adjusted Pairwise Adjusted popns P value S . E . Value popns P value S . E . P Value popns P value S . E . P Value TBxGB 0.00023 0.000 0.000368 SBxDC 0.24 0.004 0.00119 TBxPI 0 56 0.004 0.00119 GBxTI 0.00071 0.000 0.00037 GlxTI 0.25 0.004 0.00122 CCxSB 0 60 0.004 0.00122 PCxGB 0.003 0.000 0.00037 GlxUI 0.26 0.004 0.00125 SBxNM 0 60 0.004 0.00125 UlxGB 0.003 0.000 0.00038 NBxSB 0.26 0.004 0.001282 NBxPC 0 63 0.004 0.001282 CCxTI 0.003 0.000 0.00038 BMxPC 0.26 0.004 0.001316 NBxDC 0 63 0.004 0.001316 AlxGB 0.004 0.000 0.00038 SBxAl . 0.27 0.004 0.001351 DCxTB 0 63 0.004 0.001351 SBxPC 0.009 0.001 0.00038 CCxDC 0.28 0.004 0.001389 UlxTB 0 63 0.004 0.001389 CCxGB 0.011 0.001 0.00039 CCxBM 0.28 0.003 0.001429 FLxPI 0 64 0.004 0.001429 SBxTI 0.011 0.001 0.00039 AlxFL 0.28 0.004 0.001471 AlxSI 0 65 0.004 0.001471 SIxGB 0.013 0.001 0.00039 CCxAl 0.28 0.004 0.001515 CCxFL 0 65 0.004 0.001515 DCxTI 0.016 0.001 0.00040 BMxDC 0.31 0.004 0.001563 GlxJJ 0 66 0.004 0.001563 UlxTI 0.019 0.001 0.00040 AlxNM 0.32 0.004 0.001613 DCxJJ 0 68 0.004 0.001613 JJxTI 0.022 0.001 0.00040 CCxPC 0.33 0.004 0.001667 CCxTB 0 69 0.004 0.001667 SBxSI 0.033 0.001 0.00041 PCxTB 0.33 0.005 0.001724 FLxTB 0 70 0.004 0.001724 SBxGB 0.039 0.001 0.00041 AlxDC 0.34 0.004 0.001786 DCxPC 0 73 0.004 0.001786 FLxGB 0.044 0.002 0.00041 SBxGI 0.35 0.004 0.001852 NBxAlt 0 76 0.003 0.001852 CCxUI 0.047 0.002 0.00042 PCx PI 0.35 0.004 0.001923 BMxPI 0 76 0.003 0.001923 PCxTI 0.054 0.002 0.00042 GlxPC 0.37 0.005 0.002 GlxFL 0 76 0.004 0.002 PIxGB 0.055 0.002 0.00042 FLxJJ 0.37 0.005 0.002083 NMxFL 0 78 0.003 0.002083 BMxGB 0.057 0.002 0.00043 SBxJJ 0.38 0.005 0.002174 AlxJJ 0 79 0.003 0.002174 PIxTI 0.061 0.002 0.00043 PCxJJ 0.38 0.005 0.002273 GlxBM 0 79 0.003 0.002273 JJxGB 0.073 0.002 0.00043 AlxBM 0.38 0.003 0.002381 NBxBM 0 81 0.002 0.002381 AlxTI 0.074 0.002 0.00044 AlxPC 0.40 0.005 0.0025 NBxNM 0 81 0.003 0.0025 NMxTI 0.075 0.002 0.00044 FLxUI 0.41 0.004 0.002632 DCxUI 0 82 0.003 0.002632 PCxSI 0.075 0.002 0.00045 UlxJJ 0.41 0.005 0.002778 CCxPI 0 82 0.003 0.002778 FLxTI 0.079 0.002 0.00045 NMxGB 0.43 0.004 0.002941 DCxFL 0 83 0.003 0.002941 TBxTI 0.082 0.002 0.00045 SBxTB 0.44 0.005 0.003125 NMxJJ 0 84 0.003 0.003125 FLxPC 0.086 0.003 0.00046 AlxTB 0.44 0.004 0.003333 NMxGI 0 84 0.003 0.003333 SIxTI 0.10 0.003 0.00046 UlxPI 0.44 0.004 0.003571 AlxPI 0 85 0.002 0.003571 SIxJJ 0.10 ' 0.003 0.00047 BMxFL 0.45 0.004 0.003846 CCxGI 0 85 0.003 0.003846 TBxJJ 0.11 0.003 0.00047 DCxGB 0.45 0.005 0.004167 NBxPI 0 86 0.002 0.004167 UlxSI 0.11 0.002 0.00048 NMxTB 0.46 0.005 0.004545 GlxSI 0 86 0.003 0.004545 PCxUI 0.11 0.003 0.00048 BMxTB 0.46 0.004 0.005 NBxFL 0 86 0.002 0.005 CCxSI 0.11 0.003 0.00049 NMxSI 0.46 0.004 0.005556 NMxDC 0 88 0.002 0.005556 AlxUI 0.13 0.003 0.00049 NBxSI 0.47 0.004 0.00625 CCxNM 0 91 0.002 0.00625 SBxUI 0.15 0.003 0.00050 NMxUI 0.48 0.004 0.007143 NBxJJ 0 92 0.002 0.007143 TBxSI 0.15 0.003 0.00050 BMxTI 0.49 0.004 0.008333 PIxJJ 0 93 0.002 0.008333 GlxGB 0.16 0.003 0.00051 GlxTB 0.51 0.004 0.01 NMxBM 0 95 0.001 0.01 NBxGB 0.17 0.003 0.00051 SBxPI 0.51 0.005 0.0125 GlxPI 0 96 0.001 0.0125 FLxSI 0.18 0.003 0.00052 NMxPC 0.51 0.005 0.016667 NMxPI 0 97 0.001 0.016667 NBxTI 0.19 0.003 0.00052 NBxTB 0.51 0.004 0.025 NBxGI 0 98 0.001 0.025 SBxBM 0.20 0.003 0.00053 CCxNB 0.51 0.004 0.05 AlxGI 1 00 0.000 0.05 SBxFL 0.20 0.004 0.00053 BMxJJ 0.53 0.004 BMxSI 0.23 0.003 0.00054 SIxPI 0.54 0.004 BMxUI 0.23 0.003 0.00054 NBxUI 0.55 0.004 DCxSI 0.24 0.004 0.00055 CCxJJ 0.55 0.005 DCxPI 0.24 0.004 0.00056 GlxDC 0.55 0.004 Table 2.11 continued: Locus Slep 3 Pairwise popns P value S.E. Adjusted P Value Pairwise popns P value S.E Adjusted P Value Pairwise popns P value S.E Adjusted P Value PCxGB 0.00008 0 000 0.000368 GlxGB 0.063 0 002 0.00119 NMxTI 0.31 0.004 0.00119 TBxSI 0.00035 0 000 0.00037 NBxTB 0.068 0 002 0.00122 SIxTI 0.34 0 004 0.00122 PCxJJ 0.00049 0 000 0.00037 CCxGI 0.069 0 002 0.00125 CCxBM 0.39 0 004 0.00125 SBxSI 0.00151 0 000 0.00038 SBxGI 0.071 0 002 0.001282 SBxAl 0.39 0 004 0.001282 AlxJJ 0.0019 0 000 0.00038 CCxGB 0.071 0 002 0.001316 NBxSI 0.43 0 004 0.001316 SIxGB 0.0020 0 000 0.00038 NBxTI 0.081 0 002 0.001351 NMxDC 0.45 0 004 0.001351 SBxGB 0.0020 0 000 0.00038 FLxUI 0.090 0 002 0.001389 BMxJJ 0.47 0 004 0.001389 JJxGB 0.0031 0 000 0.00039 NMxGI 0.094 0 002 0.001429 BMxPI 0.48 0 004 0.001429 PCxTB 0.0034 0 000 0.00039 NBxGB 0.094 0 002 0.001471 SBxFL 0.49 0 004 0.001471 TBxJJ 0.0037 . 0 000 0.00039 CCxAl 0.097 0 002 0.001515 NMxBM 0.51 0 004 0.001515 JJxTI 0.0040 0 000 0.00040 DCxFL 0.098 0 002 0.001563 PCxPI 0.51 0 005 0.001563 DCxTI 0.0044 0 000 0.00040 GlxDC 0.10 0 002 0.001613 SIxJJ 0.52 0 005 0.001613 GlxTB 0.0077 0 001 0.00040 SBxDC 0.11 0 002 0.001667 BMxGB 0.53 0 004 0.001667 AlxPI 0.0086 0 000 0.00041 FLxJJ 0.12 0 003 0.001724 NMxSI 0.57 d 005 0.001724 SBxPC 0.0088 0 001 0.00041 PCxTI 0.12 0 003 0.001786 NBxJJ 0.58 0 005 0.001786 SBxJJ 0.0097 0 001 0.00041 UlxTI 0.12 0 002 0.001852 FLxTI 0.59 0 004 0.001852 PCxSI 0.014 0 001 0.00042 NBxAl 0.12 0 002 0.001923 UlxPI 0.61 0 004 0.001923 NBxPC 0.014 0 001 0.00042 SBxUI 0.13 0 003 0.002 PIxJJ 0.61 0 004 0.002 FLxTB 0.015 0 001 0.00042 BMxSI 0.13 0 003 0.002083 CCxDC 0.62 0 004 0.002083 AlxGB 0.016 0 001 0.00043 FLxSI 0.13 0 003 0.002174 NMxUI 0.62 0 004 0.002174 UlxGB 0.019 0 001 0.00043 TBxPI 0.13 0 003 0.002273 UlxSI 0.62 0 004 0.002273 NMxGB 0.020 0 001 0.00043 DCxSI 0.13 0 003 0.002381 NMxFL 0.62 0 004 0.002381 AlxDC 0.022 0 001 0.00044 BMxTI 0.13 0 003 0.0025 CCxNB 0.64 0 004 0.0025 TBxTI 0.022 0 001 0.00044 FLxPC 0.14 0 003 0.002632 NMxPC 0.66 0 004 0.002632 GlxJJ 0.022 0 001 0.00045 GlxUI 0.15 0 003 0.002778 CCxNM 0.70 0 004 0.002778 GlxPC 0.023 0 001 0.00045 NBxGI 0.17 0 003 0.002941 AlxGI 0.71 0 003 0.002941 AlxTB 0.025 0 001 0.00045 SIxPI 0.17 0 003 0.003125 NMxJJ 0.72 0 004 0.003125 CCxTB 0.025 0 001 0.00046 AlxUI 0.18 0 003 0.003333 BMxFL 0.75 0 003 0.003333 PIxTI 0.029 0 001 0.00046 BMxTB 0.18 0 003 0.003571 DCxJJ 0.78 0 003 0.003571 DCxTB 0.035 0 001 0.00047 GlxPI 0.18 0 003 0.003846 BMxUI 0.79 0 003 0.003846 TBxGB 0.038 0 001 0.00047 GlxTI 0.18 0 003 0.004167 DCxPI 0.81 0 003 0.004167 UlxTB 0.039 0 001 0.00048 UlxJJ 0.19 0 003 0.004545 CCxUI 0.83 0 003 0.004545 DCxPC 0.041 0 002 0.00048 CCxSB 0.20 0 004 0.005 NBxBM 0.84 0 003 0.005 FLxGB 0.041 0 002 0.00049 GlxFL 0.21 0 004 0.005556 NBxPI 0.87 0 002 0.005556 SBxTI 0.041 0 002 0.00049 AlxBM 0.22 0 003 0.00625 NBxUI 0.88 0 002 0.00625 CCxJJ 0.041 0 002 0.00050 SBxPI 0.22 0 004 0.007143 NBxNM 0.89 0 002 0.007143 BMxPC 0.042 0 002 0.00050 NBxSB 0.23 0 004 0.008333 BMxDC 0.91 0 002 0.008333 DCxGB 0.043 0 001 0.00051 NBxFL 0.23 0 004 0.01 SBxTB 0.91 0 002 0.01 CCxPC 0.044 0 002 0.00051 CCxFL 0.24 0 004 0.0125 NMxPI 0.92 0 002 0.0125 PIxGB 0.045 0 002 0.00052 CCxTI 0.24 0 004 0.016667 NBxDC 0.95 0 001 0.016667 AlxSI 0.046 0 002 0.00052 GlxBM 0.26 0 003 0.025 CCxPI 0.96 0 001 0.025 AlxPC 0.046 0 002 0.00053 CCxSI 0.26 0 004 0.05 DCxUI 0.96 0 001 0.05 AlxFL 0.047 0 002 0.00053 PCxUI 0.28 0 004 GlxSI 0.050 0 002 0.00054 AlxTI 0.28 0 003 NMxTB 0.051 0 002 0.00054 SBxNM 0.30 0 004 AlxNM 0.055 0 002 0.00055 SBxBM 0.30 0 004 GBxTI 0.062 0 002 0.00056 FLxPI 0.31 0 004 82 Table 2.12: Fisher's Exact Test P values of population differentiation of allele frequencies conducted by combining five microsatellite loci and pairwise among localities. Original nominal P values are indicated below the diagonal and significant pairwise comparisons after sequential Bonferroni corrections are indicated above the diagonal. Calculated using T F P G A (version 1.3) (Mil ler 1997). Abbreviations for localities are as given in Table 2.1 CC NB SB Al NM Gl BM DC FL PC Ul TB SI PI JJ GIB TI CC * * * * * * * * * * NB 0.37 * * SB 0.07 0.21 * * * * * Al 0.00 0.08 0.34 * NUN 0.01 0.00 0.10 0.65 * Gl 0.00 0.07 0.01 0.53 0.05 * * * BM 0.00 0.42 0.14 0.23 0.22 0.27 * DC 0.00 0.00 0.00 0.04 0.27 0.00 0.06 ^^^^^ * * * FL 0.00 0.18 0.00 0.05 ,0.26 0.02 0.44 0.27 * PC 0.00 0.00 0.00 0.33 0.15 0.02 0.09 0.00 0.00 . . • - . - . * * Ul • 0.00 0.02 0.00 0.35 0.18 0.00 0.17 0.08 0.02 0.15 * TB 0.00 0.00 0.00 0.13 0.00 0.00 0.19 0.00 0.00 0.00 0.21 * * * SI 0.00 0.02 0.00 0.11 0.13 0.00 0.03 0.02 0.18 0.00 0.16 0.00 * PI 0.00 0.21 0.10 0.11 0.41 0.05 0.69 0.02 0.24 0.21 0.29 0.14 0.06 * JJ 0.00 0.01 0.00 0.00 0.31 0.00 0.09 0.00 0.00 0.00 0.01 0.00 0.05 0.16 * GIB 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 i l l s * TI 0.00 0.01 0.00 0.34 0.11 0.00 0.07 0.00 0.03 0.04 0.02 0.00 0.01 0.02 0.00 0 FST analysis and AMOVA of regions F s T values for each locus by pooled localities for T16 was 0.007 (P = 0.001, S.E. 003); for C S L 9 was 0.008 (P = 0.001, S.E. 0.004); T12 was 0.008 (P = 0.0001, S.E. 0.003); for Slep9 was 0.001 (P = 0.039, S.E. 0.002); for Slep3 was 0.004 (P = 0.0001, S.E. 0.002); and over all five loci the F S T was 0.005 (P = 0.0001, S.E. 0.001) (Table 2.13). A s T16, C S L 9 and T12 had the highest F S T values, these three loci were combined and over all the F S T was 0.007 indicating the relative importance of these three loci in establishing a multilocus F S T estimate of genetic divergence among localities (Table 2.13). The maximum F S T detectable in this study (where F S T = 1 - H e , Hedrick 2000) based on an average heterozygosity (0.90) or unbiased estimate of average heterozygosity (0.91) over all loci and all fish is 0.1 or 0.09 respectively. 83 Table 2.13: F S T values for each locus and overall five microsatellite loci . P values indicate the probability that the values are significantly different from zero. Estimates of standard errors and 95% confidence intervals of F S T are also provided. N / A indicates values that are not available. Statistics were generated using 10,000 permutations in F S T A T 2.9.3.2 (Goudet 2001) software. ~ 1 Conf idence Locus Name Fst Va lue S . E . Interval P va lue T16 0.007 0.003 0.0001 C S L 9 0.008 0.004 0.001 T12 0.008 0.003 0.0001 S L E P 9 0.001 0.002 0.039 S L E P 3 0.004 0.002 0.0001 Overa l l 0.005 0.001 0.003-0.007 0.0001 O v e r T 1 6 , C S L 9 , T12 0.007 N/A N/A N/A F S T values for each locus by pairwise comparison of localities were also calculated (Appendix II). F S T values by pairwise localities for locus T16 ranged from -0.0009 to 0.045; for Locus C S L 9 from -0.0008 to 0.086; for locus T12 from -0.0006 to 0.034; for locus Slep9 from -0.0006 to 0.023; and for locus Slep 3 from -0.0005 to 0.019 (Appendix II). F S T values of the locus T16 revealed a trend of genetic heterogeneity of the China Creek locality and several differences involving the Gibralter Island locality (Appendix II). The locus C S L 9 revealed a trend of genetic heterogeneity of the Gibralter Island and Dodger Channel sites from all other localities (Appendix II). Only the locus T12 showed similar trends as seen in pairwise multilocus G-tests and pairwise multilocus F S T comparisons (Appendix II). Both Slep9 and Slep3 revealed genetic homogeneity among localities with a few exceptions (Appendix II). r Using all five loci, combined loci F S T values of pairwise comparisons of localities were lower than those obtained with individual loci and ranged from -0.0009 to 0.017 (Table 2.14, 2.15). Overall, after adjusted table-wide alpha levels, 14 of 136 pairwise comparisons are 84 statistically significant. At adjusted alpha levels the locality Gibralter Island is significantly different from 13 localities, China Creek significantly different from three localities, and Snug Basin different from three localities. Nine significantly different locality-pairs involved either at least one island locality and in five comparisons, both were island localities. These results were generally similar to those found with exact tests of genetic differentiation. Pairwise F S T comparisons within the fjordal habitat of the Alberni Inlet (East Barkley Sound) are low ranging from -0.0 - 0.0008. A trend of increasing differentiation, however, is evident as pairwise comparisons increase by one and two orders of magnitude in the majority of pairwise comparisons of Alberni Inlet localities with more distant localities in the southern and northern regions of Barkley Sound. Within the Broken Group Islands, localities showed a high degree of differentiation from other localities within the archipelago however appear to be genetically similar to localities within nearby Loudoun Channel (North Barkley Sound) and Trevor Channel/Deer Group area (South Barkley Sound). Of 136 locality-pair comparisons, 26 have multilocus F S T values equal to or greater than 0.01 and distances among localities ranged from 3.8 km between laques-Iarvis/Gibralter Islands and 66.0 km between China Creek/Turret Island. Of these 26 pairs, 15 had distances separating them of 3.8 - 32.9 km and 11 had distances separating them of 33.0 - 66 kilometres. 85 Table 2.14: Multi-locus F S T values between pairwise localities using five loci. F S T values are below the diagonal and geographic distances in kilometres between localities are above the diagonal in italics. Calculations were performed using GENEPOP1.3 (Raymond and Rousett 1995) with 10,000 permutations. Site abbreviations are as given in Table 2.1. CC NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI cc 11.9 32.7 32.2 36.3 55.2 54.9 58.2 49.4 45.5 42.1 83.1 76.3 65.7 .64.8 61.8 66.0 NB 0.000 20.8 18.9 23.0 41.9 4T6 46.3 37.5 33.6 28.8 71.2. 64.4 53.8 52.9 49.9 54.1 SB 0.001 -0.001 12.9 17.0 35.9 35.6 38.9 29.2 22.5 22.8 50.4 43.6 32.9 32.1 29.1 33.3 Al 0.007 0.010 0.005 4.1 23.1 22,7 26.0 16.3 12.4 9.9 63.7 52.1 37.7 37.7 37.5 35.9 NUM 0.002 0.006 0.002 -0.003 18.9 , 18.6 21.9 14.0 10.0 15.0 59.5 47.9 33.5 33.6 33.4 31.7 Gl 0.004 -0.001 0.004 0.001 0.004 5.4 8.7 9.9 11.4 29.7 45.6 40.1 33.0 18.0 15.7 20.6 BM 0.005 -0.004 0.000 0.003 0.002 -0.001 8.4 9.6 13.4 31.7 45.4 39.8 32.8 23.3 21.1 26.0 DC 0.013 0.011 0.012 0.003 -0.003 0.014 0.004 10.5 14.5 32.7 35.8 28.9 19.2 13.4 10.6 13.4 FL 0.011 0.001 0.007 0.011 0.004 0.006 0.000 0.004 3.9 22.1 34.0 27.1 15.0 11.8 9.8 15.6 PC 0.008 0.010 0.008 0.000 ' 0.001 0.004 0.003 0.006 0.006 18.3 35.3 26.7 15.9 13.4 11.4 17.5 Ul 0.008 0.009 0.009 0.001 0.002 0.010 0.001 0.000 0.009 0.002 41.2 34.3 23.6 23.7 19.9 28.0 TB 0.008 0.007 0.004 0.000 0.003 0.006 0.002 0.010 0.004 0.002 0.002 6.9 17.6 20.2 24.1 21.5 SI 0.008 0.006 0.011 -0.001 0.001 0.004 0.004 0.003 0.004 0.003 0.002 0.000 10.6 12.1 15.9 12.6 PI 0.003 -0.002 0.001 0.002 -0.001 0.002 -0.004 0.004 0.002 0.003 0.003 0.001 0.002 6.7 8.6 8.3 JJ 0.004 0.000 0.003 0.005 0.000 0.001 0.001 0.002 0.004 0.004 0.005 0.003 0.000 -0.001 3.8 8.3 GIB 0.012 0.010 0.013 0.015 0.009 0.004 0.007 0.017 0.013 0.012 0.013 0.010 0.010 0.010 o.or 9.5 TI 0.011 0.004 0.005 -0.002 0.001 0.003 -0.001 0.011 0.004 0.003 0.004 0.001 0.001 0.002 0.01 0.01 0 Table 2.15: Probability (P) values of F S T estimates between pairwise localities, for combined loci. Nominal P values.are below the diagonal and above the diagonal is noted significance of comparisons after sequential Bonferroni correction (*). P values less than or equal to 0.05 is noted with an (#). Calculations were performed in FSTAT 2.9.3.2 (Goudet 2001) with 10,000 permutations. Site abbreviations are as given in Table 2.1. oc NB SB Pi NUM a EM . DC FL PC U TB SI PI JJ GB Ti oc ' NB NB NS NS NS # # # * # # # NB NB * * NB Q82757 NB NB NS NS NB NS NB # NS NB NS NS NS * NS SB Q17425 0.43015 NB NS NS NS # NB * # NB * NS NS * # A) 0.06912 0.5739 0.31103 NS NS NB NS NS NS NS NB NS NS NB # NS NUV 0.32022 0.82206 0.42203 0.4335 NB NS NS NB NS NS NS NS NS NB # NS Gl 0.18971 0.54081 Q2864 0.8085 0.54265 NS NS NB NS NS NB NS NB NS NS NS EM 0.0133 0.80772 0.15809 0.4537 0.66765 0.80919 # NB NS NS NB NS NS NS # NB DC 0.00515 0.19154 0.00184 0.2768 0.43419 0.04081 0.38382 NB NS NS NS NS NS NS # # FL 0.01471 0.63162 0.08051 0.1956 0.64706 0.17537 0.68301 0.36287 NS NS NB NS NS NS * NS PC 0.00O37 0.02363 0.00037 0.3184 0.5636 0.33493 0.31912 0.19449 0.13897 NS NB # NS # * NS U 0.0489 0.70809 0.00368 0.1636 0.32721 02864 0.3136 0.48272 0.18971 0.1S963 NB NS NS NS * # • TB Q00735 0.07279 0.15037 0.3629 0.34926 0.14044 0.34228 0.12574 0.4239 0.38493 0.42794 NS NB NS * NS a Q00735 0.34779 0.00037 0.6816 0.51434 0.33456 0.1261 0.12279 0.09375 0.02096 0.2864 0.44154 NS NS * # PI Q21066 0.88125 0.075 0.5015 0.95551 0.59632 0.52353 0.3989 0.88676 0.47831 0.24375 0.43088 0.41912 NB # NS JJ Q05843 0.32353 0.11507 0.4268 0.57941 0.35735 0.2239 0.29412 0.0386 0.00588 0.0614 0.1 0.23787 0.6331 * # GIB Q00037 0.00037 0.00037 0.0018 0.00074 0.02059 0.00074 0.00074 0.00037 0.00037 0.00037 0.00037 0.00087 0.0018 0.00037 * TI Q00037 0.15404 0.00478 0.6243 0.17243 0.14632 0.42647 0.0011 0.18235 0.07243 0.00441 0.25478 Q0386 0.2868 0.00331 0.00037 86 Using F S T A T software, F S T values were not different among geographic regions (P = 0.58) and ranged from 0.000 for East Barkley Sound, 0.003 for South Barkley Sound, and 0.005 for North Barkley Sound (Table 2.16). Genetic variance among all island localities ( F S T = 0.007) was higher than among mainland coastal groups ( F S T = 0.003) but this difference was not statistically significant (P = 0.20)(Table 2.16). As well , dividing island groups into subgroups representing the Deer Group Islands ( F S T 0.006) and the Broken Group islands ( F S T = 0.007), relative F S T values were, again, higher than among mainland coastal localities but no statistical difference among these groups was detected (P = 0.75)(Table 2.16). Relative F S T differences suggest connectivity among mainland coastal is greater than that among island localities. Table 2.16: Comparison of F S T values of groups of localities classified into geographic regions and coastal and island habitat types as well as distinct archipelagos. Calculations performed using F S T A T 2.9.3.2 (Goudet 2001) and 15,000 permutations. Site abbreviations are as given in Table 2.1. Division Fst Two Sided P value Localities in each group Regions East Barkley Sound South Barkley Sound North Barkley Sound 0.000 0.003 0.005 CC, NB, SB Al, NUM, Gl, BAM, DC, FL, PC, Ul 0.6 TB, SI, PI, JJ, GIB, TI Habitat Coastal Islands 0.003 0.007 0.2 CC, NB, SB, Al, NUM, Gl, BAM, Ul, TB, PI DC, FL, PC, SI, JJ, GIB, TI Coastal Deer Group Islands Broken Group Islands 0.003 0.006 0.007 0.7 CC, NB, SB, Al, NUM, Gl, BAM, Ul, TB, PI DC, FL, PC JJ, GIB, TI 87 Overall, similarities between the multilocus Fisher's exact tests on allele frequencies and multilocus F S T analysis included evidence for a genetic subpopulation of pipefish located in the area of the Alberni Inlet (East Barkley Sound) and high levels of genetic differentiation involving the localities China Creek and Gibralter Island with other localities. A s well, both test sets revealed that sites within the Broken Group Island archipelago are significantly genetically differentiated from each other. Genetic homogeneity between South and North Barkley Sound was more strongly revealed with F S T analysis than exact tests of allele frequency differences yet trends were similar. Population substructure was also investigated using analysis of molecular variance to test the hypotheses of subregional allele frequency differences. Testing the hypothesis of a subpopulation located within East Barkley Sound, little microsatellite variation was detected when testing localities within East Barkley Sound and the rest of the study area (0.33% of the total variance, P = 0.074) (Table 2.17). After removing the locality Nahmint Bay from the East Barkley Sound grouping, the test neared significance (P = 0.058) and slightly more variation was attributed to this division (0.39%). The hypothesis of allele frequency differences among the three geographic regions was tested but no significant partitioning of genetic variance was detected (0.12%, P = 0.29) (Table 2.17). The hypothesis of allele frequency differences among mainland coastal habitat types and island habitat types was tested but no significant partitioning of genetic variance was detected (0.28%, P = 0.08). After further testing for allele frequency differences among mainland coastal habitat types and two archipelagos (Deer Group Islands and the Broken Group Islands) no significant partitioning of genetic variance was detected (0.26% , P = 0.14) (Table 2.17). 