Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Patterns and consequences of dispersal for Arctic Char (Pisces: Salvelinus alpinus) from the Canadian… Moore, Jean-Sébastien 2012

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

Item Metadata

Download

Media
24-ubc_2013_spring_moore_jeansebastien.pdf [ 8.94MB ]
Metadata
JSON: 24-1.0073414.json
JSON-LD: 24-1.0073414-ld.json
RDF/XML (Pretty): 24-1.0073414-rdf.xml
RDF/JSON: 24-1.0073414-rdf.json
Turtle: 24-1.0073414-turtle.txt
N-Triples: 24-1.0073414-rdf-ntriples.txt
Original Record: 24-1.0073414-source.json
Full Text
24-1.0073414-fulltext.txt
Citation
24-1.0073414.ris

Full Text

PATTERNS AND CONSEQUENCES OF DISPERSAL IN ARCTIC CHAR (PISCES: SALVELINUS ALPINUS) FROM THE CANADIAN ARCTIC by Jean-Sébastien Moore BSc., McGill University, 2004 MSc., McGill University, 2007 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (ZOOLOGY) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2012 © Jean-Sébastien Moore, 2012  !"#$%!&$' Dispersal can have a multitude of ecological and evolutionary consequences that can be either positive or negative for population fitness and persistence. In this thesis, I describe patterns of dispersal in Arctic Char (Salvelinus alpinus), and I explored some of its consequences. I first examined the consequences of post-glacial dispersal for the distribution of genetic variation across the Canadian range of the species. MtDNA sequences and microsatellite markers provided evidence that the populations of Arctic Char currently inhabiting the Arctic Archipelago probably recolonized from a small glacial refugium, most likely located in ice-free areas of the Archipelago itself. I also presented evidence that two glacial lineages of Char (an Arctic lineage and an Atlantic lineage) probably hybridized post-glacially in the eastern Arctic. Finally, the importance of contemporary dispersal in redistributing genetic variation was illustrated by the fact that anadromous populations have greater within-population genetic diversity, and are less genetically differentiated, than landlocked populations. Second, I used a genetic assignment approach to study patterns of dispersal among populations distributed around Cumberland Sound, Nunavut. Estimates of dispersal rates varied extensively depending on the analysis method used, but all were relatively high compared to other salmonid species. I also found evidence that overwintering individuals have a greater propensity to disperse to non-natal habitats than individuals destined to spawn that year. The consequences of this behaviour for local adaptation among populations was examined using a population genetic model parameterized with estimates of gene flow obtained from microsatellite data. Third, I tested alternative hypotheses for the co-existence of sympatric migratory ecotypes in three lakes of southeast Baffin Island. Microsatellite data showed that the resident and anadromous components of the population are not genetically differentiated, suggesting that migratory behavior is not a genetically fixed trait. Together, the three parts of my thesis provide a general understanding of the patterns and consequences of dispersal for Arctic Char. Since dispersal will be crucial for the response of Arctic Char to environmental change, I conclude by discussing how my work can serve as a foundation for future work on the role of dispersal in adaptation to a changing Arctic.  ii  (%)*!&)' Chapter 2 was a project I conceived and conducted in collaboration with R. Bajno, J.D. Reist, and E.B. Taylor. The samples were collected by various people working under J.D. Reist over the years. I collected all the microsatellite data and some of the mtDNA sequence data. I performed all data analyses, and wrote the original manuscript. All three other collaborators provided advice and contributed revisions. Chapter 3 was a project I conceived and conducted in collaboration with L.N. Harris, R.F. Tallman, and E.B. Taylor. Adult samples used in this study were collected by various people working under R.F. Tallman over the years, and I collected all samples of juveniles according to the sampling design I conceived. L.N. Harris and I conducted all the laboratory work. I performed all data analyses and wrote the original manuscript. All three collaborators provided advice and contributed revisions. A version of chapter 4 has been submitted as: “T.N. Loewen, L.N. Harris & R.F. Tallman (in review) Genetic analysis of sympatric migratory ecotypes of Baffin Island Arctic Char Salvelinus alpinus: alternative mating tactics or reproductively isolated strategies?” T.N. Loewen and I collected the samples, and L.N. Harris and I conducted the laboratory work. I conducted all data analyses and wrote the original version of the manuscript. All three collaborators provided advice and contributed revisions to the manuscript. Permits for Arctic Char collection were obtained from Fisheries and Oceans Canada (Licenses to fish for scientific purpose S-09/10-1007-NU and S-10/11-1030-NU). All collections were done according to the Animal Use Protocol of Fisheries and Oceans’ Freshwater Institute’s Animal Care Committee (animal use protocols # FWI-ACC-09/101007-NU and FWI-ACC-10/11-1030-NU).  iii  $!"+)',*'&,-$)-$#' !"#$%!&$'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('))! *%+,!&+'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((')))! $!"-+'.,'&./$+/$#'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((')0! -1#$'.,'$!"-+#'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('0)))! -1#$'.,',123%+#'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('4! !&5/.6-+72+8+/$#'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('4))! 9! 1:;<=>?@;)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9! 9(9! 7AB):);)=:C'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9! 9(D! $EA'@=:CAF?A:@AC'=B'>)CGA<CHI'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('J! "#$#"! %&'!()*+,-*'!'./)(*+/0-12!3-(4'4!/5!6+47'14-)!####################################################!8! "#$#$! 93/)/:+3-)!3/04';('03'4!/5!6+47'14-)!######################################################################!<! "#$#=! 9./)(*+/0-12!3/04';('03'4!/5!6+47'14-)!-06!:'0'!5)/>!##################################!?! 9(K! #;?>L'CLC;AMN'H:H><=M=?C'!<@;)@'&EH<'):';EA'&H:H>)H:'!<@;)@'((((((((((((('99! "#=#"! %-@/0/,2!##########################################################################################################################!"$! "#=#$! A+4*1+B(*+/0!#######################################################################################################################!"=! "#=#=! C+/)/:2!#################################################################################################################################!"=! "#=#D! E,7/1*-03'!-4!-!5+4&'1+'4!1'4/(13'!#########################################################################!"8! "#=#8! F!B1+'5!0/*'!/0!47'))+0:!################################################################################################!"G! 9(J! !'<A0)AO'=B'>)CGA<CHI'):'!<@;)@'@EH<'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9P! 9(Q! %ACAH<@E'F?AC;)=:CR'HGG<=H@E'H:>'S?C;)B)@H;)=:'(((((((((((((((((((((((((((((((((((((((((('DT! "#8#"! F771/-3&'4!*/!*&'!4*(62!/5!6+47'14-)!#####################################################################!$"! "#8#$! H'-4(1+0:!6+47'14-)!-06!:'0'!5)/>!########################################################################!$$! "#8#=! I(,,-12!-06!/1:-0+J-*+/0!/5!3&-7*'14!################################################################!$8! D! !'@<LG;)@'UIH@)HI'<AB?U)?MR'ELV<)>)WH;)=:R'H:>'H:H><=ML'CEHGA';EA' >)C;<)V?;)=:'=B'UA:A;)@'0H<)H;)=:'):'&H:H>)H:'!<@;)@'&EH<'(((((((((((((((((((((((((((((((((((((('KD! D(9! !VC;<H@;'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('KD! D(D! 1:;<=>?@;)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('KK!  iv  D(K! 8H;A<)HIC'H:>'MA;E=>C'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('KX! $#=#"! I-,7)'4!###############################################################################################################################!=<! $#=#$! AKF!,'*&/64!####################################################################################################################!=L! $#=#=! ,*AKF!-0-)24+4!###############################################################################################################!=L! $#=#D! M+4*/1+3-)!6',/:1-7&2!#################################################################################################!DN! $#=#8! H+31/4-*'))+*'!-0-)24+4!##################################################################################################!D"! D(J! %AC?I;C'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('JK! $#D#"! ,*AKF!.-1+-*+/0!#############################################################################################################!D=! $#D#$! M+4*/1+3-)!6',/:1-7&2!#################################################################################################!DD! $#D#=! H+31/4-*'))+*'!AKF!#########################################################################################################!D8! D(Q! 7)C@?CC)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('JY! $#8#"! 9.+6'03'!5/1!-!&+:&!F13*+3!1'5(:+(,!5/1!F13*+3!3&-1!########################################!D?! $#8#$! I'3/06-12!3/0*-3*!>+*&!*&'!F*)-0*+3!)+0'-:'!#######################################################!8=! $#8#=! 955'3*4!/5!-0-61/,2!/0!:'0'*+3!4*1(3*(1'!-06!+4/)-*+/0!B2!6+4*-03'!#######!88! $#8#D! O/03)(4+/04!-06!+,7)+3-*+/04!5/1!-6-7*-*+/0!*/!-!3&-0:+0:!F13*+3!###########!8<! K! $EA'):;A<GIHL'VA;OAA:'>)CGA<CHIR'UA:A'BI=OR'H:>'G=;A:;)HI'B=<'I=@HI' H>HG;H;)=:'):'"HBB):'1CIH:>'H:H><=M=?C'!<@;)@'&EH<'((((((((((((((((((((((((((((((((((((((((((((((('PY! K(9! !VC;<H@;'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('PY! K(D! 1:;<=>?@;)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('PZ! =#=! H-*'1+-)4!-06!,'*&/64'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('XK! =#=#"! P+')6!4-,7)+0:!##################################################################################################################!<=! =#=#$! H+31/4-*'))+*'!AKF!-0-)24+4!#######################################################################################!<=! =#=#=! O/,7-1+4/04!B'*>''0!-6()*!-06!Q(.'0+)'!4-,7)'4!##########################################!<D! =#=#D! R'0'*+3!-44+:0,'0*4!######################################################################################################!<8! =#=#8! S/*'0*+-)!5/1!)/3-)!-6-7*-*+/0!####################################################################################!<<! K(J! %AC?I;C'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('XY! =#D#"! H+31/4-*'))+*'!7/)2,/17&+4,!####################################################################################!<?! =#D#$! F44+:0,'0*!*'4*4!#############################################################################################################!<L! =#D#=! C1''6+0:!6+47'14-)!-06!/.'1T>+0*'1+0:!6+47'14-)!############################################!?"! =#D#D! U*&'1!3/11')-*'4!/5!6+47'14-)!71/7'04+*2!##############################################################!?"!  v  =#D#8! S/*'0*+-)!5/1!)/3-)!-6-7*-*+/0!####################################################################################!?$! K(Q! 7)C@?CC)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('YD! =#8#"! M+:&!6+47'14-)!+0!-0-61/,/(4!F13*+3!O&-1!##########################################################!?=! =#8#$! O/06+*+/0T6'7'06'0*!6+47'14-)!################################################################################!?D! =#8#=! S/*'0*+-)!5/1!)/3-)!-6-7*-*+/0!####################################################################################!??! J! 2A:A;)@'H:HILC)C'=B'CLMGH;<)@'M)U<H;=<L'A@=;LGAC'=B'"HBB):'1CIH:>'!<@;)@' &EH<N'HI;A<:H;)0A'MH;):U';H@;)@C'=<'<AG<=>?@;)0AIL')C=IH;A>'C;<H;AU)AC['(((((((((((('ZZ! J(9! !VC;<H@;'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('ZZ! J(D! 1:;<=>?@;)=:'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9TT! J(K! 8H;A<)HIC'H:>'MA;E=>C'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9TK! D#=#"! P+')6!4-,7)+0:!###############################################################################################################!"N=! D#=#$! H+31/4-*'))+*'!AKF!-0-)24+4!####################################################################################!"N=! J(J! %AC?I;C'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9TP! J(Q! 7)C@?CC)=:'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9TX! D#8#"! V-3W!/5!:'0'*+3!6+55'1'0*+-*+/0!+0!O(,B'1)-06!I/(06!'3/*27'4!##############!"N?! D#8#$! S1'4'03'!/5!>'-W!:'0'*+3!6+55'1'0*+-*+/0!+0!VMNN"!######################################!"NL! D#8#=! 9.+6'03'!5/1!-)*'10-*+.'!,-*+0:!*-3*+34!#############################################################!""$! D#8#D! E,7)+3-*+/04!5/1!5+4&'1+'4!,-0-:','0*!#############################################################!""D! D#8#8! O/03)(4+/04!####################################################################################################################!""G! Q! &=:@I?C)=:'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9DJ! Q(9! 1MGI)@H;)=:C'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9DQ! 8#"#"! E,7)+3-*+/04!5/1!F13*+3!O&-1!B+/)/:2!####################################################################!"$8! 8#"#$! E,7)+3-*+/04!5/1!4-),/0+6!B+/)/:2!########################################################################!"$<! 8#"#=! E,7)+3-*+/04!5/1!*&'!4*(62!/5!6+47'14-)!###############################################################!"=N! Q(D! ,?;?<A'>)<A@;)=:C'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9K9! 8#$#"! U(*4*-06+0:!;('4*+/04!###############################################################################################!"="! 8#$#$! X+))!'./)(*+/0-12!-6-7*-*+/0!&')7!F13*+3!3&-1!3/7'!>+*&!3)+,-*'!3&-0:'Y !  "=D!  Q(K! &I=C):U'<AMH<\C'((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9KP! ")VI)=U<HGEL'(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((('9KX! vi  !GGA:>)4'!N'#?GGIAMA:;H<L'MH;A<)HIC'B=<'&EHG;A<'D'(((((((((((((((((((((((((((((((((((((((((((('9X9! !GGA:>)4'"N'#?GGIAMA:;H<L'MH;A<)HIC'B=<'&EHG;A<'K'(((((((((((((((((((((((((((((((((((((((((((('9YD!  vii  +.#$',*'$!"+)#' Table 1.1 Some definitions of dispersal from recent reviews on the topic. ......................27! Table 1.2 Estimates of dispersal rates from tagging studies performed on Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma. .....................................................28! Table 1.3 Summary of evidence for increased propensity to disperse when overwintering in Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma..............................28! Table 1.4 Summary of studies examining the marine migrations of Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma. .....................................................29! Table 2.1. Results of the historical demography analyses using neutrality tests (Fu’s Fs and Tajima’s D) and mismatch distribution analysis on mitochondrial DNA sequence variation (D-loop) in Arctic Char, Salvelinus alpinus. ..............................58! Table 2.2. Differences in genetic diversity (allelic richness and expected heterozygocsity) and average pairwise FST between landlocked and anadromous populations of Arctic Char, Salvelinus alpinus, assayed at nine microsatellite DNA loci...........................58! Table 3.1. Information regarding the sampling locations used in this study, and measures of genetic diversity at each location for anadromous Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci. ............................................................91! Table 3.2. Results of discriminant function analysis testing for trait differences between philopatric and dispersing individuals of anadromous Arctic Char (Salvelinus alpinus) sample from Cumberland Sound. ................................................................93! Table 4.2. Details of the primers and PCR reactions used for each of the four multiplexes used to assess genetic variation of resident and anadromous Arctic Char (Salvelinus alpinus). ...................................................................................................................117! Table 4.3. Summary statistics of microsatellite variation for each ecotype of Arctic Char (Salvelinus alpinus) per sampling location.. ............................................................118! Table 4.4. Semi-matrix of pairwise FST (Weir & Cockerham, 1984) values between all sampling locations and ecotypes of Arctic Char (Salvelinus alpinus). ...................119! Table 4.5. Results of the analysis of molecular variance (AMOVA) examined among sampling locations and ecotypes of Arctic Char (Salvelinus alpinus). ...................119!  viii  Table A.1. Name and geographic coordinates of sampling locations where specimens of Arctic Char (ARCH, Salvelinus alpinus) and Dolly Varden (DVCH, Salvelinus malma) used for mtDNA sequencing were collected. .............................................171! Table A.2. Sampling location information for the samples used in the microsatellite analysis of Arctic Char (Salvelinus alpinus) ...........................................................174! Table A.3. Frequency of haplotypes of Arctic Char (Salvelinus alpinus) in each sampling locations. ..................................................................................................................175! Table A.4. Summary population statistics for the microsatellite analysis across nine loci in Arctic Char (Salvelinus alpinus). ........................................................................179! Table B.1. Details of the primers and PCR reactions used for each of the 4 multiplexes. .................................................................................................................................183! Table B.2. Heterozygosity (observed and expected) and FIS (Weir & Cockerham 1984) per locus in samples of Arctic Char (Salvelinus alpinus) ........................................185! Table B.3. Pairwise FST (Weir and Cockerham 1984) between each Arctic Char (Salvelinus alpinus) sample used in the study. ........................................................189! Table B.4. Summary of sibship reconstruction among samples of juvenile Arctic Char (Salvelinus alpinus) using the program COLONY..................................................192!  ix  +.#$',*'*./0%)#' Figure 1.1. Arctic Char, Salvelinus alpinus. The drawing is based on an anadromous specimen in spawning colours from Anderson Bay Lake in the Cambridge Bay Region of Nunavut, Canada.......................................................................................30! Figure 1.2 Map of the northern hemisphere showing the Holarctic distribution of Arctic Char, Salvelinus alpinus, in dark shaded areas surrounded by dashed lines. ............31! Figure 2.1. Map of the Northern Hemisphere showing the distribution and names of the five major mitochondrial DNA lineages of Arctic Char (Salvelinus alpinus) identified in the Holarctic phylogeography study of Brunner et al. (2001). .............59! Figure 2.2. Maximum likelihood phylogenetic tree of Arctic Char (Salvelinus alpinus) mtDNA haplotypes generated using GARLI (Zwickl 2006).....................................60! Figure 2.3 (Next page) Map showing the geographical distribution of the two mtDNA lineages identified in North American Arctic Char (Salvelinus alpinus) and Dolly Varden (S. malma). ....................................................................................................61! Figure 2.4. Results of the pairwise mismatch analysis done on the mtDNA data for two lineages of Arctic Char (Salvelinus alpinus). ............................................................63! Figure 2.5. Neighbour-joining tree of Cavalli-Sforza’s chord measure genetic distance from nine microsatellite loci assayed in samples of Arctic Char (Salvelinus alpinus). ...................................................................................................................................64! Figure 2.6. Variation in expected heterozygosity as a function of waterway distance from a putative glacial refugium (Banks Island) in samples of Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci. ....................................................65! Figure 2.7. Results of the STRUCTURE analysis based on the most likely (sensu Evanno et al. 2005) number of genetic cluster: K = 4 in samples of Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci. ....................................................66! Figure 2.8. Relationship between pairwise genetic distance measured as linearized FST (Rousset 1997) and geographical distance (km) in samples of Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci. ..................................67!  x  Figure 3.1. Map of Cumberland Sound, Baffin Island, Nunavut, Canada. Sampling locations for Arctic Char (Salvelinus alpinus) examined in this study are shown with a dot, and identified with a three letter code referenced in Table 3.1 .......................94! Figure 3.2. Summary of population genetic structure among Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci from locations of Cumberland Sound. ........................................................................................................................95! Figure 3.4. Results of the STRUCTURE analysis for anadromous Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci. ...........................97! Figure 3.5. The relationship between the critical value of the selection coefficient, s, and the migration rate, m, for different values of the mutation rate (both axes are on a log scale) estimated for anadromous Arctic Char (Salvelinus alpinus). ..........................98! Figure 4.1. Map showing the location of the three lakes with the sympatric migratory ecotypes of Arctic Char (Salvelinus alpinus) in southeastern Baffin Island. .........120! Figure 4.2. Results of the factorial correspondence analysis (FCA) conducted in GENETIX for samples of Arctic Char (Salvelinus alpinus). ..................................121! Figure 4.3. Results of the STRUCTURE analysis on all sampling locations of Arctic Char (Salvelinus alpinus).........................................................................................122! Figure 4.4. Results of the STRUCTURE analysis on single sampling locations of Arctic Char (Salvelinus alpinus).........................................................................................123! Figure A.1. Details of the STRUCTURE analysis for Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci are presented in the text. .........................180! Figure A.2. Bubbles plots of microsatellite allele frequencies showing introgression of Labrador Arctic Char (Salvelinus alpinus) alleles (the site at the extreme right) in other population of the Arctic Archipelago. ............................................................181! Figure B.1 Frequency distributions of gonadosomatic index of males (top) and females (bottom) Arctic Char (Salvelinus alpinus)...............................................................182! Figure B.2. Frequency histograms of fork length of juvenile Arctic Char (Salvelinus alpinus) showing that multiple cohorts (seen as different length-classes in the histograms) were collected for most (but not all) sampling locations. ....................191! Figure B.3. Biological correlates of dispersal propensity in Arctic Char (Salvelinus alpinus). ...................................................................................................................195!  xi  !&1-,2+)3/)4)-$#' First of all, I would like to thank my supervisor, Eric Taylor. His confidence and encouragement when I proposed a risky thesis topic made all of what follows possible. I am grateful for the latitude he gave me, and for the always-prompt help he provided whenever I needed it. I also thank my supervisory committee for much insightful advice: Dolph Schluter, Wayne Maddison, Ross Tallman, Jim Reist, and honorary member Michael Whitlock. Ross and Jim deserve further thanks: their forethought of preserving tissue samples collected from many years of Arctic fieldwork allowed my work to span far greater geographical and temporal scales than if I had to collect them myself. Their liberal sharing of samples, resources and knowledge greatly helped me on my way to building a career as an Arctic scientist. My office mates, Jon Mee and Aleeza Gerstein, probably deserve more than a brief mention. I hope I was able to convey in person the extent of my gratitude for everything they have done and meant for me. For now, suffice to say that when it got tough, I thought to myself: ‘we got each other, and that’s a lot’. A lot of other UBC students provided support (scientific and otherwise) and will, I am sure, remain life-long friends. I can’t name them all here, but special thanks go to Rowan Barrett, Alistair Blachford, Gwylim Blackburn, Alan Brelsford, Iain Caldwell, Gina Conte, Carla Crossman, Anne Dalziel, Florence Débarre, Simone Des Roches, Stefan Dick, Rich Fitzjohn, Jennifer Gow, Jonathan Griffiths, Frédéric Guillaume, Joel Heath, Jessica Hill, Crispin Jordan, Julie Lee-Yaw, Milica Mandic, Leithen M’Gonigle, Sara Northrup, Damon Nowasad, Jasmine Ono, Chad Ormond, Kate Ostevik, Antoine Paccard, Jess Purcell, Kieran Samuk, Alana Schick, Matt Siegle, Andrea Stephens, Patrick Tamkee, Patrick Thompson, Dave Toews, Laura Tremblay-Boyer, Kathryn Turner, Thor Veen, Gerrit Velema, Tim Vines, Patricia Woodruff, Monica Yau, and Sam Yeaman. A few non-UBC people have also contributed to my enjoyment of living in BC. First among them is Rod Docking, with whom I had what has been described as a common-law  xii  relationship for almost three years: my sanity and guitar skills benefitted greatly from our sharing an apartment. Klaus Gantner helped me out of a rough patch and provided badly needed friendship during my brief stay in Victoria. Many thanks also to Katharine Hudson for your love and support. Working at UBC has been a tremendous opportunity. I doubt I will ever again find myself surrounded by such a distinguished and friendly group of academics. Special thanks go to Mike Whitlock, Dolph Schluter, and Sally Otto for consistently showing up for discussion groups and sharing their knowledge and opinions with such enthusiasm. The tireless work of Sally to promote a collegial and fun social environment in the department has contributed immensely to my enjoying my time at UBC. Before I joined the Taylor Lab, I e-mailed a MSc student who was working on Arctic fishes with Rick. I wanted to know more about his project and about working in the Arctic. In typical Les Harris fashion, he replied: “Let’s go for a beer when you get here and we can chat about it”. Les almost instantly became a good friend, and that beer he promised turned into probably several hundreds over the years. When he joined Fisheries and Oceans as an Arctic fisheries biologist, an extremely fruitful collaboration began from which I benefitted disproportionately. Les did a lot of the lab work in this thesis, and I would probably still be optimizing PCR reactions without his help. And that is not even mentioning the considerable intellectual input he had on some of the chapters. Many thanks, and here’s to many more years of collaboration and friendship! Simon Wiley is a man’s man: he plays hockey, he shoots guns, he knows boats and small engines, he can build stuff, and he doesn’t even shop for his own clothes. These qualities (except maybe for that last bit) made him an amazing person to guide me through the challenges of Arctic fieldwork. It still baffles me that I did 4 field seasons in the Arctic and that essentially nothing serious went wrong. I credit Simon for much of this success – and for much of the fun I’ve had doing it. A bunch of other people from Fisheries and Oceans helped for field, lab and samples related issues. Tracey Loewen kindly accepted that I use her samples for chapters 3 and 4. Melanie Toyne, helped with lots of paper work and with community consultations, among many other things. Zoya Martin and Christopher Lewis from the Iqaluit office were always helpful and welcoming  xiii  when I was going through town. It’s a real shame the current government doesn’t realize how many amazing people work for the future of Canadian fisheries. Working in the Arctic has its challenges, but it also has many rewards. Chief among them was the privilege of working with Inuit resource users. Their capacity to make a living in such a harsh place is incredibly inspiring, and the depth of their knowledge of their land and its animals is humbling. The following people helped greatly in the field, and shared knowledge, stories, food, and friendship with me: Joe Akpailaluk, Jamesie Ishulutak, Jacopie Kakee, Patrick Kilabuk, Jevua Maniapik, Noah Maniapik, Norman Mike, Noah Mosesee, David Nakashuk, Tony Nauyuk, Peterosie Qappik, Jake Shoapik, Noah Shoapik, Sakiasie Sowdlooapik. To all of you, and to those I forgot here: !"#$%&! Un gros merci à ceux que j’ai laissé derrière au Québec le temps de mon doctorat. Merci aux boys de St-Césaire qui m’aident toujours à garder les pieds sur terre: Mario Bouthillier, Keaven Larose, et Sébastien Mailloux, et à Benjamin Provencher et Étienne Tremblay, sur qui je peux toujours compter. Merci finalement à ma famille pour votre support et vos visites. Vincent, Alexandre, et Marie-Claude : je pourrais difficilement demander meilleurs frères et soeurs. Pas étonnant, peut-être, avec des parents comme les nôtres. Je ne pourrai jamais assez vous remercier, papa et maman, d’avoir cru en moi et de m’avoir supporté à tous moments lors de cette belle aventure. Finally, I would like to acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada, the Fond Québécois de Recherche sur la Nature et les Technologies, the Polar Continental Shelf Project, the Department of Aboriginal Affairs and Northern Development, the Association of Canadian Universities for Northern Studies, and the University of British Columbia.  xiv  Cette thèse est dédiée à Camille et Nicole Moore  xv  5 '.6789:;<7=96' Movement is a fundamental aspect of the biology of most organisms. Yet, there is tremendous variation in dispersal capabilities among species. Plants have evolved fruits and winged seeds to allow their seeds to disperse. Many marine species have pelagic larvae that are passively carried by currents over hundreds of kilometers. Salmonid fishes (Actinopterygii; Salmonidae), on the other hand, have evolved precise homing capabilities that allow them to return to their natal habitat to spawn following extensive marine migrations, thus reducing dispersal between spawning aggregations. This formidable variation in dispersal strategies among organisms constitutes, as pointed out by Roff & Fairbairn (2001), prima facie evidence that they are the results of evolution. Not only have they evolved, but the fact that some organisms adopt strategies that increase their ability to disperse while others have strategies that minimize dispersal suggests that the costs and benefits of dispersal are not the same for all organisms. Whether they increase or decrease dispersal capabilities, the very fact that these traits evolved demonstrates that the consequences of dispersal are important enough to offset the costs of building such costly and elaborate traits that facilitate or minimize dispersal. This dissertation is concerned with dispersal, and focuses on describing patterns and consequences of dispersal for a salmonid fish –the Arctic Char, Salvelinus alpinus (Linnaeus, 1758).  5>5 3?@=6=7=96A' There are multiple definitions of dispersal currently in use in the literature. Because definitions differ on some fundamental aspects, the expected consequences of dispersal can vary according to which definition is used. It is thus important to clarify which definition I will use throughout the thesis (Table 1.1). The definitions in Table 1.1 can be organized according to whether or not they recognize potential for gene flow as a fundamental aspect of dispersal. Greenwood (1980) remarked that Howard’s (1960) definition was the most widely used at the time, and my non-exhaustive list (Table 1.1)  1  suggests that most current research still adopts a definition restricting dispersal to movement potentially leading to gene flow. This focus on gene flow is perhaps not surprising. Indeed, much work on dispersal has been conducted with the use of theoretical models, which often focus on the evolution of dispersal, or on the evolutionary consequences of dispersal (Ronce 2007). Dispersal obviously has the greatest relevance for these two factors when it has the potential to lead to gene flow. Second, botanists have been responsible for a considerable portion of the work on dispersal, and because plant dispersal mainly involves either seeds or gametes, plant dispersal must almost necessarily have the potential to lead to gene flow. As Bowler and Benton (2005) remarked, however, this definition fails to capture some of the complexities of animal behaviour, which can move over long distances and make decisions regarding settlement in a new patch that have nothing to do with breeding. As will be discussed later, Arctic Char represents a case of an animal with a behaviour too complex to be satisfactorily encapsulated by the narrower definition of dispersal. Throughout the thesis, I will therefore adopt the broad definition suggested by Bowler and Benton (2005), and define dispersal as any inter-patch movement, whether it has the potential of resulting in gene flow or not. It should be noted that despite some variation in the definitions of dispersal, most authors agree that dispersal is a three-step process involving (1) the decision to leave the current patch (emigration); (2) movement between patches (inter-patch movement); and (3) the decision to settle in a new patch (immigration) (Bowler & Benton 2005; Ronce 2007; Clobert et al. 2009). It is useful to consider these three steps when defining dispersal, as it helps make the broad definition suggested by Bowler & Benton (2005) more precise. Dispersal, then, is the movement an individual makes between suitable habitat patches that are separated by unsuitable habitat patches. Furthermore, I would argue that for this movement to constitute dispersal, the decision to settle to a new patch must involve significant costs or benefits for the individual. The choice of over-wintering habitats by Arctic Char would therefore conform to this definition. Other potential examples would be the choice of over-wintering habitats by monarch butterflies (e.g., Calvert & Brower 1986), amphibians (e.g., Holenweg & Reyer 2000), and bats (e.g., Kurta & Murray 2002). In all these cases, individuals make a choice to settle in a suitable  2  habitat patch that is surrounded by potentially unsuitable patches, they settle in that patch for an extended period of time relative to their life span, and the quality of the habitat they choose will have important consequences for their survival or their fitness. Such movement can thus be understood under the framework of dispersal, even if it does not involve breeding or the potential for gene flow. Greenwood (1980), who adopts Howard’s (1960) general definition of dispersal, also divided dispersal in sub-categories that are still widely used. First, he distinguished between natal and breeding dispersal, the former referring to a permanent movement from the birth site to the site of first breeding and the latter referring to movement between successive breeding sites. Second, he distinguished between gross dispersal, the “permanent movement of individuals to a new location irrespective of whether or not they reproduce after dispersing”, and effective dispersal, the dispersal events that result in reproduction (Greenwood 1980). While dispersal and migration are sometimes used interchangeably, they refer to fundamentally different processes. Because Arctic Char is a migratory animal, the definition of the term migration should be discussed here since it has been defined in various ways. In population genetics, the term is often equated with gene flow, and the parameter m (for migration) found in many models of population structure represents the proportion of immigrants in a given population (e.g., Wright 1931; Hedrick 2004). In this thesis, I avoid using the term migration to represent gene flow, and I instead follow Endler (1977) in defining migration as the “relatively long-distance movements made by large numbers of individuals in approximately the same direction at approximately the same time … usually followed by a regular return migration”. In the case of anadromous salmonids, the term migration therefore refers to the periodic movement of individuals from fresh water to the marine environment after smolting (or after spawning for iteroparous species), and to their return migration to fresh water for spawning (Hendry et al. 2004; Quinn 2005). In the salmonid literature, specific terms are used to describe movement and migratory behaviour that should be introduced here. Straying is the use of non-natal freshwater habitat by an individual returning from its marine migration (Quinn 1993; Hendry et al. 2004). Because most salmonid species migrate to fresh water solely for the  3  purpose of spawning, this term is usually understood as dispersal for the purpose of breeding, i.e., dispersal as defined by Howard (1960). To avoid confusion due to the complex lifecycle of Arctic Char (see below), I will generally avoid using the term straying to describe Arctic Char dispersal. In semelparous species like Pacific salmon (i.e., species that only spawn once in their lifetime), all straying events represent natal dispersal, whereas iteroparous species (i.e., species that can spawn more than once in their lifetime) can stray between spawning events and are thus capable of breeding dispersal (sensu Greenwood 1980). The individuals that do not stray (i.e., the large majority of individuals in most species) are said to home to their natal habitats (Dittman & Quinn 1996). The terms philopatry and homing are often used interchangeably in the salmonid literature (e.g., Dittman & Quinn 1996; Hendry et al. 2004). I have noticed a tendency to use the term homing when discussing the proximate physiological and behavioral mechanisms that allow individuals to find their natal habitat (e.g., Dittman & Quinn 1996; Nordeng 2009), and the term philopatry when discussing the ultimate evolutionary forces favoring the return of individuals to their natal sites (e.g., Hendry et al. 2004; Walter et al. 2009). I will use both terms interchangeably in the thesis, although my usage probably reflects the aforementioned bias.  5>B $C?'<96A?D;?6<?A'9@':=AE?8AFG' “It is difficult to imagine any ecological or evolutionary problem that would not be affected by dispersal.” - Dieckmann et al. 1999  Few biologists would dispute the importance of dispersal in ecology and evolution. The consequences of dispersal, however, are numerous, can be both positive and negative for the fitness and persistence of populations, and can interact in complex ways (Garant et al. 2007). Thankfully, there is a large body of theoretical and, but to a lesser extent, empirical work that provides a framework for understanding those consequences. I will first briefly review theories relating to the evolution of dispersal because the ultimate causes of dispersal can offer clues regarding its consequences. I will next discuss the ecological consequences of dispersal, mainly from the perspective of  4  metapopulations (i.e., spatially discontinuous populations; Hanski 1999). Finally, I will review the evolutionary consequences of dispersal, which are mainly engendered when dispersal leads to gene flow, and discuss how the evolutionary and ecological consequences of dispersal can interact.  !"#"! $%&'()*+,-*&'&./)(*+/0-12'3-(4&4'/5'6+47&14-)' There is a large and growing body of theory on the evolution of dispersal, and several recent reviews are available (Levin et al. 2003; Clobert et al. 2004; Bowler & Benton 2005; Ronce 2007). It is almost a truism to say that dispersal will evolve if the fitness benefits of dispersal exceed its fitness costs. The costs of dispersal include the risk of not finding a suitable habitat (Holt 1985), and movement itself is often costly energetically or in terms of increased mortality (Johnson & Gaines 1990). Furthermore, if local adaptation is important, dispersal leading to gene flow can introduce locally deleterious alleles (Lenormand 2002). But what are the benefits of dispersal? There are three main hypotheses that have been proposed to explain the evolution of dispersal (Bowler & Benton 2005; Ronce 2007): (1) competition, including competition among kin, (2) inbreeding avoidance, and (3) habitat variability. In situations of high population densities, dispersal can allow escape from overcrowding, but this is advantageous only if there is not a perfect match between population density and resource availability (Ronce 2007). If kin competition is introduced in the model, however, this last requirement is relaxed. Hamilton & May (1977) showed that “altruistic” dispersal, where the disperser does not gain a fitness advantage from dispersing, can evolve if inclusive fitness (Hamilton 1964) is considered. This occurs because densities are thus lowered, and competition is reduced among kin that do not disperse (Ronce 2007). Breeding between close relatives can have deleterious fitness consequences (Charlesworth & Charlesworth 1987; Keller & Waller 2002), and dispersal will be beneficial if it reduces the likelihood of such matings (Bengtsson 1978; Perrin & Mazalov 1999; Perrin & Goudet 2001). One prediction of models of dispersal evolution in response to inbreeding is that dispersal should be sex-biased (Perrin & Mazalov 1999).  