88 Table 2.17: Hierarchical analysis of the distribution of genetic diversity in the eastern Pacific bay pipefish populations under various hypotheses. Calculated using ARLEQUIN (version 2.0) (Schneider et al. 2000). Vbg represents the percentage of variance existing between groups; Vap, represents the percentage of variation among populations within groups; and Vwp is the percentage of variation existing within populations. The stated P value refers to the probability that the observed value for Vbg is equaled or exceeded by chance determined from 10,000 permutations. Localities included in groups are indicated below. Comparisons Vbg Vap Vwp P value Groups East Barkley Sound (CC, NB, SB) v. All Other localities 0.33 -0.12 99.79 0.074 East Barkley Sound (CC, SB) v. All other localities 0.37 -0.14 99.77 0.058 East Barkley Sound (CC, NB, SB) v. South v. North Barkley Sound 0.12 -0.11 99.9 0.29 Mainland Coastal Habitats v. Island Habitats 0.28 -0.18 99.89 0.081 Mainland Coastal Habitats v. 0.26 -0.22 99.96 0.14 Deer Group Islands v. Broken Group Islands Localities within each group: East Barkley Sound Alberni Inlet: C C China Creek; NB Nahmint Bay; SB Snug Basin South Barkley Sound: Trevor Channel: Al Assits; NumB Numukamis Bay; Gl Grappler Inlet; BAM Deer Group Islands: DC Dodger Channel; FL Fleming Isl; PC Port Alberni Yacht Club North Barkley Sound: Louden Channel: TB Toquart Bay; SI Stopper Islands; PINKS Pinkertons Broken Group Islands: JJ Jaques/Jarvis Island; GIB Gibralter. Island, TI Turrett Mainland Coastal Habitats: CC, NB, SB, Al, NUM, Gl, BAM, Ul, TQ, PINKS 89 Principal component analysis Projection of localities in principal component space (Figure 2.4) suggested significant differentiation between two subgroupings, the Alberni Inlet localities and all other localities. East Barkley Sound sites form a distinct cluster and show a high genetic variance from all other localities. Gibralter Island was the most distinct locality and grouped with no other locality. These results are similar to those resolved with exact tests and F S T analysis. No component axes were found to be significant using five multilocus F S T values (PCI inertia 16.02, P = 0.126, F S T = 0.003; PC2 inertia 12.3, P = 0.069, F S T = 0.002) (Figure 2.4). Clustering results using combined F S T values with the three loci T16, C S L 9 and T12 were similar but both axes were found to be significant (PCI inertia 22.78, P = 0.0.0017, F S T = 0.004; PC2 inertia 17.52, P = 0.009, F S T = 0.003). Up the fjordal arm of Alberni Inlet, the Snug Basin locality clusters more closely with China Creek than with Nahmint Bay and this configuration was also revealed with F S T analysis. Within the Broken Group Islands National Park Reserve, a trend of high genetic variance among the three island localities (Jaques/Jarvis Island, Gibralter Island, and Turret Island) indicates low connectivity among reserve localities in contrast to closer clustering of these localities with proximal localities Toquart Bay, Stopper Islands and Pinkertons all located outside the border of the marine reserve. The large cluster defined by Trevor Channel localities (Bamfield Inlet, Grappler Inlet, Numukamis Bay and Assits), Deer Group Island localities (Dodger Channel, Fleming Island and Port Alberni Yacht Club), Useless Inlet, Broken Group Islands localities and Loudoun Channel localities (Toquart Bay, Stopper Islands and Pinkerton Islands) was also revealed by exact tests and F S T analysis. Although genetic variance is low within this large cluster, there was a pattern of clustering along horizontal P C axis 1 revealing higher gene flow among some proximal localities and a 90 general relationship between genetic similarity and geographic proximity. Assits (4), Dodger Channel (8), Port Alberni Yacht Club (10), and Useless Inlet (11) form a distinct cluster and may indicate a stepping stone for movement of pipefish between islands to a coastal habitat in this geographic area of Barkley Sound. Two other clusters of spatially proximate localities are Toquart Bay (12), Stopper islands (13), and Turret Island (17); and the Pinkertons (14) and Jaques-Jarvis Islands (15). Localities such as Grappler Inlet, Bamfield Inlet, Fleming Island, and Numukamis Bay did not show a correspondence between geographical proximity and genetic similarity. BARRIER analysis The B A R R I E R analysis of multilocus F S T values and choosing two barriers, revealed that the highest order barrier, or strongest area of restricted gene flow (noted as "a") divided the Broken Group Islands/Loudoun Channel and the Gibralter Island locality from the rest of the study area by first drawing a barrier between localities Turret Island (17) and Gibralter Island (16) from localities Dodger Channel (8) and Fleming Island (9). A barrier was then constructed dividing Gibralter Island (16) from the Deer Group localities (8, 9) and from Broken Group localities (17 and 15) (Figure 2.5). The second strongest barrier (noted as "b") resulted in divisions separating the Alberni Inlet localities creating the Alberni Inlet group as well as a large group composed of localities in Trevor Channel (4, 5, 6, 7), the Deer Group Islands (8, 9, 10) and Useless Inlet (11). Overall, these results are similar to the exact tests, F S T analysis and P C A - G E N analysis. 91 Fst 0.003 Figure 2.4. Results of principal component analysis of allele frequency variation in Barkley Sound pipefish assayed at five microsatellite loci . Putative groups of subpopulations are noted within ellipses. Dotted ellipse indicates a spatially contiguous group of localities. Calculations were performed using P C A G E N 1.2 (Goudet 1999) and 15,000 permutations. 1 China Creek; 2 Nahmint Bay; 3 Snug Basin; 4 Assits; 5 Numukamis Bay; 6 Grappler Inlet; 7 Bamfield Inlet; 8 Dodger Channel; 9 Fleming Island; 10 Port Alberni Yacht Club; 11 Useless Inlet; 12 Toquart Bay; 13 Stopper Islands; 14 Pinkertons; 15 Jaques/Jarvis Islands; 16 Gibralter Island; 17 Turrett Island 92 4 9 * 20* 4 9 * 10* 4 9 * OO' 48* 50* 48* 4 0 ' - 1 2 5 * 3 0 * - 1 2 5 * 2 0 * - 1 2 5 * IO" - 1 2 5 * OO* -124* 50* - 1 2 4 * 12 Toquart Biy 13 Stopper Isk 14 pmkertonfch 15 Jaques/Jamis Isis It OteiUn&l 17 Turret! Isl llt/selcs <: "J*ih* l * * " £ NumuVamis Bay .£3'^ jf*6 Grappler Inlet (-''TEarnfieldlrOet — " V ^ 8 Dodger Channel 0 Fleming lsl 10 Port Alberni Yacht Club X 4 9 * IO* 4 9 " OO* 48* SO* 4 8 4 0 ' - 1 2 5 * 30" - 1 2 5 * 20* - 1 2 5 * IO* - 1 2 5 * OO* - 1 2 4 * 5 0 * - 1 2 4 * 4 0 * A . B. Figure 2.5: Analysis of areas of restricted gene flow as identified by BARRIER (Manni et al. 2004) set to two barriers and using pairwise multi-locus FST estimates between localities, spatial coordinates and the Monmonier algorithm (Manni et al 2004). A . represents the study sites according to spatial coordinates in latitude and longitude. B. represents a BARRIER map where the outer borders are determined by Voronoi tessellation of study site spatial coordinates in latitude and longitude (black enumerated dots [1-17]) and the inner borders the Delauney triangulations among sites. Edges outlined in heavy dashed lines represent barriers to gene flow determined by the Monmonier algorithm and (a) represents the first barrier detected and (b) the second barrier detected. 93 STRUCTURE analysis The program S T R U C T U R E using models A , B(i) and B(i i ) revealed that the most likely number of populations (K) was 1 with any of the three models (Model A , Model B i , and Model B i i ) as calculated by likelihood probabilities (0.59, 0.7, and 0.7 respectively) and log likelihood values (Table 2.18). Probabilities of K = 2 populations over the three models were 0.22; 0.27 and 0.27 respectively. Probabilities of K = 3 populations were 0.14; 0.0 and 0.0 respectively. Results of analyses run excluding the locality Gibralter Island did not change these results. Alpha values were more stable and less variable with Mode l A . Alpha values did not exceed 1 when K - 2, or 3; however exceeded 1 when K > 4. S T R U C T U R E is noted to have difficulty resolving genetic substructure when levels of genetic variance are low as indicated by fluctuating alpha levels (Pritchard et al. 2000). Individual-based analyses: Assignment tests Assignment probabilities of individuals to their sampling localities were estimated using the Rannala-Mountain model in G E N E C L A S S 2 (Rannala and Mountain 1997). Overall, assignment success was low and ranged from 0 (Numukamis Bay and Pinkertons) to 53% (Gibralter Island)(Table 2.19). Overall, 16% of individuals were assigned to their sampling locality. 94 Table 2.18: Mean likelihood scores and standard deviation of S T R U C T U R E (Pritchard et al. 2000) for hypothesized populations (K) of Barkley Sound pipefish inferred from variation at five microsatellite loci . Model parameters were: Model A admixture/correlated allele frequencies over 5 runs; Model B(i) prior population identified/correlated allele frequencies over 5 runs; and Model B(i i) prior population identified/correlated over 10 runs. The probability of ' K ' test is as set out by Pritchard et al 2000. A v e r a g e A v e r a g e Var ia t ion Probabi l i ty Rat io Tes t K Ln(P(D)) Ln(P(D)) of 'K ' Tes t Stat is t ic P v a l u e Mode l A 1 -17979 .76 71 .16 0.59 Admix tu re 2 -19062 .98 2539 .9 0.22 -0 .12 N S 3 -19555.6 3858 .5 0.14 -0 .17 N S 4 -20676.9 6282 .02 0.05 -0 .28 N S Mode l B (i) 1 -17976 .98 58.6 0 .73 Pr ior P o p n Id 2 -17984.4 93 .04 0.27 0.00 N S 5 runs 3 -18026.54 273 .06 0.00 -0.01 N S 4 -18406.4 1108.22 0.00 -0 .05 N S Mode l B (ii) 1 -17977.61 65 .53 0.73 Pr ior P o p n Id 2 -18079 .7 303.51 0.27 -0.01 N S 10 runs 3 -18706 1655.8 0.00 -0 .08 N S 95 Table 2.19: Results of G E N E C L A S S 2 (Piry et al. 2004) assignment of individual fish to their location of sampling. Calculated using the Rannala and Mountain (1997) model. N u m b e r of N u m b e r of an im a l s an im a ls se l f - P e r c e n t a g e se l f -Loca l i t y s am p led a s s i g n e d a s s i g n e d C C 37 9 24 N B 25 3 12 S B 44 6 14 A l 17 1 6 N M 39 0 0 G l 30 1 3 B M 30 1 3 D C 48 4 8 F L 45 6 13 P C 64 16 25 U l 43 3 7 T Q 50 7 14 SI 43 8 19 PI 34 0 0 J J 83 7 8 G I B 79 42 53 TI 39 6 15 To ta l 120 16 Each misassigned fish was placed in a distance (kilometer) category denoting the distance from the locality of capture to the misassigned locality. There was a statistically significant pattern of declining numbers of misassigned individuals with increasing geographic distance from locality of capture (X ( 0 . 0 5 , 8 ) 422, P < 0.001) (Figure 2.6 (A)). The cumulative percent frequency curve of this data set revealed that a high percentage (81%) of misassigned individuals were found to be most similar to fish within 3-40 kilometres of their sampling locality and 90% within 3-50 kilometres. The remaining 10% of misassigned fish were assigned to study localities 51-87 kilometres from their site of capture (Figure 2.6 (B)). The 1 geographical distances between study localities within a geographic region (intra-regional distance) are 32.7 kilometres (East Barkley Sound), 26 kilometres (South Barkley Sound), 24.1 kilometres for North Barkley Sound, and 64 kilometres for South/North Barkley Sound (Table 2.14). The maximum intra-regional geographic distances together with the cumulative percent frequency of misassignment of 81% between 3 - 5 0 kilometres or 90% between 3 - 6 0 kilometres suggests that pipefish migration occurs among individual eelgrass beds at distances less than 40 or 60 kilometres. Assignment probabilities of individuals to their geographic region of sampling (East, South, or North Barkley Sound), depending on the proximity of the sampled locality to these geographical areas, were reassessed in a separate G E N E C L A S S 2 analysis. Of all fish assigned, 62.5% were assigned to their region of sampling and 37.5% were assigned as migrants to regions (Table 2.20). East Barkley Sound had a self-assignment level of 73% and received 13% of its migrants from South Barkley Sound and 17% of its migrants from North Barkley Sound. South Barkley Sound had a self-assignment level of 62% and received 15% of its migrants from East Barkley Sound and 23% of its migrants from North Barkley Sound. North Barkley Sound had a self-assignment level of 60% and received 12% of its migrants from East Barkley Sound and 25% of its migrants from South Barkley Sound (Figure 2.7 and Table 2.20). East Barkley Sound had the highest self-recruitment level, received the lowest number of immigrants, and exported the least number of emigrants. South Barkley Sound and North Barkley Sound had similar self-recruitment levels, received the highest number of immigrants, and exported a similar number of emigrants. The highest migration level of fish between regions was between South and North Barkley Sound. These trends are similar to results of allelic frequency analysis with exact tests and F $ T analysis and principal component analysis. 97 Table 2.20: Results of G E N E C L A S S 2 (Piry et al. 2004) assignment of fish to geographic regions of sampling. Results divided into percentage of fish assigned to geographic region of capture and assignment from other regions. Calculated using the Rannala and Mountain (1997) model. Site abbreviations are as listed in Table 2.1. East %of Percentage South %of Percentage North %of Percentage Barkley regional of Total Barkley regional of Total Barkley regional of Total Sound total Study Sound total Study Sound total Study Self assigned 77 73 10 196 62 26 196 60 26 Regional Total 106 316 328 Migrants from other areas Total from EBS 42 13 56 17 Total from SBS 16 15 Total from NBS 13 12 78 25 76 23 Total Percentage Self Assigned Over All Regions 62.5 East Barkley Sound: CC, NB, SB South Barkley Sound: Al, NUM, Gl, BAM, DC, FL, PAYC, Ul North Barkley Sound: TB, SI, JJ, GIB, TI 98 * r ?f s i (fl i n •31 E -art "c n» v- C :?? E £ 200 180 160 140 120 100 80 60 40 20 175 89 130 114 59 26 30 •# >a ^ ^ £ Distance(kilometres) from locality of capture to locality of mis assignment A . I u L a CO Bi. 0) .a £• f= <*= 13 r= .SP 8-(A 1 100 90 80 70 60 50 40 30 20 10 0 99 100 TOO •J* *P -*f K# J .*•> ^ & - \ s <bs Distance (kilometres) from locality of capture to locality of misassignment B . Figure 2.6: Relationship between geographic distance (kilometres) and the number of Barkley Sound pipefish misassigned (from location of capture) to location of assignment. (A) Number of fish misassigned categorized by the distance (kilometers) between the locality of capture and the locality of misassignment. There was a statistically significant pattern of declining numbers of misassigned individuals with increasing geographic distance from the locality of capture (X 2 ( 0 .o5,8)422, P < 0.001). (B) Cumulative percent frequency curve of misassignment showing that 81% and 90% of misassigned pipefish were within 40 and 60 kilometres, respectively, of the capture locality. 99 25% Arrow thickness indicates degree of genetic exchange Sel f assignment Figure 2.7: Genetic connectivity estimates by migration levels of Barkley Sound pipefish among geographic regions of Barkley Sound. Estimated using G E N E C L A S S 2 and the Rannala and Mountain (1997) model. Localities within East Barkley Sound are C C , N B , S B ; within South Barkley Sound are A I , N U M , G l , B A M , D C , F L , P C ; and within North Barkley Sound are T B , SI, PI, JJ, G I B , TI. Site abbreviations are as given in Table 2.1. 100 Geographical distances between localities situated on either side of a geographical region (inter-regional distance) ranged considerably, from 9 km (South - North Barkley. Sound) to 83.1 km (East - North Barkley Sound). Reclassification of these inter-regional pairwise geographic distances, revealed that the majority of pair-wise distances between East -South Barkley Sound were greater than 35 kilometres and this was also true for East - North Barkley Sound (Figure 2.8). The geographical regional G E N E C L A S S 2 analysis revealed that the lowest levels of inter-regional migration occurred with East Barkley Sound (Figure 2.7 and Figure 2.8). The majority of pair-wise distances between South and North Barkley Sound were less than 35 kilometres and this was reflected in the highest levels of inter-regional migration of 24% and 25% (Figure 2.7 and Figure 2.8). Overall, inter-regional migration levels of fish increased as distances between pair-wise localities decreased. 101 = e .= o. 8-1 8 " a 1 ! = t i •35 -25 -15 10 5 0 F13-15% 14 14 • E - S E- N-34 14 S-• < • >35km Distance (kilorTietres) categories within Inter-Regional Grouping Level of j migration * Between | Inter-regional group " I Figure 2.8: The number of pair-wise localities within inter-regional groupings (N=North Barkley Sound, E=East Barkley Sound, S = South Barkley Sound) in each of two distance categories (< 35 km, > 35 km) separating these pairs of localities. The number of inter-regional, pairwise locality comparisons in each distance category is indicated above each bar. The levels of inter-regional migration were determined using G E N E C L A S S 2 (Piry et al. 2004) and are presented in Figure 2.7. Inter-regional migration levels were highest between S - N Barkley Sound which has the greatest number of pair-wise localities < 35 kilometres. 102 Isolation-by-distance A Mantel test for correlation of logio transformed genetic distance by logio transformed geographic distance over all localities failed to find a positive correlation (z = -491.21, r = 0.076, P = 0.2505)(Table 2.21 and Figure 2.9). Using partial correlation analysis of pairwise genetic distance and two independent variables of geographic distance and habitat type, there was a significant correlation for logio genetic distance by logio geographic distance after taking into account the effect of habitat type ( r = 0.19, P < 0.038, total correlation of r = 0.196) (Table 2.21 and Figure 2.9). Table 2.21: Mantel and partial Mantel tests of the correlation of genetic distance and factors of geographic distance and habitat types. Data were logio transformed and tests were conducted using I B D (Bohonak 2002 software with 10,000 randomizations. Mante l / Partial Mode l Mantel A Genetic Distance x Geographic Distance Z t P value r2 Genet ic distance x geographic d istance -491.2 Control l ing for habitat type 0.07 0.19 0.25 0.04 0.005 0.036 B Genetic Distance x Geographic Distance of specific locality categories Coas ta l x Coas ta l -142.6 Island x Island -54,42 0.33 -0.24 0.027 0.81 0.11 0.058 103 "O oo ca T— o> u- o o O 0.02 0.018 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0 •0.002 -0.004 •0.006 • t • * • • • • • • # • 20 40 60 80 Distance (kilometres) A . 1.6 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 •0.1 A £ A . A . A * / .AA A ••Mfc A * A A A t • t . A A A» • «A A Island x Island A Island x Coasta l • Coasta l x Coas ta l 4.9 5.2 5.5 5.8 6.1 Distance (Transformed Log 1 0 Kilometres) B . Figure 2.9: Isolation by Distance„of eastern Pacific bay pipefish sample localities. ( A ) represents pairwise F S T (Weir and Cockerham 1984) values and kilometers (r=0.076); and (B) represents log-o transformed F S T ar>d kilometer values. Labels of data points indicate habitat category. 104 Pipefish genetic distance sampled in continuous eelgrass bed habitats along coastlines were tested using a Mantel test for correlation of logio genetic distance by the logio of geographic distance for all pairwise comparisons involving only the sites China Creek, Nahmint Bay, Snug Basin, Assits, Numukamis Bay, Grappler Inlet, Bamfield Inlet, Useless Inlet and Toquart Bay. The Mantel test revealed a. strong correlation suggesting that gene flow along coastlines of continuous eelgrass beds follows an isolation by distance or stepping-stone model (Z = -142.60, r = 0.33, P = 0.027) (Figure 2.10 and Table 2.21). Pipefish sampled in eelgrass patches within island localities (localities separated by a water barrier) were tested using a Mantel test for correlation of the logio genetic distance by logio geographic distance among pairwise neighbouring island sites Dodger Channel, Fleming Island, Port Alberni Yacht Club, Stopper Islands, Jaques-Jarvis, Gibralter and Turret Island. The Mantel test revealed that gene flow among island localities is not explained by a stepping-stone dispersal model (Z = -54.42, r = -0.24, P = 0.81). Further subdividing the data set, Mantel tests were conducted to evaluate the relationship among genetic distance, geographic distance and deep-water channels. After logio transformations of F S T values and geographic distances, a Mantel test of genetic distance by deep-water channels effect found a significant correlation (Z = -1348.93, r - 0.235, P = 0.02); and a partial Mantel found a significant correlation (r = 0.226, P = 0.03) holding geographic distance constant. A partial Mantel test revealed no correlation between genetic distance and geographic distance (r = 0.042, P = 0.36) holding putative deep-water barrier types constant (Table 2.22 and Figure 2.11). 105 10 20 30 40 50 60 70 80 90 Distance (kilometres) between localities A . Coastal Habitat Types c CO k_ '53 in LL 00 o> CD u o o 0.02 0.018 0.016 ' 0.014 0.012 -0.01 • 0.008 -0.006 0.004 0.002 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Distance (kilometres) between localities B . Island Habitat Types Figure 2.10: Mantel tests of correlation between genetic distance and geographic distance of coastal and island habitat types. Calculations performed using I B D software and 10,000 permutations. (A) represents the Mantel correlation test of coastal habitat types (r=0.33, P = 0.027) and (B) represents island habitat types (r = -0.24, P = 0.81). 106 Table 2.22: Mantel and partial Mantel tests of the correlation of genetic distance and factors of geographic distance and putative deep-water barrier types. Data were logio transformed and tests were conducted using I B D (Bohonak 2002) software with 10,000 randomizations. P Model Z r value Mantel Genetic Distance x Water Barriers -1349 0.23 0.02 0.05 Partial Mantel Controlling for geographic distance 0.226 0.03 0.05 107 0.02 0.015 oo CD E (0 o o o TJ c (0 CO 0.01 0.005 -0.005 • • • • • * • • • • • • • • • * • * • • • • • • • • i • $ A l • • • W • i : • I • i » 1 • • • • • i « * • • • • • • • • • 1 2 3 1 = coast to coast within region 3 = island to island within regions 5 = island to island among regions 2 = coast to coast among regions 4 = coast to island within regions 6 = coast to island among regions Figure 2 .11 : Multi-locus F S T values between pairwise localities by categories used in the Mantel test for the effect of putative deep-water barriers on genetic distance. The distance of deep-water channels ranges from no deep-water channel separating localities in groups ( 1 ) and (2) to the greatest distance between localities separated by a deep-water channel in (6 ) . 