5  Indeed, dispersal of only one sex is enough to reduce inbreeding, but because costs of dispersal often differ among sexes, most of the dispersing individuals should be of the sex for which dispersal is least costly (Greenwood 1980; but see Perrin & Mazalov 1999). It should be noted that because inbreeding and kin competition are both influenced by relatedness among individuals, it is difficult to consider them independently. Models that have examined both mechanisms jointly have revealed complex interactions that make empirical tests of the relative importance of those two mechanisms difficult (Perrin & Goudet 2001). Habitats vary in quality in both space and time, and if the spatial distribution of high-quality habitats varies asynchronously through time, dispersing genotypes will be favored (McPeek & Holt 1992; Travis & Dytham 1999; Blanquart & Gandon 2011). Under this scenario, dispersal can be understood as a bet-hedging strategy, where the ‘placement’ of offspring in a variety of habitats reduces the variance in fitness over time, thereby increasing its geometric mean fitness (Venable & Brown 1988; Ronce 2007). Note that this only applies to temporal variations in habitats. If instead habitats are temporally stable but vary in space, selection should favor philopatry because offspring should stay in the habitat where their parents already had success (Balkau & Feldman 1973; Holt 1985). Models of the evolution of dispersal under temporally varying habitats typically assume that this variation is extrinsic (i.e., variation in carrying capacity or selection coefficient), but dispersal can also evolve in response to chaotic population dynamics under density dependence (Holt & McPeek 1996). Even if habitat quality does not vary through time, dispersal can be advantageous in the case where new habitats become available for colonization (Van Valen 1971). This scenario is especially relevant when considering species that recolonized previously glaciated areas. In most models discussed so far, dispersal is assumed to be a blind process with a certain random component of the population dispersing (Ronce 2007). It is clear from the previous discussion, however, that dispersal is only favorable under certain conditions. It would thus be advantageous for organisms to evolve mechanisms that would allow them to make dispersal decisions on the basis of current conditions (Ronce 2007; Clobert et al. 2009). Such informed dispersal is referred to as condition-dependent dispersal (Ims and Hjermann 2001). The ‘conditions’ used as cues by individuals making dispersal decision  6  can be intrinsic (their own phenotype) or extrinsic (their environment; Ims & Hjermann 2001; Clobert et al. 2009). Examples of extrinsic cues include dispersal in response to high population density (Matthysen 2005), food availability (Kennedy & Ward 2003), predation pressure (Sloggett & Weisser 2002), sex-ratio biases in the local population (Colwell & Naeem 1999), and relatedness of conspecifics (Wolff 1992; Sinervo & Clobert 2003). Examples of intrinsic clues include body size (Hanski et al. 1991) or body condition (Nunes et al. 1997), sex (Greenwood 1980; Hutchings & Gerber 2002), and age (Quinn 1993). Note that these intrinsic cues can be purely phenotypic (e.g., age), or can be related to an individual’s genotype (Roff & Fairbairn 2001; Bolnick et al. 2009). When dispersal and gene flow is not random with respect to genotypes, its evolutionary consequences are modified because certain alleles are disproportionately represented in dispersers (Edelaar et al. 2008; Edelaar & Bolnick 2012).  !"#"# 83/)/9+3-)'3/04&:(&03&4'/5'6+47&14-)' At the most fundamental level, dispersal (i.e., immigration and emigration) controls population size along with births and deaths and is thus a major demographic parameter. As one of the ultimate controls of population size, it is therefore not surprising that dispersal has featured prominently in some of ecology’s most foundational theories (e.g., island biogeography: MacArthur & Wilson 1967; metapopulation theory: Levins 1969; neutral theory: Hubbel 2001; metacommunity theory: Leibold et al. 2004). For example, dispersal has been suggested as a population regulation mechanism because it is often density-dependent (Hanski 2001). In classical metapopulations (sensu Levins 1969), where local populations have a substantial risk of extinction, dispersal allows the recolonization of extinct local populations, and is thus essential to prevent global extinctions (Levins 1969; Hanski & Gilpin 1997). This population turnover characteristic of classical metapopulations, however, is not necessary for dispersal to have important consequences. Dispersal (more specifically immigration) can also reduce the risk of local extinctions by increasing local population size, a phenomenon referred to as the ‘rescue effect’ (Brown & Kodric-Brown 1977; Hanski 1999). An extreme case is found in source-sink dynamics, where species are allowed to occur in patches where survival  7  would otherwise be impossible (sinks) by virtue of immigration from higher quality habitats (sources; Pulliam 1988; Dias 1996). Sink habitats can be absolute sinks (Kawecki & Holt 2002) (or ‘black hole sinks’; Gomulkiewicz et al. 1999) where local populations would deterministically go extinct in the absence of dispersal. Alternatively, sinks can be relative (Kawecki & Holt 2002) (or pseudosinks; Watkinson & Sutherland 1995) where the population is self-sustaining but where immigration exceeds emigration. In both cases, dispersal has dramatic consequences for local population dynamics. Dispersal, however, does not only have positive consequences for metapopulation persistence. For example, the flip side of the rescue effect is that emigration from small populations can lead to increased extinction risk (Hanski 2001). This can be especially important in already small populations where dominant winds or currents cause asymmetries in dispersal such that emigration is higher than immigration (for examples see Kawecki & Holt 2002). If dispersal is costly, the increased mortality suffered by dispersers could also lead to metapopulation extinction (Hanski & Zhang 1993). Furthermore, high emigration to small patches can cause populations to exceed the local carrying capacity, and lead to overcrowding and reduce the average fitness of the population (Holt 1985). Finally, high dispersal among local populations synchronizes population dynamics, which can have both beneficial and deleterious effects for persistence (Abbott 2011). The beneficial effects of synchrony arise because dispersal will dampen stochastic fluctuations in local populations (Briggs & Hoopes 2004). But this also leads to synchrony of population dynamics in the entire metapopulation, which can increase global extinction risk by limiting the potential for rescue effects (e.g., Molofsky & Ferdy 2005).  !"#"; 8./)(*+/0-12'3/04&:(&03&4'/5'6+47&14-)'-06'9&0&'5)/<' Perhaps the most widely acknowledged evolutionary consequence of gene flow is its negative effect on adaptation (Endler 1977; Slatkin 1987; Lenormand 2002). In short, heterogeneous environments should favor the evolution of local adaptation (and ultimately speciation), but gene flow opposes divergent natural selection by homogenizing allele frequencies across populations. While this is well established in  8  theory, the importance of gene flow in preventing adaptive divergence and speciation in the wild has been debated. Mayr (1942; 1963) famously argued that gene flow was the major force maintaining the homogeneity of species, and that speciation would therefore occur mainly in allopatry (i.e., in the absence of gene flow). Many authors challenged this view and argued that dispersal often did not result in gene flow (i.e., gross dispersal is higher than realized dispersal), or that natural selection was often strong enough to oppose the homogenizing effects of gene flow (e.g. Ehrlich & Raven 1969; Endler 1973). The consequences of gene flow depends on its relative importance compared to natural selection (Slatkin 1985; 1987; Barton 2001), and numerous theoretical (e.g., Slatkin 1973; García-Ramos & Kirkpatrick 1997; Hendry et al. 2001) and empirical studies (Stearns & Sage 1980; Bell & Richkind 1981; Riechert 1993; King & Lawson 1995; Dias & Blondel 1996; Storfer & Sih 1998; Nosil & Crespi 2004; Moore et al. 2007) have now shown that it can be an important force in the wild. While the negative effects of gene flow on adaptation are typically emphasized, gene flow can also constitute a creative evolutionary force (Slatkin 1987). Hybridization, which could be defined as gene flow between species, has also been widely recognized as a creative evolutionary force (Arnold 1997), but I will limit this discussion to intraspecific gene flow. Sewall Wright introduced the ‘shifting balance theory’ to describe how seemingly non-adaptive trait changes due to random drift can help populations stranded on local fitness peaks overcome fitness valleys that block access to the global maximum fitness peak (Wright 1932). Population subdivision is essential to this process, by maintaining small enough local population sizes for drift to operate. Gene flow then allows the beneficial mutations that fixed in local populations to spread to other populations in the metapopulation. The importance of the shifting balance in the process of adaptation in the wild is disputed (Coyne et al. 1997; Whitlock & Phillips 2000). Nevertheless, there is some evidence that gene flow does play a role in the spread of globally advantageous mutations among subdivided populations (Morjan & Rieseberg 2004). Indeed, theory shows that if natural selection favors an allele across subpopulations, this allele will spread at a rate that is determined by the amount of dispersal and the strength of selection on that allele (Fisher 1937; Slatkin 1976). Such an allele can spread relatively quickly even if gene flow is low, when selection is  9  sufficiently strong, and rare long-distance dispersers can considerably speed up the process (Slatkin 1976). Experimental evidence for the role of dispersal in spreading advantageous alleles through subdivided populations comes from a few microbial systems (e.g., Perron et al. 2007; Bell & Gonzalez 2011). It is often difficult to decouple the ecological (i.e. demographic) consequences of dispersal from its evolutionary consequences (Barton 2001). For instance, the widely recognized negative effects of gene flow on adaptation can under some circumstances turn into positive effects if one takes demography into account. Indeed, Holt and Gomulkiewicz (1997) showed that immigration into a sink, by increasing population size, can help the process of adaptation even if the influx of migrants from the source continually introduces locally maladapted alleles. This is an example of how the demographic effects of dispersal can influence the adaptive process, but the inverse is also true: the fitness consequences of gene flow can have demographic consequences. When gene flow leads to a distribution of allele frequencies that is different from that favored by natural selection, the populations are said to suffer from a migration load (Lenormand 2002). This reduction in average fitness can in turn have negative demographic consequences, and can under extreme circumstances lead to population extinction, a process referred to as migrational meltdown (Ronce & Kirkpatrick 2001). Such negative effects of gene flow on population persistence have even been implicated as a determinant of species ranges (Kirkpatrick & Barton 1997; Case & Taper 2000). Little evidence for this process, however, currently exists (Bridle & Vines 2007; Moore & Hendry 2009). Another negative effect of gene flow is that it can lead to outbreeding depression. This occurs when recombination disrupts advantageous epistatic interactions among loci (so-called co-adapted gene complexes), thus reducing population fitness (Edmands 2007). Gene flow, however, can also have positive demographic consequences. For instance, matings among related individuals can also lead to inbreeding depression in small populations (Charlesworth & Charlesworth 1987; Keller & Waller 2002). Gene flow, by supplying novel alleles, can replenish genetic diversity and thus alleviate some of the negative fitness consequences of inbreeding depression, a process sometimes referred to as genetic rescue (Tallmon et al. 2004). This occurs principally as a result of heterosis (or hybrid advantage) whereas deleterious recessive alleles are masked by  10  immigrant alleles (Tallmon et al. 2004). Note that both heterosis and outbreeding depression are not mutually exclusive processes and can both occur following contact among previously isolated populations. Indeed, because the effects of outbreeding depression are not expressed before the F2 generation, heterosis in F1 crosses can be followed by outbreeding depression in subsequent crosses (e.g., Edmands 1999). In summary, dispersal and gene flow have a wide variety of potential ecological and evolutionary consequences for populations. Even more complex is the fact that ecological and evolutionary consequences of dispersal interact sometimes in complex ways. Numerous theoretical studies provide a solid framework to understand those multifarious consequences, but their conclusions depend on assumptions whose validity in natural populations is often not known. Furthermore, qualitative conclusions often apply to specific portions of the parameter space, and typical parameter values are not always available. Clearly, therefore, the consequences of dispersal are highly contextdependent. Accumulating more empirical data from natural populations should therefore be considered a priority (Barton 2001; Garant et al. 2007).  5>H #7;:I'AIA7?JK'F6F:89J9;A'!8<7=<'&CF8'=6'7C?'&F6F:=F6'!8<7=<' “No higher praise can be given to a Salmonid than to say, it is a charr” - Jordan and Evermann (1896) “If any species of fish could be characterized as defying description, it is the Arctic char” -Behnke and Tomelleri (2002) p.303  While the overarching theme of this dissertation is the study of dispersal and its consequences, all of my work has focused on populations of Arctic Char (Salvelinus alpinus) from the Canadian Arctic (Figure 1.1). I will here review different aspects of the species’ biology that are relevant to the work conducted in this dissertation. Because most of my work involved anadromous populations (i.e., the ones that have potential for dispersal and gene flow), I will here focus mainly (but not exclusively) on the biology of anadromous Arctic Char.  11  !";"! $-=/0/,2' The taxonomic status of Arctic Char has been a contentious topic (McPhail 1961; Johnson 1980; Behnke 1984; Savvaitova 1995; Reist et al. 1997; Brunner et al. 2001; Jonsson & Jonsson 2001; Taylor et al. 2008). First, the species’ high level of polymorphism provides fertile ground for debates as to which forms should be recognized as formal species (Behnke 1984; Savvaitova 1995; Jonsson & Jonsson 2001). Savvaitova (1995) reported that there were up to 29 recognized species of Char in Europe, 15 in North America, and 12 in the Far East and Siberia. The prevailing view, however, is that the various forms of Arctic Char all belong to the same species complex (Jonsson & Jonsson 2001; Benhke & Tomelleri 2002). Therefore, whenever I use the words ‘Arctic Char’ in this thesis, I use the term to mean the ‘Arctic Char species complex’ as recognized by various authors (Reist 1997; Jonsson & Jonsson 2001; Benhke 2002; Alekseyev et al. 2009). Despite the wide recognition that these various forms belong to the same taxonomic unit, some authors have attempted to organize the intra-specific diversity into subspecies. In North America, three subspecies are generally recognized (Behnke 1984): S. a. oquassa in Maine, New Brunswick and Québec; S. a. erythrinus throughout the Canadian Arctic and in isolated lacustrine populations of Northern Alaska; and S. a. taranetzi, or Taranets Char, in the Bristol Bay/Gulf of Alaska area. Those subspecies more or less correspond to different mtDNA lineages, and may therefore reflect the recent evolutionary history of the species, especially in relation to glacial cycles (Wilson et al. 1996; Brunner et al. 2001; Taylor et al. 2008; Chapter 2). Another area of contention has been whether or not Arctic Char and Dolly Varden (Salvelinus malma) represent separate species. Similarities in morphology and mtDNA between Arctic Char and Dolly Varden have led to the repeated questioning of their taxonomic distinctiveness (DeLacy & Morton 1942; McPhail 1961; Behnke 1984; Reist et al. 1997; Brunner et al. 2001; Taylor et al. 2008). It is now generally accepted that Dolly Varden and Arctic Char constitute separate species (McPhail 1961; Behnke 1984; Reist et al. 1997), and good evidence exists that, when found in sympatry, the two species do not interbreed (Taylor et al. 2008). Note, however, that the two species share  12  mitochondrial DNA haplotypes in some parts of Alaska, and that hybridization almost certainly occurred between the two species at some point in their recent evolutionary history (Brunner et al. 2001; Taylor et al. 2008). !";"# >+4*1+?(*+/0' The Arctic Char has a coastal distribution across the Holarctic (Figure 1.2). It is the freshwater fish with the most northerly distribution in the world, sharing the fresh waters north of 75° with no other fish species (Johnson 1980). In Europe, the species is found as far south as the Alps, and as far north as Iceland and Svalbard, and is distributed around most of coastal Greenland. It is also found in the British Isles and throughout the Fennoscandian outer coast. In Asia, the species is mainly restricted to Siberian Russia where is found throughout the coast of the Arctic Ocean. It is found as far north as Novaya Zemlya and as far south as the Lake Baikal basin (Alekseyev et al. 2009). In North America, the species is distributed from Maine to Ellesmere Island on the Atlantic coast and extends to the South of Cook Inlet, Alaska, on the Pacific coast (Scott & Crossman 1973). The presence of Arctic Char in the Brooks Range of north slope Alaska is mentioned in many publications (McPhail 1961; Scott & Crossman 1973; Johnson 1980; Reist et al. 1997), but inquiries to several Arctic Char biologists could not locate samples or confirm the presence of the species there. The species is also distributed around the north of Hudson Bay, but does not extend south beyond James Bay.  !";"; @+/)/92' The life cycle of anadromous Char shares features with other anadromous salmonids, but several adaptations to cold, resource-poor waters distinguish it from other species in that group. Arctic Char is a mostly lacustrine species, and spawning typically occurs in the fall in the gravel bottom of lakes (Johnson 1980). Following an incubation period that typically ends in the following spring, hatched juveniles will spend anywhere from three to nine years rearing in fresh water before undergoing smoltification, a process that involves behavioural, morphological, and physiological changes necessary for a marine existence. Downstream migration of smolts and adults occur at ice break-up in the spring or early summer. Char will then spend the summer feeding at sea for a period of 30 to 60 13  days. At the end of the short summer, all individuals migrate back to fresh water to overwinter, regardless of their reproductive status. In other salmonid species, only individuals that are ready to spawn migrate upstream, while non-mature and resting fish stay in the ocean to feed. However, lower salinity tolerance and inability to survive sub-zero temperatures in salt waters constrain Arctic Char to move back to fresh water every winter (Johnson 1980). This necessary yearly energy expenditure, combined with low resource availability of Arctic waters and short feeding season, make growth slow: anadromous Char can take up to 25 years to reach sexual maturity, although ages between 10-15 years are more typical (Johnson 1980). Upstream migrations back to fresh water invariably occur in the fall, sometimes starting as early as July. Spawning occurs between September and November (Johnson 1980) and most individuals home to their natal lake, although dispersal/straying appears common in at least some populations (Gyselman 1994). Char are iteroparous (i.e., they reproduce more than once, as opposed to Pacific salmon) and seem to have the highest rate of repeat spawning of all anadromous salmonids, with up to 50% of adults spawning at least twice in some populations (Fleming 1998). This is in line with expectations of life-history theory, which predicts iteroparity to be favored as a bet-hedging strategy in the unpredictable northern environments (Charnov & Shaffer 1973; Leggett & Carscadden 1978). The spawning events, however, are separated by two to four years, the time necessary to re-build energy reserves for gonadal development (Dutil 1986). Such skipped spawning is not unique to Arctic Char and occurs in other iteroparous salmonids (Rideout & Tomkiewiscz 2011), although other anadromous salmonids typically do not migrate back to fresh water in the years where they do not spawn (but see Larsen et al. 2008 for an example in Brown Trout, Salmo trutta). Arctic Char are also characterized by an unusually long lifespan for a salmonid, and in some populations, they can live over 40 years old (Johnson 1980). The polymorphic nature of Arctic Char has made it a model system of choice in evolutionary ecology, with a majority of the work focusing on repeated adaptive radiations of multiple sympatric forms in post-glacial lakes throughout the range of the species (Hindar et al. 1986; Reist et al. 1995; Schluter 1996; Skúlason et al. 1999; Jonsson & Jonsson 2001; Knudsen et al. 2006). In those lakes, up to four different morphotypes have evolved, most likely driven by ecological opportunities afforded by  14  the species-poor post-glacial lakes (Skúlason et al. 1999). Another important axis of variation is found in the migratory tactics displayed by Char (Johnson 1980), some populations being resident/landlocked and others being anadromous (the term ‘resident’ is used to refer to populations that remain in fresh water even if access to sea exists, while landlocked populations are restricted to fresh water by geography). As a general rule, Arctic Char tend to be resident (non-migratory) at the northern and southern periphery of its range, whereas it tends to be anadromous at intermediate latitudes, if access to the ocean exists (Johnson 1980; Hammar et al. 1989; Doucett et al. 1999; Svenning & Gullestad 2002). Latitudinal variation in diadromy is a well-known phenomenon and is believed to arise because of the relative availability of food resources in ocean and freshwater habitats (McDowall 1987; Gross et al. 1988). The anadromous populations of Char have received comparatively less attention from evolutionary ecologists than landlocked populations, or than anadromous populations of other salmonids. Because only anadromous populations have the potential to disperse between different ‘patches’ of freshwater habitats, they constitute the main focus of my work.  !";"A B,7/1*-03&'-4'-'5+4%&1+&4'1&4/(13&' The Arctic Char is of tremendous cultural, subsistence and economic importance for peoples of the North. The Inuit people of Nunavut have been fishing Arctic Char for thousands of years, and it has been argued that the reliability of this resource made it a more important food source for pre-contact Inuit than even the more iconic caribou or seal (Balikci 1980). To this day, Char remains an important food source for the communities of Nunavut, and it is still the most harvested species of wildlife in that territory (Priest & Usher 2004). The species is also the target of a commercial fishery in several communities, and four federally registered fish plants process the harvest directly in the communities (Read 2000; Roux et al. 2011). The harvest and processing of fish for southern markets thus provides one of the rare sources of non-government income to the residents of Nunavut. The Government of Nunavut estimates that the commercial fishery for Char has a value of $1.4 million, while the subsistence harvest has a value of $4.4 million (Nunavut Fisheries Strategy 2005).  15  !";"C D'?1+&5'0/*&'/0'47&))+09' The fishes of the genus Salvelinus have been, and continue to be, referred to both as Char and Charr (e.g. Benhke 1984; Benhke & Tomelleri 2002). The spelling has been the object of a lively debate. In 1955, Morton was prompted to write an opinion piece in the journal Science deploring the fact that the Committee on Common Names of the American Fisheries Society favored the spelling Char (Morton 1955). This was unfortunate, according to Morton, because “the bulk of evidence seemed to indicate that charr was the better spelling” (Morton 1955). I will not discuss the nature of this evidence here. In this thesis, I use the spelling Char throughout. This does not necessarily represent personal preference, but is rather a reflection of my academic genealogical tree (McPhail 1960; Taylor et al. 2008; but see Hendry & Stearns 2004). Note that another member of my supervisory committee seem to prefer the spelling Charr (Tallman et al. 1996; Loewen et al. 2009), while yet another’s preference appears to have changed over the years (Reist 1989; Reist et al. 1997).  5>L !'8?M=?N'9@':=AE?8AFG'=6'!8<7=<'<CF8' The migratory behavior of anadromous Arctic Char is poorly understood in comparison to that of other anadromous salmonids. This undoubtedly reflects the logistical constraints associated with working in the Arctic regions inhabited by the species. Despite these challenges, there are a number of studies of Arctic char migratory behavior and dispersal that provide a basis of understanding on which to build future studies. What follows is a fairly exhaustive review of available studies of migratory behavior in anadromous Arctic Char. I have also added a few studies from anadromous populations of the closely related Dolly Varden char because the two species are generally assumed to have a similar migratory behavior. A parameter of interest in studies of dispersal in anadromous salmonids is their straying or dispersal rates. Estimate of straying rates would ideally measure the likelihood that any given individual strays away from its natal river. In reality, however,  16  most estimates are generated by measuring the proportion of fish migrating up-river that are immigrants from other rivers. While the proportion of immigrants can provide an indication of the importance of dispersal, it is not necessarily equivalent to the straying rate. For example, if a small population is in close proximity to a much larger population, we may expect the number of immigrants to be large even if the straying rate is low. This distinction should therefore be kept in mind when interpreting the results. In addition, studies of straying rates in other salmonids assume that all migrating individuals are migrating for the purpose of reproducing, an assumption that is valid in most species. As discussed earlier, this is not the case for Arctic Char, and the studies listed here do not have the capacity to make a distinction between individuals migrating for the purpose of overwintering and those that migrate for the purpose of breeding. The term dispersal rate is thus used (instead of straying rate) to mean dispersal of any individual from another river. Studies from which a dispersal rate can be extracted are listed in table 1.2. Note that two older studies (Grainger 1953; Moore 1975) were excluded from the table because movement was inferred from either direct observation (Moore 1975) or captures from gill nets without tagging (Grainger 1953), thus not allowing estimates of dispersal or migration distance. There are a few salient points that can be concluded from these various studies. First, estimates of dispersal rates vary tremendously among studies. This could reflect differences in dispersal rates among regions, which could in turn be driven by environmental differences, or differences in topography (i.e., average distance between rivers, presence of suitable spawning or overwintering habitats, etc.). Differences among studies could also reflect among-year differences, perhaps driven by varying meteorological conditions. Current studies, however, cannot help discriminate among these hypotheses because these variables are confounded with study design: each study was conducted independently in different regions, different years, and using different methodologies. Even the study by Dempson & Kristofferson (1987), which is probably the most informative because it studied two regions in a similar time period, is difficult to interpret. The proportion of tagged fish that were recaptured in a different river (dispersers) varied between 0% and 17% per year among twelve different rivers in Labrador (eastern Canadian Arctic). By contrast, in the Cambridge Bay area of Canada’s  17  central Arctic, those proportions were much higher, varying between 13% and 51% per year among four rivers (Dempson & Kristofferson, 1987). But these differences may stem from the design of the study in Labrador, where the fishery-dependent recovery of tags at sea did not allow an evaluation of dispersal between rivers, but only between subareas, i.e., bays where multiple rivers flow. Second, despite variation among studies, it appears that Arctic Char do have an increased propensity to disperse compared to other salmonids. Hendry et al. (2004) recently reviewed available estimates of straying rates in anadromous salmonids (excluding Arctic Char), and found that they ranged from 0.0% to 41.6% with a median of 4.4%. In contrast, the estimates in table 1.2 range from 3% to 51% with a median of 12% (if I take the averages of the two Dempson & Kristofferson 1987 estimates). Part of this discrepancy is likely due to the fact that a large portion of Arctic Char dispersal is restricted to non-reproducing individuals (see below). But even if that is the case, the increased dispersal propensity of overwintering individuals requires an explanation. Hendry et al. (2004) concluded from their comparative analysis of straying rates in salmonids that habitat unpredictability was a likely factor promoting the evolution of high straying rates in salmonids. The fact that Arctic Char has a high dispersal rate, and inhabits highly variable habitats (Power & Power 1995; Power 2002), lends support to this hypothesis. Third, it seems likely that the settlement decisions of overwintering Arctic Char are different than that of spawning char. The studies (Table 1.3) have all found evidence for overwintering individuals utilizing non-natal habitats. Most of the evidence, however, is indirect and comes from weir studies that inferred overwintering in alternative habitats from return of individuals to their location of capture after 1 or 2 years of absence. This kind of evidence is problematic for several reasons. First, the weir design may not have captured all individuals from the population, and there is typically a bias in that most studies will miss the first part of the downstream migration because of ice condition. If those individuals are more likely to migrate (or vice-versa) the estimates of dispersal will also be biased. Second, the reproductive status of individuals in those studies is not known, and it is therefore impossible to know if the individuals that have left the system for one or more years are indeed overwintering, or if the homing individuals are indeed in  18  reproductive condition. Third, it is assumed that the river where fish are tagged on the downstream migration is their natal river. If dispersal among river is common, this will not always be the case and will thus lead to erroneous dispersal rate estimates. The data collected in Chapter 3 addresses some of these limitations. One of the studies listed found direct evidence of over-wintering dispersal using telemetry (Beddow et al. 1998). The river system under study, however, is unusual and the results may not apply to other regions. In Reid Brook, northern Labrador, 61% of the tagged fish moved through to an adjacent river system to over-winter in the same fall after spawning in Reid Brook (Beddow et al. 1998). The overwintering river in this case empties in the same bay as the spawning river, and examples of such complex migratory behavior may be limited to areas where multiple potential overwintering sites are found in close geographical proximity. Although marine migrations do not represent dispersal per se, knowledge of the geographic extent of marine movement can help understand the potential scope of dispersal distances. The results of most tagging studies (Table 1.4) suggest that a large majority of individuals do not move far from their rivers of origin. Telemetry studies allowing fine-scale tracking of individuals further show that any movement by Arctic Char tend occur within 1 km of the coast (Bégout Anras et al. 1999; Spares et al. 2012). This suggests that most individuals would be unlikely to cross stretches of open water, for example between islands. Despite the fact that most individuals showed little tendency for long-distance movement, most studies detected one or a few individuals at or near the detection limit of the study design. The most notable example is probably the observation by DeCicco (1992) of two Dolly Varden individuals tagged in Alaska and captured by Russian biologists working in Siberia. One individual travelled more than 1,500 km across the Bering Strait over approximately 60 days, demonstrating the great capacity for movement in this species that is closely related to Arctic Char. Anadromous Dolly Varden are assumed to be similar to Arctic Char in their migratory behaviour, and this study thus provides a forceful illustration that the geographical scope of a study may greatly limit our capacity to evaluate the dispersal capabilities of Arctic Char (Porter & Dooley 1993; Nathan et al. 2003). Extending the geographical scale of studies of Arctic Char dispersal should thus be a priority for future studies. This is especially true because  19  despite being rare, long-distance migrants probably have disproportional ecological and evolutionary importance (Nichols & Hewitt 1994; Trakhtenbrot et al. 2005).  5>O %?A?F8<C'D;?A7=96AP'FEE89F<C'F6:'Q;A7=@=<F7=96' The current distribution of Arctic Char, which encompasses the northernmost reaches of land and the highest naturally colonized alpine lakes of Europe, suggests that it is a fish with tremendous dispersal and colonizing abilities. This would suggest that char should be relatively homogeneous across its range, different populations being linked by gene flow. However, Arctic Char is also well known as an extremely polymorphic species (Johnson 1980; Behnke 1984; Jonsson & Jonsson 2001). This leads to an apparent paradox: how can Arctic Char both be an excellent naturally colonizing salmonid, and at the same time be one of the most polymorphic species of the group? Part of the answer to this question involves a mix of history and physiology: Char, being well-adapted to cold water, were able to colonize post-glacial lakes quicker and thus take advantage of many empty niches before other species could colonize (Schluter 1996; Skúlason et al. 1999; Power 2002). While this constitutes a satisfactory scenario for landlocked populations, it fails to account for local adaptation and phenotypic variation observed in anadromous populations (e.g., Rikardsen et al. 2004; Power et al. 2005). Another commonly invoked explanation is that Arctic Char is more phenotypically plastic than other species of salmonid (Jonhson 1980). However, little evidence exists for this assertion. Finally, dispersal could be very high without hampering local adaptation if (1) natural selection is strong enough to counteract gene flow (Slatkin 1985; Hendry et al. 2001) or (2) if dispersal does not result in high levels of gene flow, i.e., if gross dispersal is higher than effective dispersal (sensu Greenwood 1980). By providing estimates of dispersal at different temporal and spatial scales, and by evaluating its consequences for local adaptation, my thesis will attempt to shed light on this paradox. My work also has a number of applications in a context where the Canadian Arctic will face many changes in the years to come (IPCC 2007). Data collected during this thesis will help devise better management plans for existing and emerging fisheries (Roux et al. 2011) and will help  20  predict the response of Arctic Char to climate change. Indeed, species will respond in one of three ways to future changes: (1) move their ranges pole-ward, (2) adapt to new environmental conditions, or (3) go extinct (Parmesan 2006). Dispersal will necessarily play an important role in the two most favorable of the predicted responses.  !"C"! D771/-3%&4'*/'*%&'4*(62'/5'6+47&14-)' As discussed earlier, much progress towards a better understanding of the consequences of dispersal stems from theoretical studies (reviewed in Ronce 2007). While such studies have provided a solid framework to understand patterns in the natural world, many authors have deplored the expanding disconnect between theoretical studies and empirical studies (Barton 2001; Ronce 2007; Clobert et al. 2009). Ideally, empirical tests of the theory would take the form of carefully controlled experiments that would allow mechanisms to be inferred with confidence. While empirical systems where this is possible exist (e.g., Krebs et al. 1969; Kuussaari et al. 1996; Bell & Gonzalez 2011), they tend to be from taxonomic groups (microorganisms, insects, rodents) of little relevance for conservation or management. Furthermore, I would argue that a great deal of ecological realism is lost in small-scale experiments (others have also expressed that view in relation to other areas of ecological research: e.g., Peters 1991; Schluter & Ricklefs 1993). This is especially true for studies of dispersal, which is a highly stochastic and context-dependent process (Nathan 2001). Observational studies conducted on wild populations of organisms of conservation or management concern can thus help in verifying the validity of theoretical and experimental studies and help ensure that their results are relevant for practical problems. Purely observational studies have many pitfalls and disadvantages, not the least of which being that they do not readily allow us to infer causation. Nevertheless, used in combination with a well-established theoretical framework previously supported by experimental data, I believe that observational studies can lead to considerable insight. This is the approach I have chosen for my dissertation work. Measuring dispersal and gene flow in wild populations was thus the major challenge of my research, and I briefly review some available methodologies below.  21  !"C"# E&-4(1+09'6+47&14-)'-06'9&0&'5)/<' Measuring dispersal is a notoriously difficult task (Nathan 2001). The methodologies available are numerous and increasingly sophisticated, but most have drawbacks that need to be recognized if the data are to be interpreted accurately. The physical tagging of individuals is the oldest method used to study dispersal in natural populations. Due mainly to logistical challenges, however, this family of methods was not used for my dissertation work and will thus only be briefly reviewed here. I have already discussed some of the disadvantages of such approaches in my review of the dispersal literature on Arctic Char. Most importantly, given my interests in the evolutionary consequences of dispersal, physical tagging makes the distinction between gross and realized dispersal difficult, and cannot allow for estimates of gene flow. Another potential problem with physical tagging is that the tags can alter the behavior or survival of the tagged individuals. For an example of how important mortality and tag loss can be in anadromous Arctic Char, see Berg & Berg (1990). All my dissertation work used molecular approaches to the study of dispersal. These approaches are possible because when populations are subdivided, differences in allele frequencies tend to accumulate between them. These differences can be the result of natural selection deterministically increasing or decreasing the frequency of certain alleles, or can be due to random genetic drift. Many molecular methods used to measure dispersal rely on putatively neutral markers, and therefore assume models of evolution that do not consider natural selection to make inferences about gene flow (i.e., realized dispersal). Some methods, especially assignment methods, do not make the assumption of neutrality and measure gross dispersal, not gene flow. The most widely used statistic to measure genetic divergence between populations is FST, which is a standardized measure of variance in allele frequencies among subpopulations (Hedrick 2004; Whitlock 2011). In theory, it is possible to derive estimates of gene flow from FST values because mathematical models have demonstrated a relationship between FST and the parameter Nem, the effective number of migrants (Slatkin & Barton 1989). Most famously, Wright (1931) used the infinite island model to demonstrate that, under a long list of stringent  22  assumptions, FST is approximately equal to 1/(4Nem+1). Because these assumptions are rarely met in natural populations, it is generally unadvisable to use this simplistic relationship to estimate gene flow (Whitlock & McCauley 1999). Also note that many authors have used the relationship to estimate dispersal, which is another major problem since the method can only be used to estimate long-term gene flow, or effective dispersal (sensu Greenwood 1980). The other major category of models of evolution used to interpret patterns of variation in allele frequencies are isolation by distance (IBD) models (e.g., Wright 1943; Malécot 1950). In short, these models, unlike the infinite island model, assume that nearby subpopulations exchange more migrants than distant subpopulations. Under such models, FST is expected to increase as a function of distance, and given a few corrections (see Rousset 1997), a linear relationship is expected between genetic differentiation and distance. The strength of this relationship is mainly determined by !2, i.e. the variance in dispersal distance (also known as the neighborhood size), which can be estimated from the slope (Rousset 1997). More recent methods rely on coalescent theory to make inferences about the most likely past drivers of present-day patterns of genetic variation. Some methods use maximum likelihood to estimate parameter values. To do so still requires these methods to make assumptions about specific models of evolution. The models of evolution used are more complex and realistic than the island model, but attempting to use these methods on situations that do not fit the assumptions can still be problematic (e.g., Beerli 2004). The most widely used such methods are MIGRATE (Beerli & Felsenstein 1999; Beerli & Felsenstein 2001) and the IM family of programs (Hey & Nielsen 2004; Hey 2010). MIGRATE fits an equilibrium model of population subdivision with gene flow and allows for asymmetric gene flow between populations. It thus simultaneously estimates effective population size (Ne) for each sub-populations, and an estimate of gene flow (m) in each direction. Being based on the coalescent, these estimates are scaled by the mutation rate µ, and a good estimate of µ is thus required to obtain parameter estimates. The IM family of programs does not assume equilibrium, but instead assumes a model where an ancestral population splits into two descendant populations some number of generations in the past. The programs then estimates the time of the population  23  split and the amount of gene flow between the populations (in each direction) after the split that is most likely given the data. This last method is quite revolutionary because it circumvents the difficult problem of distinguishing between current gene flow and incomplete lineage sorting (Slatkin & Maddison 1989; Funk & Omland 2003). It should be noted, however, that the performance of this method is debated (Strasburg & Rieseberg 2010; Strasburg & Rieseberg 2011; Gaggiotti 2011). Other methods not relying on maximum likelihood allow more flexibility in the models of evolution tested, but do not technically allow estimates of gene flow or dispersal. Instead, they focus on falsifying scenarios and hypotheses (e.g., vicariance vs. historical gene flow) not likely to generate the pattern observed in the data. This is an approach that tends to be used more commonly in studies of historical gene flow, such as in phylogeography. The formerly popular nested clade analysis (NCA) can be classified among these approaches (Templeton 1998). It has, however, recently fallen into disfavor (Petit 2008), and the statistical phylogeographic approach is now preferred (Knowles & Maddison 2002; Knowles 2004). In short, the statistical phylogeographic approach involves simulating the evolution of artificial DNA sequences similar to the empirically derived sequences. Those sequences are made to ‘evolve’ in silico according to an empirically-derived model of mutation and under different evolutionary scenarios. The evolutionary scenarios often involve vicariant events occurring at varying times in the past. These models are ran multiple times, and the distribution of summary statistics (e.g., S, a phylogenetic measure of gene flow; Slatkin & Maddison 1989) describing the outcomes of evolution under the different scenarios are then compared to the empiricallyderived values of these statistics. Scenarios that are deemed statistically unlikely are then eliminated (for an excellent example see Steele & Storfer 2006). Another similar approach that is more amenable to use with microsatellite data is that of approximate Bayesian computations (ABC). Because this approach was not used in my dissertation, I will refer the interested reader to the reviews by Beaumont et al. (2002) and Csilléry et al. (2010). Another major class of molecular methods to estimate dispersal are genetic assignment methods (Manel et al. 2005). Assignment methods classify individuals into their most likely populations of origin on the basis of their multilocus genotype (Paetkau  24  et al. 1995; Paetkau et al. 2004). In most cases, potential source populations are sampled and probabilities of a genotype occurring in that population are computed. The putative source population of an individual is then inferred to be that where its genotype is most likely (Manel et al. 2005). These methods have the advantage of not assuming the selective neutrality of genetic markers, but they still assume that all potential source populations are sampled randomly, that they are in Hardy-Weinberg and linkage equilibrium (Manel et al. 2005), and that they are sufficiently genetically differentiated to allow confident assignment (Paetkau et al. 2004). Another potential advantage is that assignment methods measure gross dispersal, not realized dispersal or gene flow. They can thus be combined with previously mentioned methods to jointly evaluate and compare current dispersal and long-term/historical gene flow. The program STRUCTURE (Prichard 2000), which is used heavily in this dissertation, can be classified as an assignment procedure, the main difference here being that it does not require a priori designation of source populations, but rather infers genetic clusters that maximize linkage and Hardy-Weinberg equilibrium. The probability of membership of each individual to the genetic cluster is then calculated and can be interpreted as a hybrid index or as a probability of belonging to certain cluster.  !"C"; F(,,-12'-06'/19-0+G-*+/0'/5'3%-7*&14' In Chapter 2, I revisit the phylogeography of North American Arctic Char and aim to increase our understanding of patterns of post-glacial dispersal. The North American range of the species having been almost entirely covered by ice during the last glaciation (Dyke et al. 2003), dispersal clearly had a central role in determining the current distribution of the species. Glaciations should also have had major consequences for the distribution of intraspecific genetic variation (Hewitt 2000, 2004), but existing phylogeographic studies of Arctic Char are mainly silent on the issue (Wilson et al. 1996; Brunner et al. 2001). In addition to examining the role of post-glacial dispersal on the distribution of genetic variation, I also consider the roles of hybridization between lineages and of differences in potential for gene flow among populations in determining the amount and distribution of genetic variation in the Canadian Arctic.  25  In Chapter 3, I explore patterns and consequences of contemporary dispersal on a regional scale. As discussed earlier, Arctic Char has a complex migratory behaviour and patterns of dispersal among populations are poorly understood. All existing studies of dispersal and migrations have involved tagging individuals, most often in a single river. I take advantage of the geography of the Cumberland Sound, Baffin Island, Canada, to study dispersal in a large number of populations that are all found at the end of long glacial fiords and are thus well geographically delineated. I rely on genetic assignment of individuals to their putative population of origin to quantify dispersal rates, but also to test the hypothesis that overwintering char have an increased propensity to utilize nonnatal environments. Because overwintering dispersal does not lead to gene flow, evaluating the relative importance of overwintering and breeding dispersal is crucial to making inferences regarding the consequences of dispersal for local adaptation. I therefore parameterize a population genetic model of the balance between drift, gene flow, and selection based on empirical estimates of gene flow, to explore potential for local adaptation among those populations. In Chapter 4, I explore the genetic basis for anadromy, a trait that constitutes a necessary prerequisite for dispersal. I do so by studying three populations of Arctic Char where anadromous and freshwater resident individuals co-exist in sympatry. I test two alternative hypotheses to explain this co-existence: (1) that the two components of the population are reproductively isolated and use different migratory strategies, or (2) that it is the result of a conditional mating strategy where a single genotype is able to express to alternative strategies (Gross 1996). I examine genetic differentiation among life-history ecotypes using microsatellite markers because the former hypothesis predicts genetic differentiation according to phenotype, whereas the latter does not. Together, the three parts of my dissertation aim to provide a broad description of patterns of dispersal in Arctic Char at a variety of temporal and spatial scales. Describing the consequences of this dispersal for the distribution of genetic variation and for potential for local adaption is also a major goal. In the concluding chapter, I explore different implications of my work, and attempt to unify the different chapters under the overarching goal of predicting the response of Arctic Char to a changing climate.  26  Table 1.1 Some definitions of dispersal from recent reviews on the topic. Reference  Definition  Howard 1960  “…the movement the animal makes from its point of origin to the place where it would have reproduced if it had survived and found a mate.”  Stenseth & Lidicker 1992  “Circumstances in which individuals leave their existing home ranges and do not return, at least in the short-term as they would after brief excursions.”  Clobert et al. 2001  “…the movement between the natal area or social group and the area or social group where breeding first takes place”  Nathan 2001  “the movement of organisms away from their parent source”  Bowler & Benton 2005  “…any movement between habitat patches, with habitat patches being defined as areas of suitable habitat separated in space from other such areas, irrespective of the distance between them.”  Garant et al. 2007  They distinguish between dispersal, which they define as the movement of individuals, and gene flow, which they define as the movement of genes.  Ronce 2007  “…any movement of individuals or propagules with potential consequences for gene flow across space.”  Clobert et al. 2009  “…active or passive attempt to move from a natal ⁄ breeding site to another breeding site”  27  Table 1.2 Estimates of dispersal rates from tagging studies performed on Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma. Study  Location  Type of study  Dempson & Kristofferson 1987 Dempson & Kristofferson 1987 Berg & Jonsson 1989  Labrador  Carlin tags with fishery-dependent recovery  Victoria Island, Nunavut N. Norway  Floy tags with fishery-dependent recovery  6195  7  ~500km  13-51%  Carlin tags with fishery-dependent recovery  6155  >500km  11%  Gyselman 1994  Kent Pen. Nunavut Norway  Floy and Carlin tags with weir  5753  Not indicated but >1yr 5  <100 km  47%  Carlin tags with fishery-dependent recovery  131  1  <25km  12%  Alaska  Floy tags with fishery-dependent recovery  4075  1 yr  ~2000km  21%  Nordeng & Bratland 2006 Dolly Varden, S. malma DeCicco 1992  Number of fish 7566  Duration of study (yrs) 7  Spatial scale (km) ~500km  Estimated dispersal rate 3-17%  Table 1.3 Summary of evidence for increased propensity to disperse when overwintering in Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma. Study  Location  Type of study  Finstad & Heggberget 1993 Gyselman 1994  N. Norway  Carlin tags with weir Floy and Carlin tags with weir Telemetry  Beddow et al. 1998  Kent Pen., Nunavut, Labrador  Number of fish tagged/genotyped 1000 wild (1390 hatchery-raised) 5753  Duration of study 3 yrs  Spatial scale (km) <100 km  5 yrs  <100 km  154  3 months  <50km  How was overwintering dispersal inferred? From individuals returning after 1-2 yrs absence From individuals returning after 1-2 yrs absence Radio-tracked  Inferred frequency of over-wintering dispersal "a small proportion" 7.2% 44.5-61%  Bernatchez et al. Labrador 1998 Dolly Varden, S. malma  Microsatellites  257  NA  ~300km  Wahlund's effect  NA  Armstrong 1974  Dart tags with weir  11,901  4 yrs  ~20km  Non-spawners captured in non-natal streams; no spawners captured in nonnatal streams  NA  SE Alaska  28  Table 1.4 Summary of studies examining the marine migrations of Arctic Char, Salvelinus alpinus, and Dolly Varden, S. malma. Study  Dempson & Kristofferson 1987 Dempson & Kristofferson 1987 Berg & Jonsson 1989 Finstad & Heggberget 1993 Bégout-Anras et al. 1999 Gulseth & Nilssen (2000) Spares et al. 2012  Location  Type of study  Number of fish tagged/genotyped  Duration of study (yrs)  Labrador  Carlin tags with fisherydependent recovery Floy tags with fisherydependent recovery  7566  Carlin tags with fisherydependent recovery Carlin tags with weir  Victoria Island, Nunavut N. Norway N. Norway Victoria Island, Nunavut Svalbard, Nor.  Baffin Island, Nunavut Dolly Varden, S. malma DeCicco 1992  Alaska  Maximum distance (km) 250  Range of most recaptures/observations  7 yrs  Spatial scale (km) ~500km  6195  7 yrs  ~500km  550  85% within 60km  6165  Not indicated but >1yr 3 yrs  >500km  >400km  54.9% within 3km  <100 km  100  80% within 30km  2 weeks  ~1km  ~1  <1km  95% within 70km  Telemetry  1000 wild (1390 hatchery-raised) 9  Floy tags with weir  2060  3 yrs  ~150km  150  NA  Telemetry  27  2 yrs  ~30km  26.6  <4km  Floy tags with fisherydependent recovery  4075  1 yr  ~2000k m  1690  79% returned to natal river  29  Arctic Charr (Salvelinus alpinus) Anderson Bay Lake, Cambridge Bay Region  Canada, Figure 1.1.Nunavut, Arctic Char, Salvelinus alpinus. The drawing is based on an anadromous specimen in spawning colours from Anderson Length: 730 mm FL,  Bay Lake inSpecimen the Cambridge Bay Region number: 40259of Nunavut, Canada. Drawing by Paul Vescei. © Government of Canada.  100 mm  Illustration by Paul Vecsei for Dr. Jim Reist, Fisheries and Oceans Canada Contract: Charr diversity in Northern Canada Reference Number F2402-070209 through DFO Central and Arctic Regional Office  30  Figure 1.2 Map of the northern hemisphere showing the Holarctic distribution of Arctic Char, Salvelinus alpinus, in dark shaded areas surrounded by dashed lines. The distribution is a simplification of that found in Reist and Sawatzky (2010) and omits a small number of introduced populations.  Lake Baikal  Russia  Novaya Zemlya  Alps  Svalbard UK  ka as ) Al SA (U  G re  Brooks Range  Canada  USA  31  en  la  nd  Iceland  ! "#$%&'()$#*+,$),+#%-./*)/01#2&3%)4)5,()671#,74#,7,4%60&# 82,'-#(2-#4)8(%)3/()67#6.#*-7-()$#9,%),()67#)7#:,7,4),7#"%$()$# :2,%##  !;< "38(%,$(# The geographic distribution of genetic diversity is determined by both historical and contemporary factors. I examined the roles of glaciation, hybridization, and contemporary dispersal in shaping genetic diversity in Arctic Char (Salvelinus alpinus) from the North American Arctic. I identified two highly divergent mtDNA lineages of Arctic char in the high Arctic. The geographic distribution of mtDNA haplotypes and historical demographic analyses suggest that one lineage survived in Beringia, while the other may have survived in a small high Arctic refugium. Patterns of variation at nine microsatellite loci reflected both historical and contemporary processes. Contrary to expectations, genetic diversity did not decrease with increasing distance from a putative glacial refugium. I found evidence that secondary contact with Atlantic basin Arctic Char leads to increased genetic diversity in populations farthest from the putative refugium, thus blurring the historical signatures of re-colonization. Landlocked populations had lower within-population genetic diversity, and higher pairwise genetic differentiation among populations, than anadromous populations, as predicted by the reduced potential for gene flow among landlocked populations. Patterns of isolation by distance also differed between landlocked and anadromous populations. These results illustrate the importance of considering a variety of historical and contemporary processes when making inferences regarding factors driving the geographic distribution of genetic diversity.  32  !;! =7(%64/$()67# The geographic distribution of genetic variation has important evolutionary consequences, and understanding the historical and contemporary factors shaping this distribution remains a major goal of evolutionary biology (Avise 2004). The distribution of genetic variation is ultimately shaped by the evolutionary forces of genetic drift, selection, and gene flow. The strength and direction of these forces, however, are in turn influenced by a variety of historical and contemporary ecological factors that impact population size and how easily genes move between populations. For high latitude biota of the Northern Hemisphere, the Pleistocene glaciations had a dominant and well-documented role in shaping patterns of genetic diversity (Hewitt 2000, 2004). The consensus view emerging from hundreds of phylogeographic studies is that most northern species survived in refugia south of the ice sheets during the last glacial maximum (LGM) and recolonized their current range as the ice receded (Bernatchez & Wilson 1998; Hewitt 2000; Soltis et al. 2006). Not all species, however, survived in southern refugia. In North America, one well-known exception to the general rule of survival south of the ice sheets is the Beringian refugium, which was a major icefree area that many species used as a refugium (e.g. Abbott et al. 2000; Federov & Stanseth 2002; Loehr et al. 2006; Harris & Taylor 2010; Shafer et al. 2011). There is also accumulating evidence of species having survived the LGM in smaller refugia north of the ice sheets. Such refugia are commonly referred to as ‘cryptic refugia’ or ‘microrefugia’ (Provan & Bennett 2008; Rull 2009; Stewart et al. 2010) and are defined as small areas of favorable conditions outside of the species main (macro-) refugia, and often situated at different latitudes or longitudes than would normally be expected (Rull 2009; Stewart et al. 2010). In North America for example, there is accumulating evidence that a number of species may have survived the LGM in the Arctic Archipelago itself. Indeed, some areas of the Canadian Arctic Archipelago, notably on Banks Island in the western part of the Archipelago, were ice free throughout the LGM (Fig. 2.1 and Dyke et al. 2003), and phylogeographic evidence suggests species of mammals (Federov & Stanseth 2002), birds (Holder et al. 1999), and plants (Tremblay & Schoen 1999; Abbott et al. 2000) may have survived the LGM in these ice-free areas of the high Arctic. Survival of those species in multiple previously unknown refugia leaves a signature in the 33  distribution of genetic diversity, with centers of diversity and phylogeographic breaks that are otherwise difficult to explain (Provan & Bennett 2008). Despite the obvious role played by glaciations in shaping patterns of genetic variation of northern species, a variety of other processes can act to blur the genetic signal left by range contraction/expansion cycles brought about by the glaciations. For instance, many studies have documented cases where secondary contact between different lineages leads to patterns of elevated genetic diversity (e.g. Turgeon & Bernatchez 2001; Petit et al. 2003). Such contact zones, if ignored, can lead to misleading historical inferences, because patterns of genetic diversity are often used to infer the location of putative refugium under the assumption that refugial areas should display elevated genetic diversity (Provan & Bennett 2008). In addition, contemporary processes like dispersal and gene flow will also influence geographic patterns of genetic variation, and there are well-documented effects of interspecific differences in dispersal capabilities on the geographic distribution of genetic variation (Bohonak 1999). Much less work, however, has been conducted on how intraspecific variation in dispersal propensity among populations affects the geographic distribution of genetic variation (for a few examples, see Mäkinen et al. 2006; Tonteri et al. 2007). The Arctic Char Salvelinus alpinus is a salmonid fish that shows a bewildering array of phenotypic diversity within what has become known as a “species complex”. This diversity has made the Arctic Char a model system of choice in studies of the ecology of adaptive radiation (Skúlason et al. 1999; Jonsson & Jonsson 2001; Schluter 2002). An understanding of the ecological factors driving diversification in this species, however, requires knowledge of the historical and contemporary factors that have shaped patterns of genetic variation. Arctic Char is the most northerly-distributed species of freshwater fish in the world (Johnson 1980), and ice covered large portions of its current distribution during the LGM. Our current understanding of the effects of Pleistocene glaciations on the distribution of genetic variation in Arctic Char is mainly based on the Holarctic phylogeographic study of Brunner et al. (2001; see also Wilson et al. 1996). Brunner et al. (2001) used mtDNA control region sequences to identify five distinct lineages of Arctic Char worldwide (Fig. 2.1) which were interpreted as lineages that survived in isolation in five different glacial refugia during the LGM. While global in  34  scope, the Brunner et al. (2001) study did not have the resolution to identify the location of these glacial refugia. Of particular interest is the hypothesis formulated by Crossman & McAllister (1986) that North American Arctic Char may have survived the last glaciation in a high Arctic ‘cryptic refugium’. While there currently exists no evidence of a fish species surviving in such a high Arctic refugium, Arctic Char would be an ideal candidate as it is considered well adapted to extreme cold environments (Power 2002). For example, populations of Arctic Char currently survive in freshwater environments on the narrow strip of land directly adjacent to the Greenland ice-sheet (Crossman & McAllister 1986). The presence of four highly divergent mtDNA lineages (sensu Brunner et al. 2001) in North America also suggests the possibility of secondary contact and introgression between lineages. In fact, Wilson et al. (1996) showed that two mtDNA lineages co-existed in populations of Arctic Char from northern Labrador. It remains unknown how this potential introgression shaped genetic diversity in the nuclear genome. In addition to the historical influences of isolation in glacial refugia and postglacial recolonization, I expect geographic patterns of genetic diversity in Arctic Char to be influenced by contemporary life-history variation between populations. Arctic Char populations vary tremendously in their life-history characteristics and dispersal capabilities. Many populations of Arctic Char are anadromous (Johnson 1980), i.e., they are hatched in fresh water but as adults and sub-adults undergo annual feeding migrations to the ocean before returning to fresh water to spawn and over-winter. In fact, the species’ coastal Holarctic distribution strongly suggests that it recolonized its current range mainly via marine routes. Other populations, however, are landlocked (or freshwater resident) and have no capacity for dispersal and thus low potential for gene flow. This difference in potential for gene flow between populations with different dispersal capabilities would be expected to have dramatic consequences for the distribution of genetic diversity (Bohonak 1999; DeWoody & Avise 2000). In the present study, I evaluated the relative importance of historical and contemporary drivers of genetic diversity in Arctic Char in Arctic North America (excluding the Eastern sub-Arctic regions), focusing more specifically on the effects of glaciations, secondary contact, and variation in dispersal propensity between populations. My first objective was to refine our understanding of the effects of the last glaciation,  35  with a particular emphasis on elucidating the locations of glacial refugia. To do so, I used mtDNA control region sequence data and substantially increased the number of locations sampled and the number of samples per location compared to previous studies (Wilson et al. 1996; Brunner et al. 2001). In addition, I surveyed variation in nine microsatellite markers to provide increased resolution for historical inference given that they have proved useful in many other phylogeographic studies of northern salmonid fishes from recently deglaciated areas (e.g., Angers & Bernatchez 1998; Koskinen et al. 2002; Harris & Taylor 2010). Microsatellite data also provide genetic markers independent from mtDNA to assess whether or not patterns observed in mtDNA are reflected in the nuclear genome. I assessed two alternative hypotheses regarding refugial origins of Arctic Char in the North American Arctic: that (1) the North American Arctic was recolonized only from a Beringian refugium, or (2) recolonization occurred from two separate refugia, one in Beringia and one in ice-free areas of the Arctic Archipelago. If Arctic Char did survive in a high Arctic refugium in addition to a Beringian refugium, I predicted that I would find evidence of two separate genetic lineages within the North American Arctic using both the mtDNA and microsatellite data. Based on patterns observed in previous studies that identified high Arctic refugia in other North American Arctic species (e.g., collared lemmings; Federov & Stenseth 2002), it is likely that the Beringian and high Arctic lineages would be restricted to Beringia and the Arctic Archipelago. In the Arctic Archipelago, I predicted that genetic diversity would decline with increasing distance from the location of the putative glacial refugium (Provan & Bennett 2008). Finally, because a high Arctic refugium may have been much smaller in area than a Beringian refugium, I predicted that I would find comparatively lower genetic diversity and a signature of recent population expansion in a putative high Arctic lineage. I was wary of the potential influence of introgression with other refugial lineages (i.e., lineages other than those originating from Beringia and the high Arctic). Introgression with other lineages would lead to elevated genetic diversity in areas not used as refugia, and could thus confound interpretations if ignored (e.g., Petit et al. 2003). Hence, I assessed whether or not introgression with other refugial lineages has occurred, explicitly considering the possibility of introgression between the Arctic and Atlantic lineages (sensu Brunner et al. 2001; Fig 2.1).  36  My second objective was to explicitly consider differences in life-history characteristics (i.e., anadromy) between populations as a potential explanatory variable for differences in genetic diversity and population structure. I predicted that populations that do not migrate to sea (i.e., landlocked or resident) have had no potential for gene flow since isostatic rebound isolated their current lake habitats, and should therefore have lower within-population genetic diversity, but higher genetic differentiation among populations. Anadromous populations, on the other hand, migrate to sea every year, and a portion of migrating individuals may not home to their natal streams to spawn (i.e., strays). This potential for gene flow should lead to higher genetic diversity within populations, and lower genetic differentiation between populations (Bohonak 1999; DeWoody & Avise 2000).  !;> ?,(-%),+8#,74#0-(2648# !"#"$ %&'()*+, I collected mtDNA sequence data from a total of 1,262 individuals from 99 sampling locations distributed across eastern Siberia and the North American Arctic (see Appendix A table A.1 for details). The majority of the sequences generated in this study were from archived samples collected by Fisheries and Oceans Canada over the last 30 years. Sequences generated for this study were combined with sequences from Brunner et al. (2001), Taylor et al. (2008), Power et al. (2009) and Alekseyev et al. (2009) obtained from GenBank. Those sequences represented all the available mtDNA control region sequences of Arctic Char and Dolly Varden (S. malma) accessible on GenBank for the geographic region of interest. The taxonomic status of Arctic Char and Dolly Varden as separate species has been contentious (Reist et al. 1997; Brunner et al. 2001), but recent work strongly suggests they are distinct biological species (Taylor et al. 2008). Nevertheless, I included samples of Dolly Varden to (1) be consistent with previous studies (Brunner et al. 2001; Taylor et al. 2008; Alekseyev et al. 2009), and (2) because mtDNA haplotypes have been shown to introgress between the two species (Brunner et al. 2001; Taylor et al. 2008), and their inclusion can thus help elucidate the history of  37  mtDNA lineages in this region. I used the morphological criteria in Reist et al. (1997) combined with geographic location for the taxonomic designation as Arctic Char or Dolly Varden, and genetic information was then associated with these taxa a posteriori. The focus of the present study, however, remains Arctic Char specifically, and the inclusion of Dolly Varden only serves to place the evolutionary history of Arctic Char in this broader taxonomic context and to fill geographical gaps in sample availability. I also added some sequences from the three other lineages of Arctic Char (i.e., Siberian, Atlantic, and Acadian) identified by Brunner et al. (2001) in order to place my samples within this holarctic evolutionary framework. Finally, I used homologous sequences from Lake Trout (S. namaycush), White-Spotted Char (S. leucomaenis), and Brook Trout (S. fontinalis) as outgroup taxa obtained from Taylor et al. (2008) for Lake Trout and from Brunner et al. (2001) for the other two, again so that my results can be compared to those previous studies. A different set of samples was used for the microsatellite analysis (Appendix A Table A.2). All samples were specimens that were morphologically identified as Arctic Char and no samples of Dolly Varden were included in this analysis. The samples came from a total of 37 sampling locations (or populations; I use the terms interchangeably) that covered most of the distribution of Arctic Char in the Canadian Arctic. I also included three sampling locations from Alaska and one sampling location from Labrador so that representatives of the three mtDNA lineages found in the North American Arctic (sensu Brunner et al. 2001) would be included in the microsatellite analysis. The number of individuals sampled per population varied from 7 to 56 (mean = 25.2; total = 931). Because of the difficulty of obtaining samples from the high Arctic, some populations were represented by a small number of individuals. Population-based statistics should thus be interpreted cautiously, although the STRUCTURE analysis should not be affected by small number of individuals per population, because population origin is not used as part of the analysis (Pritchard et al. 2000).  38  !"#"! -./,'*0123+, Samples consisted either of dorsal muscle tissue, liver, or fin preserved in a 20% DMSO ⁄ NaCl solution or in 95% ethanol and kept frozen until DNA extraction. The entire mtDNA control region (D-loop) was amplified with primers Tpro2 (Brunner et al. 2001) and SalpcrR (Power et al. 2009). I sequenced 502 base pairs of the D-loop left domain region according to methods outlined in Power et al. (2009). Individual genotypes were obtained at nine microsatellite loci combined in two multiplexes. The first multiplex included the following loci: Sco200, Sco215, Sco220 (DeHaan & Ardren 2005), and Smm22 (Crane et al. 2004); and the second multiplex included OMM1105, OMM1128 (Rexroad III et al. 2001), Smm24 (Crane et al. 2004), OtsG253b (Williamson et al. 2002), and SSOSL456 (Slettan et al. 1995). Qiagen (Valencia, CA, USA) Multiplex PCR Kits were used for the PCR, and both multiplex reactions used the following cycling conditions: an initial denaturation period of 15 min at 95°C, followed by 35 cycles of denaturation (94°C for 30 sec), annealing (55°C for 1m30s) and elongation (72°C for 1min), and a final elongation period of 30 min at 60°C. Concentrations of each reagent followed the guidelines provided by the manufacturer. The PCR products were run on an Applied Biosystems (Carlsbad, CA, USA) 3100 Genetic Analyzer. GeneMapper Software version 3.7 (Applied Biosystems) was used to automatically score microsatellite alleles, and all scores were manually checked for quality. !"#"# '0-./,&4&)5+6+, Control region sequences were aligned by hand using Se-Al v2.0a11 (Andrew Rambaut 2002, http://tree.bio.ed.ac.uk/software/seal/, accessed 2010) with gaps treated as single characters. I calculated haplotype (h) and nucleotide diversity (!) using DnaSP v5.0 (Librado & Rozas 2009). I used jModelTest 0.1.1 (Posada 2008) to identify the most likely model of evolution on an alignment containing one copy of all haplotypes (including outgroups). I limited the analysis to three substitution schemes (i.e., JC, HKY, GTR) because most programs we used downstream did not allow for alternative substitution models. The most likely model according to both the Akaike information criterion corrected for small sample size (AICc) and the Bayesian information criterion (BIC) criteria was the HKY + G model. I used the maximum likelihood-based genetic  39  algorithm implemented in GARLI v.0.96beta8 (Zwickl 2006) to infer phylogenetic relationships among haplotypes. The runs were terminated automatically when 10,000 generations of the genetic algorithm elapsed without improving the tree topology, with an increase of 0.01 in lnL as the criterion for significant improvement. Support for the nodes was assessed with 1,000 bootstrap replicates, also in GARLI. Percent sequence divergence estimates corrected for multiple hits (under the HKY + G model) were then obtained from the branch lengths of the tree. A rough estimate of divergence times was then obtained using a standard molecular clock of 5-10% per million years (see Brunner et al. 2001 for a detailed discussion of this choice of clock rate). !"#"7 86+0296:&),3*'2;9&(15, The hypothesis that the Arctic and Bering lineages experienced recent population expansion following the LGM was evaluated using a variety of tests. First, departures from neutrality of the mtDNA sequence data were tested using Tajima’s D (Tajima 1989) and Fu’s FS (Fu 1997). Significant negative values for both of these statistics would be indicative of recent population expansion. The statistics were estimated and their significance was assessed using 1,000 bootstrap replicates in ARLEQUIN (Excoffier et al. 2005). Second, I performed mismatch analysis (Rogers and Harpending 1992) to compare the demographic histories of the Arctic and Bering lineages. The method of Schneider & Excoffier (1999) implemented in ARLEQUIN fits a model of sudden demographic expansion to the observed distribution of pairwise nucleotide differences, estimating the population size parameters before ("0 = 2u#0) and after ("1 = 2u#1) the demographic expansion, and the time of the expansion $ = ut, where N0 and N1 are the female effective population sizes before and after the demographic expansion, u is the cumulative mutation rate per generation per gene (i.e., substitution rate per nucleotide per generation multiplied by the length of the gene in base pairs; Schenekar & Weiss 2011), and t is the number of generations since the population expansion event (Rogers & Harpending 1992). The goodness of fit of the sudden population expansion model to the observed data was measured using two test statistics: the sum of squares deviation method (SSD; Schneider & Excoffier 1999) and Harpending’s raggedness (Harpending 1994). Significance of the goodness of fit test statistics were determined using 1,000  40  bootstrap replicates in ARLEQUIN. Third, I attempted to construct Bayesian skyline plots (Drummond et al. 2005) in program BEAST (Drummond & Rambaut 2007) as an alternative to the parametric approaches to reconstructing historical demography. Unfortunately, the program failed to converge even after 5*109 MCMC repetitions. This was true for a variety of combinations of priors for clock rates, time to most recent common ancestor, and population sizes. This failure of the program is presumably due to the low information content of the S. alpinus Arctic lineage samples that have very low genetic diversity (see Results section). I therefore do not report the results of this analysis. !"#"< =6:92+&0*))60*,&4&)5+6+, FSTAT version 2.9.3.2 (Goudet 2001) was used to test for Hardy-Weinberg equilibrium and genotypic disequilibrium using 10,000 permutations for both analyses, and setting the nominal significance level at 0.05. I used MicrosatelliteToolkit (v. 3.1.1; Park 2001) to generate estimates of observed heterozygosity (HO), and expected heterozygosity (HE) corrected for samples size. FSTAT was used to calculate allelic richness (AR) and pairwise FST (Weir and Cockerham 1984) between each sample, and significance was assessed with 10,000 permutations (experiment-wide ! = 0.05 after Bonferroni correction). I used PHYLIP version 3.68 (Felsenstein 1993) to generate a neighborjoining tree of all samples based on Cavalli-Sforza’s chord measure (Cavalli-Sforza & Edwards 1967) employing 1,000 bootstrap replicates to test support for each node. The tree was visualized and formatted in FigTree v1.3.1 (Rambaut 2010). To test the prediction that genetic diversity should decline away from the putative high Arctic glacial refugium, I regressed values of expected heterozygocity (Nei’s unbiased gene diversity; Nei 1987) and allelic richness (as calculated by FSTAT; details presented above) against distance from Banks Island, the most likely high Arctic putative glacial refugium. The distances were waterway distances generated using Google Earth (v. 5.2.1.1588). When more than one route was possible, I chose the most likely route based on patterns of glacial retreat (Dyke et al. 2003). Note that the patterns would have been nearly identical if distance from the eastern boundary of Beringia had been used, because Banks Island is located relatively close to Beringia. This analysis therefore does not allow falsification of any of the alternate hypotheses regarding refugial origins (i.e.,  41  whether the lineages survived in Beringia or the high Arctic). Only samples from regions where the mtDNA Arctic lineage predominated were included in this analysis (i.e., Alaska and Labrador samples were excluded) because I am specifically interested in the re-colonization history of the Arctic Archipelago. I used the program STRUCTURE (Pritchard et al. 2000) to test the hypothesis that Arctic Char from the Canadian Arctic Archipelago are part of a single genetic group. This analysis was performed to assess the degree of concordance with the mtDNA evidence and to ensure that patterns of variation in mtDNA reflect historical demography rather than natural selection for linked mitochondrial genes (Ballard & Whitlock 2004; Toews & Brelsford 2012). The analysis included all the samples (931 individuals), and hypothesized number of genetic groups (K) ranging from one to ten were tested using 500,000 burn-in and 1,000,000 MCMC repetitions under the admixture and correlated allele frequency models, without location priors. Twenty independent runs were performed for each K, and the "K method of Evanno et al. (2005) was used to determine the most likely value of K. The program CLUMPP (v. 1.1.2; Jakobsson & Rosenberg 2007) was used to combine the results of the 20 independent runs using the Greedy algorithm, and the program DISTRUCT (v. 1.1; Rosenberg 2004) was used to visualize the results. I also tested for the effect of anadromy on patterns of genetic variation within the Arctic lineage (i.e., excluding Alaskan and Labrador samples). I first tested the hypothesis that average genetic diversity should be higher within anadromous populations than within landlocked populations using comparisons of both allelic richness and expected heterozygosity, and assessed the significance of the differences using 10,000 permutations in FSTAT. Second, I tested the hypothesis that average genetic differentiation between populations should be higher between landlocked populations than between anadromous populations by comparing average pairwise FST values among anadromous and landlocked populations, also using 10,000 permutations in FSTAT. Third, I explored the consequences of anadromy for isolation by distance. I used Mantel tests (Mantel 1967) implemented in the R package ecodist (Goslee & Urban 2007) to test for the effect of distance (log10 transformed) on genetic differentiation (i.e., linearized FST; Rousset 1997) for (1) all populations, (2) anadromous populations only,  42  and (3) landlocked populations only (excluding the Yukon samples). I used a partial Mantel test (Smouse et al. 1986), also implemented in ecodist, to test for the effect of life-history (i.e., anadromous vs. landlocked) on genetic differentiation while controlling for the effects of geographic distance. Because this test requires distance matrices, I coded the life-history data as a distance matrix by assigning a ‘0’ to the population pairs that had the same life-history and a ‘1’ to pairs that differed in life-history. For both the Mantel and partial Mantel tests, significance was determined with 105 permutations.  !;@ A-8/+(8# !"7"$ '0-./,>&96&0624, My mtDNA sequencing results uncovered many inconsistencies with the results reported by Brunner et al. (2001) (details below). Samples used by Brunner et al. (2001) were graciously made available to us by one of the authors of the study (L. Bernatchez), and re-sequencing suggested that many of the haplotypes reported were in fact sequencing artifacts. I therefore excluded all haplotypes from Brunner et al. (2001) from subsequent analyses unless the haplotypes were corroborated by another study. I did, however, keep some Atlantic haplotypes from Brunner et al. (2001) to construct the phylogenetic tree (Fig. 2.2) because no other studies provide haplotypes from this region. I found two distinct groups of mtDNA haplotypes in the North American Arctic (Fig. 2.2) that roughly corresponded to the Beringian and Arctic lineages identified in Brunner et al. (2001; Fig. 2.1). The Arctic lineage formed a well-supported (76.2% bootstrap support) monophyletic group (Fig. 2.2). The Bering group did not resolve into a reciprocally monophyletic clade or lineage, in contrast to previous studies (Brunner et al. 2001; Taylor et al. 2008; Alekseyev et al. 2009), but rather forms a paraphyletic group (Fig. 2.2). This different result probably resulted from the exclusion of the erroneous haplotypes. The Arctic and Bering haplotypes had seven fixed differences, which lead to a 2.01% divergence estimate between the Arctic lineage and the nearest Bearing haplotype after correction for multiple hits. Assuming a clock rate of 5-10% sequence divergence per million years, this leads to an estimate of divergence time of 201,000 to  43  402,000 years ago. The Atlantic-Siberia-Acadian super-group identified by Alekseyev et al. (2009) was recovered, although with low bootstrap support (<50%). Also consistent with the results of Alekseyev et al. (2009), I found the Siberian group to be paraphyletic. Although I identified twelve new haplotypes not previously described from the Canadian Arctic (Fig. 2.2), the overall level of genetic variation in mtDNA haplotypes in the Arctic lineage was very low, with 1,075 of 1,124 (95.6%) individuals sharing the ARC19 haplotype. The phylogenetic relationship between the Arctic haplotypes is a ‘star-phylogeny’ centered on the ARC19 haplotype (Fig. 2.2). The other twelve Arctic lineage haplotypes had no clear pattern of geographic distribution and tended to be found in a single sampling site, or in multiple sampling sites that were geographically proximate (see Appendix A Table A.3 for details). The Bering group contained more genetic diversity: I uncovered almost the same number of haplotypes in the Bering group (10) and the Arctic group (13) despite a much larger sample size for the Arctic group. In addition, the most common haplotype in the Bering group (BER12) was shared by only 39.8% of the individuals. Estimates of haplotype and nucleotide diversity for the Arctic lineage (h = 0.085, StDev = 0.012; ! = 0.00018, StDev = 0.00002) were both an order of magnitude lower than those for the Bering lineage (h = 0.744, StDev = 0.026; ! = 0.00303, StDev = 0.00015). The geographic distribution of the Bering and Arctic lineages (Fig. 2.3) also differed from that reported by Brunner et al. (2001; Fig. 2.1). Similar to Brunner et al. (2001) I found that the Arctic lineage was distributed throughout the Arctic Archipelago, the Canadian mainland Arctic, and Greenland (Fig. 2.3). Unlike in Brunner et al. (2001), however, the Arctic lineage was not distributed in northern Alaska and Beringia where the Beringian haplotypes predominated. Most of the Dolly Varden samples from the North Slope of Alaska had Bering group haplotypes (Fig. 2.3), except for two individuals which had Arctic lineage haplotypes (see inset Fig. 2.3). Samples of Dolly Varden and Arctic Char from eastern Siberia also had Bering group haplotypes (Fig. 2.3). !"7"! 