1 0 8 2.4 Discussion 2.4.1 Microsatellite variation within populations Overall, there was little indication of significant linkage disequilibrium between pairs of loci; and it appears that pipefish groups at the 17 localities are in Hardy-Weinberg Equilibrium. Departures from Hardy-Weinberg Equilibrium were not found at any of the loci resolved using the 3 2 P detection system. Two tests involving the locus Slep9 showed significant linkage disequilibrium. There were difficulties resolving weak allele peaks for Slep9 and Slep3 using the fluorescently labeled primers and the C E Q 8000 auto analyzer; and some known heterozygotes were scored for one allele and the others as missing. This wi l l upwardly bias the data to a heterozygote deficiency. It is l ikely that these results are not due to a true heterozygote deficiency but is due to difficulties in scoring individuals. If heterozygotes were deficient at these 10 localities, other loci should show significant deviation from Hardy-Weinberg Equilibrium but this was not the case. Another explanation is the chance that these violations of Hardy-Weinberg Equilibrium are due to non-random mating. As pipefish are likely to choose their mates, non-randomness of mating due to this behaviour may be a factor (Berglund et al. 1986). Again, other loci should also show significant deviations from Hardy-Weinberg Equilibrium. No significant difference was observed between observed and expected combined heterozygosities. Overall, these results indicate that outbreeding is l ikely occurring at each sampled locality (Hedrick 2000). F i s ranged from -0.005 (Turret Island) - 0.088 (China Creek). The low F ; s values associated with each locality also substantiates the finding of outbreeding. Within localities, the high observed heterozygosity indicates that gene diversity within 109 individuals is high. The high gene diversity found for this species (He > 0.78) is consistent with the average gene diversity reported for marine fish (DeWoody and Avise 2000). 2.4.2 Fine-scale population genetic structure ofthe eastern Pacific bay pipefish Based on the evidence provided in this study, Syngnathus leptorhynchus in Barkley Sound/Alberni Inlet do not form a single panmictic unit but instead form two genetic subpopulations with some areas of high genetic isolation within the Broken Group Islands due to restricted gene flow. Genetic differentiation was weak but statistically significant within the confines of this fine-scale spatial study where the maximum distance between localities was 83.1 km. There is evidence for closed population dynamics over the largest spatial area of the study consistent with geography. Al le l ic richness differences, allelic frequency analyses, principal component analysis of genetic variation and A M O V A results support a finding of weak genetic divergence and restricted gene flow of post-glacially founded populations. The finding of two genetic subpopulations as well as examples of extreme genetic isolation of the locality Gibralter Island promotes the idea of restricted gene flow. Due to the highly specialized habitat of the eastern Pacific bay pipefish and its unique morphological adaptations to eelgrass, I predicted that greater genetic divergence would be found. High connectivity among habitats is, however, a prediction of metapopulations where animals living in highly fragmented habitats are, generally, good dispersers (Harrison and Hastings 1996). Hardy-Weinburg Equilbrium tests, high heterozygosities among localities and low F , s estimates revealed high outbreeding by the animals sampled in this study likely due to large population sizes and perhaps the fact that these are post-glacially founded populations. Inferred high levels of gene flow among localities suggests that together they represent a patchy genetic metapopulation (Harrison and Hastings 1996). 110 From my data, I suggest that the first genetic subpopulation is situated in the eastern reaches of Barkley Sound within the Alberni Inlet in samples taken from China Creek, near Port Alberni, to Snug Basin a geographic distance spanning approximately 32.7 k m and termed the East Barkley Sound population. The second genetic subpopulation is contained within the geographic regions of South and North Barkley Sound and encompasses a maximum distance between sampled localities of 63.7 kilometres. Gene flow of the eastern Pacific bay pipefish appears to be restricted by numerous seascape features. N o difference in reproductive timing between East and South/North Barkley Sound was detected and the subpopulations appear to undergo similar trends with respect to recruitment and reproductive timing (de Graaf, unpublished data). The finding that gene flow is restricted is important (Palumbi 1994) as it indicates that there are physical or biological factors that impact gene flow at small spatial scales in this temperate, nearshore, marine fish species. A s the eastern Pacific bay pipefish is a good candidate for use as an indicator species of eelgrass ecosystem integrity, this is an important finding in view of management of local eelgrass ecosystem function and marine reserve design. As factors limiting gene flow in marine environments are not as well understood as in terrestrial or freshwater habitats, this study may provide evidence for seascape level processes restricting gene flow in nearshore, temperate marine habitats (Palumbi 1994, Gold and Turner 2002). Marked differences in pairwise comparisons of exact tests on allele frequencies and significant pairwise comparisons of F S T values, A M O V A , and principal component analysis revealed that the three East Barkley Sound localities China Creek, Nahmint Bay and Snug Basin are significantly divergent from all other localities within the study area. The genetic similarity of proximal regions South and North Barkley Sound was evident from multi-locus, pairwise comparisons of exact tests and F S T analyses. Numerous pairwise comparisons 111 involving East and North Barkley Sound, however, were significantly different from each other. Al le l ic richness tests revealed no statistical difference in this measure between the geographically proximal East and South Barkley Sound geographic regions and a statistical difference between East and North Barkley Sound. G E N E C L A S S 2 assignment tests among geographic regions also revealed decreased gene flow between East Barkley Sound and North Barkley relative to migration levels with South Barkley Sound. As well , East Barkley Sound had the highest self-assignment rate of 73%. Misassignment analysis revealed that levels of migration between East Barkley Sound and South/North Barkley Sound may be less favoured due to geographic distance as misassignment rates were lowest from 61-87, kilometres and most localities between these regions (inter-regional localities) are separated by distances greater than 35 kilometres. The B A R R I E R analysis also suggested the uniqueness of East Barkley Sound due to restricted gene flow. The A M O V A analysis of difference between East Barkley Sound and all other localities neared significance (P = 0.058) after removal of the Nahmint Bay locality. The Nahmint Bay locality was removed as this eelgrass bed was highly degraded by log booming, and during seining, few pipefish were obtained and a low sample size has likely resulted in an under-representation of allelic variation at that site. G E N E C L A S S 2 analysis of regional assignment indicated overall a high percentage of self assignment within regions (62.5%) and gene flow patterns that support the idea of greater similarity between South and North Barkley Sound relative to East and North Barkley Sound. S T R U C T U R E results, however, did not support a finding of two genetic subpopulations. S T R U C T U R E models must be chosen carefully and information generated during runs closely considered when scrutinizing results. During numerous runs, fluctuating alpha values 112 (Dirichlet parameter for the degree of admixture) were common between runs suggesting a lack of population structure (Pritchard et al. 2000). S T R U C T U R E results are highly influenced by the level of genetic divergence among samples and the number of loci used (Pritchard et al. 2000; Waples and Gaggiotti 2006). In this study global F S T was low ( F S T = 0.005), loci highly polymorphic and the number of loci was only five. Similar difficulties utilizing S T R U C T U R E with low genetic divergence and fluctuating alpha values have been encountered by other authors (Coulon et al. 2005). Ultimately the low level of differentiation found in this study may be the reason for the failure of S T R U C T U R E to detect population structure although other analyses used infer restricted gene flow. Assignment methods such as S T R U C T U R E have been demonstrated to be effective when gene flow is low to moderate ( N m < 5) and the number of loci is high (20) and the number of individuals is 50. Results drop dramatically when these ideal conditions are not met (Waples and Gaggiotti 2006); and other statistical tests need to be considered when ascertaining population substructure, a precaution noted by the authors of the S T R U C T U R E software (Pritchard et al. 2000). In some cases, more traditional multilocus contingency tests (such as Fisher's exact tests) were better able to detect true population substructure than the S T R U C T U R E program, when gene flow was moderate to high ( N m = 25) (Waples and Gaggiotti 2006). The most homogenous area of this study was the geographic regions of South and North Barkley Sound and together comprises a subpopulation of the eastern Pacific bay pipefish. The homogeneity of these two areas is revealed by exact tests, F S T analysis, non-significant A M O V A results, and the high levels of migration revealed by G E N E C L A S S 2 regional assignment tests. Misassignment analysis revealed exchange between South and North Barkley Sound is favoured due to geographic proximity as misassignment rates are highest from 3 - 4 0 km, and these two regions have the highest number of pairwise inter-113 regional localities situated less than 35 kilometres apart. In contrast, allelic richness comparisons revealed a significant difference in this measure between South and North Barkley Sound groupings as North Barkley Sound has lower allelic richness relative to both East and South Barkley Sound. As well, the B A R R I E R analysis divided North Barkley Sound and Gibralter Island from all other localities. It is possible that restricted gene flow is occurring but insufficient time has passed since the last glacial retreat to allow a complete genetic discontinuity between the areas. Imperial Eagle Channel, a large, deep-water channel, may serve to restrict gene flow while connectivity along mainland coastal sites from Toquart Bay to Useless Inlet allows sufficient exchange of individuals. The most genetically divergent localities were situated in the Broken Group archipelago. Comparisons with other taxa While marine fish species generally reveal lower relative levels of genetic differentiation than other categories of fishes (Gyllensten 1985, Ward et al. 1994, DeWoody and Avise 2000), the finding of even low levels of genetic differentiation in marine systems has implications for reduced population exchange in the absence of obvious physical barriers to gene flow (Palumbi and Warner 2003). The F S T values in my study ranged among single loci from 0.001 to 0.008 and over five combined loci of 0.005, and in individual pairwise tests higher .values such as F S T = 0.017 were found. Ward et al. (1994) surveyed 57 marine species and reported a mean and median F S T of 0.062 (+/- 0.011) and 0.020 respectively. Waples (1998) pointed out that 60% of these marine species surveyed by Ward et al. (1994) had F S T values less than 0.03 (ranging from greater than 0.000 to 0.029). These estimates of genetic structure indicate that marine species show relatively low levels of significant genetic differentiation over large spatial scales (eg between oceanic basins and along entire continental 114 coastlines) which is consistent with the assumption of high dispersal for most marine species. Among marine fish populations at small spatial scales, these lower F S T values are not unusual, and low F S T values may be significant indicating important genetic differentiation due to less obvious barriers to gene flow (Balloux and Lugon-Moulin 2002). On a larger geographic scale, the coast-wide analyses of phylogeography of Syngnathus leptorhynchus by Louie (2003) (Alaska to Baja, C A ) and Wilson (2006) (Alaska to California), reported F S T values for S. leptorhynchus that were comparable to the mean and median reported by Ward et al. (1994). Comparisons of F S T values from my study, Louie (2003) and Wilson (2006) to those reviewed by Ward et al. (1994) poses some difficulty due to the fine spatial scale of my study. Microsatellite loci used in my research were highly polymorphic with the number of alleles ranging from 10 - 46 and an average of 31.2 over five loci . Highly polymorphic loci and high heterozygosity values downwardly bias F S T values results (Hedrick 1999, Balloux and Lugon-Moul in 2002) both features present in this study; and care should be taken to apply the results of F S T analyses with other relevant analyses to deduce population substructure. The highly polymorphic nature of the markers used as well as high heterozygosities may be important contributing factors to the low degree of population substructure found in this study when greater population genetic substructure was predicted. Highly variable loci together with unequal sample sizes can increase genetic variance and decrease the ability of tests to detect population structuring (Waples 1998, Hutchinson et al. 2001, Balloux and Lugon-Moulin 2002). Balloux and Lugon-Moulin (2002) reiterate the advice of Waples (1998) with respect to high-gene flow species in relying too heavily on only statistical evaluation of population structure when biologically and geographically disjunct units may exist undetected by the molecular loci and tests used. 115 Louie (2003) reported statistically significant population structure, isolation by distance and major phylogenetic breaks for Syngnathus leptorhynchus over 3,600 kilometres from Alaska to Baja, California in an analysis using mitochondrial D N A (mtDNA). The strongest phylogenetic break was between northern and southern clades between Padilla Bay, Puget Sound, W A and Gray's Harbor, OR, a well recognized phylogenetic break between northern and southern aquatic species due to glacial refugia history of this part of the west coast (Louie 2003). This break represents a straight-line distance of approximately 200 kilometres between Padilla Bay and Gray's Harbor. In other areas of her study, however, Louie (2003) also reported high gene flow between localities located 560 kilometres apart. In a study of S. leptorhynchus with microsatellite loci including loci Slep 3 and Slep 9 used in this study, Wilson (2006), found a statistically significant FST of 0.006 in comparison of a Washington State locality and Oregon State locality, a linear distance of approximately 300 kilometres. The Australian hairy pipefish, Urocampus carinirostris, a seagrass endemic, displayed isolation by distance of genetic divergence and a clinal secondary intergradation of m t D N A clades over a 130-km stretch of coastline resulting in the separation of two distinct northern and southern clades (Chenoweth et al. 2002). Although seahorses, in general, are not capable of powerful locomotion (Lourie and Vincent 2004), a comparison of this group, of syngnathids with this study may be worth while as the literature is lacking in population level studies of both pipefishes and seahorses. Over a spatial area of approximately 6,600 k m 2 within the reefs and islands of the Camotes Sea, Philippines, the seahorse species Hippocampus comes revealed an FST of 0.0016 and expected heterozygosity of 0.928 over three microsatellite loci (Casey 1999). N o statistically significant genetic divergence among H. comes samples was found (Casey 1999). The spatial area, habitats sampled (complex reef structures) and FST values of the H. comes study are similar to those of my study. In contrast, The Knysnea seahorse, H. 116 capensis, was sampled in three spatially contiguous estuaries (26 k m and 42 km apart) and m t D N A revealed distinct populations in each estuary (Teske et al. 2003). Within Barkley Sound, several studies of population structure have spanned a wide variety of taxonomic groups. Population structure of the shrimp Pandalus jordani and Pandalopsis dispar were investigated in Imperial Eagle Channel and Trevor Channel, Barkley Sound using age composition and parasite tags (Thompson and Margolis 1987). .Evidence for discrete populations was found at small spatial scales both within and between channels (Thompson and Margolis 1987). P. jordani formed separate populations in Imperial Eagle Channel, Trevor Channel and offshore; P. dispar formed separate populations in Imperial Eagle Channel and Trevor Channel/Offshore (Thompson and Margolis 1987) distances of less than 15 kilometres. These animals are found at depths ranging from 35-137 m and shallower depths represent discontinuous habitat (Thompson and Margolis 1987). Forming a chain running northeast to southwest, the Deer Group Islands archipelago separates Trevor Channel from Imperial Eagle Channel. Passages between the islands are narrow and shallow shelves of depths of 20 metres are likely barriers to population mixing (Thompson and Margolis 1987). No apparent bathymetric barriers to movement exist to explain the finding of discrete populations within Imperial Eagle Channel (Thompson and Margolis 1987). The giant kelp Alaria marginata was surveyed within Trevor Channel and genetic subpopulations were detected at a distance of 8 kilometres (Kusumo and Druehl 2000). Localized net tidal current flow was invoked as a process to explain the genetic divergence of this nearshore kelp species as well as directionality of gene flow. The tidepool sculpin, Oligcottus maculosus, an abundant, nearshore marine fish species strongly associated with its habitat due homing behaviour showed no statistically significant population structure with 117 m t D N A restriction fragment length polymorphism (RFLP) assays throughout Trevor Channel island and coastal habitats sampled (Altman and Taylor 2003). Over wide geographic scales, marine fish of varying life-history traits display significant population structuring (Daemen et al. 2001 (European eel), Hutchinson et al. 2001 (Atlantic cod), Appleyard et al. 2001 (Yellowfin tuna), Jorgensen et al. 2005 (Atlantic herring)). Hutchinson et al. (2001) noted that in Atlantic cod, Gadus morhua, genetic connectivity was detected among coastal, locally geographically proximal localities in the United Kingdom as well as significant divergent of these localities with more distal North and Barents Sea localities. Local genetic divergence was. found to be non-significant yet significant on a larger spatial scale. Hutchinson et al. (2001) suggested that local spawning populations have limited migration on a larger spatial scale. Atlantic cod populations along the Canadian east coast were found to show significant genetic divergence among geographically distant, continental shelf localities (Ruzzante et al. 2000). In this study, sonie coastal localities revealed evidence of localized gene flow, and reduced offshore migrations (Ruzzante et al. 2000). Ruzzante et al. (2000) suggested that these Gilbert Bay fishes provided evidence of barriers to gene flow among discrete stocks and that these barriers maintained genetic divergence between coastal, bay resident-populations and nearby coastal and offshore populations. Ruzzante et a/.(2000) showed that significant population substructure can occur at only 31 km among coastal populations. Both these studies of Gadus morhua revealed restricted migration and closed population dynamics at local scales of a fish species typically associated with large-scale offshore migrations. Marine fishes with smaller scale migrations and habitat ranges also exhibit genetic population structure. Marine coastal species can be confined to habitat types due to specializations and physical limitations, and this is often reflected in population genetic structure. The benthic 118 dwelling cockle Cerastoderm glaucum is limited to lagoons, salt marshes and brackish lakes due to physiological intolerances (Mariani et al. 2002). A planktonic disperser, C. glaucum revealed high population differentiation within the Mediterranean Sea (over 1000s of kilometers) and realized larval dispersal was considerably less than predicted (Mariani et al. 2002). Analysis of genetic distance revealed a regionality of genetic population structure within various basins of the Mediterranean Sea (Mariani et al. 2002). Overall, genetic populations were structured by isolation by distance and dispersal was restricted by the lack of suitable habits and habitat discontinuity across an ocean barrier (Mariani et al. 2002). Also residing within the Mediterranean Sea, the sand smelt, Atherina boyeri, is an anadromous fish with strong natal homing behaviour during spawning to brackish lagoons discontinuous along coastlines (Congiu et al. 2002). Juveniles use these lagoons as rearing habitats (Congiu et al. 2002). Spawning adults along the Italian coast revealed high levels of genetic differentiation predicted by geographic proximity (isolation by distance) and a likely barrier to gene flow across open sea tracts with localized currents (Congiu et al. 2002). Genetic divergence among localities across tracts of open sea was significantly higher than among coastal localities revealing the importance of a barrier to gene flow in increasing genetic divergence (Congiu et al. 2002). Aphanius fasciatus is a brackish-water, live-bearing, marine kill if ish (cyprinodontid) of the Mediterranean Sea which spends most of its sedentary life-cycle in coastal estuaries and has life-history traits which would predict limited dispersal (Maltagliati 1999). Using allozyme markers, A. fasciatus regional populations were characterized by high genetic divergence among regions over a wide regional spatial scale (Central Mediterranean) and followed an isolation-by-distance model (Maltagliati 1999). Analyses revealed high genetic connectivity among neighbouring sites inferring self-recruitment at smaller spatial scales (Maltagliati 1999). 119 Pipefish and some rockfish species can be classified as nearshore, coastal marine species or species limited to a narrow, shallow-water habitat along coastlines. The brown rockfish, Sebastes auriculatus, an eastern Pacific, nearshore coastal species with demersal adults and pelagic larvae, is a coastally adapted species limited to a narrow, linear strip of shallow-water habitat (Buonaccorsi et al. 2005). Larvae and juveniles settle into bays and estuaries to depths of approximately 100 metres out of the way of prevailing offshore coastal currents (Buonaccorsi et al. 2005). These habitats are afforded hydrodynamic protection by kelp forests and rocky reefs and larvae are thought to be constrained by a proposed boundary layer of "sticky water" thought to dominate 1-3 km out from the shore (Buonaccorsi et al. 2005). S. auriculatus shows high genetic divergence among populations sampled from Puget Sound, W A to Baja, Mexico ( F S T = 0.056 over 1,500 kilometres) and after removal of the highly divergent site in Mexico, overall FST using six microsatellite loci was 0.009 (Buonaccorsi et al. 2005). In a study of the grass rockfish, Sebastes rastrelliger, a species associated with nearshore marine habitats including eelgrass beds, and sampled from outer coastal Washington State sites to California, F S T was 0.001 (over approximately 1000 km), although significant population subdivision was not detected (Buonaccorsi et al. 2004). At smaller, local spatial scales, there are numerous examples of highly divergent coral reef fish populations (Shulman and Bermingham 1995, Planes et al. 1996, Planes et al. 1998, Taylor and Hellberg 2003, Hoffman et al. 2005). Recently, Taylor and Hellberg (2003) found genetically distinct island populations of the cleaner goby, Elacatinus evelynae, separated by only 27 km, and Hoffman et al. (2005) describe significant population subdivision of the Indonesian Banggai archipelago cardinal fish, Pterapogon kauderni, at distances of 2 kilometres. 120 M y study has revealed that populations of the eastern Pacific bay pipefish, founded no more than 8,000 years B P (McMi l l an 1999) - 13,000 years B P (McPhai l and Lindsey 1986) exhibits significant genetic divergence within archipelagos at fine-spatial scales of 3.8 kilometres. While the last inter-glacial period is reported to have ended approximately 13,000 years B P (McPhail and Lindsey 1986), sea levels in Barkley Sound were approximately 90 metres higher than today and did not reset to approximately 3 metres above current levels until approximately 8,000 years B P (McMi l l an 1999). A discussion of processes driving this pattern can begin with examining isolation by distance models and local seascape features. Isolation-by-distance in the eastern Pacific bay pipefish Genetic distance was not explained by geographical distance until the factor of habitat type was controlled in a partial Mantel analysis. Genetic distance was positively correlated with geographic distance of mainland coastal habitat types as revealed by an r of 11%, but genetic distance was not correlated with geographic distance of island habitat types. Other differences between coastal and island habitat types appear evident as Fisher's Exact Tests and F S T analysis revealed that neighbouring, mainland coastal localities were genetically homologous indicating high levels of gene flow, while in contrast, neighbouring, island localities exhibited genetic heterogeneity. Relative F S T values between pairwise coastal localities are two-times lower than seen between island localities although the A M O V A analysis of these two habitat types was not significant. Mainland coastal localities peripheral to each other such as China Creek and Toquart Bay, however, were significantly divergent by F S T analysis. Even in the face of high migration rates, genetic divergence can occur i f there are factors to limit gene flow such as such as distance-dependent gene flow (Hedrick 2000). 121 Under an isolation-by-distance model ("stepping stone model") mating is non-random and adjacent localities experience higher gene flow relative to more distant localities. Stepping stone gene flow would result in neighbouring localities being genetically similar with genetic divergence occurring among geographically distal and peripheral localities (Hedrick 2000). Stepping stone dispersal maintains variation within populations and reduces among population differentiation (Taylor et al. 2003), both attributes evident in my study with pipefish having high heterozygosities within localities and weak subpopulation differentiation among localities. In my study, coastal habitats represent discontinuous eelgrass beds arranged in a linear manner along a shallow depth contour favourable to eelgrass growth; and island habitats represent sites separated by deep-water unfavourable to eelgrass growth. Coastal habitat types may facilitate gene flow due to increased connectivity and abundance of habitats organized in a linear, stepping stone pattern along the coastline within the study area. A species l iving in different landscape types can have differences in genetic distance measures due to different habitat types, and this is seen in other fish species that utilize diverse habitat types. Westslope cutthroat trout (Oncorhynchus clarkii lewtei) is a freshwater fish species l iving in semi-isolated populations within British Columbia's interior. In contrast with coastal cutthroat trout (O. c. clarkii) microsatellite-based assays showed marked differences in relative F S T measures (Taylor et al. 2003). Westslope cutthroat trout had F S T values ranging from 0.08 to 0.45 whereas the coastal cutthroat trout had consistently lower F S T values ranging from 0.03 to 0.12 (Taylor et al. 2003). Coastal cutthroat trout populations may have higher connectivity due to the nature of the habitat and their anadromous behaviour (Taylor et al. 2003). Among salmonids, it has been noted.that among isolated populations, average microsatellite based F S T was 0.27 (0.14 - 0.37) (Hendry et al. 2003). Isolation of habitats can decrease connectivity and increase relative FST values relative to habitats that are either closer together or have fewer 122 barriers to dispersal. Island habitats may tend to isolate pipefish leading to F S T values higher than predicted by geographic distance. A pattern of isolation-by-distance should be more distinct in a species distributed along a linear gradient i f dispersal is less than the geographic range of the species (Hellberg 1994). In his study of the marine brooding cup coral, Balanophylla elegans, Hellberg (1994) reported a significant pattern of isolation-by-distance along a latitudinal gradient along the west coast of North America. Geographic distance accounted for 15% of the between-locality pattern of genetic differentiation between cup coral localities samples (Hellberg 1994). Louie (2003) reported a statistically significant correlation between genetic distance and geographic distance 2 for Syngnathus leptorhynchus and an r of 12% along a coastal linear gradient from Alaska to 2 Baja, Mexico using m t D N A . The r of 12% (Louie 2003) is similar to what was found in this current study,- of 11%, using microsatellites and a much smaller spatial scale. Genetic population structure of populations of the seahorse H. comes over 140 kilometres was found to follow an isolation by distance model (Casey 1999). Patterns of isolation by distance have been detected in numerous coastal fish species. Buonaccorsi et al. (2004) investigated isolation by distance in samples of the grass rockfish, Sebastes rastrelliger, from Oregon to California and found a high correlation between genetic distance and geographic distance with an r2 of 58% (Buonaccorsi et al. 2004). For the brown rockfish, Sebastes auriculatus, an r2 of 90% was found for genetic distance by geographic distance along coastal habitats from Oregon to California (Buonaccorsi et al. 2005). For the Australian estuarine sparid, Acanthopargrus butcheri, a fish completing its entire lifecycle within estuaries, and sampled along coastal estuaries using m t D N A revealed an isolation by distance relationship and a r2 of 70% (Burridge et al. 2004). Planes et al. (1996) found that genetic subdivision of the coral reef fish, Acanthurus triostegus, could be predicted 123 by an isolation-by-distance model within an archipelago with a dominant local gyre (r 2 of 71%). Other studies of coastal marine fishes with genetic subdivision following an isolation-by-distance processes include the Mediterranean sand smelt, Atherina boyeri, and the Mediterranean killifish, Aphanius fasciatus (Congiu et al. 2002, Maltagliati 1999). Isolation-by-distance between coastal habitat types plays an important role in eastern Pacific bay pipefish gene flow among- coastal localities at both small (my study) and large spatial scales (Louie 2003). Other processes, however, are needed to explain the incomplete gene flow resulting in divergence among Broken Group Island localities as well as between the genetic subpopulations of East Barkley Sound and South/North Barkley Sound. Isolation-by-distance and fjord habitats As previously stated, the area of most genetic divergence was between the regions of Alberni Inlet (East Barkley Sound) and South/North Barkley Sound. Physical attributes of the Alberni Inlet fiord are somewhat different from Barkley Sound. The Alberni Inlet is a steep-sided fiord 37 kilometres long, with an average width of 1000 metres, depths up to 300 metres, and has major freshwater sources at its head and along its course (Stronach et al. 1993). The Alberni Inlet lacks the feature of archipelagos and where eelgrass is present, is mainly a linear, coastal eelgrass habitat. Eelgrass beds are likely more limited in a fjordal system due to steep contour lines (bathymetry) and water depths limiting the photic environment as well as limiting the transport of bottom sediments important for eelgrass establishment. The linear nature of the fjordal habitat is likely to promote a geographic distance dependent (stepping-stone) model of gene flow. Isolation-by-distance analyses of only coastal habitat types revealed the importance of geographic distance in structuring genetic variance among coastal habitat types. ^124 East Barkley Sound (Alberni Inlet) coastal localities are on the periphery of the study area, and stepping-stone gene flow may be significant in maintaining connectivity with other localities to the south as well as maintaining local homogeneity. Alberni Inlet is heavily affected by estuarine current flow (Doe 1952) and the surface, brackish water creates a lower salinity environment relative to South/North Barkley Sound (Doe 1952, Thomson 1981, Stronach et al. 1993). Recently, gene flow of Atlantic herring has been shown to be restricted between a boundary of temperature stable, high salinity marine waters in the North Sea and a temperature variable, brackish water environment in the Baltic Sea (Bekkevold et al. 2005, Jorgenson et al. 2005). While differences between these salinity environments may restrict pipefish recruitment success, pipefish species, as other estuarine fish, are generally able to tolerate wide salinity gradients (Power and Attr i l l 2003), although no specific studies on Syngnathus leptorhynchus salinity tolerances have been conducted. Thermal tolerances of pipefish species, however, have been suggested to limit pipefish abundance in estuaries with seasonally, colder water temperatures (Power and Attr i l l 2003). Not only are Alberni Inlet eelgrass beds arranged in a linear manner, eelgrass habitat is limited in this region of Barkley Sound. This attribute may account for the high self-assignment rate of 73% within this geographic region. Eelgrass beds of significant area size were surveyed and only found in areas of river outlets and small pocket coves. Small , fringing eelgrass beds likely also exist along narrow, favourable depth contours. Nahmint Bay eelgrass has been heavily impacted, by log booming operations and a low number of pipefish were found inhabiting this area. No eelgrass was found near the Somas River estuary (Port Alberni) (personal observation) as this area has been heavily industrialized. Eelgrass in other favourable areas, such as Uchucklesit Inlet near Snug Basin, has been lost due to historical canneries and wharves, log booms, private docks, resort marinas, and aquaculture 125 development. Due to these fjordal characteristics and industrial and recreational development, although this area has not been extensively mapped for eelgrass habitat, there is likely less eelgrass area in the Alberni Inlet relative to South/North Barkley Sound. 2.4.3 Effect of seascape features on genetic subdivision of the eastern Pacific bay pipefish, Syngnathus leptorhynchus Archipelagos With pairwise F S T analyses and exact tests as well as principal component analysis, all three Broken Group Island localities were found to be significantly genetically differentiated from each other and the Gibralter Island locality genetically differentiated from all other localities in this study. The three localities of the Broken Group Islands archipelago are the most genetically divergent area of this study revealing local heterogeneity at small spatial scales from 3.8 to 9.5 kilometres. This is.in stark contrast to the relative genetic homogeneity detected in South/North Barkley Sound pair-wise comparisons of localities at larger geographical distances. The principal component analysis revealed that these localities do not cluster together, but instead they cluster with geographically proximal localities within Loudoun Channel indicating low connectivity within the marine reserve and high connectivity with sites outside of the marine reserve. As well, the BARRIER result was informative in confirming this area of restricted gene flow. A l l analyses indicate a complex scenario with respect to contemporary connectivity by gene flow of pipefish within and outside of the Broken Group Islands that is, perhaps, mirrored by the complex oceanographic and land form 126 features of this archipelago. Al le l ic richness within the North Barkley Sound region was found' to be significantly lower than allelic richness within both the South and East Barkley Sound regions (Table 2.6). Pipefish populations within the reserve may rely on genetic exchange with localities outside of the marine reserve and a geographic area greater than represented by marine reserve boundaries. For example, Jaques/Jarvis Islands and the mainland, coastal locality of the Pinkertons are separated by Peacock Channel (depths 40 m) and Sechart Channel (depths up to 40-60 m) and a distance of approximately 6.7 km. Pairwise F S T between Jaques/Jarvis Islands and the Pinkertons was negative and exact tests were not significant suggesting high gene flow between these two localities. Turrett Island is well connected by gene flow to the distal, mainland coastal locality of Toquart Bay (21.5 kilometres) and Stopper Islands (12.6 kilometres). Yet Jaques/Jarvis Islands and Turrett Island are geographically proximal to each other (8.3 kilometres) and are 3.8 kilometres and 9.5 kilometres, respectively, from Gibralter Island. Pairwise F S T values between each of these population pairs were 0.010, 0.010, and 0.010. These intra-archipelago F S T values were twice that of the overall F S T value of 0.005 and up to one magnitude higher than pairwise comparisons with other localities in this study. Archipelagos can disrupt and complicate local hydrographic flows (Johnson et al. 1994; Watts and Johnson 2004) making movement among islands difficult while processes affecting coastal, mainland habitats may enhance pipefish migration and potential gene flow. Overall, the effect of archipelagos on genetic subdivision can be to create local heterogeneity at small geographic distances within a larger geographic region of genetic homogeneity (Johnson et al. 1994, Planes et al. 1996; Watts and Johnson 2004). Johnson et al. (1994) recognized that genetic subdivision of the atherind fish Craterocephalus capreoli within archipelagoes is largely independent of geographic distance and that genetic divergence among mainland 127 coastal localities over much larger spatial scales was generally eight times lower than that seen within archipelagoes. Interestingly, pipefish gene flow within the Broken Group Archipelago appears to follow a similar pattern of significant local heterogeneity at small spatial scales; and outside of the marine reserve also appears to follow a pattern of homogeneity. Genetic divergence among island localities in my study is independent of geographic distance as revealed by a negative slope of genetic variance ( F S T ) by geographic distance (Figure 2.10 B) and a non-significant Mantel test for correlation (Table 2.21). Among Broken Group Island localities, genetic divergence is likely driven by restricted gene flow resulting in the process of vicarance at micro-geographic scales. Among islands and coastal pairs, dominate currents as well as coastal connectivity of eelgrass habitats may serve to maintain large-scale homogeneity. Examples of an archipelago isolating groups of animals were also found in this study with the example of the Gibralter Island locality. The Gibralter Island locality is genetically distinct from its archipelago and from all other localities in this study. Gibralter Island individual pairwise F S T values equaled or ranged up to 10-fold higher than the global F S T estimate (0.004-0.017), and it may represent a distinct genetic subpopulation. The distinctiveness of Gibralter Island was also seen in population differentiation and F S T tests a n d morphologically with a unique, low frequency colour morph (de Graaf, unpublished data). As well , Gibralter Island has more rare alleles and has one private allele relative to the Jaques/Jarvis Islands and Turret Island localities; and Gibralter Island had the highest self-assignment rate of 52% in G E N E C L A S S 2 assignment analysis. A s the Gibralter Island site indicates, in an archipelago with an abundance of eelgrass habitat, the complex configuration of islands and passages in this area may lead to significantly reduced levels of connectivity among eelgrass beds leading to significant genetic divergence at distances of only 3.8 km. 128 The brown rockfish sampled in locations within Puget Sound showed a similar effect of isolation from other rockfish populations (Buonaccorsi et al. 2005). Pairwise F S T values of 0.013 between two Puget Sound brown rockfish populations (Buonaccorsi et al. 2005) were similar to F S T values of 0.010 between two Broken Group Island localities of the eastern Pacific bay pipefish found in my study. Buonaccorsi et al. (2005) suggest that the glacial sill situated at the mouth of Puget Sound together with depressed rockfish stocks due to over harvesting are likely factors responsible for the high level of genetic subdivision. Both Planes et al (1998) and Hoffman et al. (2005) report high pairwise F S T values 0.055 and 0.27, respectively, between several island populations of coral reef fishes. A t two - five kilometres, the Banggani cardinal fish study reports the highest degree of population subdivision between population pairs ( F S T = 0.06-0.22) measured by F S T analysis with microsatellite loci over microgeographical scales (Hoffman et al. 2005), values similar to pair-wise F S T of 0.010 between pipefish Broken Group Island localities at 3.8 kilometres and 0.017 between pipefish island localities (Dodger Channel and Gibralter Island) at 10.6 kilometres. M y study differs from that of Planes et al. (1998) in that the age of pipefish populations, due to geological instability, is on the order of thousands of years whereas Planes et al. (1996) reported geological stability of French Polynesia of millions of years. A s well , both Hoffman et al. (2005) and Planes et al. (1996) reported pairwise F S T comparisons as being significant at the nominal P = 0.05 while the current study has applied the highly conservative sequential Bonferroni correction (Hutchinson et al. 2001) to pairwise F S T values at a table-wide alpha level of 0.000368. The level of divergence of the eastern Pacific bay pipefish populations sampled in this study, however, can not be expected to reach that of many tropical fish populations primarily due to their relatively recent age of founding due to the post-glacial history of recolonization of Syngnathus leptorhynchus along the eastern Pacific coast. As 129 well, the high level of heterozygosity of S. leptorhynchus imposes a theoretical F S T limit of 0.10 that is below the level of divergence documented by Hoffman et al. (2005). M y current study is the first to document the effect of archipelagos on population subdivision in a post-glacially founded, temperate, near-shore fish species. Deep-water channels and genetic divergence Mantel and partial Mantel tests of genetic distance and putative water barrier category provided weak but significant evidence for a correlation between increasing genetic distance and the presence of deep-water channels. The B A R R I E R analysis revealed a restricted area of gene flow between North Barkley Sound and the rest of the study area. Also , relative F S T values tend to increase with increasing deep-water channel width. These results suggest that deep-water channels may act as barriers and play a role in structuring genetic differences perhaps by restricting migration and potential gene flow among regions as revealed by G E N E C L A S S 2 analysis. The effect of deep-water channels contributing to increased genetic differentiation between localities has been revealed in numerous studies of marine fishes. Deep-water tracts separated coastal and island populations of the sand smelt, Atherina boyeri, resulting in higher genetic differentiation than revealed between coastal populations (Congiu et al. 2002). A similar result was reported for populations of the Mediterranean kil l if ish, Aphanius fasciatus, separated by deep-water tracts (Maltagliati 1999). A s well , populations of the rocky, intertidal blennioid fish, Axoclinus nigricaudus, separated by deep open waters were more genetically distinct than those separated by continuous habitat (Riginos and Nachman 2001). 