86+0296:&),3*'2;9&(15, The results of all analyses of historical demography are consistent with the hypothesis that the Arctic mtDNA lineage has undergone a recent population expansion, while the  44  Bering lineage has had a more or less constant population size since deglaciation. First, the estimates of both Fu’s FS and Tajima’s D are significantly negative for the Arctic lineage, while they are not significantly different from zero for the Bering lineage (Table 2.1). Significantly negative values of those two indices are generally interpreted as an indication of either recent population expansion, or of selection on the marker. The mismatch distribution analysis indicated that the patterns of mtDNA substitutions in the Arctic lineage could not be distinguished from that expected under a model of sudden population expansion (Table 2.1 and Fig. 2.4). That was true for both measures of goodness-of-fit examined (i.e., SSD and raggedness). In contrast, the patterns of substitution in the Bering lineage were significantly different than expected under this model (again, for both SSD and raggedness; see Table 2.1). It should be noted, however, that the shape of the mismatch distribution for the Arctic lineage is also very similar to what would be expected under a constant population size model. Indeed, I also visually examined the fit to an expected distribution under a constant population size model (not , where !! is the  shown). The model I used is the following:  equilibrium probability that two random neutral genes will differ at exactly i nucleotide sites, and where " = 2Nµ = the mean observed pairwise differences (Rogers & Harpending 1992). Both distributions fit the data, because both models make similar predictions if the population expansion is very recent (Rogers & Harpending 1992). !"7"# =6:92+&0*))60*,-./, I found a high level of polymorphism at the microsatellite loci genotyped in the samples of Arctic char, with the number of alleles per locus ranging from 3 (for Omm1128) to 40 (for Sco220) with a mean of 22.9 (see Appendix A Table A.4 for population statistics). Before correction for multiple comparisons, 14 locus/population combinations displayed significant heterozygote deficit, and 8 displayed significant heterozygote excess. No population/locus combination remained significant after correction for multiple comparisons, and no marker consistently displayed departure from Hardy-Weinberg equilibrium across multiple populations. Nine pairs of loci displayed significant linkage disequilibrium before correction for multiple comparisons, but only one remained  45  significant after correction: OtsG253b x Ssosl456. The two Yukon populations, Lake 103 and Lake 104, showed considerably lower variation at the microsatellite loci than any other populations, and were fixed (or nearly fixed) for one allele at four of the nine otherwise polymorphic loci (Omm1105, Omm1128, OtsG253b, and Ssosl456). I ran those samples twice in different laboratories and confirmed that this was not due to technical errors. I do not have a good hypothesis to explain this unusually low level of genetic diversity, but I excluded them from some analyses to avoid biasing conclusions. The neighbor-joining tree of microsatellite data (Fig. 2.5) showed that most of the variation in the Arctic Archipelago and coastal mainland was distributed within populations, and that the internal branches were short and poorly supported by bootstrap analysis. The separation of the Alaska samples from the Arctic samples, on the other hand, was well supported by high bootstrap values, as was the Labrador samples, although the bootstrap value was slightly lower (Fig. 2.5). The two Yukon samples also showed some separation. The fact that they branch from within the Arctic group, not from an internal branch (i.e., a branch connecting the Alaska, Labrador and Arctic samples), suggests that this high degree of separation (bootstrap = 100%) is merely the result of the unusually low genetic variation and not because the group survived the last glaciation in a different refugium. I found little evidence for a pattern of decreasing genetic diversity away from the putative refugium on Banks Island (Fig. 2.6). Because of small sample size and missing data in some populations, FSTAT could only calculate allelic richness from a common sample size of 1 diploid individual, which leads to allelic richness estimates that are equal to expected heterozygosity + 1 (see Appendix A Table A.4 for details). The patterns of allelic richness are therefore identical to those of heterozygosity and we only report the latter. When all sampling locations were included, I found no evidence that expected heterozygosity decreased with increasing distance from Banks Island (df = 31; r2 = 0.0356; p = 0.293). There was also no evidence of decreasing heterozygosity when only anadromous populations were included (df = 20; r2 = 0.006; p = 0.743). When only landlocked populations were included (but excluding the two Yukon samples), there was a weak trend in decreasing heterozygosity (Fig. 2.6), but the relationship was not  46  significant despite explaining a moderate proportion of the variation (df = 9; r2 = 0.319; p = 0.113). Results of the STRUCTURE analysis provided evidence that genetic variation in the nuclear genome parallels that observed in the mitochondrial genome, and further that eastern Arctic populations have undergone secondary contact with Atlantic lineage Arctic Char from Labrador. The "K method of Evanno et al. (2005) identified K=4 as the most likely number of genetic clusters (see Appendix A Fig. A.1). One cluster consisted of Arctic Char from the three southern Alaska sampling locations (CL, TL, SL), a second consisted of the two Yukon sampling locations (Lakes 103, 104), the Labrador population formed a third cluster (Ikadlivik Brook), and the Arctic Archipelago and coastal mainland samples formed a fourth cluster with varying amounts of admixture with the ‘Labrador cluster’ (Fig. 2.7). The tight clustering of the Yukon samples together is probably because of their much lower genetic diversity than other Arctic lineage populations. The analysis also provides compelling evidence that Arctic lineage Arctic Char interacted genetically with Atlantic lineage Arctic Char. The genetic contribution of Labrador genetic material appeared less important in landlocked populations than in anadromous populations (Fig. 2.7). Three populations of anadromous Arctic Char from the Cumberland Sound region of southeastern Baffin Island (Kin, Iqa, and Isu) clustered with the Labrador samples, despite consisting entirely of Arctic lineage mtDNA haplotypes (Fig. 2.3 and 2.7). Bubble plots of allele frequencies (Appendix A Fig. A.2) showed that the sampling locations from the eastern Arctic contain alleles that are otherwise restricted to the Labrador sampling locations. Presence of these additional alleles is also consistent with the elevated genetic diversity displayed by these sampling locations. The microsatellite data also provided evidence that life history (anadromous or landlocked) influenced the distribution of genetic variation. As predicted, landlocked populations displayed significantly lower amounts of genetic diversity as measured by both allelic richness (p = 0.0001) and expected heterozygosity (p = 0.0002; Table 2.2). Landlocked populations also displayed significantly greater genetic differentiation between populations, as measured by average pairwise FST (p = 0.002). Only the differences in allelic richness remained significant when the tests were performed excluding the Yukon samples (Table 2.2). This could in part be due to lower sample sizes,  47  since the average pairwise FST among landlocked populations was still more than double that observed among anadromous populations (0.116 vs. 0.043). There was evidence of isolation-by-distance (Fig. 2.8) whether I analyzed all populations together (Mantel r = 0.287; p < 0.001), landlocked populations only (Mantel r = 0.409; p = 0.003), or anadromous populations only (Mantel r = 0.331; p < 0.001). There was a significant effect of life history on population differentiation even after controlling for geographical distance (partial Mantel r = 0.406; p < 0.001; Fig. 2.8)  !;B C)8$/88)67# The geographical distribution of genetic diversity across species’ ranges is influenced by a variety of historical and contemporary factors. In the present study, I evaluated the relative contributions of historical and contemporary drivers of genetic diversity in populations of Arctic Char from the North American Arctic. More specifically, I documented the influence of glaciations, post-glacial hybridization between lineages, and contemporary gene flow. Both mtDNA and microsatellite data provided evidence for the presence of multiple glacial refugia for Arctic Char in the North American Arctic. I also provided evidence of post-glacial introgression between Arctic and Atlantic lineages of Arctic Char. Finally, I documented the genetic effects of differential dispersal potential on the distribution of genetic variation between landlocked and anadromous populations of Arctic Char. !"<"$ ?>63*4:*,@29,&,16;1,/9:06:,9*@A;6A',@29,/9:06:,:1&9, Evidence for the presence of a high Arctic refugium in the Arctic Archipelago exists for a variety of taxa (birds: Holder et al. 1999; mammals: Federov & Stenseth 2002; plants: Tremblay and Schoen 1999), but my study presents the first evidence of a fish species surviving in a high Arctic refugium during the LGM. I demonstrated the presence of two highly divergent mtDNA lineages in the North American Arctic (Figs. 2.2 and 2.3). This suggests that a vicariant event separated populations of Arctic Char, which I estimated to have occurred 201,000-402,000 years ago. These estimates should obviously be interpreted with caution, but they support a divergence that predates the onset of the last glaciation (ca. 110,000 years ago; Dyke et al. 2003). The presence of  48  those two lineages as well as the amount of divergence between them is consistent with the results obtained by Brunner et al. (2001). The geographic distribution of mtDNA haplotypes, however, differs from that reported in Brunner et al. (2001), where they found the Arctic lineage distributed throughout the northern parts of Beringia. On the contrary, I found that the Beringian lineage is distributed throughout both the northern and southern parts of Beringia, while the Arctic lineage is essentially restricted to the Canadian Arctic. Despite the geographical proximity of the two lineages, I found almost no geographic overlap between them, a fact that would be difficult to reconcile with survival of both lineages in sympatry in Beringia throughout the LGM. The only sampling locations where both haplotypes were recovered were in populations of Dolly Varden from the Brooks Range, Alaska. In both cases, a single individual was identified that carried an Arctic haplotype (Ayers 2010). Owing to the striking geographic partitioning of the two mtDNA lineages, I conclude that this geographic distribution of the mitochondrial lineages is most consistent with the hypothesis that the Arctic Char that re-colonized the Canadian Arctic Archipelago and coastal mainland survived in an icefree area of the Archipelago itself. Note that this does not preclude the survival of some populations of Arctic Char in northern Beringia (a likely possibility), but rather that the lineage currently residing in the Canadian Arctic did not originate from there. The second line of evidence that supports the hypothesis of a small high Arctic refugium is the low genetic diversity observed in the Arctic lineage. This again is in contrast with the results of Brunner et al. (2001), who identified far more genetic diversity in the Arctic Archipelago, but did not recover the haplotype, ARC19, that was most common in my study. My data, however, are consistent with the results of Alekseyev et al. (2009) and Power et al. (2009), who also found that this haplotype predominated in the Canadian Arctic. Estimates of nucleotide and haplotype diversity were an order of magnitude lower in the Arctic lineage than in the Bering lineage. This is consistent with the hypothesis that the Arctic lineage would have survived as a small population in the small ice-free areas of the Arctic Archipelago throughout the LGM. The Bering lineage, however, probably survived as a much larger population given the extensive ice-free areas of Beringia in northeastern Asia and northwestern Yukon and central Alaska of North America. As the ice receded, the presumably small group of fish  49  from the Arctic refugium could have recolonized all areas of the Arctic Archipelago, leading to a sudden population expansion. Evidence for such a demographic expansion comes from the star-shaped phylogeny of the Arctic lineage, with one common haplotype (ARC19, shared by 94% of the individuals sequenced) found in high frequency in all populations, and numerous rare, geographically restricted haplotypes, thus fitting theoretical expectations (Slatkin & Hudson 1991; Excoffier et al. 2009). The results of the historical demographic analyses were also consistent with this hypothesis. In the Arctic lineage, negative values of the statistics Tajima’s D (Tajima 1989) and Fu’s FS (Fu 1997) were consistent with a recent population expansion (Excoffier et al. 2009), while values for the Bering lineage were not consistent with a population expansion. These statistics, however, are also expected to be negative if positive selection acted (Hedrick 2004), and this alternative hypothesis cannot be discounted with this analysis alone. The mismatch distribution analysis is also consistent with the hypothesis of a recent population expansion in the Arctic lineage - but not in the Bering lineage. The shapes of mismatch distributions of populations that have undergone a very recent demographic expansion and those that have had a constant population size, however, are very similar (Rogers & Harpending 1992; Excoffier et al. 2009) and both fit my data well. Unfortunately, the information content of the mtDNA dataset did not allow for more sophisticated analyses that may have rejected alternative hypotheses. Indeed, I attempted to perform a Bayesian skyline plot analysis in program BEAST (Drummond & Rambaut 2007), but the program was not able to converge despite using a large number (5*109) of MCMC repetitions. Similarly, the fact that the mtDNA lineages were completely sorted rendered approaches such as nested clade analysis (Templeton 2008) or statistical phylogeography using coalescent simulations (Knowles & Maddison 2002) unable to distinguish amongst alternative hypotheses. In summary, three different lines of evidence (star-shaped phylogeny, Tajima’s D and Fu’s FS statistics, and mismatch distribution analysis) were consistent with the hypothesis of a sudden expansion from an Arctic refugium by Arctic Char, but I was unable to exclude alternative scenarios completely. One major alternative explanation for the lack of variation in mtDNA is that it reflects positive selection for a mitochondrially-encoded gene (or genes) linked to the control region haplotypes used to define the Arctic lineage. Selection on mitochondrial  50  genes has been demonstrated in many taxa and can severely bias historical interpretations based on the assumption of neutrality (Ballard & Rand 2005). Furthermore, Arctic Char mitochondrial DNA has been shown to introgress to fixation (mitochondrial capture) into several Lake Trout (S. namaycush) and Brook Trout (S. fontinalis) populations (Bernatchez et al. 1995; Glémet et al. 1998; Wilson & Bernatchez 1998), perhaps suggesting selection on mitochondrial genes (Doiron et al. 2002). Data from nine microsatellite markers, however, are generally consistent with the mtDNA in showing one major genetic group of Arctic Char within the Canadian Arctic Archipelago. First, the neighbor-joining tree showed that within the Arctic group, internal branches are short and not well supported by bootstrap values. Second, the STRUCTURE analysis showed that, apart from populations that showed evidence for introgression from Atlantic lineage Arctic Char, most populations in the Canadian Arctic clustered together. One exception was the two Yukon populations, which formed their own cluster. Those populations, however, have dramatically lower genetic diversity than any other populations, and since STRUCTURE identifies genetic cluster by maximizing Hardy-Weinberg and linkage equilibria (Pritchard et al. 2000), it is not surprising that they would have clustered away from other populations. It seems unlikely that these Yukon populations survived in a refugium other than an Arctic refugium given that the Yukon samples are most related to populations from within the Arctic in the neighbor-joining tree. If the Yukon populations survived in a different refugium, we may expect them to share a more recent common ancestor with the lineages from the other refugium (i.e., the Alaska and Labrador populations). In summary, the microsatellite data are largely consistent with the mtDNA data. Because parallel evolution of multiple nuclear and mitochondrial markers in response to selection is unlikely, the most parsimonious explanation would then be that the low amount of variation in mtDNA reflects demographic history, not a selective sweep. Another possibility is that low haplotype diversity is the result of allele ‘surfing’ at the expanding edge of the post-glacial range expansion (Excoffier et al. 2009). It has been shown theoretically (Edmonds et al. 2004) and empirically (Excoffier & Ray 2008) that neutral alleles can increase in frequency away from their location of origin in rapidly expanding populations as a result of the two-dimensional nature of the spatial population  51  expansion (Excoffier et al. 2009). In some cases, allele surfing is expected to lead to patterns similar to that generated by vicariant events, because it will lead to the fixation of alternative alleles in different parts of the range (Excoffier et al. 2009). While allele surfing could in part explain the very low diversity of haplotypes observed in the present study, it is unlikely to explain the presence of the two lineages. Indeed, the Bering and Arctic lineages are highly divergent, and surfing of a single Beringian allele at the expanding edge of a demographic expansion from Beringia would have more likely led to the near fixation of a haplotype that would have been similar to other Beringian haplotypes. In the absence of fossil evidence, and because of the lack of resolution in the molecular markers, it is impossible to conclude with absolute confidence that the Arctic Char currently distributed in the Canadian Arctic Archipelago originated from only a high Arctic refugium. For instance, it is possible that the Arctic lineage survived the last glaciation in Beringia, and that mtDNA introgression from Dolly Varden occurred after deglaciation and replaced ‘Arctic’ haplotypes in the region. This hypothesis, however, implies that two highly divergent lineages survived in sympatry for several thousands of years without gene flow. This would require the presence of isolating barriers between the two lineages. Because most of the Beringian lineage samples from North Alaska are, based on morphological characters, actually Dolly Varden, it is a possibility that reproductive barriers between those two species maintained isolation between the two lineages. I consider this scenario to be unlikely because in areas of Alaska where Arctic Char and Dolly Varden occur in sympatry they share Beringian mtDNA, suggesting that isolating barriers are or were imperfect (Taylor et al. 2008). Furthermore, even if the two lineages had evolved in sympatry in Beringia, this hypothesis would not be consistent with the very low levels of genetic diversity observed in the Arctic lineage, because we would expect the Beringian populations to have been sizable given the geographical extent of those ice-free areas. Nevertheless, care should be taken when interpreting the mtDNA lineages as reflecting the recent evolutionary history of Arctic Char and Dolly Varden since it cannot address the issue of species delimitation and/or hybridization. The present study, therefore, can only address the history of the mtDNA lineages. In summary, based on all of the above lines of evidence, I conclude that the Arctic Char lineage  52  currently residing in the Arctic Archipelago and coastal mainland recolonized from a high Arctic refugium in the Archipelago itself, while the Arctic Char (and Dolly Varden) currently residing in Beringia recolonized from Beringia. !"<"! %*:243&95,:240&:0,B601,01*,/0)&406:,)64*&;*, Many studies have shown that genetic diversity declines as distance from putative glacial refugia increases (e.g. Turgeon & Bernatchez 2001; Costello et al. 2003; Stamford & Taylor 2004; Harris & Taylor 2010; Shafer et al. 2011) and such patterns can be useful to infer the location of refugia. In my study, both putative glacial refugia (ice-free areas of the Archipelago and Beringia) are in the western Arctic, and a gradual decrease in genetic diversity may have been expected to occur from west to east. Instead, the only evidence for such a pattern was a weak and non-significant trend in decreasing genetic diversity in landlocked populations. This lack of a decline of genetic diversity with distance from putative refugia could be explained, at least in part, by secondary contact and hybridization with Arctic Char that survived in an Atlantic refugium (Brunner et al. 2001). Indeed, populations from the eastern Arctic contained many unique alleles not shared with western populations, a pattern contrary to expectations given the locations of the most likely refugia in the western Arctic. These unique alleles were shared with samples obtained from the Labrador coast, and a STRUCTURE analysis confirmed that anadromous populations from the eastern Arctic shared a portion of their nuclear genome with ‘Atlantic lineage’ Arctic Char. Hybridization between the two lineages contributed to increased genetic diversity in areas far from the putative glacial refugium, and this was especially the case in anadromous populations. This penetration of Atlantic alleles could either have occurred in the past or be the result of ongoing gene flow. The Hudson Bay Basin where Atlantic alleles are common was recolonized by one of two routes, both allowing for contact with Atlantic lineage Arctic Char. The most likely recolonization route through the North was through Hudson Strait on the east because the Frozen Strait in the west probably remained blocked by ice for longer (Dyke et al. 2003). Such a route would have put recolonizing Arctic lineage char in close proximity to Atlantic lineage fish from the Labrador Coast. There is also a possibility that Atlantic alleles penetrated the Hudson Bay Basin via freshwater routes  53  offered by the presence of large pro-glacial lakes connecting the Atlantic Coast to the Hudson Bay Basin (Crossman & McAllister 1986). Finally, the very high prevalence of Atlantic alleles in the Cumberland Sound region of southeastern Baffin Island, the sampled region of greatest geographical proximity to Labrador, suggests the possibility for ongoing gene flow. Evidence from a tagging study suggests that char can move up to 550 km from their natal river, although such long-distance dispersal is rare (Gyselman 1984). The distance between the Cumberland Sound and the Labrador coast, however, is > 600 km and involves crossing the Hudson Strait. Acoustic telemetry work undertaken in Frobisher Bay (the bay directly south of Cumberland Sound) shows that Arctic Char rarely venture far from coastal habitats (Spares et al. 2010), suggesting the crossing of the > 100 km of open waters of the Hudson Strait is a rare event. Still, it is not impossible for Arctic Char to cover this distance, and stepping-stone patterns of gene flow along the coast may maintain the presence of the Atlantic alleles in the more distant Cumberland Sound populations. Hybridization between the two glacial lineages left no trace in the mitochondrial DNA, and Arctic populations with ‘Atlantic’ microsatellite alleles (e.g., south east Baffin Island populations) remained fixed for the Arctic lineage mtDNA. This constitutes an example of an asymmetric cyto-nuclear or mito-nuclear discordance (Wirtz 1999; Toews & Brelsford 2010). Such patterns can result from a variety of processes (Toews & Brelsford 2012), most importantly sex-biased dispersal (Wirtz 1999; Petit & Excoffier 2009) or adaptive introgression of mitochondrially encoded genes (Ballard & Rand 2005). It is uncertain which of these processes is more likely to explain the patterns observed in Arctic Char. Male-biased dispersal has been documented in salmonids (e.g., Hutchings & Gerber 2002), but my own work on Cumberland Sound Arctic Char found no evidence of a sex-bias in dispersal propensity (Chapter 3). As mentioned earlier, mtDNA introgression is common in the genus Salvelinus (e.g., Glémet et al. 1998; Redenbach & Taylor 2002), and selection on mitochondrial genes has been suggested as an important mechanism for this repeated trend (Doiron et al. 2002). Evidence for functional differences between mitochondrial types that could explain patterns of introgression, however, remains elusive (Blier et al. 2006).  54  !"<"# C1*,*@@*:0+,2@,&4&392'5,24,;*4*06:,+09A:0A9*,&43,6+2)&0624,D5,36+0&4:*, The relationship between dispersal abilities and population structure is well known, and many studies have shown that population structure tends to be more pronounced in species that have lower dispersal abilities (Bohonak 1999; DeWoody & Avise 2000). For example, studies that compare the population structure of marine animals with and without pelagic larvae often find that population structure is strongest in the latter (e.g., Kyle & Boulding 2000; Breton et al. 2003). The genetic effects of life-history variation in dispersal propensity within species, however, have received less attention. Fish species that are found as both anadromous and non-anadromous populations offer an interesting opportunity to test for the intraspecific effects of dispersal propensity on population structure. My study is one of a relatively small number of studies that have directly compared population structure and genetic diversity among anadromous and nonanadromous populations of fish (for other examples, see Mäkinen et al. 2006 and Tonteri et al. 2007) As expected, I found that landlocked populations of Arctic Char had lower amounts of genetic diversity within populations, and were more genetically differentiated from each other, than anadromous populations. Lower genetic diversity within landlocked populations is most likely explained by post-glacial genetic drift that would have been stronger in small, isolated landlocked populations than in anadromous populations connected by gene flow. Indeed, theory predicts that subdivided populations should have reduced effective population size, and thus experience increased drift, compared to populations exchanging genes freely (e.g., Whitlock & Barton 1997). Evidence for increased population differentiation among landlocked populations is simply the result of the well-known inverse relationship between FST and the effective number of migrants (Nem) exchanged by subpopulations (Hedrick 2004). I also found evidence that dispersal potential shaped patterns of isolation by distance (IBD). A partial Mantel test determined that life-history characteristics had an effect on population differentiation even after accounting for geographical distance, such that the intercept of the IBD relationship between landlocked populations was greater  55  than that between anadromous populations. In a two dimensional stepping-stone model, theory shows that the IBD relationship should be approximately linear, with a slope of 1/4N!%2 and an intercept of –ln(%) + &e – ln(2) + 2!A2 (Rousset 1997), where N is the population size, %2 is the variance in parent-offspring distance, &e is Euler’s constant, and A2 is a constant related to features of the dispersal distribution. Because the slope of the IBD is inversely proportional to dispersal (%2), we may have expected a stronger slope among landlocked populations. The slope is indeed slightly higher among landlocked populations, but not significantly so. The interpretation of the differences in intercept, however, is less straightforward because biological interpretations of constant A2 are difficult (Rousset 1997). Still, the value of the intercept is also inversely related to the dispersal parameter %. Furthermore, because average genetic differentiation between populations not exchanging alleles should be higher, it makes intuitive sense that the intercept of the IBD for landlocked populations should be higher. These theoretical predictions, however, assume a stepping-stone model of gene flow where migration between demes is non-zero; when there is no dispersal (i.e. %2 = 0) the slope of the relationship should be zero because each isolated population would accumulate random differences through drift (Rousset 1997). Assuming that landlocked population have no potential for gene flow, why then are we still observing significant IBD? A likely explanation would be that equilibrium in this system has not been reached following the population expansion. In this case, it may be useful to interpret the IBD pattern observed in Arctic char not according to a stepping-stone model of gene flow at equilibrium, but instead according to Good’s non-equilibrium model of unidirectional stepwise population expansion (Slatkin 1993). Using this model, Slatkin (1993) showed that IBD patterns would be evident in non-equilibrium populations by virtue of the colonization process, and any subsequent gene flow would only obscure the relationship. Because the expansion is recent in Arctic char (i.e., within the last 10,000 years), and because no contemporary gene flow occurs between landlocked populations, the signatures of this initial colonization are still apparent. We would then expect the pattern of IBD to gradually decrease with time as allele frequencies in the newly isolated populations evolve through random drift. This interpretation is also consistent with evidence for nonequilibrium processes influencing IBD patterns in North American anadromous Brook 56  Trout (Castric & Bernatchez 2003), a close relative of Arctic Char. Because Arctic Char typically has a longer generation time than Brook Trout (Johnson 1980), and because its current range was covered by ice more recently than that of Brook Trout (Dyke et al. 2003), it is no surprise that the signature of non-equilibrium processes would be at least as strong in Arctic Char as in Brook Trout. !"<"7 E24:)A+624+,&43,6'()6:&0624+,@29,&3&(0&0624,02,&,:1&4;64;,/9:06:, My data contributes to a growing number of studies that show the importance of ‘cryptic’ or ‘microrefugia’ for Northern taxa (Provan & Bennett 2008; Stewart et al. 2010), and provides the first phylogeographic evidence that the Arctic Archipelago was used as a refugium by fish during the LGM. In addition to cryptic refugia, I showed that gene flow, either historical or contemporary, can further complicate patterns of genetic variation, and thus needs to be taken into account when making historical inferences. These results also have implications for conservation and management of Arctic Char. Indeed, the Arctic is expected to be among the regions most affected by climate change, and a few authors have pointed out that understanding how species responded to past climatic change may help us understand their response to future climate change (Davis & Shaw 2001; Provan & Bennett 2008; Williams et al. 2008). The supply of genetic variation through gene flow could also help anadromous Arctic Char adapt to a changing environment (Bell & Gonzalez 2011). Lower genetic diversity and the limited gene flow in landlocked populations, however, may render these populations more vulnerable to the effects of climate change. Furthermore, populations of Arctic Char containing alleles from the Atlantic lineage may be at an advantage when adapting to rapid environmental change because of increased standing genetic variation (Reusch & Wood 2007). The fact that this variation comes from populations that are presumably adapted to a more southerly climate reinforces this possibility. For instance, it will be interesting to determine if some alleles preferentially introgress from the Atlantic to the Arctic lineage, which would provide evidence of selective advantage of some portions of the genome (Nosil et al. 2009). Future genomic studies in char may thus shed some light on the capacity for adaptation to change in those populations, and will benefit from the framework for understanding genetic variation that I established in the present study.  57  Table 2.1. Results of the historical demography analyses using neutrality tests (Fu’s Fs and Tajima’s D) and mismatch distribution analysis on mitochondrial DNA sequence variation (D-loop) in Arctic Char, Salvelinus alpinus. !  "#$%&'()%*!%#+%+!  -%4"5!  ,)+-'%./!0)+%&)1$%)23!'3'(*+)+!  !"#$%!&%  "#'()"*%  +(,-.(#$%$%  "#'()"*%  ! /001%234  "5%/001%234%  "6/001%234%  "#'()"*%/&&74%  8(99*:;*$$%  "#'()"*%/8<=4%  5&.%).!  >?@AB0%  C%5A55D%  >5A65?%  C%5A55D%  EA5%/5AF>EAD4%  5%/5>5A?4%  5A6%/5>G4%  5A?B%  5AH%  5AHD%  8#&)37!  >6A@F%  5A?@%  >5A5F6%  5ADH%  ?A6%/5A5>FAE4%  5%/5>6AH4%  0A?%/6A@>G4%  5A5?0%  5A?H%  C%5A55D%  6)3#'7#!  Table 2.2. Differences in genetic diversity (allelic richness and expected heterozygocsity) and average pairwise FST between landlocked and anadromous populations of Arctic Char, Salvelinus alpinus, assayed at nine microsatellite DNA loci. %%  %  3;I)":-;9%J"KL;%$(.M)*$%  NOI)":-;9%J"KL;%$(.M)*$%  %  <;(:PL.L"$%  Q(;:)LIK*:%  ">'()"*%  Q(;:)LIK*:%  ">'()"*%  <))*)-I%8-IR;*$$%  6AHEB%  6A@E6%  5A5556%  6A@H@%  5A5F%  %$%  5AHEB%  5A@6E%  5A555?%  5A@B5%  5A?5%  !$S%  5A5FE%  5A6B@%  5A55?5%  5A66@%  5A?F%  58  Figure 2.1. Map of the Northern Hemisphere showing the distribution and names of the five major mitochondrial DNA lineages of Arctic Char (Salvelinus alpinus) identified in the Holarctic phylogeography study of Brunner et al. (2001). Dashed lines show the approximate extent of the three major ice-sheets that covered North America at the last glacial maximum (Dyke et al. 2003). The locations of some of the major islands of the Canadian Arctic Archipelago discussed in the text are also shown.  59  Figure 2.2. Maximum likelihood phylogenetic tree of Arctic Char (Salvelinus alpinus) mtDNA haplotypes generated using GARLI (Zwickl 2006). Bootstrap support (1,000 replicates) of more than 50% are shown in italics. Asterisks denote previously unpublished haplotypes. All other haplotypes were identified in previous studies.  Bering  =?7$$ 51.1 A-BC(D =?7$: !"#$%"&'() =?7$) A-BC(# -')%"&'( =?7$6 =?7$@9 =?7$E9  85.2  *+,-.-/0123 *+CI10(.-, 0.0080  Figure 2  60  Atlantic-SiberiaAcadia  *<=)6 *<=); *<=8 *<=)> *<=@ 4"5$) 53.6 4"5$6 4"5$> 59.9 63.34"58 4"5J 4#K;  Arctic  47#)89 47#);9 50.8 47#:69 47#):9 47#))9 47#)@9 47#)E9 47#$; 76.2 47#)69 47#:$9 47#)$9 47#)J9 47#)>9  *+F(,GH,-CH2  Figure 2.3 (Next page) Map showing the geographical distribution of the two mtDNA lineages identified in North American Arctic Char (Salvelinus alpinus) and Dolly Varden (S. malma). Each point is a sampling location. The colours denote the haplotype composition of the samples: blue if only Arctic lineage haplotypes were present and red if only Bering haplotype were present. Circles with a black outline represent locations where Arctic Char were sampled and circles with a white outline represent locations where Dolly Varden were sampled. The sizes of the circles are proportional to the sample size (square-root corrected). The inset shows the north slope of Yukon-Alaska (outlined with black rectangle) where the two lineages overlap. The asterisk denotes the two locations (they are too close to appear distinct at this scale) where Arctic and Bering lineages co-existed. Note that the sizes of the circles in the inset are not drawn to the same scale as in the full map (the smallest circles are N = 1). A simplified haplotype network is drawn over the sampling locations showing the number of mutations separating the two lineages.  61  62  Figure 2.4. Results of the pairwise mismatch analysis done on the mtDNA data for two lineages of Arctic Char (Salvelinus alpinus). The grey bars represent the empirically determined distribution of pairwise sequence differences in the Arctic (a) and Bering (b) lineages. The dashed lines represent theoretical expectations for a model of sudden population expansion (Rogers & Harpending 2002).  63  Figure 2.5. Neighbour-joining tree of Cavalli-Sforza’s chord measure genetic distance from nine microsatellite loci assayed in samples of Arctic Char (Salvelinus alpinus) Bootstrap values higher than 50% are shown (based on 1000 bootstrap replicates). Samples from Alaska, Labrador, and the Yukon are shown. All other samples are from the Canadian Arctic Archipelago and coastal mainland.  64  Figure 2.6. Variation in expected heterozygosity as a function of waterway distance from a putative glacial refugium (Banks Island) in samples of Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci. Landlocked populations are shown in black, and anadromous populations in grey.  65  Figure 2.7. Results of the STRUCTURE  Individual assignment probability (q)  analysis based on the most likely (sensu cluster: K = 4 in samples of Arctic Char  SL  (Salvelinus alpinus) assayed at nine  TL Lk103  bar represents the probability of Lk104  assignment of each individual to the four  Rad  genetic clusters. The location code for each sampling location is shown on the  Mid  Banks Is.  left, along with regional groupings to help interpretation. The sampling locations are arranged approximately  on the STRUCTURE results are shown on the right. Asterisks denote landlocked populations.  Kent Pen.  Nauy  Cornwallis Is.  Res  Boothia Pen. Melville Pen. SW Baffin Is.  Ami Lor Tou Arr Bec Pri Gri Hal Gif Rav  Ellesmere Is.  LakeB Mur Mus  * * *  Alex  Arctic Archipelago and coastal mainland  scale regional groupings based roughly  Sac Tho Kuu Kag Nal Kuuj  * **  from west (top) to east (bottom). Larger  Victoria Is.  Cap  Yukon  microsatellite DNA loci. Each coloured  Alaska  CL  * * * * * ** *  Evanno et al. 2005) number of genetic  Kin  SE Baffin Is.  Iqa Isu Nalu Kas Pet IB  66  Labrador  Hudson Bay  Figure 2.8. Relationship between pairwise genetic distance measured as linearized FST (Rousset 1997) and geographical distance (km) in samples of Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci. Landlocked (black) and anadromous (white) populations are analyzed separately. Significant (Mantel tests) isolation by distance relationships are shown by a full line for landlocked and a dashed line for anadromous populations. Note that the analyses were all performed on log10-transformed distances (according to Rousset 1997) but actual distances are shown here for improved visualization. 0.25 Anadromous Landlocked  Genetic distance (Fst/(1-Fst))  0.20  0.15  0.10  0.05  0.00 0  500  1000  1500 Distance (km)  Figure 8 67  2000  2500  3000  ! "#$%&'($)*+,-%.$(/$$'%0&1*$)1,+2%3$'$%4+5/2%,'0%*5($'(&,+%45)%+56,+% ,0,*(,(&5'%&'%7,44&'%81+,'0%,',0)595:1%;)6(&6%<#,)%% !=> ;.1(),6(% Dispersal can influence the process of local adaptation within populations, particularly when the dispersers successfully breed in the non-natal habitat. Anadromous Arctic Char (Salvelinus alpinus) display a complex migratory behaviour that makes the distinction between breeding and non-breeding dispersal especially important. This species does not reproduce every year, but individuals must migrate to fresh water each fall to over-winter, such that a large proportion of fish migrating up-river are not in breeding condition and do not contribute to gene flow when overwintering. I used two genetic assignment approaches to identify dispersers between populations of Arctic Char from Baffin Island, Canada. I found that a large proportion of Arctic Char assigned as dispersers were not in reproductive condition and therefore did not contribute to gene flow between localities. Furthermore, we found evidence that non-breeding individuals were more likely to use non-natal habitats than breeding individuals. Other biological traits (sex, age, fork length, weight, gonad weight, and condition factor) were not good predictors of dispersal propensity. Finally, I generated estimates of gene flow to parameterize a population genetic model of the balance between migration, selection, and drift, which showed that gene flow among localities is probably low enough to allow for local adaptation among populations, given that selection is sufficiently heterogeneous amongst these same localities. The greater propensity of non-breeding individuals to disperse has interesting implications for the study of the evolution of dispersal rates in salmonids, and suggests that inbreeding avoidance is not an important force favoring dispersal in Arctic Char.  68  !=? 