130 Spatial area size of breeding groups ofthe eastern Pacific bay pipefish A useful measure of the function of ecologically connectivity of eelgrass beds, from the perspective of the pipefish, would be a measure of the spatial area required to maintain breeding groups and the geographic distance of inter-patch gene flow. A t smaller spatial scales, local demes within a genetic subpopulation support lower numbers of fishes relative to the total number of fishes throughout a larger spatial area; and interbreeding could result in a finding of higher genetic similarity locally than in the larger study area (Gold and Turner 2002). Gold and Turner (2002) stated that smaller groupings of fishes can represent "geographic neighbourhoods" of interest to conservation planning due to the importance of maintaining functional breeding groups of animals. Misassignment tests revealed that pipefish are most likely to be assigned to a locality within 40-60 kilometres of its site of capture. Principal component and B A R R I E R analysis confirmed the genetic distinctiveness of the East Barkley Sound region. Within a geographic region, distances between sampled localities range from 32.7 km (East Barkley Sound), 26 km (South Barkley Sound), and 24.1 km (North Barkley Sound). The distance between localities in South/North Barkley Sound range from 9.8-63.7 kilometres. According to assignment and misassignment analysis, migration occurs within the spatial range of a region as self-recruitment rates range from 60-73% and inter-patch dispersal is most likely to occur at distances between 3-60 kilometres. Assuming fish breed after migration, localized groups of genetically homogenous pipefish most likely to interbreed can be characterized by a spatial distance of up to 60 kilometres. Migration levels between regions and inter-regional distances (distances separating pairs of localities where each locality is in a different region) vary widely. The highest levels of migration, 24-25%, as determined by G E N E C L A S S 2 analysis was found between South and North Barkley Sound with the majority of its inter-regional distances being 131 less than 35 kilometres again that the potential for gene flow occurs most frequently at distances up to 40-60 kilometres. Levels of migration were lowest between the regions of East-South Barkley Sound and the regions of East-North Barkley Sound likely as a result of fewer inter-regional pair-wise localities separated by distances less than 35 kilometres (Figure 2.8). Principal component analysis also revealed a genetically homologous, spatially contiguous grouping of Dodger Channel, Port Alberni Yacht Club, Useless Inlet and Assits that may constitute a geographic neighbourhood (Figure 2.4). Distances between localities in this grouping range from 12.4 - 32.7 kilometres being similar to the maximum distance between East Barkley Sound localities. These results suggest that inter-patch dispersal resulting in genetic neighbourhoods within a genetic subpopulation may be best characterized at 40-60 kilometres. M y data also suggest that migration does occur at geographic distances greater than 35-40 kilometres, such as between East-North Barkley Sound and South-North Barkley Sound; but migration levels decrease at distances greater than 40-60 kilometres as did the probability of misassignment between regions. Evidence for smaller areas of high localized retention and reduced migration was shown for the red drum, Sciaenops ocellatus, (Gold and Turner 2002). Aphanius fasciatus, the live-bearing, brackish water killifish within the Mediterranean Sea also revealed evidence of clustering of genetically similar and geographically contiguous localities (Maltagliati 1999). The anadromous Mediterranean sand smelt, Atherina boyeri, followed a similar pattern of geographically contiguous localities and genetic homogeneity (Congiu et al. 2002). Although similar analyses are lacking in the syngnathid literature, Louie (2003) found that Syngnathus leptorhynchus gene flow neared panmixia between closely spaced localities and was restricted among other localities. Louie (2003) suggested that eelgrass habitats at geographic distances 132 from 10-20 km would best maintain eastern Pacific bay pipefish breeding dynamics, gene flow and chance dispersal among linear habitats although. This conclusion, however, may differ depending on localized conditions and the extent of discontinuous habitat along open coastlines as is common along the Washington and Oregon coastlines. In my study there is evidence of smaller, geographic neighbourhoods of fish with higher genetic connectivity and reduced migration within the greater spatial area of South/North Barkley Sound. Overall, breeding groups are arranged along geographic distances of 32.7 km within East Barkley Sound and a smaller, localized breeding group within South/North Barkley Sound encompassing a distance of 12 - 33 kilometres. Both my study and that of Louie (2003), however, provide similar evidence for smaller, localized breeding groups of the eastern Pacific bay pipefish within larger genetic subpopulations. Abundances of Syngnathus leptorhynchus within coastal eelgrass beds in Barkley Sound declines in colder, winter months (de Graaf, unpublished data) likely due to freezing and near freezing temperatures of surface fresh water run off. Lazzari and Able (1990) deduced from abundance data, that Syngnathus fucus dispersed from coastal estuaries to deeper waters (approximately 10 - 20 metres depth) 20 kilometres from shore. While no study has been conducted to determine i f the decline in seasonal abundance of S. leptorhynchus is due to migration to a thermal refuge (Power and Attr i l l 2003), a seasonal migration may have implications for the spatial area of gene flow. A s well as seasonal variation in abundance of pipefish, there is variation in the use of eelgrass beds by female and male pipefish. In several studies of different pipefish species, males have been noted to be more site faithful, less active and remain more cryptic in eelgrass beds likely due to increased predation risk and the energetic investment required while brooding the young (Svensson 1988, Steffe et al. 1989, Roelke and Sogard 1993, Vincent et al.. 133 1995). Syngnathus leptorhynchus caught in recapture studies were both male and females (Hornbeck 2004, de Graaf, unpublished data). Females compete for access to males (Berglund et al. 1986, Vincent et al. 1995), are more active swimmers and travel farther distances to find mates (Roelke and Sogard 1993, Vincent et al. 1995). S. leptorhynchus females mate with 2 -3 males (de Graaf, unpublished data) and are more energetic and aggressive toward conspecifics than males (personal observation). Genetic homogeneity and population substructure may be affected by female mediated gene flow (Louie 2003) but this prediction has not been tested. Local currents and inter-patch gene flow Gene flow is wide spread throughout the entire study area maintaining essential connections even between genetic subpopulations throughout a highly fragmented habitat matrix. Pipefish and other syngnathids are thought to be limited in their realized dispersal potential (Vincent et al. 1995, Lourie and Vincent 2004) due to their modified morphology. Within a population, however, movement among eelgrass habitat, patches is essential when resources become limited or when habitat is lost due to storms and disease events known to devastate Z. marina beds (Burdick et al. 1993), and to allow sufficient outbreeding to promote genetic diversity. Further, animals l iving in patchy or fragmented habitats may require a dispersal strategy to allow successful movement among favourable habitats (Harrison and Hastings 1996). Inter-patch migration was found to occur most frequently within 40-60 kilometres. Favourable depth contours connecting coastal habitats promote a stepping-stone model of genetic connectivity among habitats while seascape features such as archipelagos and deep-water channels were found to be important in promoting areas of genetic divergence at micro-geographic scales. Effective mechanisms to allow individuals to migrate may include 134 utilization of localized currents while drift vegetation may provide effective mobile corridors across inhospitable habitats (Highsmith 1985). Eelgrass beds can dampen advection and mixing effects of wind and tidal currents (Worchester 1995). Mark-recapture experiments of this species show that eelgrass bed retention ranges from two to three weeks (Hornbeck 2004) to 2.3 months (de Graaf, unpublished'data). Ut i l iz ing localized currents and rafting with drifting vegetation may be an effective mechanism facilitating either stepping-stone dispersal or longer distance migration to keep patches connected both within a subpopulation and between subpopulations. Pipefish species are highly associated with habitats that allow them to remain inconspicuous to predators (Kendrick and Hyndes 2003). Camouflage during dispersal is important, and many species of pipefishes have evolved to resemble seagrasses for both predator avoidance and prey capture (Howard and Koehn 1985). Louie (2003) noted individual Syngnathus leptorhynchus associated with drift eelgrass blades and this has been reported by Fritzsche (1980) and Dawson (1985). Female pipefish have also been noted vertically aligned with floating eelgrass blades in close proximity to eelgrass beds (personal observation). Evidence for rafting of pipefish in vegetation more distant from shore, however, is lacking. As the majority of S. leptorhynchus individuals are uniformly brown or green (Herald 1941) or have a dorsal stripe and lateral colouration of brown or green (de Graaf, unpublished data), pipefish would blend in well with drift vegetation mainly composed of seagrasses and brown seaweeds. Another important species utilizing currents to promote gene flow is eelgrass, Zostera marina. Genetic subpopulation structure of Zostera marina ranges from 12-42 kilometres (Ruesch et al. (2000) and 54 kilometers (Ruesch et al. 2002) to 150 kilometres (Olsen et al. 2004) due to passive transport of Z. marina seeds. It is intriguing that spatial distance of inter-patch migration and potential gene flow found in my study for the eastern Pacific bay pipefish 135 was within the range of 12-42 kilometres reported for Z. marina. Perhaps this similarity in gene flow estimates between the eelgrass dependent pipefish and Z. marina supports the idea of use of rafting within detached Z. marina blades for camouflage. In areas of Barkley Sound where currents are not disrupted, as along coastlines, Syngnathus leptorhynchus gene flow is high as evidenced by low FST values between localities (eg East Barkley Sound FST 0.00). Principal component analysis revealed that several locations on either side of deep-water channels, including the widest deep-water channel of Imperial Eagle Channel, were genetically similar to each other. Localities within the Deer Group may be connected to coastal localities by stepping stone dispersal, as evidenced from a clustering of island and proximal mainland coastal localities (Riginos and Nachman 2001). In addition, movement across inhospitable areas may be promoted by strong currents regardless of the width of the water barrier. B y contrast, I observed high F S T values among archipelago habitats, but this may be due to disruption of localized currents leading to local retention areas and complex routes to migration. Local currents should bias migration rates to favour movement from East Barkley Sound to South/North Barkley Sound. Levels of migration to and from East Barkley Sound as determined by G E N E C L A S S 2 assignment were, however, similar in both directions. Possibly, migration patterns cannot be predicted by currents because pipefish do not move out of the protective environment of shallower depth counters and into current flows, or that their movements are independent of local currents direction due to swimming abilities and the need to chose favourable patches. Pipefish are hardly passive particles and likely seek their optimal habitats (Monteiro et al. 2005) when placed outside of an eelgrass environment for numerous reasons including evading predators and finding mates (Kendrick and Hyndes 2003). This is further illustrated by the similarity in levels of migration between South and North Barkley 136 Sound. Predominant swell and winds from the west and the northwest should bias drift vegetation or drifting pipefish moving toward southern areas of Barkley Sound. Pipefish migration inferred by G E N E C L A S S 2 analysis between South and North Barkley Sound are, however, similar. While Kusomo and Druehl (2000) stated that kelp spore movement in Barkley Sound was predicted by northwestward net tidal flow (Doe 1952), I did not test the directionality of movement from the genetic data of my study; and there is a paucity of knowledge of tidal and wind driven currents in Barkley Sound. If pipefish do not utilize currents and drift vegetation for long distance movement, a low-risk migration route linking populations is still possible. Pipefish effective movement may reflect that of coastal gene flow models of isolation by distance where favourable depth contours (to approximately 20 metres) limit exposure time to predators in deep-water habitats.. Pipefish may move among favourable habitats in a stepping stone manner (Parson 1996, Barber et al. 2002, Burridge et al. 2004). Eelgrass beds are located along the mainland coast of Vancouver Island throughout Barkley Sound and individual islands in both archipelagos are located close to numerous coastal sites. Principal component analysis revealed the clustering of Deer Group Island localities Dodger Channel and Port Alberni Yacht Club with mainland coastal localities of Useless Inlet and Assits Island, A similar result was found by Riginos and Nachman (2001) where islands facilitated gene flow between blennioid fishes on either side of the northern region of the Gulf of California. 2.5 Conclusion Genetic subpopulations of the eastern Pacific bay pipefish in Barkley Sound are weakly differentiated and still well connected by migration resembling a patchy genetic "mosaic" of populations (Harrison and Hastings 1996). Gene flow and genetic population structure of 137 pipefish over fine-spatial scales are influenced by numerous factors including the spatial arrangement of eelgrass beds within the local seascape. Fjordal environments may restrict gene flow and promote genetic divergence due to features of this environment, reduced habitat sizes and mainly isolation by distance processes. Among coastal habitat types, the process of isolation-by-distance significantly affects genetic divergence among sampled localities. Island habitats promote genetic divergence among localities. The Broken Group archipelago serves to disrupt gene flow creating a region of local heterogeneity at micro-spatial scales relative to genetic homogeneity at larger spatial scales. Archipelagos also serve to isolate individual localities and deep-water channels act to restrict gene flow. Connections between island habitats and coastal habitats promote genetic homogeneity and maintain connectivity between subpopulations. Overall, it appears that knowledge of the influence of seascape features and habitat types on gene flow of the eastern Pacific bay pipefish is valuable for predicting genetic population substructure and connectivity. Migration levels appeared to occur at the highest frequencies at distances from 3 - 6 0 kilometres. Also , G E N E C L A S S 2 data would refute a passive, drifting model of dispersal. It may be that pipefish movements are not predictable by localized currents or that the data from this study are not sufficient to test these hypotheses. Genetic connectivity among localities within the Broken Group Islands, Pacific R i m National Park Reserve (BGI), is restricted and has lead to genetic isolation of localities. Gene flow between these island and mainland coastal habitats outside of the reserve is high. M y results, therefore, strongly suggest that contemporary connectivity of the eastern Pacific bay pipefish in the B G I requires a geographic area greater than currently provided, and there is no protection of coastal habitats geographically proximal to the B G I . Maintaining genetic connectivity of the eastern Pacific bay pipefish within reserve localities to proximal coastal habitats outside of the reserve likely requires maintenance of vital coastal migration corridors. 138 Areas of coastal eelgrass habitat in Toquart Bay and near the Pinkertons should be protected to maintain connectivity between island and coastal habitats. Currently, eastern Pacific bay pipefish gene flow is high between localities inside and outside of the reserve and this should be maintained on the assumption that genetic exchange across reserve boundaries is important to the genetic fitness of population units on both sides of the reserve. High levels of gene flow between North and South Barkley Sound is another important finding of my study. Protecting the genetic resources of pipefish within the B G I wi l l require continued exchange with pipefish located in South Barkley Sound and protection of eelgrass habitats in this southern region is important. Protection of individual genetic subpopulations is important for the maintenance of genetic diversity for local adaptation and the genetic potential of a species. The eelgrass habitats of East Barkley have no formal protection except under sections of the Fisheries Act and, perhaps, face a greater threat of loss due to human disturbance and alteration. Due to the genetic distinctiveness of pipefish in this region, and the unique physical characteristics of these fjordal, coastal eelgrass habitats, protection of these eelgrass habitats should be considered. Genetic studies at fine-scale spatial areas to test the influence of seascape features are important to understand local processes contributing to population substructure and gene flow. Local populations are most at risk from habitat loss due to human interference or natural environmental perturbations. Understanding the role of local oceanography, archipelagos, habitat types and barriers to dispersal on genetic population connectivity (seascape genetics), wi l l greatly aid in designing marine reserves to protect genetic diversity and promote stewardship of surrounding areas outside of reserve boundaries. 139 Chapter 3: General Conclusions and Conservation Implications 3.1 General Conclusions In total, 156 alleles were detected with five microsatellite loci and the average over the five loci was 31.2 ( S T D V 13.9). Genetic diversity was high with an observed multi-locus heterozygosity of 0.91. Both genetic diversity and allelic richness was lowest in the geographic region of North Barkley Sound. Overall, F s t was 0.005 over five loci and 0.007 over three loci. Individual pairwise Fst comparisons ranged from -0.00 to 0.017 I found that the eastern Pacific bay pipefish, Syngnathus leptorhynchus, within Barkley Sound formed two genetic subpopulations with high self-recruitment rates, but also some gene flow between them. Seascape features and habitat types both influenced population genetic substructure. The unique features of fjordal environments appeared to restrict gene flow and promote genetic divergence between the subpopulations. Coastal habitats served to maintain genetic connectivity of the eastern Pacific bay pipefish among eelgrass beds and were important in promoting an isolation-by-distance pattern of genetic differentiation. Archipelagos and deep-water channels appeared to restrict gene flow resulting in increased genetic differentiation. Effective mechanisms allowing individuals to migrate may include utilization of local currents while drift vegetation may provide effective mobile corridors across inhospitable habitats. Levels of inter-patch gene migration occurred at the highest frequencies at distances from 3-60 kilometres and genetic neighbourhoods were estimated to range from 32.7 kilometres to 63 kilometres. M y data suggest that genetic connectivity of the eastern Pacific bay fish among localities of the Broken Group Islands is restricted and has led to increased genetic differentiation in this archipelago. Two Broken Group Island localities, Turrett Island and Jacques/Jarvis Islands, however, are genetically similar to neighbouring outslide of the reserve. 140 The high number of alleles detected may bias FST estimates generated in my study. As well , only five microsatellites were available although pipefish sample sizes were high. In order to increase the power of my study to detect significant population subdivision, increasing the number of microsatellite loci would have been desirable. Due to the recent stabilization of sea levels in Barkley Sound, it is possible that eastern Pacific bay pipefish populations have not reached genetic equilibrium (McMi l l i an 1999). This would affect the isolation-by-distance analysis due to underestimation of genetic differentiation (Hellberg et al. 2002). Although microsatellite loci are generally neutral (Chapter 1), the possibility does exist that selection may be acting on the loci in my study and affect interpretation of the data and hypotheses (Neigel 1997). If selection is responsible for the genetic differentiation between two major subpopulations, the hypothesis of population panmixia and open population dynamics cannot be rejected. 3.2 Conservation Implications Marine reserve design, like terrestrial reserve design, is a complicated task; and objectives of reserves vary depending on the unit to be conserved: the population, the community, or the ecosystem (Zacharias and Roff 2000). A difficult task in designing marine reserve networks is estimating whether or not animals among reserves are connected or form independent populations (Allison et al. 1998). Many marine species, particularly those with planktonic larvae, use different habitats throughout their life cycles and these habitats may have different temporal and spatial scales. For example, planktonic dispersers require habitats for feeding that may be different from habitats of spawning or resident adults. As well, for 141 some fish species, males and females may also use areas differently. Some advocates of marine reserve design propose that marine reserves be set out as a network of reserves spaced close enough together to allow for larval connections to adult habitats (Allison et al. 1998; Palumbi 2003, 2006). As both larval and adult forms need protection, planning reserve sizes and spacing between networks is a challenging endeavour. The science of marine reserve design is a vast topic and a review is outside of the scope of this thesis. M y data can, however, address a few basic principles of reserve design. Marine reserves should encompass source populations and corridors for movement should be spaced appropriately to allow for connectivity among reserves (All ison et al. 1998). Reserves should also encompass home ranges or spatial neighbourhoods of populations. Source populations are those that have a high self-recruitment rate and are not reliant on outside populations for immigrants yet provide emigrants to other areas (All ison et al. 1998). Home ranges are those areas used by a species for a high percentage of activities (Kramer and Chapman 1999) and spatial neighbourhoods, as defined by Palumbi (2004), are areas encompassing juveniles and adults. Home ranges of fish have been determined by direct tagging (Kramer and Chapman (1999) and references therein) while spatial neighbourhoods can also be defined using population genetic data (Palumbi 2004). Obviously, home range sizes and spatial neighbourhoods wi l l vary depending on the species involved and management w i l l have to reflect this variation. Marine reserves designed to protect ecosystems may benefit from using an indicator species approach (Zacharias and Roff 2000). A n indicator species is one deemed to be representative of a community of organisms l iving within the same system (compositional indicator) (Zacharias and Roff 2000) and whose population dynamics are significantly affected by the health of the ecosystem such that they can serve as an indicator of 142 ecosystem integrity (conditional indicator). Fish species have been recognized as appropriate for use as indicators of estuarine health (Fausch et al. 1990). The eastern Pacific bay pipefish, Syngnathus leptorhynchus, is an excellent candidate as an indicator species for monitoring not only because of its dependency on vibrant eelgrass habitats but also because it is a major component of the biomass and functional ecology of eelgrass beds (Adams 1976; Kendrick and Hyndes 2003). Whi le no experimental study has tested for correlations between pipefish abundance and eelgrass quality, long-term monitoring of an estuary in Britain demonstrated that pipefish abundances declined with changes in water temperature and increased pollution levels (Power and Attr i l l 2003); and in a Californian estuary, S. leptorhynchus abundances decreased with decreasing eelgrass bed area (Swift et al. 1993). The eastern Pacific bay pipefish was found to be absent or in low abundance in eelgrass habitats with levels of suspended sediments high enough to reduce visibility in the water column (personal observation). As pipefish are visual predators, a reduction in visibility either due to naturally turbid waters or waters rendered cloudy due to pollution would decrease foraging success for these sit-and-wait predators. In eelgrass beds known to have pipefish present, a significant decline in abundance may indicate a change in the eelgrass environment. Another interesting aspect of choosing an indicator species is the value of it to raise conservation concern among the general public or the "charisma" factor (Zacharias and Roff 2000). The charismatic approach has worked well to raise awareness of the plight of such endangered animals as panda bears, whales, and seahorses. Pipefish and seahorses certainly do capture the imagination of behavioural and evolutionary scientists and of the general public (Foster and Vincent 2004). 143 3.3 Marine reserves for eelgrass ecosystem integrity M y study can address the issue of maintaining functioning source populations of the eastern Pacific bay pipefish. The eastern Pacific bay pipefish shows population genetic substructure within my study area with a maximum distance between localities of 83 kilometres. Within Barkley Sound, I found that pipefish genetic subpopulations have high self-recruitment rates, exhibit high gene flow rates as well as sufficient levels of inter-population migration to maintain the population dynamics most desired of source population networks (Allison et al. 1998). Considering each genetic subpopulation as a source population, the spacing between East and South/North Barkley Sound source populations is approximately 13 kilometres. The home range size estimates to maintain connectivity within a subpopulation (generated by calculating the distance between patches within each subpopulation) ranged from 32.7 kilometres (East Barkley Sound) and 35 - 63 kilometres (South/North Barkley Sound). A home range size estimate generated from misassignment analysis was 40-60 kilometres. These spatial area estimates include island and coastal habitat types. A s with terrestrial and freshwater species, features of the landscape are important considerations when planning to maintain connections among populations as well as protecting corridors. M y data also provide useful information regarding the influence of seascape features in Barkley Sound on the genetic diversity and genetic variation of the eastern Pacific bay pipefish. Habitat heterogeneity is important in shaping genetic differentiation among localities of the eastern Pacific bay pipefish. Some seascape features promoted genetic connectivity while other features served to restrict connectivity. Coastal eelgrass habitats maintain genetic homogeneity and loss of this habitat type may disrupt gene flow potentially isolating breeding 144 groups by fragmenting movement corridors. Island eelgrass habitats benefited from stepping-stone connections to coastal habitats promoting genetic connectivity throughout southern and northern Barkley Sound. Features such as archipelagos and deep-water barriers were found to be important in promoting genetic divergence at micro-geographic scales. Archipelagos can disrupt and complicate local hydrographic flows (Johnson et al. 1994; Watts and Johnson 2004) making movement among islands difficult. Overall, the effect of archipelagos on genetic subdivision is to create local heterogeneity at small geographic distances within a larger geographic region of genetic homogeneity (Johnson et al. 1994; Planes et al. 1996). Fjordal areas, either due to their limited capacity to promote eelgrass beds or due to estuarine circulation, were also important in promoting genetic divergence among subpopulations of the eastern Pacific bay pipefish. Deep-water channels restrict gene flow and enhance genetic variation. Isolating features can allow for the persistence and generation of novel genetic diversity (Johnson and et al. 1994; Watts and Johnson 2004; Hoffman et al. 2005) which may be important to the adaptability of a species to changing environments. Within Barkley Sound, the only protection of eelgrass beds, outside of protective policies under the Fisheries Act, lies within the Broken Group Islands. There is no protection of eelgrass beds within the Alberni Inlet area of East Barkley Sound and industrial activities (log booming, fish farms, resort marinas) is high. M y data suggest that the source population located in the Alberni Inlet region of East Barkley Sound provides recruits to the greater area of Barkley Sound. Maintaining population processes of the eastern Pacific bay pipefish in East Barkley Sound is important due to the increased genetic diversity and allelic richness in this region relative to North Barkley Sound. Although it is difficult with neutral loci to make assumptions about selection pressures and adaptive value of alleles, the physical oceanography 145 of the fjordal habitat may promote novel genetic variability to deal with this different environment. Unlike other localities of the eastern Pacific bay pipefish, genetic connectivity within the Broken Group Islands archipelago was restricted (high intra-patch genetic variation), and together with other localities in the northern geographic region, showed the lowest allelic richness and genetic diversity of any geographic region. Gibralter Island is isolated from all localities assessed in this study. From the data in my study, therefore, connectivity among the eastern Pacific bay pipefish among localities within reserve boundaries is restricted. Further genetic sampling within the reserve boundaries, however, may reveal that these localities are more connected by gene flow than was resolvable in my study. Conversely, genetic connectivity of the eastern Pacific bay pipefish between localities inside the reserve and outside of the reserve appears to be high. For instance within the reserve, two of the localities, Turret Island and Jaques/Jarvis Islands show high levels of connectivity with both coastal and island habitats outside of reserve boundaries. Eastern Pacific bay pipefish within the Broken Group Islands form part of the larger source population of South/North Barkley Sound. The particular attributes of this archipelago (Chapter 1) may influence the estimates of genetic diversity and genetic variation resolved in my study. The eelgrass environment within the Broken Group Islands lacks mainland coastal habitats and is separated by the widest deep water channel, Imperial Eagle Channel, from habitats in east Barkley Sound. Connections to mainland, coastal habitats may provide essential corridors for gene flow. Contemporary connectivity of the eastern Pacific bay pipefish may require habitat attributes not found in the Broken Group Islands and a geographic area greater than provided by present reserve boundaries. To ensure the conservation of genetic diversity of the eastern Pacific bay pipefish within the boundaries of the reserve, it would be 146 advantageous to protect neighbouring eelgrass habitats outside of the reserve from activities detrimental to the persistence of eelgrass meadows. In British Columbia, extant populations of a variety of species reveal the signature of glacial founding populations often resulting in large-scale genetic homogeneity among localities (McLean 1999; Hickerson and Ross 2001; Louie 2003). Knowing the influence of the glacial history of British Columbia on genetic population structure at the large geographic scale, investigating processes influencing genetic structure at smaller spatial scales, as evidenced from this study, is important in accurately assessing genetic resources. Local populations are at risk from habitat loss due to human activities or natural environmental perturbations. Protected areas are generally small in size and affected by local physical processes as well as local impacts of human development. M y study has shown the importance of understanding and managing for the influence of local, seascape features on genetic variation and genetic diversity. Numerous studies of genetic variation in freshwater fish species have also shown the importance of landscape features to shape genetic variance and connectivity among populations (e.g., Costello et al. 2003). Similarly, marine species often show a correlation of population genetic substructure with physical features of the environment other than isolation by distance. A s seascape features and habitat spatial configuration influence gene flow of the eastern Pacific bay pipefish, it may be that other marine organisms are also influenced by these factors. Gene flow is not a species-specific or a fixed trait, and genetic variation may reflect more strongly the local environment (Watts and Johnson 2004) and historical events (Sotka et al. 2004). Population genetic substructure and genetic diversity maybe more heavily influenced by the interaction of individuals with the spatial arrangement of habitats within diverse environments than by their dispersal potential. Management of eelgrass habitats as a coastal resource should include protection of the genetic 147 resources of taxa inhabiting eelgrass habitats and requires a full inventory of variation including that owing to environmental heterogeneity. 148 Literature Cited Adams S M (1976) The ecology of eelgrass Zostera marina L , fish communities. I. Structural analysis. Journal of Experimental Marine Biology and Ecology, 22, 269-291. Ahnesjo I (1992) Fewer newborn result in superior juveniles in the paternally brooding pipefish Syngnathus typhle L . Journal of Fish Biology, 41, Supplement B , 53-63. Al l ison G W , Lubchenco J, Carr M H (1998) Marine reserves are necessary but not sufficient for marine conservation. Ecological Applications, 8(1) Supplement, S79-S92. Altman A D , Taylor E B (2003) A molecular assessment of homing in the tidepool sculpin (Oligocottus maculosus) Girard. Journal of Fish Biology, 62(3), 623-640. Appleyard S A , Grewe P M , Innes B H , Ward R D (2001) Population structure of yellowfin tuna (Thunnus albacares) in the western Pacific Ocean, inferred from microsatellite loci . Marine Biology, 139, 383-393. Arndt A , Smith M J (1998) Genetic diversity and population structure in two species of sea cucumber: differing patterns according to mode of development. MolecularEcology, 7, 1053-1064. Avise J C (1998) Conservation Genetics in the Marine Realm. The Journal of Heredity, 89(5), 377-382. Balloux F, Lugon-Moulin N (2002) The estimation of population differentiation with microsatellite markers. Molecular Ecology, 11, 155-165. Barber P H , Palumbi SR, Erdmann M V , Moosa M K (2002) Sharp genetic breaks among populations of Haptosquilla pulchella (Stomatopoda) indicate limits to larval transport: patterns, causes, and consequences. Molecular Ecology, 11, 659-674. Bayer R D (1980) Size, seasonality, and sex ratios of the bay pipefish {Syngnathus leptorhynchus) in Oregon. Northwest Science, 54(3), 161-167. Bekkevold D , Andre C, Dahlgren T G , Clausen L A W , Torstensen E , Mosegaard H , Carvalho G R , Christensen T B , Norlinder E , Ruzzante D E (2005) Environmental correlates of population differentiation in Atlantic herring. Evolution, 59(12), 2656-2668. Bel l L J , Moyer JT, Numachi K (1982) Morphological and genetic variation in Japanese populations of the anemone fish Amphiprion clarkii. Marine Biology, 72, 99-108. Berglund A , Rosenqvist G , Svensson, I (1986) Mate choice, fecundity, and sexual dimorphism in two pipefish species. Behavioural Ecology and Sociobiology, 19, 301-307. Berglund A , Rosenqvist G , Svensson I (1986) Reversed sex roles and parental energy investment in zygotes of two pipefish (Syngnathidae) species. Marine Ecology Progress Series, 29, 209-215. 149 Bohonak A J (2002) I B D (Isolation by Distance): a program for analyses of isolation by distance. Journal of Heredity, 93, 153-154. Booth J, Mason B , Berry M , Zacharias M , Richer S (2000) Pacific Coastal Resources Atlas for British Columbia. Available from http://www.shim.bc.ca/coastal/manual/PCRAtlas Info.htm Bortone S A (ed) (1998) Seagrasses: monitoring, ecology, physiology and management. C R C Press, New York. Buonaccorsi V P , Westerman M , Stannard J, Kimbrel l C , Lynn E , Vetter R D (2004) Molecular genetic structure suggests limited larval dispersal in grass rockfish, Sebastes rastrelliger. Marine Biology, 145, 779-788. Buonaccorsi V P , Kimbrel l C A , Lynn E A , Vetter R D (2005) Limited realized dispersal and introgressive hybridization influence genetic structure and conservation strategies for brown rockfish, Sebastes auriculatus. Conservation Genetics, 6, 697-713. Burdick, D M , Short FT , W o l f J (1993) A n index to assess and monitor the progression of wasting disease in eelgrass Zostera marina. Marine Ecology Progress Series, 94, 83-90. Burridge C P , Hurt A C , Farrington L W , Coutin P C , Austin C M (2004) Stepping stone gene flow in an estuarine-dwelling sparid from south-east Australia. Journal of Fish Biology, 64, 805-819. Carcupino M , Baldacci A , Mazz in i M , Franzoi P (2002). Functional significance of the male brood pouch in the reproductive strategies of pipefishes and seahorses: a morphological and ultra-structural comparative study on three anatomically different pouches. Journal of Fish Biology, 61, 1465-1480. Casey SP (1999) Conservation Genetics of the seahorse (Hippocampus species). PhD Thesis, University of London, Queen Mary and Westfield College, London, U K Castric V , Bonney F, Bernatchez L (2001) Landscape structure and hierarchical genetic diversity in the brook char, Salvelinus fontinalis. Evolution, 55(5), 1016-1028. Castric V , Bernatchez L (2004) Individual assignment test reveals differential restriction to dispersal between two salmonids despite no increase of genetic differences with distance. Molecular Ecology, 13, 1299-1312. Chenoweth SF, Hughes J M , Connolly R C (2002) Phylogeography of the pipefish, Urocampus carinirostris, suggests intergradation of ancient lineages. Marine Biology, 141, 541-547. Clemens W A , Wilby, G V (1949) Fishes of the Pacific Coast of Canada. Bulletin No. L X V I I I Fisheries Research Board of Canada, Ottawa. Congiu L , Rossi R, Colombo G (2002) Population analysis of the sand smelt Atherina boyeri (Teleostei Atherinidae), from Italian coastal lagoons by random amplified polymorphic D N A . Marine Ecology Progress Series, 229: 279-289. 150 Connolly R M , Melv i l le A J , Preston K M (2002) Patterns of movement and habitat use by leafy seadragons tracked ultrasonically. Journal of Fish Biology, 61, 684-695. Cornuet J - M , Piry S, Luikart G , Estoup A , Solignac, M (1999) New Methods Employing Multilocus Genotypes to Select or Exclude Populations as Origins of Individuals. Genetics, 153, 1989-2000. Costello A B , Down T E , Pollard S M , Pacas C J , Taylor E B (2003) The influence of history and contemporary stream hydrology on the evolution of genetic diversity within species: an examination of microsatellite D N A variation in bull trout, Salvelinus confluentus (Pisces: Salmonidae). Evolution, 57(2), 328-344. Coulon A , Guillot G , Cosson JF, Angibault M A , Aulagnier S, Cargnelutti B , Galan M , Hewison A J M (2005) Genetic structure is influenced by landscape features: empirical evidence from a roe deer population. Molecular Ecology, 13, 2841-2850. Cowen R K , L w i z K M M , Sponaugle S, Paris C B , Olson D B (2000) Connectivity of Marine Populations: Open or Closed? Science, 287(4), 857-859. Cowen R K , Paris C B , Srinivasan A (2006) Scaling of Connectivity in marine populations. Science, 311, 522-527. Dawson C E (1985) Indo-Pacific Pipefishes Gulf Coast Research Laboratory, Ocean Springs, Mississippi. Dawson M N , Louie K D , Barlow M , Jacobs, D K , Swift C C (2002) Comparative phylogeography of sympatric sister species, Clevelandia ios and Eucyclogobius newberryi (Teleostei, Gobiidae), across the California Transition Zone. Molecular Ecology, 11, 1065-1075. Daemen E , Cross T, Ollevier F, Volckaert F A M (2001) Analysis of the genetic structure of the European eel (Anguilla anguilla) using microsatellite D N A and m t D N A markers. Marine Biology, 139, 755-764. Davis M B , Shaw R C (2001) Range shifts and adaptive responses to Quaternary climate change. Science, 292, 673-679. Dennison W C (1997) Effects of light on seagrass photosynthesis, growth and depth distribution. Aquatic Botany 27 ',15-26. DeWoody J A , Avise JC (2000) Microsatellite variation in marine, freshwater and anadromous fishes compared with other animals. Journal of Fish Biology, 56, 461-473. Doherty PJ, Planes S, Mather P (1995) Gene flow and larval duration in seven species of fish from the Great Barrier Reef. Ecology, 76, 2373-2391. E S R I (2005) A R C V I E W GIS (version 9.1) GIS and Mapping Software, Redlands, C A . 