8'()50:6(&5'% Dispersal has important ecological and evolutionary consequences for the fitness of individuals and populations (Clobert et al. 2001; Hanski & Gaggiotti 2004). Ecologically, immigrants can have positive demographic consequences by increasing population size, thereby either buffering the effects of demographic stochasticity or maintaining demographic sinks (Pulliam 1988). Evolutionarily, dispersal allows the distribution of potentially beneficial alleles among populations (Holt & Gomulkiewicz 1997), but may also introduce locally deleterious alleles that can constrain local adaptation (Lenormand 2002; Garant et al. 2007). The evolutionary consequences of dispersal, however, are mainly realized when dispersers successfully reproduce in the non-natal habitat, i.e. if dispersal results in gene flow (Garant et al. 2007). Many dispersal events, however, do not result in gene flow (Ehrlich & Raven 1969), and distinguishing between dispersal events that result in gene flow from those that do not is therefore critical to understand its effects on the distribution of genetic variation across populations and on the evolution of local adaptation (Slatkin 1987; Garant et al. 1997). There are several processes that can cause an apparent discrepancy between rates of dispersal and rates of gene flow, and examples of such disconnect are numerous (e.g. Mallet 1986; Tallman & Healey 1994; Mallet 1986; Via 1999; Nosil 2004; Dionne et al. 2008). For instance, dispersers may have lower reproductive success due to natural or sexual selection against migrants (Nosil et al. 2005). Individuals can also move for purposes other than breeding (Bowler & Benton 2005), which can make the perceived rate of dispersal higher than gene flow. Knowledge of the causes of dispersal, therefore, can often help understand its consequences (Clobert et al. 2004). A great deal of progress in our understanding of the causes and consequences of dispersal comes from theoretical models (Ronce 2007). Such models often require simplifying assumptions to be made, most notably regarding the nature of dispersers, which are typically assumed to be a random subset of the population from which they emigrate (Bowler & Benton 2005; Ronce, 2007; Travis et al. 2012). This, however, is rarely the case in natural populations, and there is accumulating evidence showing that dispersers are not a random subset of the population they are emigrating from, but instead have specific traits that increase their propensity to disperse or to settle in a specific habitat (Ims & Hjermann 2001; Bowler & Benton 2005; Clobert et al. 2009). Such condition-dependent dispersal can considerably influence the consequences of dispersal for local adaptation (e.g., Garant et al. 2005; Bolnick et al. 2009). For instance, dispersal is often 69  sex-biased, probably because the costs and benefits of dispersal often differ between sexes (Perrin & Mazalov 2000), and such sex biases can influence adaptation (Kawecki 2003). Body size or condition has also often been identified as a correlate of dispersal propensity (Ims & Hjermann 2001; Clobert et al. 2009), and examples where dispersers are larger or smaller than residents both exist (see Table 1 in Clobert et al. 2009). Because these traits are often the target of sexual selection, differences in body size between local individuals and immigrants could have an impact on the reproductive success of the dispersers (Nosil et al. 2005). An understanding of a species’ dispersal ecology is therefore necessary to make predictions regarding the consequences of dispersal (Clobert et al. 2004; Garant et al. 2007; Travis et al. 2012). Anadromous salmonid fishes have provided important insights into our understanding of dispersal (Hendry et al. 2004). Anadromy refers to the behaviour wherein individuals hatch and grow as juveniles in fresh water, but forage as adults in the marine environment before returning to fresh water to reproduce (McDowall 1997). Anadromous salmonids have evolved a complex series of behavioural and physiological attributes that allow them to home to their natal freshwater habitats, often following extensive marine migrations (Hendry et al. 2004; Quinn 2005). This suggests strong selection for reduced dispersal rates, although the precise nature of the selective pressures favoring low dispersal is still debated (Hendry et al. 2004). The homing strategy of salmonids typically leads to low gene flow among populations, and is largely responsible for the high levels of genetic differentiation and local adaptation commonly observed among populations even over small spatial scales (Taylor 1991; Fraser et al. 2011). Despite the general homing abilities of salmonids, a component of many populations will disperse from their non-natal rivers, a phenomenon referred to as ‘straying’ (Quinn 1993). Straying is an important behaviour because it allows the colonization of new habitats, and may constitute a bet-hedging strategy used in response to unpredictable habitats (Hendry et al. 2004). Those dispersers have also been found to be a non-random sample of the populations they are emigrating from: sexbiased dispersal appears common (e.g., Hutchings & Gerber 2002; Bekkevold et al. 2004) and some studies report evidence that older individuals are more likely to stray (Quinn 1993). The Arctic Char (Salvelinus alpinus) is a salmonid with a Holarctic distribution, and many populations at higher latitudes are anadromous (Johnson 1980). Like most other anadromous salmonids, Arctic Char have a tendency to home to their natal habitats (Johnson 1980; Hendry et al. 2004). Tagging studies in the Canadian Arctic, however, suggest that  70  dispersal (or straying) is fairly high in Arctic Char populations compared to other salmonids (Gyselman 1994). The extent to which this dispersal results in gene flow, however, remains unknown. In addition, some unusual aspects of the anadromous Arctic Char life cycle make distinguishing between dispersal and gene flow particularly important for this species. Many anadromous salmonids spend several years at sea before returning to fresh water to spawn (Fleming 1998; Quinn 2005). Arctic Char, however, are constrained to return to fresh water yearly because they have relatively low salinity tolerance and are not able to survive the sub-zero temperatures of the Arctic Ocean in winter (Johnson 1980). This yearly energy expenditure associated with the migration, combined with a short feeding season (~30-100 days; Johnson 1980; Dempson & Kristofferson 1987), results in slow growth and little energy to invest in gonadal development (Dutil 1986). Consequently, Arctic Char typically do not reproduce every year, and a substantial proportion of returning adults are migrating solely for the purpose of overwintering (Dutil 1986). As such, these individuals have no potential to introduce their alleles in non-natal habitats. Furthermore, patterns of tag recovery from movement studies suggest that the individuals migrating for the purpose of over-wintering may have a greater propensity to use non-natal habitats in comparison to individuals migrating for reproduction (Dempson & Kristofferson 1987; Gyselman 1994), although direct evidence for this behaviour is still lacking. Together, these observations suggest that a substantial proportion of dispersal events will not result in gene flow in anadromous Arctic Char. It is therefore necessary to distinguish between two types of dispersal in Arctic Char: breeding dispersal, where a mature individual enters a non-natal habitat for the purpose of reproduction; and over-wintering dispersal, where a non-breeding individual uses a non-natal habitat to over-winter. Note that this latter form of movement does not represent dispersal according to all definitions (see Chapter 1). According to some definitions, dispersal simply refers to the movement of individuals from one habitat patch to another (e.g. Bowler & Benton 2005; Garant et al. 2007), while other definitions restrict the term to movement of individuals for the purpose of breeding (whether breeding is successful or not; Howard 1960; Clobert et al. 2001; Ronce 2007). Regardless of the definition used, dispersal is always understood as a threestep process: (1) decision to leave the natal habitat (emigration); (2) inter-patch movement; and (3) settlement or immigration (Bowler & Benton 2005). Because over-wintering dispersal also  71  involves those three steps commonly associated with dispersal, I argue that the former definition is more appropriate to the current system, and use it throughout the present study. To better understand the migratory behaviour of anadromous Arctic Char and the interplay between dispersal and gene flow in determining, in part, the potential for local adaptation in this species, I conducted a microsatellite DNA survey of char from the Cumberland Sound region of Baffin Island, in the eastern Canadian Arctic. The sampling scheme included samples of adult fish collected as they enter fresh water during the fall migration, and samples of juvenile fish still rearing in freshwater habitats that have therefore not had an opportunity to disperse yet (i.e., they have not left their natal lakes). It was designed to test three hypotheses. First, I hypothesized that many dispersal events do not result in gene flow, and predicted that under this hypothesis average genetic differentiation (i.e., as measured by FST) between samples of juveniles would be higher than average genetic differentiation between samples of adults from the same system, because many of these adults represent non-breeding, over-wintering adults. Second, I tested the hypothesis that over-wintering adults are more likely to use non-natal habitats than breeding fish. To do so, I used genetic assignment tests to provide an estimate of total dispersal (i.e., breeding and over-wintering dispersal), which I then partitioned into breeding dispersal and over-wintering dispersal using information on the reproductive status of the genetically assigned fish. Third, using the same genetic assignment procedure, I tested the hypothesis that dispersers not only differ from philopatric individuals in their reproductive status, but also in other traits such as sex, age, and body size. In addition to testing these three hypotheses relating to the patterns of dispersal, I explored the potential consequences of gene flow for local adaptation in this system by parameterizing a population genetic model of the balance between migration, drift and selection with estimates of gene flow generated from the microsatellite data. By describing patterns of dispersal, and by evaluating their potential consequences, the data in this study may help shed light on the evolutionary forces promoting dispersal and philopatry in Arctic Char.  72  !"! #$%&'($)*+$,-+.&%/0-*+ !"!"# $%&'()*+,-'%./) Arctic Char were sampled from Cumberland Sound, in the southeastern part of Baffin Island, Nunavut, in Canada’s eastern Arctic (Fig. 3.1; Table 1). Adult fish were collected from a total of 10 localities between 2003 and 2009, using gill nets (140 mm stretched mesh and multi-mesh 38102 mm) set in saltwater close to the mouth of the rivers that drain the lakes in which Arctic Char spawn. Sampling was conducted from mid-July to September when fish aggregate close to those rivers prior to undergoing upstream migrations. For a few localities (QAS and IQ1), collections were made in fresh water shortly after the fish had migrated upstream, because sampling was initiated after fish had already moved into fresh water. At two localities, (KIN and NAU), fish were also collected in the winter. The following measurements were taken in the field: fork length (mm), body weight (g), and gonad weight (g). Sex and reproductive status (immature, over-wintering, i.e., not spawning that year, or spawning) was determined based on the gonadosomatic index (GSI) of each individuals, assuming that GSI should be bimodal with two non-overlapping distributions corresponding to non-breeding and breeding individuals (see online Supplementary Materials for details). Fin clips were preserved in 95% ethanol for genetic analysis and sagittal otoliths were removed and the age of most individuals was later determined in the laboratory (see Loewen et al. (2009) for full description of aging methods). Juvenile samples were collected in July 2008 and August 2009. Methods of capture differed between localities, but seine net (10mm stretched mesh) tows were most commonly used. When seining was not possible, I used a combination of dip nets, minnow traps and electrofishing. Whenever possible, I sampled juveniles from different areas of the rearing lake to minimize the likelihood of sampling related individuals. I sacrificed juvenile fish with an overdose of MS222, measured fork length (mm), and preserved the whole fish or a fin clip in 95% ethanol. !"!"1 #(2'0*$%&))(%&+345+$,$)6*(*+ Individual genotypes were obtained at 18 microsatellite loci. Polymerase chain reaction (PCR) protocols and primer information can be found in the online Supplementary Materials (Table S1). The PCR products were run on an Applied Biosystems (Carlsbad, CA, USA) 3100 genetic analyzer. GeneMapper Software version 3.7 (Applied Biosystems, Carlsbad, CA, USA) was then  73  used to automatically score microsatellite alleles, and all scores were then manually checked for quality. Ninety-six individuals were randomly selected, re-amplified at all loci, and re-genotyped to assess scoring error rates associated with most aspects of the genotyping procedure (except the DNA extraction; DeWoody & Nason, 2006). MICRO-CHECKER version 2.2.3 (van Oosterhout et al. 2004) was used to test for the presence of null alleles and large allele drop-out. I used FSTAT version 2.9.3.2 (Goudet 2001) to test for Hardy-Weinberg equilibrium and genotypic disequilibrium using default values for the number of permutations. For both tests, I set the nominal significance level at 0.05 (using a Bonferroni correction for multiple comparisons). I used GENETIX version 4.05 (Belkhir et al. 1996) to generate estimates of observed heterozygosity (HO), and expected heterozygosity (HE) corrected for samples size. FSTAT was used to calculate allelic richness (AR) and pairwise FST between each sample, and the significance of differences between samples was assessed with 10,000 permutations (experiment-wide ! = 0.05 after Bonferroni correction). I used PHYLIP version 3.68 (Felsenstein 1993) to generate a Neighbor-joining tree of all samples based on Cavalli-Sforza’s chord measure (Cavalli-Sforza & Edwards 1967) employing 1,000 bootstrap replicates to test support for each node. The tree was visualized and formatted in FigTree v1.3.1 (Rambaut 2009).  !"!"! 01,-+2%*1.*)3&45&&.)+(6'4)+.()768&.%'&)*+,-'&*) )) Sampling juvenile salmonids for the purpose of characterizing population structure is potentially problematic because of the increased risk of sampling several individuals from the same family, a phenomenon referred to as the Allendorf-Phelps effect (Waples 1998). In addition to sampling juveniles from spatially discrete locations in the lakes (whenever logistically feasible), I used the software COLONY (Wang & Santure 2009) to identify likely siblings on the basis of individual multi-locus genotypes (see Appendix B Table B.4 for details). Once full-sibs were identified (“BestML” full-sibs only), all but one randomly selected individual from each full-sib group was removed. All analyses described below were then run on the datasets with and without the fullsibs removed. I used FSTAT to assess whether there was increased genetic differentiation among samples of juveniles compared to samples of adults. The permutation procedure was used with  74  10,000 permutations to test for significant differences between FST for adults and juveniles. For this analysis, I used only the juvenile samples from locations for which I had an analogous adult sample. This was done to avoid the possibility that a juvenile sample without an analogous adult sample could bias the results. This should make the test more conservative because it reduces sample size, and hence power. !"!"9 :&.&4%;)+**%/.,&.4*) I used the program GENECLASS2 (Piry et al. 2004) for assignment tests to identify dispersers in the samples of adults. Before conducting the analysis, I evaluated whether I had sufficient power to detect dispersal using the juvenile samples as a reference. To do so, I followed the guidelines provided by Paetkau et al. (2004) performing a run of GENECLASS2 with the juvenile samples only (i.e., self-assignment). The assignment scores from that run were used to compute values of DLR (mean genotype likelihood ratios) for pairs of samples. I did not look at all pairs of samples, but focused instead on those that were geographically proximate and that were identified by FST estimates as being genetically similar. For those samples, DLR was always greater than 5, which was found to provide near maximum power by Paetkau et al., (2004). For the assignment tests, juvenile samples were used as the reference samples, and I only assigned samples of adults that had corresponding juvenile samples. All analyses were also run using the juvenile samples with the full-sibs removed as reference samples. Samples were assigned or excluded using the Bayesian computation method of Rannala & Mountain (1997) and the Monte-Carlo resampling algorithm of Paetkau et al. (2004) to simulate 100,000 individuals with a 0.05 type I error rate. In order to avoid type-I errors (i.e., individuals identified as dispersers who are actually not), I followed Hauser et al. (2006) and used the assignment score of individuals (i.e., scorei,l = Li,l/ Lij where Li,l is the likelihood of individual i belonging to sample l) as calculated by GENECLASS2 to eliminate all individuals that had a score lower than 95%. As mentioned above, the GENECLASS2+analysis was also performed on the data set from which all full-sibs identified by COLONY were removed. I also complemented the previous approach with assignment tests performed in STRUCTURE (Pritchard et al., 2000), which allows the identification of genetic clusters (K) without defining populations a priori, and also allows us to distinguish hybrid individuals. I ran STRUCTURE on the entire data set (1,290 individuals) under the admixture model with  75  independent allele frequencies. I varied K from 1 to 15 and ran ten independent runs for each value of K, employing 100,000 burnin and 100,000 MCMC replicates per run. The results were first visualized using STRUCTURE HARVESTER (Earl 2011), which implements the "K method of Evanno et al. (2005) to infer the most likely number of clusters, and combines the results of many independent runs of the program for each K value. Individual admixture coefficients (Q) were calculated for each individual after the results of the ten independent runs for the most likely K were combined in program CLUMPP (Jakobsson & Rosenberg, 2007) using 1,000 permutations under the LargeKGreedy algorithm. Each individual in the analysis has a Qvalue for each of the identified genetic clusters, and the values of Q range from 0.0 to 1.0, with higher values meaning that a greater proportion of this individual’s genome assigns to a particular cluster. The sum of the Q values for any individuals across all genetic clusters adds to 1.0. I therefore eliminated putative hybrids from this analysis by removing individuals that did not have at least one Q value above 0.5. All other individuals were assumed to unambiguously assign to the genetic cluster for which they had the highest Q. Based on the results of the assignment tests, adults were classified as dispersers (i.e., if the individual was assigned to a river different from where it was captured) or as philopatric (i.e., if the individual was assigned to the river from where it was captured). Frequency histograms of the distance (coastal distances - km) between the site where a fish was captured and where it was assigned (i.e., dispersal distance) were plotted. I then tested for differences in frequencies of dispersing vs. philopatric individuals in the groups of fish classified as breeding fish (i.e., fish whose gonads were ripe when collected in the field) and overwintering fish (i.e., fish whose gonads were mature but resting or fish whose gonads were immature) using two-tailed Fisher exact tests. This procedure was repeated on the results of the GENECLASS2 analysis including all samples, that where fish with low confidence of assignment were removed, that from which putative full-sibs were removed, and considering the results of the analysis done in STRUCTURE. Two-tailed Fisher exact tests were also used to ask whether dispersal was sexbiased, and the analysis was again repeated on the results of all four previously mentioned assignment test procedures. To investigate the effect of other biological characteristics on dispersal propensity, I tested for differences in mean trait values between dispersing and philopatric individuals. More specifically, I looked at the effect of age (years), fork length (mm), total body weight (g), gonad  76  weight (g), and condition-factor (calculated as K = (W*105)/L3, where K is the condition factor, W is the total body weight (g), and L is the length (mm; Anderson & Neuman 1996). Only gonad weight departed from normality and was thus log10 transformed. Because many of these variables may be correlated with each other, I used discriminant function analysis to test whether the biological variables predicted membership to the ‘disperser’ or ‘philopatric’ groups using the MASS library in R (R Development Core Team 2010), and omitting any individual for which data were missing at one or more variables. Significance was assessed using Wilk’s lambda. Because some traits may show an association with dispersal propensity without differing in means (for example, Gyselman (1994) found that smaller and larger Arctic char were more likely to disperse than intermediate sized Arctic char) I also visually compared the distributions of trait values of dispersing and philopatric individuals (see Appendix B for details). This analysis was repeated for all four methods of genetic assignment. !"!"< =14&.4%+')>12)'1;+')+(+-4+4%1.)) I evaluated the potential for selection to drive divergence at the scale of the river for populations of Arctic Char distributed around Cumberland Sound. Although the approach advocated by Adkison (1995) has been used extensively for anadromous fishes (e.g., Hansen et al. 2002; McCairns & Bernatchez 2008), the circular geography of Cumberland Sound violates many of its assumptions. I opted for the more conservative approach of using a continent-island model, where we assume that each river receives migrants from all the other rivers (the ‘continent’). This simple approach probably poorly reflects the complexities of dispersal and gene flow in the present system. It is, however, appropriate to address the present question because it conservatively assumes that gene flow comes from all rivers, not only the nearby rivers with similar selective environments. The continent-island model I used is derived from the two-patch model of Yeaman & Otto (2011). According to this model, selection overcomes gene flow and drift if s, the selection coefficient, is larger than (4Nem + 1)/(4Nem + 2Ne + 1), where Ne is the effective population size and m is the rate of migration. I used the program MIGRATE-n ver. 3.2.15 (Beerli & Felsenstein 2001) to obtain estimates of m and Ne. I ran the analysis only on the juvenile samples since they should better reflect long-term gene flow than the adult samples, which may contain dispersers that will not contribute genetically to the populations they disperse to. Note that MIGRATE-n assumes that Ne and m remained constant over the last 4Ne generations,  77  an assumption probably violated in the anadromous char system whose range was only recently recolonized post-glacially (i.e., in the last 10,000 years). Recent population splitting, however, leads to inflated estimates of migration rates (P. Beerli, pers. comm.), which for the current application will lead to conservative estimates of the potential for selection to drive divergence. The mean values of ! and M obtained from MIGRATE (averaged over all populations) were used to parameterize the model and explore whether local adaptation is likely for biologically realistic values of parameter s (see Supplementary Materials for details of the search strategy). For diploid data, ! can be converted to Ne using the following equation: Ne = ! / 4µ, and M can be converted to m using m = M*µ, where µ is the mutation rate. Because there is no widely accepted estimate of mutation rate for microsatellites in salmonids (Steinberg et al., 2002), I varied the mutation rate between 10-3 and 10-5 for the parameter conversions. Total gene flow into each sampling location was calculated by summing the effective number of immigrants, Nem, from all other samples, and then dividing by the Ne of the local sample.  !=@ A$1:+(1% !"9"# ?%;21*+4&''%4&)-1'@,12-A%*,) I found high levels of polymorphisms at most of the 18 microsatellite loci examined and the number of alleles per locus varied between 1 (for Smm21) and 56 (average 20.4 alleles/locus). The results of the MICROCHECKER+analysis consistently identified three loci potentially suffering from null alleles: Sco109, Sco212 and Sco218. Those loci were thus eliminated (along with the monomorphic Smm21) from all subsequent analyses, leaving a total of 14 informative loci. The average scoring error rate over all loci (excluding the three aforementioned loci) was of 2.0% (see Appendix B for details and locus-specific error rates). FSTAT identified 35 locus/sampling location pairs that had significant (P < 0.05) heterozygote deficit and four that had significant heterozygote excess. Only two locus/sampling location pairs, however, remained significant after Bonferroni correction: locus Sco202 in sample KAN03ad, and Sco216 in KIP03ad, both with heterozygote deficits. FSTAT identified 37 pairs of loci that were in significant genotypic disequilibrium, but none remained significant after Bonferroni correction.  78  I found evidence of weak, but significant, genetic differentiation among sampling locations of Arctic Char distributed around Cumberland Sound. First, the Neighbor-joining tree showed that paired samples of juveniles and adults from the same location tended to group together (Fig. 3.2a). This was supported by bootstrap values of more than 80% for all groupings of samples for the same site (not shown). Second, pairwise FST values between localities tended to be small (global FST = 0.038) but significant (Bonferroni-corrected for experiment-wide " = 0.05; Table B.3 Appendix B). None of the samples collected from the same site, but in different years, differed significantly suggesting that temporal variation in population genetic structure is minimal and that comparisons of samples collected in different years is valid. Both ISU adult samples (2003-2004) and the IQA adult sample, however, differed significantly from their respective juvenile sample. The COLONY analysis identified pairs of full-sibs in the juvenile samples (Table B.4 in Appendix B). While most juvenile samples were found to contain from zero to three pairs of fullsibs, three localities (i.e., ISU, KEK, AUN) contained more than 10 pairs of full-sibs each. For all three localities where many full-sibs were identified, the juveniles were difficult to collect and were thus collected from only one short suitable stretch of shoreline. The samples of juveniles were more genetically differentiated from each other than samples of adults collected from the same localities (Fig. 3.2b). The average pairwise FST between samples of juveniles and adults was 0.045 (95% CI: 0.034-0.059), and 0.029 (95% CI: 0.019-0.041) respectively, and this difference was statistically significant (P = 0.036). The FST values remained virtually unchanged after removing putative sibs from the analysis (juveniles: FST = 0.044; adults: 0.029), and the difference remained statistically significant (P = 0.044). !"9"B C**%/.,&.4)4&*4*) Out of the 359 individuals assayed using GENECLASS2, a total of 192 individuals were assigned to a locality where they were captured, and were inferred to be philopatric individuals. The remaining 167 individuals were classified to a locality where they were not captured and thus were inferred to be dispersers, which resulted in an initial estimate of total dispersal of 46.5%. Individuals that were identified as dispersers were typically assigned to proximate localities, as indicated by the dispersal distance frequency distribution (Fig. 3.3a). The average assignment score over all adult samples was 83.9%, and the average probability of assignment was only  79  33.9%. After eliminating all individuals that had an assignment score lower than 95%, 157 individuals were left, 40 of which were dispersers and 117 of which were philopatric, leading to an estimate of total dispersal of 25.5%. Those dispersers were also more likely to be from geographically proximate localities (Fig. 3.3b). Removing full-sibs from the analysis improved the probability of assignments (average maximum probability of assignment 49.0%, compared with 33.9% with the sibs included), but the vast majority (96.7%) of the fish identified as dispersers in this analysis were the same as those identified in the analysis from which the sibs were not removed. Therefore, I only report the results of the latter analysis in subsequent sections. Each independent run of the STRUCTURE analysis for any one K value showed consistency in the lnP(D|K) (i.e. loge probability of the data given K) values (not shown), indicating that the runs typically converged. The analysis with all the individuals returned a peak in lnP(D|K) value at K = 11 (Fig. 3.4a). Using the "K statistic, however, returned a multi-modal distribution of "K values, with peaks at K = 2, 4, and 11 ("K values of 29.2, 24.8 and 3.1 respectively). Such multi-modal distributions of "K values are expected when hierarchical population structure is present (Evanno et al., 2005; Coulon et al., 2008). Because the level of population structure most appropriate for my current question is the population (i.e., lake) level, and because the lnP(D|K) criteria has been shown to perform better than the Evanno et al. (2005) method when genetic differentiation among populations is low (Waples & Gaggiotti, 2006), I chose to use the value of K with the highest lnP(D|K) (i.e., K = 11) for all further analyses. Under this assumption, the samples of juveniles generally clustered by sampling location (Fig. 3.4d), although the results also show a number of admixed individuals in each population. The adult samples, however, had a higher frequency of admixed individuals and of individuals that were assigned to a cluster that was different than the consensus cluster for that capture location (i.e., the cluster to which most other fish from that location, including juveniles, were assigned; Fig. 3.4d). This was reflected in the average Q-value (admixture coefficient) of adult fish (average Q = 0.53) being 15% smaller than that of juvenile fish (average Q = 0.68) for the cluster in which they were captured (t = 7.26, df = 789, P < 0.0001). I eliminated all adult samples with a maximum Q value lower than 0.5 which reduced my sample size from 359 to 259. Of the 259 individuals used in the analysis, only 41 were assigned to a genetic cluster different than that from which they were caught, leading to a total dispersal estimate of 15.8%. As for the  80  GENECLASS2 analysis, the dispersers appeared to be most likely from a geographically proximate locality, although this pattern was less pronounced with this analysis (Fig. 3.4c). !"9"! D2&&(%./)(%*-&2*+')+.()18&2E5%.4&2%./)(%*-&2*+') The distribution of GSI was clearly bimodal with two almost non-overlapping distributions for each sex, thus making the distinction between breeding and non-breeding individuals accurate (Fig. B.1 in Appendix B). All analyses conducted supported the hypothesis that overwintering fish were more likely to disperse than breeding fish (Fig. 3.3c and d). In the analysis including all individuals, I found that 116 of the 167 (69.4%) individuals assigned as dispersers were overwintering dispersers. Breeding individuals, on the other hand, were classified as philopatric more often (85 individuals) than they were classified as dispersing (51 individuals; two-tailed Fisher exact test, df = 358, P = 0.009; Fig. 3.3c). This pattern held after removing the individuals with an assignment score of less than 95% (spawning: 54 philopatric individuals, 11 dispersing individuals; over-wintering: 63 philopatric individuals, 29 dispersing individuals; two-tailed Fisher exact test, df = 156, P = 0.043; Fig. 3.3d), although proportionally fewer overwintering individuals were found to disperse. This result also held when assignment tests were performed on the data set from which the full-sibs were removed (spawning: 86 philopatric individuals, 50 dispersing individuals; over-wintering: 110 philopatric individuals, 113 dispersing individuals; two-tailed Fisher exact test, df = 358, P = 0.012). Finally, the results of the STRUCTURE analysis also support the conclusion that over-wintering individuals are more likely to disperse, even if it generally estimated lower dispersal rates (spawning: 98 philopatric individuals, 10 dispersing individuals; over-wintering: 120 philopatric individuals, 31 dispersing individuals; two-tailed Fisher exact test, df = 258, P = 0.016; Fig. 3.4c). !"9"9 F4A&2);122&'+4&*)1>)(%*-&2*+')-21-&.*%4@) I found no evidence of sex-biased dispersal in any of the three analyses used. In the GENECLASS2 analysis including all individuals, I found that 55.1% of dispersers and 45.8% of philopatric individuals were females (P = 0.0907). For the GENECLASS2 analysis including only individuals with confident assignment, 65.0% of dispersers and 48.7% of philopatric individuals were females (P = 0.0985). Finally, in the analysis using STRUCTURE, 42.9% of dispersers and 50.5% of philopatric individuals were females (P = 0.3480). None of the other biological traits I  81  examined could discriminate significantly between the dispersing and philopatric individuals (Table 3.2 and Fig. B.3 in Appendix B). !"9"< =14&.4%+')>12)'1;+')+(+-4+4%1.) Parameterization of the continent-island migration population genetic model for Arctic Char suggested that local adaptation is likely for biologically realistic values of the selection coefficient s (Fig. 3.5). The MIGRATE runs returned a mean (averaged over all sampling locations) ! and M values of 1.24 (variance = 0.15) and 1.99 (variance = 0.77), respectively (see supplementary materials for details). Depending on the mutation rate used, this led to estimates of Ne varying from 310 to 31,000 and estimates of m varying from 0.00034 to 0.034 (where m was obtained by adding the effective number of migrants coming from all samples, Nem, and dividing by the local effective population size). Parameterization of the model with these estimates yielded critical values for s (i.e., values above which local adaptation is expected) of 0.065, 0.0069, and 0.0007 assuming mutation rates of 10-3, 10-4, and 10-5 respectively.  !=B C&16:11&5'% In order to evaluate the evolutionary consequences of dispersal, it is important to resolve whether dispersal typically results in gene flow (Garant et al., 2007). The main objective of my study was to use genetic assignment methods to better understand the dispersal behaviour of Arctic Char by explicitly taking into account the reproductive status of the dispersers, thus allowing an evaluation of the proportion of dispersal events that have the potential to result in gene flow. I found that dispersal among sampling locations was very common, more common in fact than reported in most other species of salmonids (Hendry et al., 2004). I also found, however, that a large proportion of dispersers were non-reproducing individuals using non-natal habitats for over-wintering. This behaviour, whereby over-wintering individuals have a greater propensity to use non-natal habitats than fish destined to spawn that year, could have important implications for the process of local adaptation in those populations.  82  !"<"# G%/A)(%*-&2*+')%.)+.+(21,16*)C2;4%;)0A+2) In this study, I used genetic assignment methods to estimate total dispersal between sampling locations. I used three different approaches (two in GENECLASS2 and one in STRUCTURE) that led to considerably different estimates, varying from 15.8% to 46.5%. The absolute values of dispersal rates calculated in my study should therefore be interpreted with caution. Furthermore, the assignment tests performed in GENECLASS2 suffered from generally low assignment scores and low probabilities of assignment. In fact, 56% of all adult samples did not meet the > 95% criterion for confident assignment. The low confidence of assignment may stem from several factors. First, the power of assignment tests increases with increasing genetic differentiation (Paetkau et al., 2004; Manel et al., 2005). The populations used in this study are only weakly differentiated (average pairwise FST of 0.045 between the samples of juveniles) and this could be due either to the fact that those populations were only very recently recolonized following the last glaciation (<10,000 years ago) or because of ongoing gene flow. Other studies, however, have achieved greater confidence in assignment despite lower levels of genetic differentiation (e.g., FST 0.02 in Hauser et al. 2006). In addition, I used sample sizes and a number of loci that simulations have shown to allow for confident assignment despite low genetic differentiation (Paetkau et al. 2004) and I verified that I had sufficient power before conducting the analyses. A second reason for low confidence in assignment could be that the individuals with low assignment probability come from un-sampled populations (i.e., so-called “ghost” populations, Beerli 2004). I cannot fully discount this potential problem, but even if all adults with low assignment probability were in fact dispersers from un-sampled populations (which seems unlikely because I sampled most populations of Cumberland Sound), they would still be dispersers, therefore not affecting the estimates of dispersal rates. Finally, low confidence in assignment may be the result of the presence of sibs in the samples of juveniles used as reference samples. The increase in average probability of assignment after the removal of sibs from the juvenile samples (the average increased from 34% to 49%) suggests that this constitutes at least part of the explanation. The proportions and identity of the dispersers, however, did not change appreciably after the removal of the sibs, suggesting the conclusions of my study are not influenced by sibling relationships. Despite considerable differences in the estimates of dispersal rates depending on the method used, all estimates remain fairly high: in the order of 15.8-46.5% per year. High rates of  83  dispersal are consistent with previously published estimates from tagging studies of Arctic Char. For example, Gyselman (1994) estimated a dispersal rate of 47% per year, with substantial variation among years (33-66% per year). In another study, Dempson & Kristofferson (1987) reported rates of dispersal that varied tremendously among regions. The proportion of tagged fish that were subsequently recaptured in a different river (dispersers) varied between 0% and 17% per year among twelve different rivers in Labrador (eastern Canadian Arctic). By contrast, in the Cambridge Bay area of Canada’s central Arctic, those proportions were much higher, varying between 13% and 51% per year among four rivers (Dempson & Kristofferson 1987). My estimates based on genetic assignment tests therefore fall well within the range of previously reported values for Arctic Char. They are, however, at the upper end of the scale of dispersal rates reported in salmonids. Indeed, a recent compilation of straying rates from tagging studies performed on salmonids (but excluding Arctic Char), documented straying rates varying from 0.0% to 41.6% per year, with a median of 4.4% (Hendry et al., 2004). The studies reviewed in Hendry et al. (2004) were conducted at a variety of spatial scales, and my study was conducted on a relatively small spatial scale (max. distance between sites < 400km), which presumably would be associated with higher dispersal amongst localities. Furthermore, all the species reviewed in Hendry et al. (2004) are species that only migrate for the purpose of breeding. I found over-wintering dispersal to be higher than breeding dispersal (next section), and this could explain why dispersal rates are generally higher in Arctic Char than in other salmonids. !"<"B 01.(%4%1.E(&-&.(&.4)(%*-&2*+')) I presented several lines of evidence that, although total dispersal was high among sampling locations, the number of dispersal events that may result in gene flow is considerably lower. I found that the average FST between pairs of juvenile samples is significantly higher than between pairs of adult samples from the same location. The weaker genetic differentiation observed among samples of adults would arise if they contained dispersing fish from non-natal populations. Greater genetic differentiation among samples of juveniles, however, suggests that dispersal of adults to other system does not always result in gene flow. One alternative explanation for this pattern is that the juveniles sampled were more often from the same family group – a phenomenon referred to as the Allendorf-Phelps effect (Waples 1998). My sibship reconstructions using the program COLONY, however, suggested that this alternative in unlikely.  84  First, the number of full-sib groups in any sample was typically small. Second, when putative sibs were removed from the analysis, my findings remained unchanged. Another explanation could be that genetic differentiation in juveniles increases through drift because of high variance in reproductive success of adults (Vitalis 2002). While I cannot discount this alternative, I note that its effect should be fairly small given that I sampled multiple cohorts of adults and juveniles. The lower genetic differentiation between samples of adults is consistent with the idea that the large proportion of dispersing individuals are not in reproductive condition, and therefore do not contribute to gene flow. In fact, of the total number of dispersers identified by the genetic assignment tests, between 69.4% (for the analysis including all individuals) and 75.6% (for the analysis conducted in STRUCTURE) were fish whose non-reproductive status suggests that they were migrating to over-winter, and have thus no potential for gene flow. These estimates of the proportion of dispersers not in reproductive condition are very consistent across analyses, even if the analyses returned considerably different total rates of dispersal. Such a high proportion of non-breeding fish is probably the result of three life-history characteristics of Arctic Char. First, Arctic Char do not breed every year (Dutil 1986). In fact, roughly two thirds of all adult fish collected for this study were not in reproductive condition, suggesting that skipping two years between reproductive events may be common in Cumberland Sound Arctic Char. Such ‘skipped breeding’ is not unique to Arctic Char, and has been documented in many species of fish, including Atlantic Salmon, Salmo salar (Rideout & Tomkiewicz 2011). Second, and contrary to many iteroparous species, the Arctic Char that skip a spawning season must return to fresh water to over-winter (Johnson 1980). This is unusual for salmonids, the great majority of them only returning to fresh water to spawn. It is not unique, however, and is a behaviour also displayed by some non-mature Brown Trout (Salmo trutta) which cannot tolerate high salinity for long periods (Larsen et al. 2008), and by the closely related Dolly Varden (Salvelinus malma, Armstrong, 1974). Third, I have provided evidence that fish migrating to fresh water for the purpose of overwintering are more likely to utilize non-natal habitats. This unusual form of conditiondependent dispersal (Ims & Hjermann 2001) is also thought to occur in Dolly Varden (Armstrong 1974) and had been posited to occur in Arctic Char (Johnson 1980), although evidence for this remained indirect. For instance, Gyselman (1994) operated a weir on the Nauyuk River and tagged fish on their downriver migration after spawning. He found that many post-spawning tagged fish did not return the following year, but that 7.2% of the tagged fish did return one or  85  two years later in spawning condition (Gyselman 1994), suggesting that they used a non-natal habitat in the intervening year. The evidence from my genetic assignment tests strongly suggests that overwintering char are more likely to use non-natal habitats and that such behaviour may be widespread among anadromous Arctic Char populations. This behaviour has some interesting implications and sheds some light on the relative costs and benefits of dispersal. Assuming that it evolved in response to natural selection, it would suggest that the costs of dispersal are greater during the years when an individual will spawn. This could be the case if suitable spawning habitats are more limited than suitable over-wintering habitats (Power 2002). It has also been suggested that homing in salmonids evolved in order to return locally adapted individuals to the conditions to which they are adapted (Hendry et al. 2004). If local adaptation is important in anadromous populations of Arctic Char – a fact that remains to be established (but see next section) – this could explain why individuals home to their natal habitats to spawn but are less discriminatory when over-wintering. These hypotheses, however, only explain why breeding individuals home, and fail to provide an explanation for why over-wintering individuals may disperse at a higher rate. This increased dispersal when overwintering is unlikely the result of mere mistakes in orientation, since breeding individuals appear capable of increased homing capabilities. It thus seems likely that this increased propensity to disperse reflects some benefits of dispersal. There are several commonly reported benefits of dispersal that may apply to over-wintering Arctic Char. Johnson (1980) suggested this behaviour may allow char to explore the available habitat before settling to spawn. This intriguing possibility would assume that char evaluate their over-wintering habitats for their suitability as a spawning habitat and then retain this information for their next migration. There is no evidence that Arctic char are able to do this, but many animals appear able to evaluate the quality of a habitat patch based on certain environmental cues (Clobert et al. 2009). Dispersal is also often regarded as a bet hedging strategy favorable in unpredictable environments (McPeek & Holt 1992). The Arctic environment inhabited by Arctic Char is subject to pronounced seasonal changes and varies unpredictably from year to year (Power 2002; Power & Power 1995). If severe conditions affect the different streams around Cumberland Sound asynchronously, there may be an advantage to utilizing different habitats in different years. This last explanation is especially compelling since other commonly recognized benefits of dispersal, i.e., kin competition and inbreeding avoidance, are unlikely to apply to over-wintering dispersal (Bowler  86  & Benton 2005). Indeed, we would not expect breeding dispersal to be lower than over-wintering dispersal if it evolved in response to inbreeding avoidance because over-wintering dispersal does not introduce novel alleles. Similarly, if dispersal evolved in response to kin competition, we would not expect over-wintering dispersers to return to their natal habitat for spawning. Regardless of the mechanisms driving the differences between breeding and over-wintering dispersal, the differences themselves strongly suggest that the costs and benefits of dispersal differ depending on the reproductive status of individuals. More work is necessary to decipher the selective forces that drive this behavior, but this pattern suggests that Arctic Char may provide an interesting model system to test some ideas regarding the evolution of dispersal in salmonids (Hendry et al. 2004). The dispersers did not differ from philopatric individuals in a number of other biological traits. First, I found no evidence of sex-biased dispersal. Sex-biased dispersal is expected to evolve when the costs and benefits of dispersal are not equal between sexes (Perrin & Mazalov 2000). The promiscuous mating system of salmonids, the intense competition for females among males, and the fact that a female’s reproductive success is mainly limited by her ability to produce eggs, led to the prediction that dispersal should be male-biased in salmonids (Hutchings & Gerber 2002). There are in fact several studies that have reported evidence of male-biased dispersal in other species of salmonids (Hutchings & Gerber 2002; Bekkevold et al. 2004; Fraser et al. 2004; Neville et al. 2006). The pattern, however, is not universal, and Consuegra & Garcia de Leaniz (2002) failed to find evidence for sex-biased dispersal in Atlantic Salmon (Salmo salar). They interpreted these patterns as a result of sex-ratio differences due to the existence of non-migratory males in some populations. Non-migratory males (“mature parr”) are known to exist in some populations of Arctic Char from Cumberland Sound (Loewen et al. 2009), but they did not lead to any bias in sex ratio in the samples used in the present study (Fig. 3.3). It is thus unclear whether this factor contributed to the lack of evidence for sex-biased dispersal in Cumberland Sound Arctic Char. Comparatively little is known about the effect of other traits on individual dispersal propensity in salmonids. Indeed, while juvenile dispersal away from the nesting site has been shown to be dependent on body condition in Atlantic Salmon (e.g., Einum et al. 2012), the traits associated with an increased propensity to stray to a different spawning site are not well known. One variable that has been found to be associated with dispersal propensity is age, with studies  87  variously reporting that older (Quinn 1993) or younger (Hard & Heard 1999) fish have increased dispersal propensity. In this study, I found no evidence of trait-dependent dispersal, and none of the biological variables I examined, including age, differed between dispersing and philopatric individuals. Lack of evidence for condition-dependent dispersal in my study, and the weak and inconsistent evidence for condition-dependent dispersal in other salmonid species, is in contrast with the widespread nature of this phenomenon in other taxa (Ims & Hjermann 2001; Clobert et al. 2004; Bowler & Benton 2005). It is possible that this is a reflection of the evolutionary forces that favor dispersal in salmonids. Indeed, body condition may be a good predictor of dispersal propensity when it evolves in response to local competition, because the costs and benefits of dispersal differ among individuals. But if dispersal instead evolves in response to unpredictable environmental conditions such that all individuals are equally affected, condition-dependent dispersal may not be expected. As discussed earlier, this is a likely selective force driving dispersal in Arctic Char and in other salmonids (Hendry et al. 2004). !"<"! =14&.4%+')>12)'1;+')+(+-4+4%1.) Dispersal can have important consequences for local adaptation, but dispersers need to successfully reproduce in the non-natal habitat. My study showed that only 24% to 31% of dispersers are in reproductive condition, which, when multiplied by the total dispersal rates calculated (i.e., 15.6% to 46.5%, depending on the analysis method), led to an estimate of breeding dispersal rate of 4-14% per year (for the analyses conducted in STRUCTURE and that conducted in GENECLASS2 including all individuals, respectively). In other words, only 4-14% of all individuals migrating up-river in a given year are dispersers that are in reproductive condition and thus have the potential to lead to gene flow. While this is considerably lower than total dispersal, it still has the potential to translate into levels of gene flow high enough to prevent local adaptation. It is possible, however, that not all breeding dispersers successfully reproduce in the non-natal environment. Indeed, heterogeneous habitats can lead to selection against migrants, and many studies have found evidence for this process in salmonids (Tallman & Healey 1994; Hendry et al. 2000; Dionne et al. 2008) and other taxa (for a review: Nosil et al. 2005). While I do not have evidence for selection against migrants in this system, the generally lower rates of gene flow calculated with MIGRATE (0.034% to 3.4%) suggest that breeding dispersers may  88  indeed have lower reproductive success than local fish. More work would be necessary to understand the cause of this discrepancy between the rates of breeding dispersal and gene flow. The estimates of gene flow generated by MIGRATE were also used to parameterize a population genetic model of the balance between selection, gene flow and drift (Adkison 1995; Yeaman & Otto 2011). Similar approaches have been used extensively in other studies of anadromous fishes as a tool to explore the potential for local adaptation (Hansen et al. 2002; McCairns & Bernatchez 2008). The goal of such analyses is to demonstrate that local adaptation is a possible outcome given biologically realistic parameter values, not to demonstrate the existence of local adaptation. Based on the estimates of gene flow from MIGRATE, my results suggest that local adaptation in Cumberland Sound Arctic Char is possible, assuming that selective environments are sufficiently heterogeneous. The model suggested that, depending on the mutation rate assumed, selection coefficients ranging from 0.0007 to 0.065 may be sufficient to drive local adaptation. A review of estimates of the strength of s on major QTLs in nature suggests that values ranging from 0.01 to 0.05 are most typical (Morjan & Rieseberg 2004). All but the highest estimate of s derived from the parameterization, therefore, could potentially lead to local adaptation under biologically realistic conditions. In addition, the highest critical s value I obtained assumed a mutation rate (10-3) that is probably higher than that of most salmonid microsatellites (Steinberg et al. 2002). This approach, however, makes several assumptions that are likely violated in Cumberland Sound. Indeed, the MIGRATE analysis, which assumes equilibrium conditions, is potentially problematic in the present case because the populations used were only recently established following the last glaciation (< 10,000 years ago). Furthermore, the model that I parameterized makes many simplifying assumptions regarding the spatial scale of gene flow. The assumptions violated both in the MIGRATE analysis and the simple model, however, are likely to lead to underestimated potential for local adaptation, and therefore make my conclusions conservative (discussed in the Materials and Methods section). In short, the parameterization of the population genetic model is a decidedly exploratory approach and cannot be substituted for a formal study of local adaptation in anadromous Arctic Char. The results do suggest, however, that local adaptation is likely in anadromous Arctic Char, given that selective environments differ sufficiently between habitats. Local adaptation is commonly reported among populations of anadromous salmonids, in part due to their homing behavior (Taylor 1991; Fraser et al. 2011), and evidence for local adaptation among landlocked  89  populations of Arctic Char exists (Larsson et al. 2005; Conejeros et al. 2008). A more formal examination of the extent of local adaptation in this species would seem important, since it will influence the capacity of this species to adapt to climate change, a threat that will be felt disproportionately by Arctic species (Arctic Climate Impact Assessment 2004).  90  Table 3.1. Summary of sampling locations used in this study, and measures of genetic diversity at each location for anadromous Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci. (juv = juveniles) Site Site name  Coastal  Nei’s  distance from  Unbiased  Allelic  code  Year  Life stage  N  NAU (km)  Lon  Lat  Obs. Het.  Exp. Het.  richness  Naulineavik  NAU  2010  Adults*  46  0  63°50'W  65°10'N  0.656  0.688  4.91  Iqaluggarjuit 2  IQ2  2009  Juv  48  145  64°46'W  65°44'N  0.684  0.707  4.85  Kingnait  KIN  2009  Juv  58  225  64°19'W  66°23'N  0.682  0.696  4.81  2010  Adults*  37  0.665  0.702  4.88  2007  Adults  15  0.67  0.699  4.98  2006  Adults  29  0.658  0.703  4.91  Pang Fiord  PAN  2009  Adults  19  200  65°42'W  66°09'N  0.649  0.710  5.13  Avataktoo  AVA  2009  Juv  65  215  66°09'W  66°19'N  0.695  0.695  5.08  Iqaluqjuaat  IAT  2009  Juv  48  240  66°32'W  66°26'N  0.700  0.695  4.97  Kekertelung  KEK  2010  Juv  54  240  66°46'W  66°20'N  0.702  0.659  4.25  Iqaluggarjuit 1  IQ1  2009  Juv  56  270  66°43'W  66°34'N  0.698  0.694  4.74  2004  Adults  10  0.684  0.700  4.85  2009  Juv  40  0.690  0.667  4.61  2004  Adults  44  0.682  0.690  4.89  2003  Adults  57  0.651  0.684  4.88  2010  Juv  70  0.671  0.691  4.83  2010  Adults  70  0.684  0.707  5.04  2003  Adults  45  0.634  0.694  4.96  2009  Juv  61  0.696  0.707  5.08  2003  Adults  48  0.658  0.699  5.02  2004  Adults  32  0.678  0.708  5.05  Isuituq  Kangitujuak  Kipisa  ISU  KAN  KIP  300  68°12'W  250  67°27'W  300  67°55'W  91  66°50'N  66°26'N  66°33'N  Table 3.1 cont’d  Site Site name  Life  Coastal  Nei's  distance from  unbiased  Allelic  code  Year  stage  N  0 (km)  Lon  Lat  Obs. Het.  Exp. Het.  richness  Aunktavik  AUN  2010  Juv  43  325  67°26'W  65°58'N  0.634  0.674  4.48  Irvine  IRV  2010  Juv  5  360  68°21'W  65°24'N  0.693  0.699  4.87  Ikpit  IKP  2009  Juv  60  350  67°25'W  65°26'N  0.654  0.668  4.79  Kanayuktuk  KAK  2003  Adults  48  360  67°21' W  65°20'N  0.666  0.680  4.95  Opingnavik  OPI  2009  Juv  26  370  67°15'W  65°14'N  0.658  0.686  5.07  Iqaluit  IQA  2009  Juv  65  385  67°07'W  65°02'N  0.689  0.685  4.93  2005  Adults  42  0.672  0.685  4.94  2009  Juv  65  0.657  0.668  4.74  2009  Adults  41  0.655  0.666  4.73  Qasigiat  QAS  460  66°19'W  92  64°37'N  Table 3.2. Results of discriminant function analysis testing for trait differences between philopatric and dispersing individuals of anadromous Arctic Char (Salvelinus alpinus) sample from Cumberland Sound. Mean value  Trait Age (yr) Fork Length (cm) Weight (g) Gonad Weight (log10[g]) Condition factor Age (yr) Fork Length (cm) Weight (g) Gonad Weight (log10[g]) Condition factor Age (yr) Fork Length (cm) Weight (g) Gonad Weight (log10[g]) Condition factor  Coefficients of linear Dispersers Philopatric discriminants GENECLASS2 analysis with all samples 12.1 11.9 -0.217 561.6 575.5 -0.020 2336.8 2564.6 0.002 1.01 1.13 0.697 1.23 1.24 -3.871 GENECLASS2 assignment scores >95% 11.1 11.7 0.191 569.3 569.9 -0.038 2371.4 2548.7 0.003 0.99 1.13 0.791 1.24 1.25 -6.708 STRUCTURE analysis 11.6 12.0 0.053 565.7 574.0 -0.022 2498.3 2513.5 0.001 1.05 1.17 0.879 1.28 1.22 -9.560  93  Wilk's lambda  F  P  0.969  1.758  0.122  0.906  2.141  0.066  0.958  1.654  0.148  Figure 3.1. Map of Cumberland Sound, Baffin Island, Nunavut, Canada. Sampling locations for Arctic Char (Salvelinus alpinus) examined in this study are shown with a dot, and identified with a three letter code referenced in Table 1. The location of the nearest town, Pangnirtung, is shown with a star. The inset map shows the location of Baffin Island in North America.  ± ISU  Gr  Baffin Island KIN  !  KIP KAN !  ! IAT ! AVA  KEK !  !  IKP IRV  !  be rl an d  !  QAS 0  20  40  80  !  NAU  ! KAK ! OPI !  IQA  QC  ON  !Pangnirtung [  Cu m  !  NFD  !  IQ2  AUN  Hudson Bay  SA MA  PAN  Kilometers 120  94  nd  NU  AB !  nl a  ! NWT  IQ1  ee  !  So un d  !  Figure 3.2. Summary of population genetic structure among Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci from locations of Cumberland Sound. (a) Neighborjoining tree of Cavalli-Sforza’s chord distance, showing that paired samples of juveniles and adults from the same localities group together (encircled groups of samples). (b) Pairwise FST IRV10  values between Arctic char sampling localities when juvenile samples (left) and adult samples are jv  compared (right). The grey circles are the actual observed pairwise FST values, and the horizontal black bars show the means across all comparisons (random jitter was added to avoid visual overlap).  jv 09 09ad AS S Q QA  T0 9  3a d KAN10jv  d 4a  0.05  IQ  10  ad  KI  P KIP 03a 09j d v  0jv  IQ104  P0 KI  09jv  K1  jv  PAN09ad  IAT  KE  KI  9jv 20 IQ win 10ad A N U AVA09jv  v  ISU03ad 4ad ISU0  09 ISU  P  K1  d  KIN0  4a  10  7ad  KE  IQ  0.0060  9j v  0.0  06  _  _  0.02  N0  9a IK d  09j  KIN KIN10 06ad adwin KIN09jv  d  KA  N0  IQ1  10a  PA  jv 09  0.01  IA  09jv  0.04  AVA  v  NAU10adw in  FFst ST  0j  jv  KAN  d  N1  jv I09 OP d a 3 KAK0  AU  d 04a KIP ad 3 jv 0 P KI 09 P KI  IQA09jv IQA0 5ad QA S QA 09j S0 v 9a d  IQ209jv  ISU09jv  7a  0.06  IK  0jv ISU 03a ISU04ad d  N0  (b)  jv P09  N1  KIN06ad adwin v KIN10 09j KIN  d KA  0.03  0jv  0a  0jv 3ad V1 N0 IR KA  N1  IQ IQA A0 09jv 5a d  N1  OPI09jv KAK0 3ad  AU  (a)  KA  0  Figure 2 95  0jv  P = 0.036  Juveniles  Adults  Figure 3.3. Results of the assignment tests performed with GENECLASS2 for anadromous Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci. The left panels (a and c) show the results of the analysis including all adults that were assigned, while the right panels (b and d) show the results of the analysis including only the individuals that were assigned with an assignment score of more than 95% (see text for details). Panels (a) and (b) show the distribution of dispersal distances for both analyses. In both cases a large proportion of individuals are classified as philopatric (dispersal distance = 0 km). The individuals classified as dispersers (i.e., all non-zero dispersal distances) tend to be classified to nearby localities. Panels (c) and (d) are mosaic plots showing the reproductive status and sex of the dispersing and philopatric individuals, respectively (note that the height and width of the bars, not the area of the squares, show the proportions of individuals in each category).  96  Figure 3.4. Results of the STRUCTURE analysis for anadromous Arctic Char (Salvelinus alpinus) assayed at fourteen microsatellite DNA loci. (a) Ln probability of the data given each K averaged over ten independent runs. The arrow shows the most likely number of genetic clusters (K = 11) according to the Pritchard et al. (2000) method. (b) Distribution of dispersal distances. A large proportion of individuals are classified as philopatric (dispersal distance = 0 km). The individuals classified as dispersers (i.e., all non-zero dispersal distances) tend to be classified to nearby localities. (c) Mosaic plot showing the reproductive status and sex of the individuals that were classified as dispersing or philopatric according to the STRUCTURE analysis. (d) Individual admixture coefficients (Q) for all adults used in the analysis (bottom). The top part of the graph shows the individual admixture coefficient of juvenile samples for comparison.  97  Figure 3.5. The relationship between the critical value of the selection coefficient, s, and the migration rate, m, for different values of the mutation rate (both axes are on a log scale) estimated for anadromous Arctic Char (Salvelinus alpinus). The three lines show the relationship between s and m for the three values of Ne obtained from MIGRATE assuming a mutation rate of 10-3 (solid), 10-4 (dashed), and 10-5 (dotted). The arrows and the shading show the calculated values of the critical s if both m and Ne obtained from MIGRATE are used to parameterize the continent-island model, again depending on the mutation rate assumed: 10-3 (top), 10-4 (middle), and 10-5 (bottom). For all values of s  Critical s value  above the arrows and shading, local adaptation is expected.  Ne = 310.5  105 50 10 3 = e  Ne = 3  N  Migration rate m  Figure 5  98  ! "#$#%&'()$)*+,&,(-.(,+/0)%1&'(/&21)%-1+(#'-%+0#,(-.(3)..&$( 4,*)$5(61'%&'(78)19()*%#1$)%&:#(/)%&$2(%)'%&',(-1( 1#01-5;'%&:#*+(&,-*)%#5(,%1)%#2&#,<( !=> 6?,%1)'%( Many populations of salmonids display polymorphisms in migratory phenotypes, with anadromous and resident individuals coexisting in sympatry. Two alternative hypotheses are generally proposed to explain the co-existence of these life-history ecotypes. First, the two ecotypes could originate from a single population, and be the result of a conditional mating tactic. Second, migratory ecotypes may be reproductively isolated populations utilizing alternative migratory strategies. Three populations of Arctic Char Salvelinus alpinus from southern Baffin Island were previously identified to display variable migratory phenotypes (freshwater resident and anadromous). The genetic mechanism driving their co-existence, however, is unknown. In this study, fourteen microsatellite markers were used to investigate whether the two ecotypes from these lakes are reproductively isolated. In two of the three systems, FST values between the resident and anadromous components of the population were non-significant, while they were significant in a third system. Bayesian clustering analysis implemented in STRUCTURE, however, failed to identify any within-lake clustering that correlated with life-history ecotypes in all three sampling locations. It is concluded from these analyses that the lifehistory ecotypes are most likely conditional mating tactics, rather than reproductively isolated populations. Other evidence in favor of the alternative mating tactic hypothesis is briefly reviewed, and implications for management of those populations are discussed.  99  !=@ 4$%1-5;'%&-$( Many populations of anadromous salmonids display variable life history, such that a portion of the population matures at an early age, opting to forego migrations to marine habitats, while another portion undergoes anadromous migrations (Fleming 1998; Hendry et al. 2004). This within-population variation in anadromy is widespread, and Fleming (1998) remarked that the presence of sympatric alternative migratory ecotypes was known to occur in all species in the subgenus Salmoninae except for Pink Salmon, Oncorhynchus gorbuscha, and Lake Trout, Salvelinus namaycush, although populations of lake trout with anadromous individuals have recently been discovered (Swanson et al. 2010). Such dimorphisms have traditionally been understood under the framework of conditional mating tactics (Hutchings & Myers 1994; Gross 1996; Thorpe et al. 1998). Under this framework, one genotype is able to give rise to the two alternative mating tactics (residency vs. anadromy) and the ‘decision’ to follow a tactic is conditional on the individual’s status (most commonly its size or growth rate in salmonids; Hendry et al. 2004) such that the tactic will maximize the individual’s fitness given its status (Gross 1996). There is considerable empirical data that supports this interpretation of mating dimorphism in salmonids (e.g., Thomaz et al. 1997; Thorpe et al. 1998; Hendry et al. 2004; Thériault et al. 2007b). An alternative explanation for intra-population variation in migratory behaviour is that it could represent two alternative, genetically-fixed strategies (Gross 1996). Under this model, differences in migratory behaviour are the result of a genetic polymorphism, and are maintained either because of frequency-dependent selection or because the two behaviours are expressed by two reproductively isolated populations (Gross 1996). There is no evidence for the former model in salmonids, but there are well-documented cases of reproductively isolated populations co-existing sympatrically exhibiting different migratory strategies. For example, Kokanee Salmon are a freshwater-resident form of the usually anadromous Sockeye Salmon O. nerka. Both of these forms live in sympatry, are genetically differentiated (Wood & Foote 1996), and have evolved independently in multiple lakes from Russia, Alaska, British Columbia, and Washington (Taylor et al. 1996). Other examples of reproductively isolated migratory strategies occur in Atlantic  100  Salmon, Salmo salar (Verspoor & Cole 1989; Tessier & Bernatchez 2000), and Rainbow Trout, O. mykiss (Docker & Heath 2003). The Arctic Char Salvelinus alpinus displays a bewildering array of phenotypes throughout its Holarctic distribution. Most well-known, perhaps, are systems where ecotypes of non-anadromous Arctic Char co-exist sympatrically in post-glacial lakes (Skúlason et al. 1999; Jonsson & Jonsson 2001). Another axis of diversification that is common in populations of Arctic Char is that displayed in migratory behaviour. Arctic Char populations at the southern and northern extremes of the range are more likely to reside in fresh water year-round (i.e., resident populations), whereas populations at intermediate latitudes are more likely to be anadromous if access to marine habitats exists (Johnson 1980). Anadromy, however, can also vary within a population and Arctic Char populations composed of anadromous and resident individuals have been identified in Norway (Nordeng 1983; Rikardsen & Elliot 2000), and in many parts of the Canadian Arctic (Reist 1989; Babaluk et al. 1997; Loewen et al. 2009). Some evidence exists arguing that life-history variation in Arctic Char populations conforms to the alternative mating tactics framework. For example, Nordeng (1983) studied a polymorphic population of Arctic Char from the Salangen River system in northern Norway that exhibited variable life-history strategies. By rearing the offspring of both anadromous and resident parents, he was able to demonstrate that resident parents were able to produce anadromous offspring and vice-versa, suggesting that the behavior is not genetically fixed in the population. He did find, however, that resident parents were more likely to give rise to resident offspring (and likewise for anadromous parents), suggesting that the determination of life-history tactic does have a heritable component (Nordeng 1983). This result is also consistent with work in another closely related species, Brook Trout, Salvelinus fontinalis, further suggesting that individual tactics have a heritable component (Thériault et al. 2007b). In southeast Baffin Island, in Canada’s Nunavut Territory, three populations of Arctic Char are known to display variable migratory phenotypes (Loewen et al. 2009; Loewen et al. 2010; Fig 4.1). Loewen et al. (2009) showed that in Qinngu, Iqaluggarjuit, and Qasigiat lakes, anadromous individuals co-existed with individuals that matured at a much smaller size and did not undergo migrations to the marine environment. Otolith  101  strontium profiles showed that the small-maturing fish spent all their lives in the freshwater environment - except in Qasigiat, where the small-maturing fish probably used the tidal habitat (Loewen et al. 2009). Small-maturing fish were also found to differ from non-mature fish of similar size in a variety of morphological traits (e.g., eye diameter, pectoral fin and pelvic fin lengths). The two life-history ecotypes were also shown to differ in growth patterns and mean age (Loewen et al. 2010). The available evidence suggests that this variation in life-history characteristics represents a case of alternative mating tactics. Indeed, Loewen (2008) found that the sex ratio of the resident component of the population was heavily skewed towards males. This observation is consistent with the precocious male life-history strategy commonly observed in salmonids (Fleming 1998; Fleming & Reynolds 2004). The fact that precocious females were observed at all, however, raises the possibility that the resident segment of the population is reproductively isolated from the anadromous segment. Therefore, the hypothesis that the co-existence of life-history types represents alternative mating tactics should be tested against the alternative hypothesis of reproductive isolation between two populations exploiting different niches. In the present paper, data from microsatellite markers were used to test for the presence of genetic differentiation among migratory ecotypes of Baffin Island Arctic Char in order to help evaluate two alternative hypotheses explaining the co-existence of the forms. The ‘conditional mating tactic hypothesis’ makes the prediction that within each population, the resident and anadromous forms should not be genetically differentiated. Alternatively, the ‘reproductively isolated ecotypes hypothesis’ makes the prediction that within each population, the resident and anadromous forms should be at least in part genetically differentiated. This last prediction, however, only holds if the two forms have been reproductively isolated for enough time for neutral genetic differences to accumulate by genetic drift. The use of microsatellites, one of the fastest mutating markers available (Ellengren 2000), for this study increases the likelihood of detecting genetic differences between life-history ecotypes if it is indeed present.  102  !=A B)%#1&)*,()$5(/#%8-5,( !"#"$ %&'()*+,-.(&/0* Loewen et al. (2009) provide extensive details regarding the study area and the field collection methods. In short, specimens of Arctic Char were collected from three lake systems in southern Baffin Island: Qinngu, Blanford Bay (LH001) close to the community of Kimmirut on the Meta Incognita Peninsula, and Qasigiat, Ptarmigan Fiord (PG015) and Iqaluggarjuit Lake, Shark Fiord (PG082) both close to the community of Pangnirtung in the Cumberland Sound area (Figure 4.1; Table 4.1). Sampling was conducted from late-August to mid-September in 2003-2009 when most anadromous Arctic Char had migrated back to fresh water to overwinter and spawn. Fish were captured with multimesh (38–102 mm stretched mesh), 38 mm stretched mesh, and 140 mm stretched mesh gillnets. For each fish collected in the field, fork length (mm), round weight (g), sex and maturity were recorded and the sagittal otoliths were removed and preserved for age and isotope analysis. The non-migratory nature of the small-maturing fish from the 2002-2005 sampling years were all confirmed using otoliths microchemistry (strontium distributions) described in Loewen et al. (2009). The migratory behaviour of the fish sampled in 2009, however, was not confirmed with otolith microchemistry. Given that all small-maturing fish in Loewen et al. (2009) had strontium profiles consistent with fresh water residency, all the small maturing fish sampled in 2009 were considered residents. The size cutoff used to define a small maturing fish was 400mm fork length, as in Loewen et al. (2009). !"#"1 2&345+,6'((&6'*789*,/,(:+&+* Individual genotypes were obtained at 18 microsatellite loci combined in 4 multiplexes (Table 4.2). For each locus, the forward primer was labeled with a fluorescent dye, and the reverse primer was PIG-tailed to reduce stutter and facilitate genotyping (Brownstein et al., 1996). PCR amplifications were carried out in 10µL volume reactions (see Table 4.2 for details). The PCR cycles were as follows: an initial denaturation step of 10 minutes at 95°C, 35 cycles of denaturation (45 seconds at 94°C), annealing (45 seconds at 55°C) and extension (45 seconds at 72°C), and a final extension  103  cycle of 30 minutes at 72°C. The PCR products were run on an Applied Biosystems (Carlsbad, CA, USA) 3100 genetic analyzer. GeneMapper Software version 3.7 (Applied Biosystems, Carlsbad, CA, USA) was then used to automatically score microsatellite alleles, and all scores were then manually checked for quality. Basic descriptive statistics were calculated for each samples in MICROSATELLITE TOOLKIT (Park 2001): number of alleles (NA), observed heterozygosity (HO), and Nei’s (1987) unbiased expected heterozygosity (HE) controlling for sample size. FSTAT version 2.9.3.2 (Goudet 2001) was used to test for HardyWeinberg equilibrium and genotypic disequilibrium using default values for the number of permutations. For both tests, I set the nominal significance level at 0.05 (using a Bonferroni correction for multiple comparisons). FSTAT was used to calculate allelic richness (AR, i.e., number of alleles controlling for sample size). Genetic differentiation between samples was also computed in FSTAT using Weir & Cockerham’s (1984) ! estimator of pairwise FST between each sample, and significance was assessed with 10,000 permutations (experiment-wide ! = 0.05 after Bonferroni correction). The data set was first analyzed with each sampling year kept separate. None of the FST values between years within sampling location, however, were significant (not shown), and the years were therefore combined for all subsequent analyses. An analysis of molecular variance (AMOVA) was conducted in ARLEQUIN v. 3.0 (Excoffier et al., 2005) to test for the presence of hierarchical population structure. The hierarchical levels of population structure were defined with the sampling locations (PG082, PG015, and LH001) as the ‘groups’, and the life-history ecotypes as the ‘populations within groups’. A factorial correspondence analysis (FCA) was conducted in GENETIX v. 4.05 (Belkhir et al., 2004) to identify genetic discontinuities among samples. FCA is a multivariate approach for categorical variables (here diploid genotypes at different loci) that identifies a number of orthogonal axes that best explain variation in the data. I visualized variation in individual genotypes along the first three axes explaining the most variation in the FCA. The Bayesian clustering algorithm implemented in STRUCTURE v. 2.3.2 (Pritchard et al. 2000) was used to determine the most likely number of genetic clusters (K) among samples. The main advantage of the STRUCTURE algorithm is that it allows the  104  identification of genetic discontinuities among samples without having to specify the sampling locations, or ecotypes, a priori. The algorithm accomplishes this by grouping together samples to maximize Hardy-Weinberg and linkage equilibrium within a genetic cluster. Because the null hypothesis to falsify with this analysis is that of panmixia, a modification to this algorithm was used to maximize the likelihood of finding any genetic discontinuities in the samples. The LOCPRIOR model (Hubisz et al. 2009) implemented in STRUCTURE v. 2.3.2 allows for the use of sample group information to aid in the clustering process. It does so by using the location of origin of each individual as a prior in the Bayesian algorithm. This model has been found to detect genetic structure at lower levels of divergence than the regular model, but does not detect structure when it is not present (Hubisz et al. 2009). STRUCTURE with the LOCPRIOR model was run under the admixture and correlated allele frequencies models, with 250,000 burnin and 500,000 Markov chain Monte Carlo (MCMC) replicates. Values of K (i.e. numbers of genetic clusters) ranging from one to ten were tested, and 20 independent runs were conducted under each K value. The most likely number of genetic cluster was determined using the post-hoc delta K method advocated by Evanno et al. (2005). The software CLUMPP v. 1.1.2 (Jakobsson & Rosenberg 2007) was used to combine the results of the 20 independent runs for the most likely K under the Greedy algorithm with 1,000 replicates, and the results of the STRUCTURE analysis were visualized using DISTRUCT v. 1.1 (Rosenberg 2004). When hierarchical population structure is present, STRUCTURE often only identifies the higher level of structure (e.g. Coulon et al. 2008). To ensure that withinsampling-location population structure is not obscured by higher-level structure, the STRUCTURE analysis was repeated on each sampling location separately. The LOCPRIOR model was again used, with life-history ecotypes the specified “population of origin”. All other run parameters were identical as to the analysis on all sampling locations at once (i.e. admixture and correlated allele frequencies models, with 250,000 burnin and 500,000 MCMC replicates, K from one to ten tested in 20 independent runs). The Evanno et al. (2005) method was used, and CLUMPP and DISTRUCT were used to combine and visualize the results according to the parameters specified above.  105  !=! C#,;*%,( Two of the 18 microsatellite loci scored were monomorphic across all samples (Smm21 and OMM1128) and were thus removed from all analyses. The remaining 16 loci were highly polymorphic and the number of alleles per locus ranged from 4 (for Sco215) to 77 (for Sco216; mean = 22.6). After correction for multiple comparisons (adj. " = 0.0002), four loci showed heterozygote deficits in at least one sampling location (here defined as samples of one ecotypes collected in one year in one sampling location): Sco218 in five, Sco109 in three, Sco202 in one and Sco212 in one sampling location. Loci Sco218 and Sco109 have been found to be problematic in other datasets as well (see Chapter 3) and may suffer from null alleles or large allele dropouts. Those loci were therefore removed from all subsequent analyses, leaving a total of 14 informative markers (see Table 4.3 for population statistics). Twelve pairs of loci displayed significant linkage disequilibrium before correction for multiple comparisons, but only two remained significant after correction: Sco200 x Sco216 and Smm22 x Sfo18. Various measures of genetic differentiation revealed significant population structure among the samples. Values of FST between LH001 and the two Cumberland Sound sampling location varied between 0.064 and 0.079 and were all significant after correction for multiple comparisons (adj. " = 0.003; Table 4.4). The two Cumberland Sound sampling locations were also significantly differentiated but FST values were smaller, ranging from 0.025 to 0.031 (Table 4.4). Genetic differentiation between ecotypes was not significant in PG015 (FST = -0.001) and in PG082 (FST = 0.004), but was significant between ecotypes in LH001 (FST = 0.039). The analysis of molecular variance (AMOVA) found that most of the genetic diversity was distributed within populations (94% of the variance explained), while sampling location explained 4.88% of the variance, and ecotypes only 0.47% (Table 4.5). Nevertheless, the ecotypes were found to explain a significant portion of the variance (P = 0.012) while sampling location did not (P = 0.063). The results of the factorial correspondence analysis (FCA) are in general agreement with the FST values and the AMOVA (Fig. 4.2). The three most important axes explained only a total of 5.67% of the variation, with axis 1, 2 and 3 explaining 2.43%, 1.74%, and 1.49% of the variation, respectively (Fig 4.2a). There is a clear distinction 106  between LH001 and the two Cumberland Sound sampling locations along axis 1, while the two Cumberland Sound locations are separated, albeit to a lesser extent, along axis 2 (Fig. 4.2a). Within sampling locations, resident and anadromous individuals are essentially indistinguishable, especially in PG015 and PG082. In LH001, however, there is a slight differentiation between ecotypes along axis 3 (Fig. 4.2b). This result is consistent with the significant FST values between ecotypes in LH001. The results of the STRUCTURE analysis suggested that ecotypes are not genetically differentiated. Both the delta K method (Evanno et al. 2005) and the lnProbK|data method (Pritchard et al. 2000) concluded that the most likely number of genetic clusters in the data was three (Fig. 4.3a,b). Examination of the individual assignment probability to each cluster revealed that each sampling location was differentiated from the others (although the two locations from the Cumberland Sound were not strongly differentiated), but that sympatric ecotypes within sampling locations were not differentiated (Fig. 4.3c). Those results were also supported by the STRUCTURE analyses conducted on each sampling location separately (Fig. 4.4). In Qasigiat (PG015), the most likely number of genetic cluster was one according to the lnProbK|data method (Fig. 4.4a). Because delta K is not defined for K = 1, it is concluded that there is no population structure within Qasigiat based on the lnProbK|data method alone. In contrast, the most likely number of genetic clusters in the other two populations according to both methods was two (Fig. 4.4a). While population structure was identified in both locations, the individuals did not cluster according to life-history ecotype (Fig. 4.4b).  !=D E&,';,,&-$( Sympatric migratory ecotypes are commonly observed in many species of salmonids (Wood & Foote 1996; Fleming 1998; Fleming & Reynolds 2004; Hendry et al. 2004). The co-existence of these phenotypes is generally explained under the framework of alternative mating tactics or of reproductive isolation between genetically-fixed mating strategies (Gross 1996). In the present study, analysis of microsatellite DNA data was used to help determine the most likely mechanism for the co-existence of sympatric  107  ecotypes of resident and anadromous Arctic Char from three lakes in southern Baffin Island (Loewen et al. 2009). The literature on migratory ecotypes in salmonids (with an emphasis on Arctic Char) is briefly reviewed to help argue that the genetic results of the present study, along with other sources of evidence, suggest that the migratory ecotypes in southern Baffin Island populations of Arctic Char represent a case of alternative mating tactics. Implications of such an interpretation for fisheries management are discussed. !";"$ <,3=*5>*0'/'6&3*)&>>'4'/6&,6&5/*&/*?@-A'4(,/)*B5@/)*'356:.'