151 Fausch K D , Lyons J, Karr, JR, Anergmeir (1990) Fish communities as indicators of environmental degradation. American Fisheries Society Symposium 8, 123-144. Feral J-P (2002) How useful are genetic markers in attempts to understand and manage marine biodiversity? Journal of Experimental Marine Biology and Ecology, 268, 121-145. Foster SJ, Vincent A C J (2004) Review Paper: Life history and ecology of seahorses: implications for conservation and management. Journal of Fish Biology, 65, 1-61. Fritzsche R A (1984) Gasterosteiformes: Development and Relationships. Ontogeny and systematics of fishes. Special publication number 1. Moser H G , Richards W J , Cohen D M , Fahay M P , Kendall A W , Richardson S L eds. American Society of Ichthyologists and Herpetologists. Al len Press, Inc., Lawrence, K S . pp398-405. In Louie K D (2003) Genetic Structure and Phylogeography of Four Eastern Pacific Estuarine Species, with an Emphasis on Eelgrass Dependent Taxa. PhD Thesis, University of California, Los Angeles. Fritzche R A (1980) Revision of the eastern Pacific Syngnathidae (Pisces: Syngnathiformes), including both recent and fossil forms. Proceedings ofthe California Academy of Sciences, 42(6), 181-227. Garmin Corporation (1995-2002) Nobeltec Visual Navigation Suite (version 7), Salem, Oregon. Glenn T C (1995). Microsatellite Manual, Version 5. Molecular Systematics laboratory, Smithsonian Institution, Washington, D C . Gold JR, Turner T F (2002) Population structure of red drum (Sciaenops ocellatus) in the northern Gulf of Mexico,as inferred from variation in nuclear-encoded microsatellites. Marine Biology, 140, 249-265. Goudet J (1999) P C A - G E N (version 1.2). Institute of Ecology, U N I L , CH-1015, Lausanne, Switzerland. Available from http://www.unil.ch/izea/softwares/pcagen.html. Goudet J (2001) F S T A T , a program to estimate and test gene diversities and fixation indices (version 2.9.3.). Institute of Ecology, U N I L , CH-1015, Lausanne, Switzerland. Available from http://www.unil.ch/zea/softwares/fstat.html. Green E P , Short F T (2003) World Atlas of Seagrasses. University of California Press, Berkeley. Gronell A M (1984) Courtship, spawning and social organization of the pipefish, Corythoichthys intestinalis (Pisces: Syngnathidae) with notes on two congeneric species. Zeitschrift fiir Tierpsychologie, 65, 1-24. In Vincent A C J , Berglund A , Ahnesjo I (1995) Reproductive ecology of five pipefish species in one eelgrass meadow. Environmental Biology of Fishes, 44, 347-361. Gyllensten U (1985) The genetic structure of fish: differences in the intraspecific distribution of biochemical genetic variation between marine, anadromous, and freshwater species. Journal of Fish Biology, 26, 691-699. 152 Hanski I (1999) Metapopulation Ecology. Oxford University Press, Oxford Hartl D L , Clark A G (1997) Principles of Population Genetics. Sinauer Asociates, Inc., Sunderland, Massachusetts. Harrison S, Hastings A (1996) Genetic and evolutionary consequences of metapopulation structure. Trends Ecology and Evolution, 11(4), 180-183. Hedrick P W (1999) Highly variable loci and their interpretation in evolution and conservation. Evolution 53, 313-318. Hedrick P W (2000) Genetics of Populations, Jones and Bartlett Publishers, Inc. London. Hellberg M E (1994) Relationships between inferred levels of gene flow and geographic distance in a philopatric coral, Balanophyllia elegans. Evolution 48, 1829-1854. Hellberg M A , Burton R S , Neigel JE , Palumbi SR (2002) Genetic assessment of connectivity among marine populations. Bulletin of Marine Science, 70(1), S U P P L 273-290. Hemminga M A , Duarte C M (2000) Seagrass ecology: an introduction. Cambridge University Press, New York, N Y . Hendry A P , Castric V , Kinnison M T , Quinn T P (2003) The Evolution of philopatry and dispersal: homing vs. straying in salmonids. In: Evolution of Salmonids (eds Hendry A P , Stearns SE) in press. Oxford University Press, Oxford, U K in Taylor Taylor E B , Stamford M D , Baxter JS (2003) Population subdivision in westslope cutthroat trout (Oncorhynchus clarki lewisi) at the northern periphery of its range: evolutionary inferences and conservation implications. Molecular Ecology, 12, 2609-2622. Herald ES (1941) A systematic analysis of variation in the western American pipefish, Syngnathus californiensis. Stanford Ichthyological Bulletin. 2(3), 49-73. Hickerson M J , Ross J R P (2001) Post-glacial population history and genetic structure of the northern clingfish (Gobbiesox maeandricus), revealed by m t D N A analysis. Marine Biology, 138, 407-419. Highsmith R C (1985) Floating and algal rafting as potential dispersal mechanisms in brooding invertebrates. Marine Ecology Progress Series, 25, 169-179. Hi l ls D M , Mortiz C , Mable B K (eds) (1996) Molecular Systematics, Second Edition, Sinauer Associates, Inc, Massachusetts. Hoarau, G , Rijnsdorp, A D , van der Veer, H W , Stam, W T and Olsen, J L (2002) Population structure of plaice (Pleuronectes platessa L . ) in northern Europe: microsatellites revealed large-scale spatial and temporal homogeneity. Molecular Ecology 11, 1165-1176. Hoelzel A R (1998) Genetic structure of cetacean populations in sympatry, parapatry, and mixed assemblages: implications for conservation policy. The Journal of Heredity 89(5), 451-458. 153 Hoffman E A , K o l m N , Berglund A , Arguello JR, Jones A G (2005) Genetic structure in the coral-reef associated Banggai cardinalfish, Pterapogon kauderni. Molecular Ecology, 14, 1367-1375. Holmquist J G (1994) Benthic macroalgae as a dispersal mechanism for fauna: influence of a marine tumbleweed. Journal of Experimental Marine Biology and Ecology, 180(2), 235-251. Hornbeck, T (2004). Abundance estimate and site fidelity in the bay pipefish, Syngnathus leptorhynchus, in the Bamfield Inlet, British Columbia. Student Reports Bamfield Marine Sciences Centre, 1-18. Hovel K A , Fonseca M S , Myer D L , Ken worthy W J , Whitfield P E (2002) Effects of seagrass landscape structure, structural complexity and hydrodynamic regime on macrofaunal densities in North Carolina seagrass beds. Marine Ecology Progress Series, 243, 11-24. Howard R K , Koehn JD (1985) Population dynamics and feeding ecology of pipefish (Syngnathidae) associated with eelgrass beds of Western Port, Victoria. Australian Journal of Marine and -Freshwater Research, 36, 361-370. Hubbs C L (1922) Variations in the number of vertebrae and other meristic characters of fishes correlated with the temperature of water during development. American Naturalist, 56, 360-372. In Louie K D (2003) Genetic Structure and Phylogeography of Four Eastern Pacific Estuarine Species, with an Emphasis on Eelgrass Dependent Taxa. PhD Thesis, University of California, Los Angeles. Hutchinson W F , Carvalho G R , Rogers SI (2001) Marked genetic structuring in localized spawning populations of cod Gadus morhua in the North Sea and adjoining waters, as revealed by microsatellites. Marine Ecology Progress Series, 223, 251-260. Hutchison D W , Templeton A R (1999) Correlation of pairwise genetic and geographic distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution, 53(6), 1898-1914. Jarne P, Lagoda P J L (1996) Microsatellites, from molecules to populations and back. Trends in Ecology and Evolution, 11(10), 424-429. Johnson M S , Watts RJ , Black R (1994) High levels of genetic subdivision in the peripherally isolated populations of the atherind fish Craterocephalus capreoli in the Houtman Abrolhos Islands, Western Australia. Marine Biology, 119, 179-184. Johnson M S , Black R (1996) Geographic cohesiveness verses associations with habitat: genetic subdivision of Bembicium vittatum Philippi (Gastropoda: Littorinidae) in the Houtman Abrolhos Islands. Biological Journal of the Linnean Society, 58, 57-74. Johnson M S , Bentley S L , Ford SS, Ladyman M T , Lambert G J (2001) Effects of a complex archipelago on genetic subdivision of the intertidal limpet Siphonaria kurracheensis, Marine Biology, 139(6), 1087-1094. 154 Jones A G , Avise JC (1997) Microsatellite analysis of maternity and the mating system in the Gulf pipefish Syngnathus scovelli, a species with male pregnancy and sex-role reversal. Molecular Ecology, 6, 203-213. Jones A G , Avise JC (1997) Polygynandry in the dusky pipefish Syngnathus floridae revealed by microsatellite D N A markers. Evolution 51(5): 1611-1622. Jones A G , Kvarnemo C, Moore G l , Simmons L W , Avise JC (1998) Microsatellite evidence for monogamy and sex-biased recombination in the Western Australian seahorse Hippocampus angustus. Molecular Ecology, 7, 1497-1505. Jones A G , Rosenqvist G , Berglund A , Avise JC (1999) The genetic mating system of a sex-role-reversed pipefish (Syngnathus typhle): a molecular inquiry. Behavioural Ecology and Sociobiology, 46, 357-365. Jones A G , Avise JC (2001) Mating systems and sexual selection in male-pregnant pipefishes and seahorses: insights from microsatellite-based studies of maternity. Journal of Heredity, 92, 150-158. Jones G P , Mi l i c i ch M J , Emslie M J , Lunow C (1999) Self-recruitment in a coral reef population. Nature, 402, 802-804. Jorgensen H B H , Hansen M M , Bekkevold D , Ruzzante D E , Loeschcke V (2005) Marine landscapes and population genetic structure of herring (Clupea harengus L . ) in the Baltic Sea. Molecular Ecology, 14,3219-3234. Kenchington E L , Patwary M U , Zouros E , B i rd C J (2006) Genetic differentiation in relation to marine landscape in a broadcast spawning bivalve mollusc (Placopecten magellanicus), Molecular Ecology, 15, 1781-1796. Kendrick A J , Hyndes G A (2003) Patterns in the abundance and size-distribution of syngnathid fishes among habitats in a seagrass-dominated marine environment. Estuarine, Coastal and Shelf Science, 57, 631-640. K i m I, Phillips CJ , Monjeau J A , Birney E C , Noack K , Pumo D E , Sikes R S , Dole J A (1998) Habitat islands, genetic diversity, and gene flow in a Patagonian rodent. Molecular Ecology, 7, 667-678. Kingsford M J , Leis J M , Shanks A , Lindeman K C , Morgan S G , Pineda J (2002) Sensory environments, larval abilities and local self-recruitment. Bulletin of Marine Science, 70(1) Supplement, 309-340. Kinlan B , Gaines S (2003) Propagule dispersal in marine and terrestrial environments: a community perspective. Ecology, 84, 2007-2020. Kramer D L , Chapman M R (1999) Implications of home range size and relocation for marine reserve function. Environmental Biology of Fishes, 55(1-2), 65-79. 155 Kuiter R H (2000) Seahorses, Pipefishes, and their relatives: A comprehensive guide to syngnathiformes. Chorleywood, U K , T M C Publishing. Kulczycki G R , Virnstein R W , Nelson W G (1981) The relationship between fish abundance and algal biomass in a seagrass drift algae community. Estuarine Coastal Shelf Science, 12, 341-347. Kusomo H T , Druehl L D (2000) Variability over space and time in the genetic structure of the winged kelp Alaria marginata. Marine Biology, 136, 397-409. La l l i C M , Parsons T R (1993) Biological Oceanography: An introduction. Oxford, Pergamon Press Ltd. Lazzari M A , Able K W (1990) Northern pipefish, Syngnathus fuscus, occurrences over the mid-Atlantic bight continental shelf: evidence of seasonal migration. Environmental Biology of Fishes, 27, 111'-185. Louie K D (2003) Genetic Structure and Phylogeography of Four Eastern Pacific Estuarine Species, with an Emphasis on Eelgrass Dependent Taxa. PhD Thesis, University of California, Los Angeles. Lourie S A , Pritchard JC , Casey SP, K y TS, Hal l H J , Vincent A C J (1999) The taxonomy of Vietnam's exploited seahorses (family Syngnathidae). Biological Journal of the Linnean Society, 66, 231-256. Lourie S A , Vincent A C J (2004) A marine fish follows Wallace's Line: the phylogeography of the three-spot seahorse (Hippocampus trimaculatus, Syngnathidae, Teleostei) in Southeast Asia . Journal of Bio geography, 31, 1975-1985. Maltagliati, F (1999) Genetic divergence in natural populations of the Mediterranean brackish-water killifish Aphanius fasciatus. Marine Ecology Progress Series, 179, 155-162. Manel S, Schwartz M K , Luikart G , Taberlet P (2003) Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology and Evolution, 18(4), 189-197. Manni F, Guerard E , Heyer E (2004) Geographic patterns of (genetic, morphologic, linguistic) variation: how barriers can be detected by "Monmonier's algorithm'. Human Biology, 76, 173-190. Mantel N (1967) The detection of disease clustering and a generalized regression approach, Cancer Research, 27, 209-220. McLean JE (1999) Marine Population Structure in an Anadromous Fish: Life-history Influences Patterns of Neutral Molecular Variation in the Eulachon, Thaleichthys pacificus. M S c Thesis, University of British Columbia. M c M i l l a n A D (1999). Since the Time of the Transformers: The Ancient Heritage of the Nuu-chah-nulth, Ditidaht, and Makah, U B C Press, Vancouver, B C . 156 McPhai l JD, Lindsey C C (1986) Zoogeography of the freshwater fishes of Cascadia (the Columbia River and rivers north to the Stikine). In: The Zoogeography of North American Freshwater Fishes (eds Hocutt C H , Wiley E O ) , pp. 615-637. John Wiley and Sons, New York. Meffe G K , Vrijenhoek R C (1988) Conservation genetics in the management of desert fishes. Conservation Biology 2(2), 157-169. Mi l le r M (1997) Tools for Population Genetic Analysis ( T F P G A ) version 1.3. Department of Biological Sciences, Northern Arizona University, Flagstaff, A Z . Monteiro N M , da Natividade Viei ra M , Almada V (2005) Homing behaviour and individual identification of the pipefish Nerophis lumbriciformis (Pisces; Syngnathidae): a true intertidal resident? Estuarine, Coastal and Shelf Science^?), 93-99. Moreau M - A , Vincent A C J (2004) Social structure and space use in a wi ld population of the Australian short-headed seahorse Hippocampus breviceps Peters, 1989. Marine and Freshwater Research, 55, 231-239. Neigel JE (1997) A comparison of alternative strategies for estimating gene flow from genetic markers. Annual Review of Ecological Systematics, 28, 105-128. Olsen J L , Stam W T , Coyer J A , Reusch T B H , Bill ingham M , Bostroem C , Calvert E , Christie H , Granger S, L a Lumie're R, Milchakova N , Oudot-LeSecq M - P , Procaccini G , Sanjabi B , Serrao E , Veldsink J, Widdicombe S, Wyllie-Eschevarria S (2004) North Atlantic phylogeography and large-scale population differentiation of the seagrass Zostera marina L . Molecular Ecology, 13: 1923-1941. Orsi J A , Gish R K , Wing B L (1991) Northern range expansions of four nearshore marine fishes in Alaska. Canadian Field-Naturalist, 105, 82-86. Orth R J , Heck K L Jr, van Montfrans J (1984) Faunal communities in seagrass beds: a review of the influence of plant structure and prey characteristics in predator-prey relationships, Estuaries, 7, 339-350. Paetkau D , Waits L P , Clarkson P L , Craighead L , Vyse E , Ward R, Strobeck C (1998). Variation in genetic diversity across the range of North American Brown bears. Conservation Biology, 12(2), 418-429. Palumbi SR (1994) Reproductive isolation, genetic divergence, and speciation in the sea. Annual Review ofEcology and Systematics, 25, 547-572. Palumbi S (2002). Marine Reserves. Pew Oceans Commission, New York; Palumbi SR, Warner R R (2003) Why Gobies are Like Hobbits. Science, 299, 51-52. Palumbi SR (2004) Marine reserves and ocean neighbourhoods: The spatial scale of marine populations and their management. Annual Review of Environment and Resources, 29, 31-68. 157 Parsons K E (1996) The genetic effects of larval dispersal depend on spatial scale and habitat characteristics. Marine Biology, 126, 403-414. Perante N C , Pajaro M G , Meeuwig JJ, Vincent A C J (2002) Biology of a seahorse species, Hippocampus comes in the central Philippines. Journal of Fish Biology, 60, 821- 837. Phillips R C (1969) Temperate grass flats, in Coastal ecological systems of the United States, V o l . 2, edited by H T Odum, B J Copeland & E A McMahan, The Conservation Foundation, Washington D C , p 244-299. Piry S, Alapetite A , Cornuet J - M , Paetkau D , Baudouin L , Estoup A (2004) G E N E C L A S S 2 : a software for genetic assignment and first-generation migrant detection. Journal of Heredity, 95,536-539. Planes S, Galzin R, Bonhomme F (1996) A genetic metapopulation model for reef fishes in oceanic islands: the case of the surgeonfish, Acanthurus triostegus. Journal of Evolutionary Biology, -9,103-117. Planes S, Romans P, Lecomte-Finiger R (1998) Genetic evidence of closed life cycles for some coral reef fishes within Taiaro lagoon (Tuamotu Archipelago, French Polynesia). Coral Reefs, 17, 9-14. Pogson G H , Fevolden S E (2003) Natural selection and the genetic differentiation of coastal and Arctic populations of the Atlantic cod in northern Norway: a test involving nucleotide sequence variation at the pantophysin (PanI) locus. Molecular Ecology 12(1), 63-74. Power M , Attr i l l M J (2003) Long-term trends in the estuarine abundance of Nilsson's pipefish (Syngnathus rostellatus Nilsson). Estuarine, Coastal and Shelf Science, 57, 325-333. Precision Idenification (2002) Field Methods for Mapping and Monitoring Eelgrass Beds in British Columbia. Environment Canada and Canadian Wildlife Service. Pritchard J K , Stephens M , Donnelly P (2000) Inference of Population Structure using multilocus genotype data. Genetics, 155, 945-959 Raymond M , Rousset F (1995) G E N E P O P (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity, 86, 248-249. Raymond M , Rousset F (1995) A n Exact Test for Population Differentiation. Evolution, 49(6), 1 2 8 0 - 1283. Roelke D L , Sogard S M (1993) Gender based differences in habitat selection and activity level in the northern pipefish (Syngnathus fuscus). Copeia, 2, 528-532. Reusch T B H , Stam W T , Olsen J L (2000). A microsatellite-based estimation of clonal diversity and population subdivision in Zostera marina, a marine flowering plant. Molecular Ecology, 9, 127-140. 158 Rice W R (1989) Analyzing Tables of Statistical Tests. Evolution 43(1), 223-225. Rico C, Zadworny D , Kuhnlein U , Fitzgerald G J (1993). Characterization of hypervariable microsatellite loci in the threespined stickleback Gasterostus aculeatus. Molecular Ecology, 2, 271-272. Riginos C, Nachman M W (2001) Population subdivision in marine environments: the contributions of biogeography, geographical distance and discontinuous habitat to genetic differentiation in a blennioid fish, Axoclinus nigricaudus. Molecular Ecology, 10, 1439-1453. Roberts T E (2006) History, ocean channels, and distance determine phylogeographic patterns in three widespread Philippine fruit bats (Pteropodidae). Molecular Ecology, 15, 2183-2199. Rozen S, Skaletsky H (1998) Primer3. Code available at http://wwwgenome.wi.mit.edu/genome software/other/primer3.html. Ruckelshaus M H (1998) Spatial scale of genetic structure and an indirect estimate of gene flow in eelgrass, Zostera marina. Evolution, 52(2), 330 - 343. Sanchez-Camara J, Booth D J (2004) Movement, home range and site fidelity of the weedy sea dragon, Phyllopteryx taeniolatus (Teleostei: Syngnathidae). Environmental Biology of Fishes, Id, 31-41. Schneider S, Roessli D , Excoffier L (2000) Arlequin version 2.000: a software package for population genetics data analysis. Genetics and Biometry Laboratory, University of Geneva, Switzerland. Shaklee JB , Samollow P B (1980) Genetic aspects of population structure of four species in the north-western Hawaiian Islands. In: Grigg, R W , Pfund, R T (eds) Proceedings of the Symposium on the Status of Resource Investigations in the Northwestern Hawaiian Islands, 24-25 Apr i l 1980, University of Hawaii , Honolulu. Sea Grant miscellaneous report, U N I F f l - S E A G R A N T -MR-80-04. University of Hawaii Sea Grant College Program, Honolulu, 264-277. Short FT, Wyllie-Escheverria S (1996) Natural and human-induced disturbance of seagrasses. Environmental Conservation, 25, 17-27. Shulman M J , Bermingham E (1995) Early life histories, ocean currents, and the population genetics of Caribbean reef fishes. Evolution, 49(5), 897-910. Skilleter G A , Olds A , Loneragan Nr, Zharikov Y (2005) The value of patches of intertidal seagrass to prawns depends on their proximity to mangroves. Marine Biology, 147, 353-365. Slatkin M (1987) Gene Flow and the Geographic Structure of Natural Populations, Science, 236, 787-792. Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139, 457-462. 159 Smith PJ, Fujio Y(1982) Genetic variation in marine teleosts: High variability in habitat specialists and low variability in habitat generalists. Marine Biology, 69, 7 - 20. Sogard S (1989) Colonization of artificial seagrass by fishes and decapod crustaceans: importance of proximity to natural eelgrass. Journal of Experimental Marine Biology and Ecology, 133, 15-37. Sork V L , Nason J , Campbell D R , Fernandez JF (1999) Landscape approaches to historical and contemporary gene flow In plants. Trends in Ecology and Evolution, 14, 219-224. Sotka E E , Wares JP, Barth J A , Grosberg R K , Palumbi SR (2004) Strong genetic clines and geographical variation in gene flow in the rocky intertidal barnacle Balanus glandula. Moleculary Ecology, 13, 2143-2156. Sponaugle S, Cowen R K , Shanks A , Morgan S G , Leis J M , Pineda J, Boehlert G W , Kingsford M J , Lindeman K C , Grimes C, Munro J L (2002) Predicting self- recruitment in marine populations: biophysical correlates and mechanisms. Bulletin of Marine Science, 70(1), Supplement, 341-375. Steffe A S , Westoby M , Be l l JD (1989) Habitat selection and diet in two species of pipefish from seagrass: sex differences. Marine Ecology Progress Series, 55, 23-30. Stepien C A , Rosenblatt R H (1991) Patterns of gene flow and genetic divergence in the northeastern Pacific Clinidae (Teleostei: Blennioidei), based on allozyme and morphological data. Copeia, 4, 873-896. Stronach J A , N g M K , Foreman M G , Murty TS (1993) Tides and Currents in Barkley Sound and Alberni Inlet. Marine Geodesy, 16, 1-41. Svensson I (1988) Reproductive costs in two sex-role reversed pipefish species (Syngnathidae). Journal of Animal Ecology, 57, 929-942. Swearer SE, Caselle JE , Lea D W , Warner R R (1999) Larval retention and recruitment in an island population of a coral-reef fish. Nature, 402, 799-802. Swift C C , Haglund T R , Ruiz M , Fisher R N (1993) The status and distribution of the freshwater fishes of Southern California. Bulletin of Southern California Academy of Sciences, 92(3), 101-167. Taylor E B (1998) Microsatellites isolated from the threespined stickleback Qasterostus aculeatus. Molecular Ecology, 7, 930-931. Taylor E B , Stamford M D , Baxter JS (2003) Population subdivision in westslope cutthroat trout (Oncorhynchus clarki lewisi) at the northern periphery of its range: evolutionary inferences and conservation implications. Molecular Ecology, 12, 2609-2622. Taylor E B , Boughman J W , Groenenboom M , Sniatynski M , Schluter D , G o w J L (2006) Speciation in reverse: morphological and genetic evidence of the collapse of a threespine stickleback (Gasterostus aculeatus) species pair. Molecular Ecology, 15(2), 343-355. 