+* The results of all analyses strongly suggest a lack of genetic differentiation between resident and anadromous fish in the two Cumberland Sound sampling locations. First, FST values between ecotypes in those populations were not significantly different from zero, and a factorial correspondence analysis showed both ecotypes were indistinguishable along the first three axes of divergence. The results of the AMOVA over all populations showed that differences between ecotypes explained a significant amount of genetic variation, but this is likely due to the presence of small amounts of genetic differences between the ecotypes in the LH001 sampling location (see next section). Finally, the STRUCTURE analysis failed to identify the presence of genetic clusters within sampling location, further suggesting that the ecotypes are not genetically differentiated. When each sampling location was analyzed separately, STRUCTURE identified population structure but it did not correlate with life-history ecotype. These results were obtained despite using ecotypes as a prior in the Bayesian model. Indeed, the LOCPRIOR model used in the STRUCTURE analysis was specifically designed to identify weak genetic differentiation (Hubisz et al. 2009), and the fact that it failed to do so suggests that ecotypes are indeed not genetically differentiated. In summary, the lack of genetic differentiation suggests that the resident and anadromous sympatric ecotypes are not reproductively isolated within those lakes. An alternative explanation would be that the ecotypes are indeed reproductively isolated, but that the divergence is too recent for genetic drift to have led to the accumulation of significant genetic differences between the forms. This is possible given that the habitats currently occupied by those populations were covered by ice until at least  108  10,000 years ago (Dyke et al. 2003), and were thus only recently colonized (Brunner et al. 2001). The microsatellite markers used, however, evolve rapidly (Ellengren 2000) and would be expected to accumulate differences in allele frequency through drift in a few generations. Indeed, there are many documented cases of salmonid populations that have accumulated significant genetic differences in a few generations. In Arctic Char, sympatric ecotypes from three post-glacial lakes in Iceland display significant genetic differences at five microsatellite loci despite having evolved over the last 10,000 years (Gíslasson et al. 1999). On an even more recent timescale, reproductively isolated sympatric ecotypes of sockeye salmon that differ in their spawning habitats (beach vs. river) have accumulated significant genetic differences in only 13 generations (Hendry et al. 2000). Populations of Chinook Salmon O. tshawytscha recently introduced to New Zealand have independently colonized different stream systems and have evolved significant genetic differentiation between streams at microsatellite loci in less than 30 generations (Quinn et al. 2001). In short, reproductive isolation between the ecotypes would have had to occur very recently to leave no trace in the microsatellite allele frequencies. This interpretation thus appears unlikely. !";"1 C4'+'/3'*5>*D',=*0'/'6&3*)&>>'4'/6&,6&5/*&/*<EFF$* Contrary to the observation of genetic homogeneity in the two Cumberland Sound sampling locations, the two sympatric ecotypes from Qinngu Lake (LH001) were found to be weakly genetically differentiated in some analyses. Indeed, the pairwise FST value between the ecotypes was significantly different from zero, and the AMOVA analysis suggested that ecotypes explained a significant amount of the genetic variation, presumably only in LH001. Finally, the FCA results show that there is moderate clustering of the ecotypes in LH001 along the third axis of variation. The STRUCTURE results, in contrast, do not support the presence of genetic differences between the ecotypes. Four alternative (but not necessarily mutually exclusive or exhaustive) interpretations of these results are possible. First, it is possible that the presence of significant genetic differentiation is simply an artifact of the small sample size or of the analysis methods used. The number of samples analyzed from LH001 is indeed quite small (residents: N = 21; anadromous N =  109  29), and it is possible that a few outlier individuals would have a disproportionate effect on average allele frequencies within samples, and thus bias estimates of genetic differentiation. The FCA analysis does indeed suggest that <5 individuals are responsible for most of the differentiation between ecotypes (Fig. 4.2b). When analyzed with methods that define ‘populations’ a priori (the AMOVA and the FST analysis) the presence of a few outlier individuals could then bias the results towards finding more genetic differentiation. The STRUCTURE analysis, however, does not rely on a priori definition of ‘populations’, and did not identify significant genetic discontinuity between the two ecotypes in LH001 (Fig. 4.3). This was true even if the analysis was conducted with Bayesian priors favoring clustering among ecotypes (Hubisz et al. 2009), and therefore suggests that the data do not support the presence of genetic differentiation among ecotypes. Second, the LH001 sampling site differs from the other two sites by the presence of a large lake upstream of the smaller lake where the samples for the present study were collected (see Loewen et al. 2009 for a map; no samples were collected from the upper lake). The two lakes are connected, and anadromous Char are known to utilize the upper lake for spawning. Migration from the ocean to the upper lake, however, requires that they first pass through the lower lake. It is therefore possible that spawning site fidelity leads to genetic differentiation between the individuals spawning in the lower and upper lakes. Sampling of anadromous individuals in the lower lake at the time of the upstream migration (Loewen et al. 2009) would lead to the capture of some individuals from this other putative upstream population. If all the resident individuals sampled, however, originate from the lower lake, this would lead to a pattern of apparent genetic differentiation between ecotypes, when it is in fact the result of genetic differentiation between spawning aggregations. The results of the STRUCTURE analysis conducted on LH001 only support this hypothesis. Indeed, STRUCTURE identified many anadromous individuals that belong to a second genetic cluster, while all resident individuals belong to a single cluster. Samples of anadromous individuals known to originate in the upper and lower lakes would be required to test this hypothesis, but unfortunately were not available for the analyses.  110  Third, it is possible that the two ecotypes from LH001 are reproductively isolated and represent ecotypes using two alternative, genetically-fixed reproductive strategies. This interpretation seems implausible for several reasons. One would expect that if there were two reproductively isolated populations within a lake, the sex ratio of both populations would be close to 1:1 (Fisher 1930). The sex ratio of the resident individuals of LH001, however, is strongly male-biased (Loewen et al. 2010). Additionally, such an interpretation would imply that similar ecological and migratory phenotypes would have evolved in close geographic proximity, but under completely different evolutionary mechanisms (alternative mating tactics vs. genetically-fixed strategies). In most cases of adaptive radiations in post-glacial fishes, however, similar phenotypes do tend to evolve in response to similar ecological pressures (see review by Schluter 1996), but in those systems that have been examined, this parallel evolution tends to be underlain by very similar genetic mechanisms (e.g., for Gasterosteus see Colosimo et al. 2004). While not impossible, the evolution of similar phenotypes via different genetic routes in Arctic Char would thus not seem to constitute the most parsimonious explanation. Fourth, genetic differences observed between ecotypes could result from assortative mating or selection against hybrids if the migratory tactics are partly genetically determined. Available evidence suggests that migratory ecotypes in the genus Salvelinus often have a heritable component. In his extensive study of Arctic Char populations of the Salangen river system in northern Norway, Nordeng (1983) reared the offspring of crosses of resident and anadromous individuals in hatchery conditions. Anadromous parents and resident parents were able to generate offspring of both ecotypes. Anadromous parents, however, had a slightly higher chance of giving rise to anadromous offspring (and vice-versa for resident parents), thus suggesting that the ‘choice’ of migratory tactic has a heritable component. Thériault et al. (2007b) used genetic pedigree reconstructions in natural conditions to estimate the heritability of migratory tactics in Brook Trout. Heritability of life-history tactics was estimated to be between 0.52 and 0.56, indicating a considerable additive genetic variance component for the trait (Thériault et al. 2007b). A heritable component to the migratory tactic could then lead to the accumulation of genetic differences between morphs, but only if a certain degree of assortative mating or selection against hybrids existed (Schluter 2002). There is  111  some evidence from other polymorphic populations of Arctic Char that small-maturing forms and large-maturing forms mate assortatively, although not completely (e.g., Jonsson & Hindar 1982; Gíslasson et al. 1999). This evidence, however, only comes from sympatric lacustrine populations, and I know of no evidence from Arctic Char population of assortative mating according to migratory ecotype. Note also that strong assortative mating would be difficult given the heavily male-biased sex ratio observed in the populations under study. There is no direct evidence of selection against hybrids in polymorphic populations of char, but Jonsson & Jonsson (2001) have suggested that patterns of morphological variation are consistent with this hypothesis (Snorrason et al. 1994; Forseth et al. 2003). Again, however, all evidence comes from sympatric, landlocked ecotypes. In summary, this fourth interpretation is possible, but assumes that (1) mating tactics have a heritable component, and (2) that assortative mating or selection against hybrids exist. Both of those assumptions lack strong empirical support in the present system. !";"# GH&)'/3'*>54*,(6'4/,6&H'*-,6&/0*6,36&3+* The existence of alternative mating tactics in salmonids is well supported by empirical evidence (Fleming 1998; Fleming & Reynolds 2004; Hendry et al. 2004). In Arctic Char, there is also extensive evidence of alternative mating tactics, but most examples are from European populations (Nordeng 1983; Svenning et al. 1992; Rikardsen & Elliot 2000; Klementsen et al. 2003). Populations of Arctic Char with variable migratory phenotypes from the North American Arctic, however, have received less attention. Indeed, while variable migratory phenotypes have been documented in a few locations (Reist 1989; Babaluk et al. 1997), the ecological and evolutionary mechanisms driving the coexistence of the two phenotypes remain elusive (but see Papst 1994). In a previous study of the migratory ecotypes of Arctic Char from south Baffin, Loewen et al. (2010) suggested that this polymorphism could best be understood under the framework of alternative migratory tactics, but that genetic evidence for, or against, reproductive isolation was warranted to strengthen this conclusion. The present study provides such evidence, and concludes that, at least in the two Cumberland Sound populations, the two migratory ecotypes are not reproductively isolated. This is consistent  112  with evidence from other populations of salmonids with variable migratory phenotypes where resident and anadromous individuals do not show significant genetic differences (e.g. Hindar et al. 1991; Thériault et al. 2007b). It is also consistent with the results of Reist (1989) who showed that migratory ecotypes of Arctic Char from the western Canadian Arctic did not differ significantly at two allozyme markers. Power to detect genetic differences using such a low number of loci of a marker system that evolves relatively slowly, however, is probably limited. Failure to detect significant differences at the 14 polymorphic microsatellite markers used in the present study constitutes stronger evidence that the two forms are indeed inter-breeding. The heavily male-biased sex ratio observed in the present populations also constitutes strong evidence of alternative mating tactics. Given the differences in selective pressures associated with vastly different gamete size in both sexes, it is expected that males will predominate in the resident component of the population (Fleming & Reynolds 2004). Female fecundity is strongly affected by size, with larger females producing more and bigger eggs whereas male sperm production may not be related to body size (Hendry et al. 2004). Larger body sizes can help males secure females during competition for mates (Fleming & Reynolds 2004), but smaller nonmigratory males often successfully fertilize eggs by utilizing a ‘sneaking’ tactic (Myers & Hutchings 1987; Thériault et al. 2007a). In most salmonid species, therefore, the ‘resident’ tactic appears to be restricted to males, and reports of females utilizing this tactic are rare (for a few examples from Atlantic Salmon, see Power 1969; Prouzet 1981). This, however, is not the case in the south Baffin Island populations of Arctic Char, where female resident fish are observed despite being considerably less common than their male counterparts: the sex ratio is approximately 20:1 in favor of males (Loewen et al. 2010). This is consistent with observations from European populations of Arctic Char with alternative mating tactics (Nordeng 1983). In the Salangen River system, for example, the sex ratio of males to females is approximately 4:1 (Nordeng 1983). Female residents are also observed in the partially migrating populations of Arctic Char of lakes Storvatn and Rungavatn in northern Norway (Rikardsen & Elliot 2000). The increased prevalence of females utilizing the resident tactic in Arctic Char relative to other salmonids is intriguing but has received little attention. This observation indeed suggests that the balance  113  between the costs and benefits of migration for female Arctic Char differs from that of other salmonids (Hendry et al. 2004). In many populations of salmonids with variable migratory tactics, the larger individuals, or those with higher growth rates, have a tendency to forego anadromous migrations and to attain sexual maturity at an early age (Morinville & Rasmussen 2003; Aubin-Horth & Dodson 2004; Hendry et al. 2004). This, however, is not typically the case in Arctic Char, and studies have demonstrated that large individuals instead have a greater tendency to undergo anadromous migrations (Rikardsen & Elliott 2000). Evidence from back-calculated size at age data for the south Baffin Island populations are consistent with this model for Arctic Char, showing that residents appear to grow slower than their anadromous counterparts (Loewen et al. 2010). Despite the available evidence suggesting that the migratory ecotypes of southern Baffin Island are indeed alternative mating tactics, it should be noted that direct evidence is still lacking. The strongest test for such a hypothesis would be to observe the offspring of anadromous parents developing into resident individuals, or anadromous offspring developing from resident parents. Such observations have been made in European populations of Arctic Char (Nordeng 1983), and in hatchery populations from Labrador (Papst 1994), but not from the present populations. Alternatively, the observation of resident individuals switching to an anadromous tactic later in life would also constitute strong evidence in favor of the alternative mating tactic strategy. Resident individuals have indeed been shown to switch to the anadromous tactic later in life in some Arctic Char populations (Nordeng 1983; Rikardsen & Elliot 2000). The observation that the resident fish from all three south Baffin lakes were on average younger than anadromous individuals (Loewen et al. 2010) suggests that it may indeed be the case. Such patterns, however, could also be the result of increased mortality or faster senescence in the resident fish - although mortality may be expected to be higher for anadromous fish facing more severe predation and fishing pressures in the marine environment. !";"! I-.(&3,6&5/+*>54*>&+J'4&'+*-,/,0'-'/6* Jonsson & Jonsson (2001) suggested that sympatric Arctic Char morphs, which are typically genetically distinct from one another, should be managed as separate species.  114  The genetic results of the present study argue instead that the sympatric ecotypes of Baffin Island Arctic Char are not separate populations, but instead are two components of the same population. Recognizing the presence of this life-history diversity for management is important, but they should not be considered different stocks. Anadromous Arctic Char populations of southern Baffin Island are currently the target of subsistence fishing by the Inuit from the communities of Kimmirut and Pangnirtung (Priest and Usher 2004). In addition, the anadromous stocks of Arctic Char from the Cumberland Sound support a small-scale commercial fishery (Read 2000). Assuming that the sympatric ecotypes do indeed constitute a case of alternative mating tactics, it can probably be safely assumed that the resident component of the population reproduces and thus contributes demographically to the overall population. Indeed, small-maturing males have been shown to reproduce with anadromous females in many populations of salmonids (Myers & Hutchings 1987; Thériault et al. 2007a). If, however, the resident component of the population has a tendency to produce more resident offspring, this could reduce their contribution to the recruitment of the fishery, which targets anadromous individuals only, and could therefore have negative economic consequences for the fishery (Myers 1984). Mating with residents would still contribute to maintaining genetic diversity in the populations, and the presence of variation in life history traits between exploited populations has been shown to have a positive effect on their longterm viability (Schindler et al. 2010). An interesting possibility is that increased fishing pressure on the anadromous component of the population may over the long-term increase the relative fitness of the resident tactic thus resulting in a higher frequency of this tactic in the population. Fisheries-induced evolutionary changes are now well documented in a large number of species (Stokes & Law 2000; Hutchings & Fraser 2008), including salmonids (Quinn et al. 2007; Hard et al. 2008). In the case of migratory tactics, theory shows that the frequency of alternative mating phenotypes in a given population represents an evolutionary stable continuum determined by the relative fitness of the two tactics (Hutchings & Myers 1994). Fishing the anadromous individuals exclusively, by increasing the costs of marine migrations, would thus increase the relative fitness of the resident tactic. Under such a scenario, the reaction norm underlying the migratory tactics  115  could evolve to change the growth threshold determining the switch in tactic such that residency is a more common outcome. This possibility was specifically explored in a model built based on empirical data from exploited populations of brook trout that display variable migratory tactics (Thériault et al. 2008). The model found that after 100 years of fishing pressures targeting exclusively the anadromous component of the population, the probability of migrating decreased significantly (Thériault et al. 2008). In conclusion, exploitation of anadromous Arctic Char in south Baffin Island could have long-term consequences for the relative frequency of the two ecotypes. !";"; ?5/3(@+&5/+* The results of this study show that reproductive isolation among sympatric migratory ecotypes of Arctic Char from South Baffin Island is weak to non-existent. This genetic evidence, when combined with other sources of evidence from previous studies, suggests that the observation of sympatric migratory ecotypes of Arctic Char is consistent with the alternative mating tactics framework. Because the fishery for Arctic Char solely targets the anadromous component of the populations, understanding the evolutionary and genetic processes responsible for this variation has some management implications. Future work should extend exploration of the genetic mechanisms of migratory ecotypes to other variable Arctic Char populations throughout the Canadian Arctic to verify the generality of the present conclusions. Gaining a better understanding of the relative abundance of the two migratory tactics would also appear useful, as they may respond to fishing pressures or other environmental changes.  116  Table 4.1. Sampling locations with geographic coordinates and sample sizes for each ecotype of Arctic Char (Salvelinus alpinus) per year for each location. Sampling location Iqaluggarjuit  Geographical coordinates Latitude  Longitude  Year  66° 43.2'  66° 34.7'  (PG082) Qasigiat (PG015)  Qinngu, Blandford  Sample size  66° 19.1'  63! 31" N  64° 37.5'  71! 18" W  Bay (LH001)  Resident  Anadromous  2002  4  41  2004  18  47  2009  11  0  2003  12  33  2004  8  30  2009  6  42  2004  6  0  2005  15  29  Table 4.2. Details of the primers and PCR reactions used for each of the four multiplexes used to assess genetic variation of resident and anadromous Arctic Char (Salvelinus alpinus). Multiplex  mpAC1  mpAC2a  mpAC2b  mpAC3  mpAC4  b  b  [Primer]  [MgCl2]  [dNTPs]  (!M)  (mM)  (!M)  1.50  200  1.00  2.00  200  0.50  1.50  200  0.50  2.00  200  0.50  2.00  200  1.00  Locus  Dye  Reference  Sco200  VIC  DeHaan & Ardren, 2005  0.40  Smm22  NED  Crane et al., 2004  0.40  Sco220  6-FAM  DeHaan & Ardren, 2005  0.50  Sco215  PET  DeHaan & Ardren, 2005  0.30  Sco212  6-FAM  DeHaan & Ardren, 2005  0.50  Sco218  VIC  DeHaan & Ardren, 2005  0.50  Sfo18  NED  Angers et al., 1995  0.20  Sco202  PET  DeHaan & Ardren, 2005  0.20  Smm21  VIC  Crane et al., 2004  0.16  OMM1128  VIC  Rexroad et al., 2001  0.16  Smm24  NED  Crane et al., 2005  0.20  OtsG253b  VIC  Williamson et al., 2002  0.12  SSOSL456  6-FAM  Slettan et al., 1997  0.50  OMM1105  PET  Rexroad et al., 2002  0.20  OtsG83b  6-FAM  Williamson et al., 2002  0.50  Smm17  NED  Crane et al., 2004  0.16  Sco109  VIC  Shaklee, 2003  0.50  Sco216  PET  DeHaan & Ardren, 2005  0.40  a  AmpliTaq Gold® DNA Polymerase with Gold Buffer and MgCl2 solution from Applied Biosystems  b  mpAC2a and mpAC2b were combined post-PCR for the genotyping  Taq  a  117  Table 4.3. Summary statistics of microsatellite variation for each ecotype of Arctic Char (Salvelinus alpinus) per sampling location. The data were analyzed first for each year separatly, but the years were found not to differ and were thus combined in all subsequent analyses. Shown is the number of alleles (NA), allelic richness (AR , i.e. number of alleles corrected for differences in sample size only calculated for the ‘years combined’ samples), observed heterozygosity (HO), and Nei’s (1987) unbiased expected heterozygosity (HE). Years separate Sampling location Iqaluggarjuit (PG082)  Life-history ecotype Anadromous  Resident Qasigiat (PG015)  Anadromous  Resident Qinngu, Blandford Bay (LH001)  Anadromous Resident  Years combined  Year  NA  HO  HE  2002  12.71  0.74  0.80  2004  12.79  0.71  0.82  2002  4.79  0.74  0.83  2004  9.64  0.70  0.81  2009  8.57  0.78  0.81  2003  10.14  0.82  0.79  2004  11.29  0.78  0.79  2009  12.14  0.77  0.78  2003  7.86  0.78  0.78  2004  8.36  0.82  0.83  2009  5.29  0.68  0.70  2005  10.93  0.65  0.75  2004  5.86  0.74  0.76  2005  8.86  0.78  0.78  NA  HO  HE  AR  15.29  0.72  0.81  10.71  11.79  0.74  0.81  10.61  14.50  0.79  0.79  9.76  11.07  0.77  0.79  10.26  10.93  0.65  0.75  9.82  9.93  0.77  0.78  9.84  118  Table 4.4. Semi-matrix of pairwise FST (Weir & Cockerham, 1984) values between all sampling locations and ecotypes of Arctic Char (Salvelinus alpinus). Asterisks denote significant population differentiation after correction for multiple comparisons (adjusted  ! = 0.003). PG082anad  PG082res  PG015anad  PG015res  LH001anad  LH001res  0  0.004  0.027*  0.025*  0.075*  0.064*  0  0.031*  0.026*  0.068*  0.068*  0  -0.001  0.079*  0.072*  0  0.079*  0.072*  0  0.039*  PG082anad PG082res PG015anad PG015res LH001anad LH001res  0  Table 4.5. Results of the analysis of molecular variance (AMOVA) examined among sampling locations and ecotypes of Arctic Char (Salvelinus alpinus). Sum of  % variance  Source of variation  df  Squares  explained  P-value  Among sampling locations  2  72.3  4.88  0.063  Between ecotypes within sampling location  3  13.2  0.47  0.012  598  1915.3  94.65  <0.0005  Within ecotypes  119  Figure 4.1. Map showing the location of the three lakes with the sympatric migratory ecotypes of Arctic Char (Salvelinus alpinus) in southeastern Baffin Island: (1) Iqaluggarjuit Lake, Shark Fiord, PG082; (2) Qasigiat Lake, Ptarmigan Fiord, PG015; (3) Qinngu Lake, Blandford Bay, LH001.  Figure 1  120  Figure 4.2. Results of the factorial correspondence analysis (FCA) conducted in GENETIX for samples of Arctic Char (Salvelinus alpinus). (a) Clustering of all the samples along the three most important axes of variation. The triangles are individuals from LH001, the circles individuals from PG015, and the squares individuals from PG082. The empty symbols represent anadromous individuals and the filled symbols are resident individuals. (b) Results of the same FCA, with only the LH001 sampling locations shown. The axes have been re-organized to help visualization of the slight differentiation between anadromous and resident individuals along axis 3 in LH001. (a)  (b) PG082  LH001 only Residents  Axis 2  Axis 2 (1.75%)  LH001  Anadromous  PG015 Axis 1 (2.43%)  Axi  s  % .49 3 (1  )  Axis 3  Ax  is  1  Figure 2  121  Figure 4.3. Results of the STRUCTURE analysis on all sampling locations of Arctic Char (Salvelinus alpinus). (a) Mean ln probability of the data given each value of K (i.e. number of genetic cluster). The arrow indicates the most likely number of genetic clusters (3) according to this method (Pritchard et al., 2000). (b) Values of delta K (Evanno et al., 2005) for each value of K. The arrow indicates the most likely number of genetic clusters (3) according to this method. (c) Probability of assignment (q) to each of the three genetic clusters for each individual. (a)  (b)  K  (c)  K  Probability of assignment q  Iqaluggarjuit PG082  anadromous Anadromous  residents Residents  Qasigiat PG015  anadromous Anadromous  Qinngu LH001  residents anadromous residents Residents Anad. Res.  Figure 3  122  Figure 4.4. Results of the STRUCTURE analysis on single sampling locations of Arctic Char (Salvelinus alpinus). (a) Mean ln probability of the data (top panel) and delta K (bottom panel; Evanno et al., 2005) values for each value of K (i.e. number of genetic cluster) in each sampling location. The arrows indicate the most likely number of genetic clusters according to each method (Pritchard et al., 2000; Evanno et al., 2005). Note that because delta K is not defined for K = 1, there is no arrow for most likely delta K in Qasigiat. (b) Probability of assignment (q) to each of the three genetic clusters for each individual. This is shown only for two sampling locations, because the most likely number of genetic clusters (K) in Qasigiat was one. In both cases, life-history ecotype is  7  9  1  3  5  7  9  1  3  5  7  9  -2700  1  3  5  7  9  1  3  5  7  9  1  3  5  7  9  0.1 0.5  0.3  K  5  7  Iqaluggarjuit PG082 9 1 3  K  1  3  5  7  9  1  3  5  7  9  1  3  5  7  9  K  0  0.1  3  Qinngu LH001  -3300  0.7 0.5 0.3 0.7  20  3010 200 10  K  20 40 600 8020 40 60 80  5  -3100 -3300 -2900 -3100 -2700 -2900  -8000 -7600 3  0  delta K  delta K  Qasigiat PG015  -8000  1  1  Probability of assignment q  (b)  -7600  -6800  -7200-7600-6800-7200  Iqaluggarjuit PG082  30  (a)  -7600  Mean LnP(K) Mean LnP(K)  not a good predictor of cluster assignment.  5  7  9  Qinngu 3 5 LH001 7 9  1  K  Anadromous 1  Residents 2  K  Anad. 1  Res. 2  123  ! "#$%&'()#$* The work presented in the previous chapters spanned several temporal and spatial scales, dealing with processes of post-glacial dispersal over a continental scale, to contemporary dispersal on a regional scale, to the evolution of the potential for dispersal within populations. The two unifying themes of the research were the study of dispersal, and the model organism used: the Arctic Char. Still, I hope that some of my work will have implications that transcend the taxonomic boundaries of this one species. In most of the previous chapters, I have discussed some implications of my work. The chapters, however, are written with the goal of being published, and there is a limit beyond which one is not allowed to wave his/her hands in scientific journals. In the discussion that follows, I offer some other implications of my work, taking more liberties to broaden the discussion even if, in some cases, the implications go beyond the currently available data. These should therefore be treated as hypotheses, not conclusions. I have organized the discussion about the implications of my work from the more specific to the more general, rather than by chapter, to further emphasize the conceptual unity of the work. It is perhaps a cliché to conclude a dissertation by asserting that it generated more questions than it answered. In the case of the present thesis, however, this happened almost by design. When I started my work, there was very limited information on Canadian anadromous Arctic Char into which an evolutionary/molecular ecologist could sink his/her teeth. Indeed, much of the molecular work that had been done prior to mine was concerned with taxonomy (e.g., Taylor et al. 2008) or with lacustrine populations (e.g., Power et al. 2009). Basic phylogeographic work had also been conducted (Wilson et al. 1996; Brunner et al. 2001), but my work has shown that the phylogeographic work was incorrect in some cases (see Chapter 2). Much of my work, therefore, has been setting a basis for future work on Canadian Arctic Char utilizing molecular markers. In the last section of this conclusion, therefore, I discuss some questions that were raised by my work, and I outline how future work may answer them. Finally, I attempt to unify the various aspects of my work under the overarching goal of predicting the evolutionary response of Arctic Char to climate change. This, I hope, will form the guiding principle  124  of my research program for the coming years. Dispersal and gene flow will have an important role to play in adaptation to climate change, but other factors will also need to be considered. I therefore discuss some of those other factors in the last section. I also hope that such a discussion will make it apparent that the research in the present dissertation has built some of the foundations required for answering this difficult but important question.  !+, -./&)%01)#$(* !"#"# $%&'()*+(,-./0,1/21)+()/34*1/5(,',67/ As discussed in the Introduction, the Arctic Char is a species of great scientific interest (Skúlasson et al. 1999; Jonsson & Jonsson 2001). It is also a species of tremendous subsistence, commercial, and cultural importance to the people of northern Canada (Balikci 1980; Priest & Usher 2004; Roux et al. 2011). Some of the findings of my work therefore have broad implications. The findings presented in Chapter 2, by revealing inaccuracies in previous studies (Brunner et al. 2001), have largely re-written the phylogeographic history of Arctic Char in North America. This first led to a re-evaluation of the amount of mitochondrial DNA variation present in the species, previously believed to be much higher (Brunner et al. 2001). Assuming that mtDNA reflects variation at other loci, this has implications for forecasting the evolutionary response of the species to environmental change (see below). The results have also re-drawn the maps of the distribution of the Beringian and Arctic ‘clades’ in the western Arctic. This is especially important because the western North American Arctic is the area where Dolly Varden and Arctic Char co-exist. While the debate regarding the taxonomic distinctiveness of those two species is probably settled (Reist et al. 1997; Taylor 2008), much work remains to be done to understand the frequency with which those two species hybridize in the wild. Indeed, others are currently working on such questions (S. May-McNally, Univ. of BC, unpubl. data; M. Kowalchuk, Univ. of Manitoba, unpub. data). The understanding of the recent evolutionary history of Arctic Char in the region gained through my results will help set  125  the stage for an understanding of contemporary hybridization between these closely related species. In Chapter 3, I have presented the most direct evidence to date that Arctic Char have an increased propensity to disperse during the years that they do not reproduce. This behaviour had been previously inferred from patterns of tag recovery, but tagging studies have two limitations that my study did not suffer from. First, it is difficult to assess the reproductive status of tagged fish because this usually requires direct examination of gonads, a lethal procedure. My work had access to this data because all genetic samples were collected as part of fishery management survey that collected this data for each individual. This allowed us to more directly test the hypothesis that dispersal propensity was dependent on reproductive status. Second, it is impossible to determine the natal river of a fish in a tagging study, and it is often assumed that the location of first tagging is the ‘natal’ river. I circumvented this problem by assigning fish on the basis of their multilocus genotypes to their most likely river of origin, as determined with reference samples of juveniles of known origin. Furthermore, the logistically less challenging approach using molecular tools compared to physical tagging allowed us to more or less systematically sample most habitats around Cumberland Sound. This limited the problems experienced with other studies, which relied on fisheries-dependent tag recoveries and could thus not control the spatial extent of the sampling. My study therefore generated more accurate dispersal kernels. It should be noted, however, that the precision of the assignment procedure based on microsatellites was less than ideal, and estimates of dispersal rates and dispersal kernels depended strongly on the analysis method used. Still, the study demonstrated the promise of utilizing molecular approaches in the logistically challenging conditions of the Canadian Arctic. Another contribution of this work was the exploration of the potential consequences of this behaviour. The potential consequences of increased dispersal during overwintering years had not escaped biologists, but their explanations emphasized the importance of this ‘exploratory behaviour’ for the colonization of novel habitats (e.g. Johnson 1980). While this may very well have been important following the last glaciation, it is unclear why such behaviour would be maintained in the interglacial (unless evolutionary stasis is invoked). Instead, I propose that this behaviour has  126  interesting implications for local adaptation. Traditional ecological knowledge shared by the Inuit fishermen of Pangnirtung indicates that there is tremendous phenotypic variation among populations of Arctic Char from around Cumberland Sound. Whether this variation constitutes local adaptation remains to be tested, but my work shows that this is a possibility given a realistic range of parameter values employed in the gene flowselection modeling (Chapter 3). The existence of local adaptation on such small spatial scales would be difficult to reconcile with the high dispersal rates documented from my study and from other tagging studies. The demonstration that most observed dispersal events involve individuals that are not in reproductive condition provides the solution to this apparent paradox. In Chapter 4, I tested alternative hypotheses regarding the maintenance of alternative migratory ecotypes in sympatry. My results suggest that this polymorphism is the result of alternative mating tactics rather than genetically fixed alternative strategies. While this is a widespread and well-documented phenomenon in salmonids (Fleming 1998; Fleming & Reynolds 2004) and in European populations of Arctic Char (e.g. Nordeng 1983; Rikardsen & Elliot 2000), little was known about this polymorphism in the North American range of Arctic Char. The evidence presented in this chapter therefore highlights the possibility that this plastic trait shares a common genetic mechanism across the species’ range – something that future work should test. !"#"8 $%&'()*+(,-./0,1/.*'%,-(9/5(,',67/ Anadromous salmonids are known for their homing abilities, but straying among populations also occurs to varying degree in all species. Variation in straying among species is striking, and some species, like the pink salmon, have an elevated propensity to stray. Assuming, therefore, that straying evolves in response to natural selection, this variation in straying rates offers the possibility of using a comparative approach to infer selective factors that promote the evolution of dispersal and philopatry in salmonids (Hendry et al. 2004). The work presented in this thesis contributes to this endeavor by providing otherwise rare estimates of levels of dispersal in anadromous Arctic Char. Indeed, the rarity of such estimates precluded their inclusion in previous comparative analysis (it was excluded from the study of Hendry et al. 2004). This is unfortunate  127  because the species inhabits extreme environments and would therefore make it a very useful data point to add in a comparative analysis. For instance, Hendry et al. (2004) concluded from their comparative analysis that one of the most likely drivers of increased propensity to stray in salmonids was a temporally unpredictable environment. The evidence of high dispersal in anadromous Arctic Char from the Cumberland Sound - and indeed from all the studies reviewed in the introduction - therefore lends support for this hypothesis, because the Arctic rivers inhabited by Arctic Char are probably more unpredictable than most other habitats inhabited by other species of salmonids (Power & Power 1995; Power 2002). Further, I would like to formulate the following hypothesis: the fact that overwintering individuals disperse at a higher frequency implies that natural selection favors such a behaviour. Indeed, the fact that spawning individuals home with higher precision demonstrates that the species possesses the machinery for precise homing and that this increased dispersal is not merely the result of orientation errors. What then would be the selective factors that would promote such behaviour? The hypothesis I favor is that this behaviour allows Arctic Char to hedge their bets against the unpredictable Arctic environment through the utilization of multiple habitats throughout their lives. An added benefit of this behaviour would be that individuals are then familiar with the surrounding environments, which may afford them an advantage during the years where they find access to their natal river is blocked for a reason or another. This idea assumes, among other things, that char retain a memory of the places they visited in the past. While there are no data supporting this interpretation of the behaviour, I was reminded of this quote from Tom Quinn speaking about homing salmon: “one will seldom win money betting against the sensory systems of animals” (Quinn 2005, p. 102). I have also pointed out on a few occasions that such behaviour would allow Arctic Char to hedge their bets against an unpredictable environment while also minimizing the impact of dispersal on local adaptation – believed to be a key factor promoting homing in salmonids (Hendry et al. 2004). Again, this assumes that populations are indeed locally adapted – a claim not yet supported by empirical evidence – but would suggest that the overwintering dispersal behaviour of non-breeding char may be a way for them to  128  optimize habitat use in the face of potentially competing demands of finding habitats best suited for overwintering, versus those best suited for spawning and juvenile rearing. There remains an alternative hypothesis that would explain the evolution of this behaviour without invoking a benefit for the utilization of non-natal habitats, but rather simply in terms of reduced costs of migration. It is possible that individuals foraging in the ocean occasionally venture at greater distances from their natal river in search of suitable prey. It may then be less costly for those individuals to overwinter in a nearby lake than to migrate back to their river of origin. The costs of spawning in a non-natal river, however, may be much higher, and the costs of migrating back to the natal river may be less than those of spawning in a non-natal habitat for individuals destined to spawn that year. What is known about Arctic Char marine migrations (see Introduction for a complete review), however, suggests that only a very small number of individuals venture at great distances each year. My own data on dispersal suggest that most dispersal events occur among adjacent rivers. It thus seems unlikely that the costs of returning to a natal habitat would exceed the costs of overwintering in an unfamiliar habitat. On the other hand, the costs of overwintering in unfamiliar habitats may not actually be that high and Arctic Char may rely on the fact that great numbers of individuals are already migrating up those other rivers. Such use of ‘public information’ has never been demonstrated in Arctic Char, but other animals can use such cues to inform dispersal behaviour (Valone & Templeton 2002; Danchin et al. 2004). In Chapter 4, I pointed out that females utilizing the resident tactic are more common in European (Nordeng 1983) and Canadian (Loewen 2008) populations of Arctic Char than in other salmonids species with variable migratory tactic (Fleming 1998). The question of why female Arctic Char choose the residency tactic at an increased level is unknown. I can see two alternative, but not necessarily mutually exclusive, explanations. First, assuming that the trait is a threshold trait with some heritable component (Roff 1996), it is possible that this heritable component is strong enough that some females adopt the residency tactic despite it being associated with lower fitness. According to this explanation, therefore, the trait is not the result of selective pressures, but of maladaptation perhaps due to genetic constraints. Second, it is possible that the relative costs and benefits of the residency tactic for females are  129  different in Arctic Char compared to other salmonids, and that some females may indeed maximize their fitness by adopting the resident tactic. Indeed, there exist many differences between Arctic and temperate freshwater systems that may lead to decreased costs or increased benefits to the residency tactic. A smaller-bodied female may, for example, use less energy while over-wintering during the long ice-cover period of most over-wintering lakes. The examination of those relative costs and benefits in Arctic char may also shed some light onto the evolution of alternative migratory tactics in other salmonids. !"#": $%&'()*+(,-./0,1/+4;/.+<97/,0/9(.&;1.*'/ The ability of organisms to disperse across the landscape is an important determinant of the geographic distribution of genetic variation, and studies that compare patterns of genetic diversity between species that differ in dispersal propensity have provided forceful empirical evidence for this assertion (Bohonak 1999). Comparisons of patterns of genetic diversity in anadromous and non-anadromous freshwater fish species have also confirmed the hypothesis that overall genetic diversity would be higher in the former, but genetic divergence among populations would be higher in the latter species (DeWoody & Avise 2000). Few studies, however, have compared the patterns of genetic variation among anadromous and non-anadromous populations within species (for a few exceptions, see Mäkinen et al. 2006 and Tonteri et al. 2007). Intraspecific comparisons provide the advantage of ‘controlling’ for confounding variables that interspecific comparisons may necessarily suffer from. Indeed, species differ not only in their dispersal propensity, but also in their biology and their evolutionary history, in ways that could potentially confound such interspecific comparisons. While the confirmation of the patterns previously observed in interspecific comparisons is not a surprise, the data collected in Chapter 1 strengthen the conclusion that dispersal propensity is one of the major forces determining the distribution of genetic variation in wild species, by showing that the patterns hold in intraspecific comparisons. Further, the fact that the Arctic char populations under study here are evolutionarily very young shows that these differences were established rapidly.  