160 Taylor M S , Hellberg M E (2003) Genetic evidence for local retention of pelagic larvae in a Caribbean reef fish. Science, 299, 107-109 Taylor M S , Hellberg M E (2005) Marine radiations at small geographical scales: speciation in neotropical reef gobies (Elacatinus). Evolution, 59(2), 374-385. Teske PR, Cherry M I , Matthee C A (2003) Population genetics of the endangered Knysna seahorse, Hippocampus capensis. Molecular Ecology, 12, 1703-1716. Thompson A B , Margolis L , (1987) Determination of population discreteness in two species of shrimp, Pandalus jordani and Pandalopsis dispar, from Coastal British Columbia using parasite tags and other population characteristics. Canadian Journal of Fisheries and Aquatic Science, 44, 982-989. Thomson R E (1981) Oceanography ofthe British Columbia Coast. Canadian Special Publication of Fisheries and Aquatic Sciences, 56, 261 p. Trizio I, Crestanello B , Galbusera P, Wauters L A , Tosi G , Matthysen E , Hauffe H C (2005) Geographic distance and physical barriers shape the genetic structure of Eurasian red squirrels (Sciurus vulgaris) in the Italian Alps. Molecular Ecology, 14, 469-481. Turner SJ, Hewitt JE, Wilkinson M R , Morrisey D J , Thrush SF, Cummings V J , Funnell G (1999) Seagrass Patches and Landscapes: The influence of wind-wave dynamics and hierarchical arrangements of spatial structure on macrofaunal seagrass communities. Estuaries, 22(4), 1016-1032, Vincent A C J , Berglund A , Ahnesjo I (1995) Reproductive ecology of five pipefish species in one eelgrass meadow. Environmental Biology of Fishes, 44, 347-361. Vincent A C J , Evans K L , Marsen A D (2005) Home range behaviour of the monogamous Australian seahorse, Hippocampus whitei. Environmental biology of Fishes, 72, 1-12. Waples RS (1998) Separating the Wheat from the Chaff: Patterns of genetic differentiation in high gene flow species. The Journal of Heredity, 89(5), 438-450. Waples R S , Gaggiotti O (2006) What is a population? A n empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Molecular Ecology, 15, 1419-1439. Ward R D , Woodwark M , Skibinksi D O F (1994) A comparison of genetic diversity levels in marine, freshwater, and anadromous fishes. Journal of Fish Biology, 44, 213-232. Warner R R , Swearer SE , Casell J E (2000) Larval accumrnulation and retention: implications for the design of marine reserves and essential fish habitat. Bulletin of Marine Science, 66(3), 821-830. 161 Watts RJ , Johnson M S (2004) Estuaries, lagoons and enclosed embayments: habitats that enhance population subdivision of inshore fishes. Marine and Freshwater Research, 55(7), 641-651. Webber H H , Thurman H V (1991) Marine Biolgy, Second Edition. New York, Harper Collins Publishers Inc. Webber JL , Wong C (1993) Mutation of human short tandem repeats. Human Molecular Genetics, 2, 1123-1128. Weir B S , Cockerham C C (1984) Estimating F-statistics for the analysis of population structure. Evolution, 38, 1358-1370. Williams SE , Bolitho E E , Fox S (2003) Climate change in Australian tropical rainforests: an impending environmental catastrophe. Proceedings of the Royal Society B: Biological Sciences, 270 (1527), 1887-1892. Wilson A B (2006) Genetic signature of recent glaciation on populations of a near-shore marine fish {Syngnathus leptorhynchus), Molecular Ecology, 15, 1857-1871. Winans G A (1980) Geographic variation in the milkfish, Chanos chanos. I. Biochemical evidence. Evolution, 34, 558-574. Withler R E , Beacham T D , Schulze A D , Richards U , M i l l e r K M (2001) Co-existing populations of Pacific ocean perch, Sebastes alutus, in Queen Charlotte Sound, British Columbia. Marine Biology, 139, 1-12. Worchester S E (1995) Effects of eelgrass beds on advection and turbulent mixing in low current and low shoot density environments, Marine Ecology Progress Series, 126, 223-232. Wright S (1943)isolation by distance. Genetics, 28, 139-156. Zacharias M A , Roff JC (2000) A hierarchical ecological approach to conserving marine biodiversity. Conservation Biology, 14(5), 1327-1334. 162 Appendices Appendix I: Allele frequencies by locality for each individual locus. T16 Size cc NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI 282 0 0 0 0 0 0 0 0 0 0.016 0 0 0 0 0 0.032 0 286 0.041 0 0 0 0 0 0 0.045 0.025 0.024 0.023 0.022 0.013 0.015 0 0 0 290 0.027 0.045 0.062 0.094 0.09 0.036 0.04 0.045 0.086 0.087 0.081 0.122 0.114 0.104 0.052 0.096 0.117 294 0 0 0.013 0 0.051 0.018 0 0 0 0.039 0.012 0.011 0.025 0.015 0.026 0.025 0.013 298 0 0 0.013 0 0 0 0 0.011 0 0.008 0.012 0.022 0 0 0.009 0 0.013 302 0 0.023 0 0 o. 0 0 0 0.012 0.008 0.012 0 0 0.015 0 0 0 306 0 0 0.025 0 0.013 0.036 0 0 0 0 0 0.011 0 0 0.035 0 0 310 0 0 0 0.031 0.013 0.018 0 0 0 0 0 0.011 0.013 0.015 0 0 0 ^ 314 0 0 0 0 0.013 0 0.06 0.011 0.037 0.016 0 0.033 0 0 0.009 0 0 318 0 0 0.013 0 0.013 0.036 0.04 0.022 0 0.008 0 0.011 0 0.03 0.043 0.051 0.026 322 0 0 0.025 0.031 0.013 0.036 0 0.034 0.025 0.024 0.023 0.022 0.025 0.03 0 0.032 0 326 0.027 0 0 0 0.038 0.036 0.04 0.045 0.025 0.055 0.058, 0.044 0.013 0.015 0.087 0 0 330 0 0 0.038 0.031 0.013 0.091 0.08 0.034 0.037 0.087 0 0.011 0.051 0.03 0.043 0.083 0.039 334 • 0.014 0.023 0.062 0.031 0.038 0 0.02 0.034 0.037 0.024 0.023 0.011 0.013 0.045 0.061 0.045 0.039 338 0.068 0.045 0.05 .0.062 0.077 0.036 0.04 0.067 0.062 0.024 0.058 0.022 0.063 0.015 0.043 0.019 0.052 342 0.137 0.068 0.087 0.031 0.038 0.018 0.1 0 0.037 0.016 0.035 0.078 0.063 0.045 0.043 0.057 0 346 0.014 0.023 0.038 0.062 0.077 0.073 0.06 0.045 0.025 0.063 0.047 0.056 0.051 0.045 0.061 0.127 0.117 350 0.014 0.091 0.025 0.031 0.077 0.055 0.06 0.067 0.049 0.055 0.058 0.078 0.051 0.075 0.052 0.121 0.065 354 0.014 0.068 0.038 0.062 0.077 0.109 0.08 0.079 0.136 0.094 0.093 0.056 0.025 0.075 0.035 0.025 0.091 358 0.014 0.068 0.075 0.156 0.077 0.091 0.16 0.124 0.062 0.047 0.14 0.089 0.101 0.104 0.061 0.038 0.156 362 0.096 0.068 01062 0.125 0.103 0.036 0.02 0.146 0.062 0.079 0.047 0.122 0.127 0.075 0.07 0.045 0.065 366 0.123 0.205 0.062 0.094 0.026 0.018 0.06 0.101 0.111 0.11 0.128 0.078 0.152 0.075 0.07 0.051 0.13 370 0.082 0.068 0.087 0.094 0.051 0.091 0.08 0.022 0.062 0.039 0.035 0.056 0.038 0.045 0.096 0.07 0.052 374 0.164 0.045 0.075 0 0.026 0.109 0.04 0.034 0.037 0.016 0.035 0 0.013 0.06 0.043 0.07 0 378 0.041 0.045 0.087 0 0.013 0.036 0.02 0.022 0.037 0.024 0.035 0 0.013 0.06 0.009 0.013 0.013 382 0.027 0.045 0.025 0.031 0.038 0.018 0 0 0.012 0.016 0.023 0.022 0.025 0.015 0.052 0 0 386 0.068 0.023 0 0 0.026 0 0 0.011 0.012 0.008 0 0.011 0.013 0 0 0 0 390 0 0 0.013 0.031 0 0 0 0 0 0.016 0 0 0 0 0 0 0.013 394 0.014 0.023 0 0 0 0 0 0 0 0 0.023 0 0 0 0 0 0 398 0.014 0.023 0.025 0 0 0 0 0 0.012 0 0 0 0 0 0 0 0 163 Appendix I continued: i CSL9 Allele Size CC NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI 100 0.014 0.02 0.023 0 0 0 0 0 0 0.008 0 0 0 0 0 0.007 0 104 0.194 0.2 0.116 0.176 0.194 0.305 0.145 0.124 0.221 0.183 0.165 0.25 0.253 0.118 0.217 0.279 0.176 106 0 0 0.012 0 0 0 0.036 0.011 0 0.024 0 0.011 0 0.015 0 • 0.007 0.015 108 0.389 0.22 0.337 0.529 0.514 0.288 0.327 0.573 0.39 0.413 0.506 0.409 0.443 0.382 0.425 0.367 0.426 110 0.208 0.32 0.209 0.059 0.111 0.186 0.255 0.191 0.286 0.111 0.165 0.159 0.215 0.25 0.225 0.231 0.191 112 0.167 0.16 0.186 0.206 0.167 0.136 0.145 0.067 0.065 0.143 0.141 0.136 0.089 0.221 0.083 0.095 0.132 114 0.028 0.04 0.047 0.029 0.014 0.068 0.018 0.034 0.026 0.079 0.012 0.011 0 0.015 0.033 0 0.015 116 0 0.02 0.023 0 0 0 . 0.055 0 0 0.024 0.012 0.011 0 0 0.008 0.014 0.044 118 0 0.02 0.023 0 0 0 0 0 0 . 0 0 0 0 0 0 0 0 120 0 0 0.023 0 0 0.017 0.018 0 0.013 0.016 0 0.011 0 0 0.008 0 0 T12 Allele Size CC NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI 254 0 0 0 0 0 0 0 0 0 0 0 0.021 0 0 0.017 0.112 0 266 0 0 0 0 0 0 0 0 0.022 0 0 0 0 0 0 0 0 268 0 0 0 0 0 0.019 0 0 0 0 0 0 0.015 0 0 0 0 270 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.007 0 274 0 0 0 0.069 0.032 0 0 0.044 0.022 0.027 0.044 0.011 0.015 0.048 0.026 0 0.041 276 0.014 0 0.012 0 0 0 0 0.015 0 0.009 0 0 0.029 0 0.017 0 0 278 0 0 0.048 0.069 0.097 0.019 0.096 0.162 0.152 0.036 0.044 0.053 0.074 0.065 0.017 0.02 0.041 280 0 0.02 0 0.069 0 0 0 0.029 0.043 0 0.022 0.011 0.015 0.016 0 0 0 282 0.014 0.08 0 0.034 0.016 0.038 0.096 0.074 0.065 0.08 0.089 0.021 0.044 0 0.009 0 0.027 284 0.027 0.1 0.036 0.034 0.065 0.058 0.038 0.074 0.065 0.062 0.067 0.063 0.118 0.048 0.095 0.013 0.041 286 0.014 0.02 0.072 0 0.032 0.019 0.058 0.029 0 0.062 0.022 0.021 0.059 0.032 0.052 0 0.014 288 0.041 0.04 0.036 0.069 0.032 0.115 0.058 0.074 0.087 0.125 0.067 0.126 0.162 0.113 0.095 0.092 0.081 290 0.027 0.02 0.036 0.034 0.048 0.058 0.038 0.074 0.065 0.054 0.022 0.063 0.015 0.048 0.06 0.02 0.054 292 0.095 0.06 0.048 0.172 0.065 0.154 0.154 0.044 0.087 0.143 0.2 0.168 0.147 0.21 0.129 0.211 0.122 294 0.068 0.06 0.06 0.069 0.097 0.038 0.058 0.059 0.152 0.045 0 0.011 0.029 0.065 0.06 0.059 0.054 296 0.122 0.06 0.12 0.069 0.065 0.077 0.077 0.029 0.065 0.125 0.222 0.126 0.059 0.081 0.095 0.092 0.135 298 0.027 0.06 0.072 0.069 0.032 0 0.038 0.044 0.022 0.018 0.022 0.063 0.059 0.032 0.052 0.007 0.014 300 0.189 0.14 0.145 0.138 0.145 0.173 0.058 0.074 0.022 0.062 0.044 0.032 0.044 0.065 0.103 0.138 0.108 302 0.041 0.02 0.036 0.034 0.032 0.019 0.058 0.059 0.022 0.036 0.022 0.021 0.029 0.032 0.034 0.02 0.014 304 0.122 0.04 0.036 0.034 0.065 0.058 0.077 0.029 0 0.036 0.089 0.053 0.044 0 0.026 0.079 0.014 306 0.027 0.04 0.072 0 0.032 0 0 0.015 0.022 0.009 0 0.021 0 . 0.016 0.017 0.053 0.054 308 0.068 0.02 0.048 0 0.048 0.038 0.038 0.015 0 0.036 0.022 0.021 0.015 0.016 0.043 0.053 0.027 310 0 0.02 0.024 0 0.016 0.038 0.019 0.029 0.022 0 0 0.011 0.015 0 0.009 0 0.014 312 0.041 0.06 0.036 0.034 0.016 0.019 0.038 0.015 0.022 0.018 0 0.011 0 0.081 0.034 0 0.041 314 0.014 0.04 0.012 0 0.032 0 0 0 0.022 0 0 0.011 0 0.016 0 0.013 0.014 316 0.027 0.02 0.036 0 0 0.038 0 0 0 0.009 0 0.032 0 0 0 0 0.054 318 0.014 0 0.012 0' 0 0 0 0.015 0 0 0 0.011 0 0 0.009 0 0 320 0.014 0.06 0 0 0.016 0.019 0 0 0 0.009 0 0.011 0 0 0 0.007 0.027 324 0 ' 0.02 0 0 0.016 0 0 0 0.022 0 0 0.011 0.015 0.016 0 0.007 0.014 164 Appendix I continued: Slep9 Allele Size CC NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI . 267 0 0 0 0 0 0 0 0 0.014 0 0 0 0 0 0 0 0 273 0 0 0 0 0 0.018 0 0 0 0 0 0 0 0' 0 0 0 277 0.015 0 0 0 0 0.018 0 0 0 0 0 0 0 0 0 0 0 279 0 0 0 0.029 0 0 0 0 0 0.008 0 0 0 0.016 0.007 0 0.016 281 0.015 0.04 0 0 0 0.018 0 0 0 0.016 0 0 0 0 0 0.009 0 283 0 0 0.023 0.029 0 0.018 0 0.038 0.054 0 0 0 0 0 0.02 0.009 0 285 0.077 0.02 0.023 0 0.041 0.035 0 0.019 0.041 0.016 0.037 0.027 0.025 0.048 0.041 0.009 0 287 0.031 0 0 0 0.014 0.018 0 0.019 0.014 0.016 0.025 0.014 0.025 0.016 0.041 0.019 0.016 289 0.015 0.02 0.012 0.088 0.027 0.035 0 0.038 0.014 0.041 0.074 0.055 0.075 0.048 0.02 0 0 291 0.031 0.02 0.012 0.029 0.068 0.035 0.05 0.019 0.041 0.065 0.012 0.014 0.087 0.065 0.041 0.028 0.094 293 0.031 0.08 0.035 0.059 0.027 0.035 0.025 0.057 0.027 0.057 0.049 0.041 0.038 0.016 0.061 0.028 0.078 295 0.062 0.06 0.081 0.118 0.041 0.053 0 0.075 0.041 0.057 0.062 0.027 0.025 0.048 0.088 0.037 0 297 0.046 0.1 0.047 0.147 0.041 0.088 0.075 0 0.054 0.024 0.012 0.027 0.1 0.065 0.054 0.037 0.094 299 0.154 0.06 0.105 0.059 0.054 0.07 0.05 0.038 0.068 0.106 0.037 0.137 0.038 0.097 0.068 0.037 0.047 301 0.077 0.06 0.07 0.029 0.122 0.053 0.075 0.113 0.054 0.057 0.062 0.055 0.087 0.065 0.048 0.102 0.031 303 0.062 0.08 0.093 0.088 0.095 0.053 0.175 0.057 0.041 0.033 0.074 0.11 0.038 0.081 0.054 0.019 0.125 305 0.062 0.06 0.035 0.088 0.068 0.053 0.075 0.075 0.054 0.089 0.025 0.055 0.062 0.048 0.048 0.037 0.016 307 0.031 0.1 0.023 0.029 0.054 0.018 0.075 0.075 0.122 0.041 0.099 0.082 0.025 0.081 0.054 0.065 0.016 309 0.015 0.08 0.047 0 0.081 0.053 0.05 0.094 0.068 0.065 0.086 0.014 0.038 0.048 0.068 0.157 0.094 311 0.031 0.02 0.07 0.029 0.041 0 0.075 0.057 0.027 0.073 0.074 0.041 0.025 0.016 0.041 0.019 0.062 313 0.031 0.02 0.035 0 0.014 0.018 0 0.038 0.081 0.024 0.062 0.082 0.05 0 0.007 0.009 0.094 315 0.031 0.06 0.105 0.029 0.027 0.035 0.05 0 0.027 0.008 0.062 0.027 0.05 0,081 0.054 0.037 0.031 317 0.015 0.06 0 0.059 0.014 0.053 0.05 0.038 0.014 0.033 0.037 0.014 0.062 0.032 0.027 0.019 0.031 319 0.046 0 0.047 0.029 0.041 0.018 0.025 0.019 0.027 0 0.025 0.014 0.013 0.048 0.014 0.028 0.016 321 0 0.02 0 0.029 0.014 0.035 0.075 0.019 0.014 0.024 0.037 0.027 0.013 0.016 0.027 0.009 0.031 323 0.077 0.04 0.012 0 0.041 0.018 0.05 0.019 0.041 0.033 0 0.014 0 0.016 0.02 0.028 0.016 325 0.015 0 0.012 0 0.014 0.018 0.025 0 0 0.008 0.012 0 0.013 0.016 0.007 0.019 0 327 0.015 0 0.012 0 0.014 0.035 0 0 0.014 0.016 0.012 0 0.013 0 0.027 0.037 0 329 0 0 0.012 0.029 0 0.018 0 0 0 0.041 0 0.014 0 0.032 0.014 0.028 0.047 331 0.015 0 0.035 0 0.014 0 0 0.038 0 0.008 0 0.027 0.025 0 0.027 0.083 0 333 0 0 0.012 0 0 0.018 0 0 0.014 0 0 0.014 0.038 0 0 0.009 0.016 335 0 0 0 0 0 0 0 0.019 0 0.016 0.012 0.027 0 0 0.007 0 0.016 339 ' 0 0 0.023 0 0.014 0.035 0 0 0 0.008 0.012 0.014 0 0 0 0.009 0 343 0 0 0.012 0 0.027 0 0 0 0 0 0 0 0 0 0.007 0.019 0 345 0 0 0 0 0 0.018 0 0 0 0 0 0 0 0 0.007 0.019 0.016 347 0 0 0.012 0 0 0 0 0.019 0 0.008 0 0 0.025 0 0 0.037 0 349 0 0 0 0 0 0.018 0 0 0.014 0 0 0.014 0 0 0 0 0 351 0 0 0 0 0 0 0 0.019 0 0 0 0 0 0 0 0 0 353 0 0 0 0 0 0 0 0 0 0.008 0 0 0 0 0 0 0 355 0 0 0 0 0 0 0 0 0.014 0 0 0.014 0 0 0 0 0 359 0 0 0 0 0 0 0 0 0.014 0 0 0 0 0 0 0 0 367 0 0 .0 0 0 0 0 0 0 0 0 0 0.013 0 0 0 0 165 Appendix I continued: Slep3 Allele Size CC NB SB Al NUM Gl BM DC FL PC Ul TB SI PI JJ GIB TI 111 0.029 0 0.024 0.067 0 0 0 0 0 0 0 0 0 0 0 0 0 113 0.014 0 0 0 0 0 0 0 0 0 0 , 0 0 0 0 0 0 115 0.014 0 0 0 0 0 0 0 0 0 0 ' 0.018 0 0 0 0 0 121 0 0 0 0 0.013 0 0 0 0 0.008 0 0 0 0 0 0 0 123 0 0 0.012 0 0 0 0 0 0.014 0.008 0.016 0 0.012 0 0.01 0.011 0 125 0 0.02 0 0 0 0 0 0 0 0 0 0 0 0 0 0.022 0 127 0 0.02 0 0 0 0 0.043 0.032 0 0.008 0.048 0.018 0.025 0 0.01 0 0 129 0 0 0.024 0.067 0.027 0.055 0.022 0 0 0.008 0 0.018 0 0 0 0 0 131 0.029 0 0 0 0.013 0.018 0.065 0 0.029 0 0 0 0 0 0.01 0.067 0.055 133 0.014 0.02 0.012 0.033 0.013 0.036 0.022 0.032 0 0 0.016 0.055 0 0 0.02 0.1 0 135 0.058 0 0.012 0.033 0.013 0.091 0 0 0.057 0.025 0.016 0 0.099 0.023 0.029 0.033 0.073 137 0.014 0.061 0.024 0 0.053 0.018 0.022 0.065 0 0.008 0.048 0.018 0.049 0 0.069 0.011 0.018 139 0.101 0.061 0.048 0 0.053 0 0.065 0.129 0.043 0.008 0.032 0.036 0.062 0.091 0.059 0.044 0.018 141 0 0.041 0.024 0.167 0 0.127 0.087 0.032 0.029 0.017 0.032 0.018 0.037 0 0.029 0.078 0.055 143 0.058 0.082 0.06 0.033 0.08 0.036 0.043 0 0.086 0.025 0.048 0 0.062 0.023 0.049 0.033 0.055 145 0 0.02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 147 0.029 0.061 0.084 0.067 0.067 0.036 0.087 0.065 0.143 0.051 0.032 0.091 0.037 0.023 0.059 0.033 0.091 149 0.058 0.041 0.06 0.1 0.04 0.036 0.022 0.065 0.014 0.059 0.048 0.055 0.049 0.045 0.039 0.022 0.036 151 0.043 0.061 0.012 0 0.067 0.018 0.087 0.129 0.043 0.042 0.111 0.018 0.111 0.068 0.157 0.056 0 153 0.014 0.041 0.036 0 0.067 0.109 0.065 0.065 0.114 0.085 0.016 0.055 0.049 0.114 0.088 0.056 0.055 155 0.087 0.041 0.084 0.033 0.027 0.036 0.087 0.032 0.057 0.085 0.079 0.073 0.025 0.045 0.01 0.067 0.036 157 0.029 0.082 0.12 0.067 0.053 0.073 0.022 0.032 0.043 0.034 0.032 0.091 0.037 0.068 0.059 0.033 0.018 159 0.029 0 0.024 0.1 0.067 0.018 0.022 0 0.043 0.144 0.032 0.018 0.049 0.023 0.02 0.044 0.055 161 0.043 0.041 0.084 0.1 0.067 0.036 0.043 0 0.071 0.042 0.032 0.127 0.099 0.045 0.069 0.033 0.164 163 0.043 0 0.036 0.033 0 0.055 0 0.032 0.014 0.008 0.048 0.036 0.012 0.023 0.01 0.011 0.018 165 0.072 0.061 0 0.033 0.08 0.073 0.022 0.032 0.014 0.068 0.079 0 0.062 0.114 0.039 0.022 0.055 167 0.014 0.02 0.06 0 0.013 0.018 0.043 0.032 0.014 0.017 0.016 0.127 0 0.068 0.029 0.044 0 169 0.029 0.041 0 0 0.027 0 0 0 0 0.025 0.016 0.055 0.037 0.023 0.01 0.067 0.073 171 0.043 0.041 0.012 0 0.027 0.018 0.022 0.065 0.014 0.034 0 0 0 0.068 0.049 0.011 0 173 0.014 0.041 0.012 0.067 0.013 0.018 0.043 0.065 0.029 0.017 0.032 0 0.025 0.023 0.02 0 0.018 175 0.029 0 0.024 0 0.013 0 0.022 0 0.029 0.008 0.016 0.018 0 0.023 0.02 0.044 0.018 177 0.014 0 0.024 0 0.027 0 0.022 0 0.014 0.051 0.032 0.036 0.012 0.023 0 0 0.018 179 0.014 0.02 0.012 0 0.013 0.018 0 0.032 0.029 0.017 0.016 0 0.012 0 0 0.022 0.018 181 0.014 0 0.012 0 0.013 0.018 0 0.032 0.014 0.025 0.032 0 0 0 0 0 0.018 183 0 0 0.012 0 0 0 0 0.032 0 0.017 0.032 0 0 0 0.01 0 0.036 185 0.014 0 0.024 0 0.027 0.018 0 0 0.014 0.025 0.016 0.018 0 0.023 0.02 0 0 187 0 0.02 0.012 0 0.013 0 0 0 0.014 0.008 0 0 0 . 0.023 0 0 0 189 0.014 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 191 0.014 0.041 or o' 0 0 0 0 0 0 0 0.016 0 0.012 0.023 0 0 0 193 0 0 0 0 0.018 0 0 0 0 0 0 0.012 0 0 0 0 195 0 0 0.012 0 0.013 0 0 0 0 0 0 0 0 0 0 0.022 0 197 0 0.02 0 0 0 0 0 0 0.014 0 0 0 0 0 0 0.011 0 199 0 0 0 0 0 0 0 0 0 0.008 0 0 0 0 0 0 0 203 0 0 0 0 0 0 0.022 0 0 0 0 0 0.012 0 0.01 0 0 215 0 0 0 0, 0 0 0 0 0 0 0.016 0 0 0 0 0 0 219 0 0 0 0 0 0 0 0 0 0.008 0 0 0 0 0 0 0 166 Appendix II: F s t values of pairwise comparisons among localities by locus. Locus T16 CC NB SB Al NM Gl BM DC FL PC Ul TB SI PI J J GIB TI cc NB 0.002 SB 0.009 -0.005 Al 0.021 -0.009 -0.005 NUM 0.023 -0.001 0.002 -0.008 Gl 0.022 0.004 0.000 0.002 0.001 BM 0.024 -0.001 0.000 -0.008 0.004 -0.004 DC 0.026 -0.001 0.010 -0.007 0.001 0.008 0.007 FL 0.019 -0.008 0.002 -0.003 0.001 0.002 0.002 0.001 PC 0.028 0.001 0.009 0.000 0.002 0.005 0.006 0.003 -0.002 Ul 0.024 -0.007 0.006 -0.008 0.002 0.006 0.001 -0.001 -0.003 0.003 TB 0.023 -0.003 0.006 -0.010 -0.004 0.011 0.003 0.002 0.001 0.002 0.001 SI 0.019 -0.005 0.006 -0.014 0.001 0.016 0.008 0.000 0.003 0.002 0.001 -0.005 PI 0.016 -0.010 -0.006 -0.011 -0.005 -0.004 -0.004 -0.003 -0.005 -0.002 -0.006 -0.004 -0.002 JJ 0.016 0.002 0.001 -0.001 0.001 0.004 0.004 0.010 0.007 0.005 0.008 0.004 0.008 0.002 GIB 0.032 0.010 0.010 0.009 0.007 0.006 0.010 0.021 0.016 0.011 0.020 0.012 0.015 0.004 0.010 TI 0.045 0.003 0.013 -0.012 0.003 0.011 0.004 0.004 0.004 0.005 -0.001 0.004 0.002 -0.001 0.015 0.014 Locus C S L 9 cc NB S B Al N M Gl B M D C FL P C Ul T B SI PI J J GIB TI cc NB 0.010 S B -0.005 0.007 Al . 0.008 0.082 0.024 N U M 0.001 0.063 0.018 -0.018 Gl -0.002 0.002 0.010 0.036 0.025 B M -0.006 -0.006 -0.008 0.039 0.027 0.003 D C 0.021 0.086 0.036 0.010 0.007 0.060 0.040 F L -0.002 0.008 0.012 0.044 0.025 0.002 -0.001 0.027 P C -0.002 0.039 0.005 -0.004 -0.002 0.008 0.011 0.019 0.017 Ul -0.003 0.053 0.012 -0.010 -0.011 0.026 0.016 -0.003 0.014 0.000 T B -0.008 0.027 0.009 0.003 -0.003 -0.002 0.006 0.022 0.002 -0.003 -0.002 SI -0.005 0.033 0.018 0.011 0.000 0.005 0.011 0.011 -0.005 0.006 -0.003 -0.009 PI -0.007 0.012 -0.007 0.021 0.015 0.018 -0.007 0.031 0.011 0.012 0.007 0.009 0.015 J J -0.006 0.017 0.008 0.024 0.011 0.004 -0.001 0.014 -0.010 0.008 0.003 -0.003 -0.008 0.007 GIB -0.002 0.013 0.015 0.035 0.019 -0.003 0.004 0.036 -0.006 0.015 0.015 -0.003 -0.005 0.016 -0.004 TI -0.011 0.019 -0.002 0.009 0.000 0.005 -0.005 0.013 -0.001 -0.003 -0.006 -0.008 -0.006 -0.002 -0.007 -0.001 l' Appendix II continued: Locus T12 CC NB SB Al NM Gl BM DC FL PC Ul TB SI PI JJ GIB TI cc NB -0.001 S B -0.002 -0.002 Al -0.004 -0.007 0.001 N U M -0.003 -0.005 -0.006 -0.006 Gl -0.004 0.001 0.005 -0.009 0.000 B M 0.007 0.005 0.006 -0.008 -0.001 0.003 D C 0.019 0.004 0.009 0.000 -0.002 0.011 -0.006 FL 0.033 0.014 0.017 0.005 0.004 0.024 0.003 -0.006 P C 0.012 0.007 0.009 -0.004 0.008 0.001 -0.006 0.006 0.011 Ul 0.013 0.021 0.019 0.005 0.020 0.013 0.000 0.025 0.030 -0.002 T B 0.018 0.014 0.013 -0.003 0.014 0.004 -0.001 0.013 0.016 -0.003 0.000 SI 0.021 0.011 0.014 -0.002 0.011 0.003 -0.002 0.003 0.012 -0.002 0.011 -0.005 PI 0.016 0.010 0.010 -0.014 0.006 0.004 -0.003 0.008 0.006 -0.001 0.013 -0.003 -0.002 J J 0.005 -0.001 0.001 -0.009 -0.002 -0.003 0.000 0.007 0.013 -0.001 0.012 0.002 0.000 -0.003 GIB 0.011 0.023 0.021 0.002 0.017 0.005 0.013 0.032 0.039 0.015 0.020 0.013 0.020 0.013 0.014 TI 0.001 -0.002 -0.003 -0.010 -0.001 -0.005 0.001 0.008 0.011 -0.003 0.006 -0.001 0.006 -0.004 -0.003 0.010 Locus Slep 9 CC NB SB Al NM Gl BM DC FL PC Ul TB SI PI JJ GIB TI cc NB -0.002 S B -0.002 -0.001 Al 0.001 -0.011 0.001 N U M -0.005 -0.006 -0.001 0.002 G l -0.005 -0.011 -0.002 -0.013 -0.006 B M 0.001 -0.008 0.000 -0.002 -0.010 -0.003 D C 0.001 -0.006 0.002 0.003 -0.009 -0.003 -0.001 FL -0.001 -0.008 0.004 0.004 -0.002 -0.003 0.004 -0.006 P C -0.003 -0.002 0.005 0.000 -0.002 -0.003 0.003 -0.005 0.002 Ul 0.008 -0.005 0.001 0.004 -0.001 -0.001 0.001 -0.008 -0.003 0.003 TB -0.005 -0.001 :0.001 0.002 0.001 0.000 0.000 0.000 -0.003 0.000 -0.001 SI 0.005 -0.003 0.007 -0.005 -0.001 -0.008 0.004 0.002 0.004 0.003 0.004 0.006 PI -0.007 -0.010 -0.005 -0.009 -0.008 -0.010 -0.007 -0.001 -0.004 -0.002 -0.003 -0.004 -0.004 J J -0.001 -0.006 -0.002 -0.004 -0.002 -0.007 0.003 -0.004 0.001 -0.001 0.000 0.004 0.002 -0.005 GIB 0.014 0.006 0.010 0.023 0.002 0.004 0.017 -0.002 0.007 0.009 0.009 0.019 0.010 0.008 0.004 TI 0.012 -0.001 0.008 0.005 0.003 0.000 -0.005 0.008 0.006 , 0.007 0.006 0.006 0.003 0.004 0.007 0.017 Locus Slep 3 CC NB SB Al NM Gl BM DC FL PC Ul TB SI cc NB -0.006 S B 0.004 -0.003 Al 0.011 0.004 0.006 N U M -0.006 -0.012 0.001 0 008 Gl 0.007 0.002 0.009 -0 005 0 004 B M -0.002 -0.011 -0.001 -0 003 -0 006 -0 002 D C 0.000 -0.012 0.011 0 013 -0 007 0 004 -0.010 FL 0.004 -0.001 0.002 0 011 -0 003 0 003 -0.009 0.008 P C 0.005 0.007 0.011 0 008 -0 002 0 011 0.004 0.008 0.006 Ul -0.005 -0.009 0.006 0 012 -0 004 0 008 -0.007 -0.015 0.009 0.005 T B 0.010 0.003 -0.006 0 010 0 005 0 014 0.000 0.015 0.005 0.012 0.012 SI -0.003 -0.005 0.011 0 009 -0 004 0 005 0.001 0.000 0.004 0.008 -0.001 0.012 PI -0.005 -0.006 0.010 0 023 -0 007 0 003 0.002 -0.008 0.004 0.007 0.005 0.007 0.004 J J 0.003 -0.007 0.007 0 016 -0 005 0 006 -0.004 -0.016 0.006 0.011 0.000 0.007 -0.002 GIB 0.001 0.000 0.011 0 009 0 004 0 005 -0.007 0.003 0.007 0.011 0.004 0.006 0.007 TI 0.004 0.004 0.008 -0 001 0 001 0 006 0.001 0.019 -0.002 0.007 0.013 0.000 -0.002 GIB 0.009 0.006 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0093068/manifest

Comment

Related Items