130  Some of the most important developments in dispersal research (and indeed in much of ecology and evolution) have ultimately been brought about by observations of the complex behaviour of species in natural environments. These observations contribute to refining hypotheses and to ensuring that available theoretical models have enough generality to be useful in complex natural systems. In terms of complexity, the Arctic Char does not disappoint. Most importantly, I think that my research illustrates well the inadequacy of the simplistic definitions of dispersal used by many researchers (e.g., Ronce 2007; Clobert et al. 2009; but see Bowler & Benton 2005). Indeed, animals disperse for purposes other than breeding, and the evolutionary forces driving these behaviors can, I would argue, be understood under the already solid theoretical framework provided by ‘conventional’ dispersal research. For instance, many species migrate to overwintering sites (monarch butterflies: Calvert & Brower 1986; amphibians: Holenweg & Reyer 2000; bats: Kurta & Murray 2002) and little work has been done on dispersal between overwintering sites. This is unfortunate because such dispersal would provide a great opportunity to test some ideas regarding the ultimate causes of dispersal. Indeed, alternative hypotheses like inbreeding avoidance are automatically eliminated with such behaviors, leaving fewer alternatives to be tested. Hopefully, my work will serve as an illustration of how these alternative dispersal behaviors can ultimately help us understand the factors selecting for movement and dispersal.  !+2 3'1'45*6)45%1)#$(* I will point out here some outstanding questions raised by my results, and provide suggestions of how they may be addressed. I will conclude by integrating the various aspects of my current and proposed research under the overarching framework of predicting the response of Arctic Char to global environmental change. !"8"# =<+.+*-9(-6/><;.+(,-./ In Chapter 1, I have provided evidence that not only did Arctic Char recolonize its current range very recently, but that it probably did so from a very small refugial population. These two facts would therefore make the Arctic Char an interesting species  131  to explore questions regarding the pace at which local adaptation can evolve. There are countless examples of evolution happening on very short time scales (e.g., Hendry & Kinnison 1999; Hairston et al. 2005), but can it do so when little genetic variation is present and in species with long generation time and where populations are linked by gene flow? This is a crucial question to answer if we are to rely on evolutionary change to save Arctic species from extinction in the face of climate change (Lavergne et al. 2010; Hoffman & Sgrò 2011). There are good indications that anadromous populations of Arctic Char vary phenotypically in ways that appear adaptive (Power et al. 2005), but I am aware of no studies that have demonstrated the presence of local adaptation in them (for examples from landlocked populations, see Larsson et al. 2005; Conejeros et al. 2008). The best tests of local adaptation involve reciprocal transplant experiments (Kawecki & Ebert 2004). Tests of local adaptation in anadromous salmonids have been conducted in aquaculture settings (Taylor 1991; Fraser et al. 2011), but they often vary a single environmental variable (e.g., pH, temperature) and thus fail to reproduce the environmental complexities faced by wild populations. The advent of genomics tools, and in particular of next-generation sequencing has made it possible to investigate regions of the genome that are divergent among populations (Nosil et al. 2009). Correlating these regions of genomic divergence with environmental characteristics may thus allow inference of local adaptation (Allendorf et al. 2010; Manel et al. 2010). Genomic approaches would also allow further exploration of another interesting result stemming from Chapter 2: the hybridization between the Arctic and Atlantic glacial races in eastern North America. My work involved a very limited number of markers, but suggests that Atlantic lineage alleles have penetrated the Arctic Archipelago as far as Hudson Bay and Southern Ellesmere Island. Furthermore, there are clear differences between the distribution of mtDNA and nuclear markers. What can be learned from such patterns of differential introgression? For instance, one would expect that neutral parts of the genome would flow freely among populations while other parts would remain differentiated as a result of divergent selection (Wu 2001; Nosil et al. 2009). The size, number, and distribution of such ‘islands of divergence’ can offer clues as to the nature of the evolutionary forces acting upon the genome. In some circumstances, it may be possible to identify candidate genes under selection, that are found in the regions of  132  genomic divergence (Allendorf et al. 2010). While natural selection may promote or maintain divergence in parts of the genome, it is also possible that globally advantageous alleles flow faster than neutral alleles among populations. Such enhanced introgression of globally favorable alleles has been demonstrated for growth genes during hybridization between hatchery-produced and wild Brook Trout (Lamaze et al. 2012). This is an especially interesting scenario in the present case. Indeed, the penetration of alleles from a more southern part of the species’ range may provide adaptive genetic variation to more northerly populations trying to adapt to warming conditions. In short, the presence of this natural hybrid zone, if studied using the modern tools of next-generation sequencing, offers a tremendous opportunity to study the evolutionary forces promoting genomic divergence in this species. In the introductory chapter of this dissertation, I reviewed the existing knowledge on the dispersal and migratory behaviour of anadromous Arctic Char, and pointed out several limitations with existing studies. I have already discussed above how the data in Chapter 3 addressed some of those issues, but several gaps remain in our understanding of Arctic Char migratory behaviour. First, while there are good indications that time spent at sea and movement at sea is at least in part determined by ocean temperatures (Berg & Berg 1993; Rikardsen et al. 2007), it remains unclear how among-year variability in climate influences distances moved and overwintering decisions. Multi-year tagging studies that follow a similar protocol in different geographical regions would appear extremely valuable in deciphering the influence of environmental conditions on Arctic Char movement. This will constitute crucial information for the task of predicting the effects of global change on Arctic Char dispersal, which will in turn greatly influence the overall response of the species. Second, while my work provides good indication that Arctic Char disperse more than most other salmonids (especially in the years that they overwinter), the assignment methods I used lead to a considerable degree of uncertainty regarding the absolute value of dispersal. It would therefore be useful to confirm those estimates using tagging studies that would be specifically designed to answer this question. I have had the good fortune of being approached to work on the Ocean Tracking Network, a multi-institution project that will allocate funds to do large-scale telemetry work on Arctic Char in several parts of the Canadian Arctic. This exciting  133  opportunity will allow me to close some of those gaps in our understanding of Arctic Char migrations in the coming years. As mentioned earlier, my work has shown that the use of genetic markers offers great promise in the study of the complex migratory behaviour of Arctic Char in the logistically challenging environments it inhabits, but clearly our power of inference is imperfect. Indeed, future studies would benefit from the use of new technologies that may allow more confidence in discriminating the genetically similar populations of Arctic Char. Next-generation sequencing now offers the possibility of performing ‘genotyping by sequencing’, or in other words, rapidly scoring individuals at thousands of single nucleotide polymorphisms (SNPs). Microsatellites have traditionally been the marker of choice for assignment studies (including parental assignment) because each locus can have many alleles (up to 40 in the case of the markers I have used). While each SNP is far less informative than a microsatellite locus, studies have shown that a large number of SNPs (>100) can perform better than a typical number of microsatellite loci (e.g., Glover et al. 2010). The combination of telemetry work with precise genetic assignment relying on next-generation sequencing data offers the promise of generating precise estimates of dispersal rates for anadromous Arctic Char. Such data will be sorely needed if we are ever to even attempt predicting the fate of Arctic Char in the face of a changing climate. This data will also address more immediate concerns regarding the management of this precious subsistence and commercial resource.  !"8"8 ?(''/;@,'<+(,-*17/*9*&+*+(,-/4;'&/21)+()/)4*1/),&;/A(+4/)'(%*+;/)4*-6;B/ Climate change is rapidly modifying the earth’s ecosystems (Parmesan & Yohe 2003; Parmesan 2006), and nowhere is this change occurring faster than in the Arctic regions (IPCC 2007). Faced with such change, species can respond in one of three ways: move their range pole-ward, adapt to new conditions, or go extinct. For Arctic species already inhabiting the northernmost tips of land, adaptation is thus imperative if they are to avoid extinction. Predicting the fate of Arctic species in the face of climate change thus offers a daunting challenge for evolutionary biologists (Williams et al. 2008; Hoffman & Sgrò 2011). Is it unrealistic to think that we may even be able to accomplish such a feat?  134  And even before we can attempt to predict the evolutionary response of Arctic species to climate change, is it even realistic to think that Arctic species will be able to adapt to such rapidly changing conditions? Recent reviews have shown that anthropogenic environmental disturbances, including climate change, have resulted in adaptive evolution in numerous species (Hendry et al. 2008; Hoffman & Sgrò 2011), and this is even true of a few Arctic species (Réale et al. 2003). In the Introduction of this dissertation, I reviewed the various ways in which dispersal and gene flow can help or hinder adaptation. I will therefore not repeat in details the myriad ways dispersal and gene flow interact with the adaptive process (Garant et al. 2007). Those theoretical considerations will help guide our understanding of the potential effects of dispersal. Empirically, however, a more detailed understanding of dispersal will be required to really predict how it will influence the adaptive process. For example, are there important barriers to dispersal in the species range? Is there a directional bias in dispersal? If dispersal is biased towards the south (e.g., if individuals tend to follow warm water) this could swamp pre-existing adaptations in southern populations. If, on the other hand, dispersal is biased northward, dispersal could help move beneficial alleles across the species range. More importantly, however, directional biases in dispersal will not necessarily lead to biases in gene flow. For instance, we may expect adaptive alleles to flow directionally across the species range even in the absence of a bias in dispersal. And as the range of the species will shift northward, we would expect dispersal to have different consequences for the leading and trailing edges of the range (Hampe & Petit 2005). The data presented in this thesis will provide a foundation on which to build future work on this problem, and the combination of telemetry work with genomics tools, as proposed earlier, may provide useful insight. While dispersal and gene flow will determine the response of specie to climate change, climate change may also change ‘natural’ patterns of dispersal of some species and put previously isolated lineages in contact (e.g. Garroway et al. 2010) – a phenomenon that may be particularly important in the Arctic regions (Kelly et al. 2010) and may be further influenced by human-induced changes to patterns of dispersal and gene flow in natural populations (Crispo et al. 2011). Because of the important effect of dispersal and gene flow in determining the adaptive response of species to environmental  135  changes, understanding how these environmental changes will in turn affect dispersal and gene flow is also essential. The study of the environmental correlates of dispersal that I proposed earlier when describing future telemetry work addresses part of this goal. Understanding how anadromy may evolve and increase in frequency in response to changes in oceanic conditions will also be crucial, since increased anadromy will also increase potential for dispersal in parts of the species range where anadromy is currently rare.  !+7 "&#()$8*45.049(* The Arctic Char is a fascinatingly complex species of great subsistence, commercial, cultural, and scientific importance. As a quintessentially cold-adapted species, it will face important threats from global change in the near future. Because there is a fundamental limit to how far Arctic species can move their range pole ward, predicting their potential evolutionary responses to climate change is a priority. The goal of predicting the adaptive response of the species to climate change, however, should be considered with a healthy dose of skepticism. Thankfully, Arctic freshwater systems are relatively species-poor, and understanding community changes brought about by climate change is therefore within the realm of the possible. In addition, the Arctic environment remains relatively pristine, and the confounding effects of other anthropogenic stressors are therefore less important. The main drawback of Arctic systems, therefore, is the absence of baseline data. My dissertation, I think, provided some of this badly needed baseline data and has built the foundations of what I hope to be years of work to come.  136  :);&)#840/<=* Abbott, K. C. 2011. A dispersal-induced paradox: synchrony and stability in stochastic metapopulations. Ecology letters 14: 1858-1169. Abbott, R. J., Smith, L. C., Milne, R. I., Crawford, R. M. M., Wolff, K. & Balfour, J. 2000. Molecular analysis of plant migration and refugia in the Arctic. Science 289: 1343-1346. Adkison, M. D. 1995. Population differentiation in Pacific salmons: local adaptation genetic drift, or the environment? Can. J. Fish. Aqu. Sci. 52: 2762-2777. Alekseyev, S. S., Bajno, R., Gordeeva, N. V., Reist, J. D., Power, M., Kirillov, A. F., Samusenok, V. P. & Matveev, A. N. 2009. Phylogeography and sympatric differentiation of the Arctic charr Salvelinus alpinus (L.) complex in Siberia as revealed by mtDNA sequence analysis. Journal of Fish Biology 75: 368-392. Allendorf, F. W., Hohenlohe, P. A. & Luikart, G. 2010. Genomics and the future of conservation genetics. Nature Reviews Genetics 11: 697-709. Anderson, R. & Neuman, R. M. 1996. Length, weigth, and associated structural indices. In: Fisheries techniques, (Murphy, B. R. & Neuman, R. M., eds.). pp. American Fisheries Society, Bethesda, Maryland. Angers, B. & Bernatchez, L. 1998. Combined use of SMM an non-SMM methods to infer fine structure and evolutionary history of closely related brook charr (Salvelinus fontinalis, Salmonidae) populations from microsatellites. Molecular Biology and Evolution 15: 143-159. Angers, B., Bernatchez, L., Angers, A. & Desgroseillers, L. 1995. Specific microsatellite loci for brook charr reveal strong population subdivision on a microgeographic scale. Journal of Fish Biology 47: 177-185. Aubin Horth, N. & Dodson, J. J. 2004. Influence of individual body size and variable thresholds on the incidence of a sneaker male reproductive tactic in Atlantic salmon. Evolution 58: 136-144. Arctic Climate Impact Assessment. 2004. Impacts of a Warming Arctic. Cambridge  137  University Press. Cambridge, UK. Armstrong, R. H. 1974. Migration of anadromous Dolly Varden (Salvelinus malma) in southeastern Alaska. Journal of the Fisheries Board of Canada 31: 435-444. Arnold, M. L. 1997. Natural hybridization and evolution. Oxford University Press, USA. Avise, J. C. 2004. Molecular Markers, Natural History and Evolution, second ed. Sinauer associates, Sunderland, Mass. Ayers, S. D. 2010. A review of the species status of the Angayukaksurak charr (Salvelinus anaktuvudensis) of Northern Alaska: perspectives from molecular and morphological data. Vol. M.Sc. pp. 83. University of Alaska Fairbanks, Fairbanks, Alaska. Babaluk, J. A., Halden, N. M., Reist, J. D., Kristofferson, A. H., Campbell, J. L. & Teesdale, W. J. 1997. Evidence for non-anadromous behaviour of arctic charr (Salvelinus alpinus) from Lake Hazen, Ellesmere Island, Northwest Territories, Canada, based on scanning proton microprobe analysis of otolith strontium distribution. Arctic 50: 224-233. Babaluk, J. A., Campbell, J. L., Evans, C. L., Halden, N. M., Mejia, S. R., Nejedly, Z., Reist, J. D. & Teesdale, W. J. 2002. Micro-PIXE analysis of strontium in Arctic char, Salvelinus alpinus, otoliths from Quttinirpaaq National Park, Nunavut, Canada. Nuclear Inst. and Methods in Physics Research, B 189: 190-195. Balikci, A. 1980. Charr fishing among the Arviligjuarmiut. In: Charrs: Salmonid fishes of the genus Salvelinus, (Balon, E. K., ed.). pp. 7-9. Dr. W. Junk bv Publishers, The Hague. Balkau, B. J. & Feldman, M. W. 1973. Selection for migration modification. Genetics 74: 171. Ballard, J. W. & Whitlock, M. C. 2004. The incomplete natural history of mitochondria. Molecular Ecology 13: 729-744. Ballard, J. W. O. & Rand, D. M. 2005. The population biology of mitochondrial DNA and its phylogenetic implications. Annual Review of Ecology, Evolution, and Systematics 36:621-642. Barton, N. H. (2001) The evolutionary consequences of gene flow and local adaptation: future approaches. In: Dispersal, (Clobert, J., Danchin, E., Dhondt, A. A. &  138  Nichols, J. D., eds.). pp. Oxford Univ. Press, Oxford. Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N. & Bonhomme, F. 2004. GENETIX 4.05, logiciel sous Windows pour la génétique des populations. Laboratoire Génome, Populations, Interactions, CNRS UMR 5000. Université de Montpellier II, Montpellier. Beaumont, M. A., Zhang, W. & Balding, D. J. 2002. Approximate Bayesian computation in population genetics. Genetics 162: 2025-2035. Beddow, T. A., Deary, C. & McKinley, R. S. 1998. Migratory and reproductive activity of radio-tagged Arctic char (Salvelinus alpinus L.) in northern Labrador. Hydrobiologia 371: 249-262. Beerli, P. 2004. Effect of unsampled populations on the estimation of population sizes and migration rates between sampled populations. Molecular Ecology 13: 827836. Beerli, P. & Felsenstein, J. 1999. Maximum-likelihood estimation of migration rates and effective population numbers in two populations using a coalescent approach. Genetics 152: 763-773. Beerli, P. & Felsenstein, J. 2001. Maximum likelihood estimation of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach. Proceedings of the National Academy of Science 98: 4563-4568. Bégout Anras, M. L., Gyselman, E. C., Jorgenson, J. K., Kristofferson, A. H. & Anras, L. 1999. Habitat preferences and residence time for the freshwater to ocean transition stage in Arctic charr. Journal of the Marine Biological Association of the United Kingdom 79: 153-160. Behnke, R. J. (1984) Organizing the diversity of the Arctic charr complex. In: Biology of the Arctic charr: proceedings of the international symposium on Arctic charr, (Johnson, L. & Burns, B., eds.). pp. University of Manitoba Press, Winnipeg. Behnke, R. J. & Tomelleri, J. R. 2002. Trout and salmon of North America. Free Press, New York. Bekkevold, D., Hansen, M. M. & Mensberg, K. L. D. 2004. Genetic detection of sexspecific dispersal in historical and contemporary populations of anadromous brown trout Salmo trutta. Molecular Ecology 13: 1707-1712.  139  Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N. & Bonhomme, F. 2004. GENETIX 4.05, logiciel sous Windows pour la génétique des populations. Laboratoire Génome, Populations, Interactions, CNRS UMR 5000. Université de Montpellier II, Montpellier. Bell, G. & Gonzalez, A. 2011. Adaptation and evolutionary rescue in metapopulations experiencing environmental deterioration. Science 332: 1327. Bell, M. A. & Richkind, K. E. 1981. Clinal variation of lateral plates in threespine stickleback fish. The American Naturalist 117: 113-132. Bengtsson, B. O. 1978. Avoiding inbreeding: at what cost? Journal of Theoretical Biology 73: 439-444. Berg, O. K. & Berg, M. 1990. Effects of Carlin tagging on the mortality and growth of anadromous Arctic char, Salvelinus alpinus (L.). Aquaculture Research 21: 221228. Berg, O. K. & Berg, M. 1993. Duration of sea and freshwater residence of Arctic char (Salvelinus alpinus), from the Vardnes River in northern Norway. Aquaculture 110: 129-140. Berg, O. K. & Jonsson, B. (1989) Migratory patterns of anadromous Atlantic salmon, brown trout and Arctic charr from the Vardnes River in northern Norway. pp. 106-115. Bernatchez, L., Dempson, J. B. & Martin, S. 1998. Microsatellite gene diversity analysis in anadromous arctic char, Salvelinus alpinus, from Labrador, Canada. Canadian Journal of Fisheries and Aquatic Sciences 55: 1264-1272. Bernatchez, L., Glemet, H., Wilson, C. C. & Danzmann, R. G. 1995. Introgression and fixation of Arctic char (Salvelinus alpinus) mitochondrial genome in an allopatric population of brook trout (S. fontinalis). Canadian Journal of Fisheries and Aquatic Sciences 52: 179-185. Bernatchez, L. & Wilson, C. C. 1998. Comparative phylogeography of Nearctic and Palearctic fishes. Molecular Ecology 7: 431-452. Blanquart, F. & Gandon, S. 2011. Evolution of migration in a periodically changing environment. The American Naturalist 177: 188-201. Blier, P. U., Breton, S., Desrosiers, V. & Lemieux, H. 2006. Functional conservatism in  140  mitochondrial evolution: insight from hybridization of Arctic and Brook Charrs. Journal of Experimental Zoology (Mol Dev Evol) 306B: 425-432. Bohonak, A. J. 1999. Dispersal, Gene Flow, and Population Structure. The Quarterly Review of Biology 74: 21. Bolnick, D. I., Snowberg, L. K., Patenia, C., Stutz, W. E., Ingram, T. & Lau, O. L. 2009. Phenotype-dependent native habitat preference facilitates divergence between parapatric lake and stream sticklebacks. Evolution 63: 2004-2016. Bowler, D. E. & Benton, T. G. 2005. Causes and consequences of animal dispersal strategies: relating individual behaviour to spatial dynamics. Biological Reviews 80: 205-225. Breton, S., Dufresne, F., Desrosiers, G. & Blier, P. U. 2003. Population structure of two northern hemisphere polychaetes, Neanthes virens and Hediste diversicolor (Nereididae), with different life-history traits. Marine Biology 142: 707-715. Bridle, J. R. & Vines, T. H. 2007. Limits to evolution at range margins: when and why does adaptation fail? Trends in Ecology and Evolution 22: 140-147. Briggs, C. J. & Hoopes, M. F. 2004. Stabilizing effects in spatial parasitoid-host and predator-prey models: a review. Theoretical Population Biology 65: 299-315. Brown, J. H. & Kodric-Brown, A. 1977. Turnover rates in insular biogeography: effect of immigration on extinction. Ecology 58: 445-449. Brownstein, M., Carpten, J. D. & Smith, J. R. 1996. Modulation of non-templated nucleotide addition by Taq DNA polymerase: primer modifications that facilitate genotyping. BioTechniques 20. Brunner, P. C., Douglas, M. R., Osinov, A., Wilson, C. C. & Bernatchez, L. 2001. Holarctic phylogeography of Arctic charr (Salvelinus alpinus L.) inferred from mitochondrial DNA sequences. Evolution 55: 573-586. Calvert, W. H. & Brower, L. P. 1986. The location of monarch butterfly (Danaus plexippus L.) overwintering colonies in Mexico in relation to topography and climate. Journal of the Lepidopterists Society 40: 164-187. Case, T. J. & Taper, M. L. 2000. Interspecific competition, environmental gradients, gene flow, and the coevolution of species' borders. The American Naturalist 155: 583605.  141  Castric, V. & Bernatchez, L. 2003. The rise and fall of isolation by distance in the anadromous brook charr (Salvelinus fontinalis Mitchill). Genetics 163: 983-996. Cavalli-Sforza, L. L. & Edwards, A. W. F. 1967. Phylogenetic analysis. Models and estimation procedures. American Journal of Human Genetics 19: 233. Charlesworth, D. & Charlesworth, B. 1987. Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18: 237-268. Charnov, E. L. & Schaffer, W. M. 1973. Life-history consequences of natural selection: Cole's result revisited. The American Naturalist 107: 791. Clobert, J., Wolff, J. O., Nichols, J. D., Danchin, E. & Dhondt, A. A. 2001. Introduction. In: Dispersal, (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.). pp. Oxford University Press, Oxford. Clobert, J., Ims, R. A. & Rousset, F. 2004. Causes, mechanisms and consequences of dispersal. In: Ecology, genetics and evolution of metapopulation, (Hanski, I. & Gaggiotti, O., eds.). pp. 307-335. Elsevier Academic Press. Clobert, J., Le Galliard, J.-F., Cote, J., Meylan, S. & Massot, M. 2009. Informed dispersal, heterogeneity in animal dispersal syndromes and the dynamics of spatially structured populations. Ecology Letters 12: 197-209. Colosimo, P. F., Peichel, C. L., Nereng, K., Blackman, B. K., Shapiro, M. D., Schluter, D. & Kingsley, D. M. 2004. The genetic architecture of parallel armor plate reduction in threespine sticklebacks. PLoS Biology 2: 635-641. Colwell, R. K. & Naeem, S. 1999. Sexual sorting in hummingbird flower mites (Mesostigmata: Ascidae). Annals of the Entomological Society of America 92: 952-959. Conejeros, P., Phan, A., Power, M., Alekseyev, S., O'Connell, M., Dempson, B. & Dixon, B. 2008. MH class IIa polymorphism in local and global adaptation of Arctic charr (Salvelinus alpinus L.). Immunogenetics 60: 325-337. Consuegra, S. & Garcia de Leaniz, C. 2007. Fluctuating sex ratios, but no sex-biased dispersal, in a promiscuous fish. Evolutionary Ecology 21: 229-245. Coulon, A., Fitzpatrick, J. W., Bowman, R., Stith, B. M., Makarewich, C. A., Stenzler, L. M. & Lovette, I. J. 2008. Congruent population structure inferred from dispersal behaviour and intensive genetic surveys of the threatened Florida scrub-jay  142  (Aphelocoma cirulescens). Molecular Ecology 17: 1685-1701. Coyne, J. A., Barton, N. H. & Turelli, M. 1997. Perspective: a critique of Sewall Wright's shifting balance theory of evolution. Evolution: 643-671. Crane, P. A., Lewis, C. J., Kretschmer, E. J., Miller, S. J., Spearman, W. J., DeCicco, A. L., Lisac, M. J. & Wenburg, J. K. 2004. Characterization and inheritance of seven microsatellite loci from Dolly Varden, Salvelinus malma, and cross-species amplification in Arctic char, S. alpinus. Conservation Genetics 5: 737-741. Crispo, E., Moore, J.-S., Lee-Yaw, J. A., Gray, S. M. & Haller, B. C. 2011. Broken barriers: Human-induced changes to gene flow and introgression in animals. BioEssays 33: 508-518. Crossman, E. J. & McAllister, D. E. 1986. Zoogeography of freshwater fishes of the Hudson Bay drainage, Ungava Bay and the Arctic Archipelago. In: The Zoogeography of North American Freshwater Fishes, (Hocutt, C. H. & Wiley, E. O., eds.). pp. John Wiley and Sons, New York. Csilléry, K., Blum, M. G. B., Gaggiotti, O. E. & François, O. 2010. Approximate Bayesian computation (ABC) in practice. Trends in Ecology and Evolution 25: 410-418. Danchin, E., Giraldeau, L. A., Valone, T. J. & Wagner, R. H. 2004. Public information: from nosy neighbors to cultural evolution. Science 305: 487-491. Davis, M. B. & Shaw, R. G. 2001. Range shifts and adaptive responses to quaternary climate change. Science 292: 673. DeCicco, A. L. 1992. Long-distance movements of anadromous Dolly Varden between Alaska and the U.S.S.R. Arctic 45: 120-123. DeHaan, P. W. & Ardren, W. R. 2005. Characterization of 20 highly variable tetranucleotide microsatellite loci for bull trout (Salvelinus confluentus) and crossamplification in other Salvelinus species. Molecular Ecology Notes 5: 582-585. DeLacy, A. C. & Morton, W. M. 1942. Taxonomy and habits of the chars, Salvelinus alpinus and Salvelinus malma of the Karluk Drainage System. Transactions of the American Fisheries Society 72: 79-91. Dempson, J. B. & Kristofferson, A. H. (1987) Spatial and temporal aspects of the ocean migration of anadromous Arctic char. In: Common Strategies of Anadromous and  143  Catadromous Fishes, American Fisheries Society Symposium, (Dadswell, M. J., Klauda, R. J., Moffitt, C. M., Saunders, R. L., Rulifson, R. A. & Cooper, J. E., eds.). pp. 340-357. DeWoody, J. A. & Avise, J. C. 2000. Microsatellite variation in marine, freshwater and anadromous fishes compared with other animals. Journal of Fish Biology 56: 461473. Dias, P. C. 1996. Sources and sinks in population biology. Trends in Ecology and Evolution 11: 326-330. Dias, P. C. & Blondel, J. 1996. Local specialization and maladaptation in the Mediterranean blue tit (Parus caeruleus). Oecologia 107: 79-86. Dieckmann, U., O'Hara, B. & Weisser, W. 1999. The evolutionary ecology of dispersal. Trends in Ecology and Evolution 14: 88-90. Dionne, M., Caron, F., Dodson, J. & Bernatchez, L. 2008. Landscape genetics and hierarchical genetics structure in Atlantic salmon: the interaction of gene flow and local adaptation. Molecular Ecology 17: 2382-2396. Dittman, A. & Quinn, T. 1996. Homing in Pacific salmon: mechanisms and ecological basis. Journal of Experimental Biology 199: 83-91. Docker, M. F. & Heath, D. D. 2003. Genetic comparison between sympatric anadromous steelhead and freshwater resident rainbow trout in British Columbia, Canada. Conservation Genetics 4: 227-231. Doiron, S., Bernatchez, L. & Blier, P. 2002. A comparative mitogenomic analysis of the potential adaptive value of Arctic charr mtDNA introgression in brook charr populations (Salvelinus fontinalis Mitchill). Molecular Biology and Evolution 19: 1902-1909. Doucett, R. R., Power, M., Power, G., Caron, F. & Reist, J. D. 1999. Evidence for anadromy in a southern relict population of Arctic charr from North America. Journal of Fish Biology 55: 84-93. Drummond, A. J., Rambaut, A., Shapiro, B. & Pybus, O. G. 2005. Bayesian coalescent inference of past population dynamics from molecular sequences. Molecular Biology and Evolution 22: 1185-1192. Drummond, A. & Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by  144  sampling trees. BMC Evolutionary Biology 7: 214. Dutil, J.-D. 1986. Energetic constraints and spawning interval in the anadromous Arctic charr (Salvelinus alpinus). Copeia 4: 945-955. Dyke, A. S., Moore, A. & Roberston, L. (2003) Deglaciation of North America. pp. Open file 1574. Geological Survey of Canada. Earl, D. A. 2009. Structure Harvester v0. 3. Edelaar, P., Siepielski, A. M. & Clobert, J. 2008. Matching habitat choice causes directed gene flow: a neglected dimension in evolution and ecology. Evolution 62: 24622472. Edelaar, P. & Bolnick, D. I. 2012. Non-random gene flow: an underappreciated force in evolution and ecology. Trends in Ecology and Evolution. Edmands, S. 1999. Heterosis and outbreeding depression in interpopulation crosses spanning a wide range of divergence. Evolution: 1757-1768. Edmands, S. 2007. Between and rock and a hard place: evaluating the relative risks of inbreeding and outbreeding for conservation and managment. Molecular Ecology 16: 463-475. Ehrlich, P. R. & Raven, P. H. 1969. Differentiation of populations. Science 165: 1228-32. Einum, S., Finstad, A. G., Robertsen, G., Nislow, K. H., McKelvey, S. & Armstrong, J. D. 2012. Natal movement in juvenile Atlantic salmon: a body size-dependent strategy? Population Ecology 54: 285-294. Ellengren, H. 2000. Microsatellite mutations in the germline: implications for evolutionary inference. Trends in Genetics 16: 551-558. Endler, J. A. 1973. Gene flow and population differentiation: Studies of clines suggest that differentiation along environmental gradients may be independent of gene flow. Science 179: 243. Endler, J. A. 1977. Geographic variation, speciation and clines. Princeton University Press, Princeton. Evanno, G., Regnaut, S. & Goudet, J. 2005. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14: 2611-2620. Excoffier, L., Foll, M. & Petit, R. J. 2009. Genetic consequences of range expansions.  145  Annual Review of Ecology, Evolution, and Systematics 40: 481-501. Excoffier, L., Laval, G. & Schneider, S. 2005. Arlequin (version 3.0): an integrated software package for population genetics data analysis. Evolutionary bioinformatics online 1: 47. Excoffier, L. & Ray, N. 2008. Surfing during population expansions promotes genetic revolutions and structuration. Trends in Ecology and Evolution 23: 347-351. Fedorov, V. B. & Stenseth, N. C. 2002. Multiple glacial refugia in the North American Arctic: inference from phylogeography of the collared lemming (Dicrostonyx groenlandicus). Proceedings of the Royal Society B: Biological Sciences 269: 2071-2077. Felsenstein, J. 1993. PHYLIP: phylogenetic inference package, version 3.5 c. Finstad, B. & Heggberget, T. G. 1993. Migration, growth and survival of wild and hatchery-reared anadromous Arctic charr (Salvelinus alpinus) in Finnmark, Northern Norway. Journal of Fish Biology 43: 303-312. Fisher, R. A. 1930. The genetical theory of natural selection. Oxford University Press, Oxford, UK. Fisher, R. A. 1937. The wave of advance of advantageous genes. Annals of Human Genetics 7: 355-369. Fleming, I. A. 1998. Pattern and variability in the breeding system of Atlantic salmon (Salmo salar), with comparisons to other salmonids. Canadian Journal of Fisheries and Aquatic Sciences 55(Suppl. 1): 59-76. Fleming, I. A. & Reynolds, J. D. 2004. Salmonid breeding systems. In: Evolution Illuminated: Salmon and Their Relatives, (Hendry, A. P. & Stearns, S. C., eds.). pp. 265-294. Oxford University Press, Oxford, UK. Forseth, T., Ugedal, O., Jonsson, B. & Fleming, I. A. 2003. Selection on Arctic charr generated by competition from brown trout. Oikos 101: 467-478. Fraser, D. J., Lippé, C. & Bernatchez, L. 2004. Consequences of unequal population size, asymmetric gene flow and sex-biased dispersal on population structure in brook charr (Salvelinus fontinalis). Molecular Ecology 13: 67-80. Fraser, D. J., Weir, L. K., Bernatchez, L., Hansen, M. M. & Taylor, E. B. 2011. Extent and scale of local adaptation in salmonid fishes: review and meta-analysis.  146  Heredity 106: 404-420. Fu, Y. X. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915. Funk, D. J. & Omland, K. E. 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annual Reviews of Ecology, Evolution, and Systematics: 397-423. Gaggiotti, O. E. 2011. Making inferences about speciation using sophisticated statistical genetics methods: look before you leap. Molecular Ecology 20: 2229-2232. Garant, D., Forde, S. E. & Hendry, A. P. 2007. The multifarious effects of dispersal and gene flow on contemporary adaptation. Functional Ecology 21: 434-443. Garant, D., Kruuk, L. E. B., Wilkin, T. A., McCleery, R. H. & Sheldon, B. C. 2005. Evolution driven by differential dispersal within a wild bird population. Nature 433: 60-65. García-Ramos, G. & Kirkpatrick, M. 1997. Genetic models of adaptation and gene flow in peripheral populations. Evolution 51: 21-28. Garroway, C. J., Bowman, J., Cascaden, T. J., Holloway, G. L., Mahan, C. G., Malcolm, J., Steele, M. A., Turner, G. & Wilson, P. J. 2010. Climate change induced hybridization in flying squirrels. Global Change Biology 16: 113-121. Gomulkiewicz, R., Holt, R. D. & Barfield, M. 1999. The effects of density dependence and immigration on local adaptation and niche evolution in a black-hole sink environment. Theoretical Population Biology 55: 283-296. Gíslason, D., Ferguson, M. M., Skulason, S. & Snorrason, S. S. 1999. Rapid and coupled phenotypic and genetic divergence in Icelandic Arctic char (Salvelinus alpinus). Canadian Journal of Fisheries and Aquatic Sciences 56: 2229-2234. Glémet, H., Blier, P. & Bernatchez, L. 1998. Geographical extent of Arctic char (Salvelinus alpinus) mtDNA introgression in brook char populations (S. fontinalis) from eastern Québec, Canada. Molecular Ecology 7: 1655-1662. Glover, K. A., Hansen, M. M., Lien, S., Als, T. D., Høyheim, B. & Skaala, Ø. 2010. A comparison of SNP and STR loci for delineating population structure and performing individual genetic assignment. BMC genetics 11: 2. Goslee, S. C. & Urban, D. L. 2007. The ecodist package for dissimilarity-based analysis  147  of ecological data. Journal of Statistical Software 22: 1-19. Goudet, J. (2001) FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9. 3). pp. Grainger, E. H. 1953. On the age, growth, migration, reproductive potential and feeding habits of the Arctic char (Salvelinus alpinus) of Frobisher Bay, Baffin Island. Journal of the Fisheries Board of Canada 10: 326-370. Greenwood, P. J. 1980. Mating systems, philopatry and dispersal in birds and mammals. Animal behaviour 28: 1140-1162. Gross, M. R., Coleman, R. M. & McDowall, R. M. 1988. Aquatic productivity and the evolution of diadromous fish migration. Science 239: 1291-1293. Gross, M. R. 1996. Alternative reproductive strategies and tactics: diversity within sexes. Trends in Ecology & Evolution 11: 92-98. Gulseth, O. A. & Nilssen, K. J. 2000. The brief period of spring migration, short marine residence, and high return rate of a northern Svalbard population of arctic char. Transactions of the American Fisheries Society 129: 782-796. Gyselman, E. 1994. Fidelity of anadromous Arctic char (Salvelinus alpinus) to Nauyuk Lake, N.W.T., Canada. Canadian Journal of Fisheries and Aquatic Sciences 51: 1927-1934. Hairston, N. G., Ellner, S. P., Geber, M. A., Yoshida, T. & Fox, J. A. 2005. Rapid evolution and the convergence of ecological and evolutionary time. Ecology Letters 8: 1114-1127. Hamilton, W. D. 1964. The genetical evolution of social behaviour. II. Journal of theoretical biology 7: 17-52. Hamilton, W. D. & May, R. M. 1977. Dispersal in stable habitats. Nature 269: 578-581. Hammar, J., Porter, R., Skjold, E. & Verspoor, E. 1989. Mixed or reversed anadromy life-history strategies of Arctic charr in southern Labrador. Physiology and Ecology Japan, Special Volume 1: 169. Hampe, A. & Petit, R. J. 2005. Conserving biodiversity under climate change: the rear edge matters. Ecology Letters 8: 461-467. Hansen, M. M., Ruzzante, D. E., Nielsen, E. E., Bekkevold, D. & Mensberg, K. L. D. 2002. Long-term effective population sizes, temporal stability of genetic  148  composition and potential for local adaptation in anadromous brown trout (Salmo trutta) populations. Mol. Ecol. 11: 2523-2535. Hanski, I. 1999. Metapopulation ecology. Oxford University Press Oxford, UK. Hanski, I. 2001. Population dynamic consequences of dispersal in local populations and in metapopulations. In: Dispersal, (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.). pp. 283-298. Oxford University Press, Oxford, UK. Hanski, I. & Gaggiotti, O. E. 2004. Ecology, genetics, and evolution of metapopulations. Elsevier Amsterdam, The Netherlands. Hanski, I. & Gilpin, M. E. (Eds.) 1997. Metapopulation biology: ecology, genetics and evolution, San Diego, CA, Academic Press. Hanski, I., Peltonen, A. & Kaski, L. 1991. Natal dispersal and social dominance in the common shrew Sorex araneus. Oikos: 48-58. Hanski, I. & Zhang, D. Y. 1993. Migration, metapopulation dynamics and fugitive coexistence. Journal of Theoretical Biology 163: 491-504. Hard, J. J., Gross, M. R., Heino, M., Hilborn, R., Kope, R. G., Law, R. & Reynolds, J. D. 2008. Evolutionary consequences of fishing and their implications for salmon. Evolutionary Applications 1: 388-408. Hard, J. J. & Heard, W. R. 1999. Analysis of straying variation in Alaskan hatchery chinook salmon (Oncorhynchus tshawytscha) following transplantation. Canadian Journal of Fisheries and Aquatic Sciences 56: 578-589. Harpending, H. C. 1994. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Human biology 66: 591. Harris, L. N. & Taylor, E. B. 2010. Pleistocene glaciations and contemporary genetic diversity in a Beringian fish, the broad whitefish, Coregonus nasus (Pallas): inferences from microsatellite DNA variation. Journal of Evolutionary Biology 23: 72-86. Hauser, L., Seamons, T. R., Dauer, M., Naish, K. A. & Quinn, T. P. 2006. An empirical verification of population assignment methods by marking and parentage data: hatchery and wild steelhead (Onchorhynchus mykiss) in Forks Creek, Washington, USA. Molecular Ecology15: 3157-3173. Hedrick, P. W. 2004. Genetics of populations, Third ed. Jones & Bartlett Learning.  149  Hendry, A. P., Bohlin, T., Jonsson, B. & Berg, O. K. 2004. To sea or not to sea? Anadromy versus non-anadromy in salmonids. In: Evolution Illuminated: Salmon and their Relatives, (Hendry, A. P. & Stearns, S. C., eds.). pp. 92–125. Oxford University Press, Oxford. Hendry, A. P., Castric, V., Kinnison, M. T. & Quinn, T. P. 2004. The evolution of phylopatry and dispersal: homing versus straying in salmonids. In: Evolution illuminated: salmon and their relatives, (Hendry, A. P. & Stearns, S. C., eds.). pp. Oxford University Press, Oxford, UK. Hendry, A. P., Day, T. & Taylor, E. B. 2001. Population mixing and the adaptive divergence of quantitative traits in discrete populations: A theoretical framework for empirical tests. Evolution 55: 459-466. Hendry, A. P., Farrugia, T. J. & Kinnison, M. T. 2008. Human influences on rates of phenotypic change in wild animal populations. Molecular Ecology 17: 20-29. Hendry, A. P. & Kinnison, M. T. 1999. The pace of modern life: measuring rates of contemporary microevolution. Evolution 53: 1637-1653. Hendry, A. P. & Stearns, S. C. (Eds.) 2004. Evolution Illuminated: Salmon and Their Relatives, Oxford University Press. Hendry, A. P., Wenburg, J. K., Bentzen, P., Volk, E. C. & Quinn, T. P. 2000. Rapid evolution of reproductive isolation in the wild: evidence from introduced salmon. Science 290: 516-518. Hewitt, G. M. 2000. The genetic legacy of the ice ages. Nature 405: 907-913. Hewitt, G. M. 2004. Genetic consequences of climatic changes in the Quaternary. Phil. Trans. Roy. Soc. Lond. B 359: 183-195. Hey, J. & Nielsen, R. 2004. Multilocus methods for estimating population sizes, migration rates and divergence time, with applications to the divergence of Drosophila pseudoobscura and D. persimilis. Genetics 167: 747-760. Hey, J. 2010. Isolation with migration models for more than two populations. Molecular biology and evolution 27: 905-920. Hindar, K. & Jonsson, B. 1993. Ecological polymorphism in Arctic charr. Biological Journal of the Linnean Society 48: 63-74. Hindar, K., Jonsson, B., Ryman, N. & Stahl, G. 1991. Genetic relationships among  150  landlocked, resident, and anadromous brown trout, Salmo trutta L. Heredity 66: 83-91. Hindar, K., Ryman, N. & Stahl, G. 1986. Genetic differentiation among local populations and morphotypes of Arctic charr, Salvelinus alpinus. Biological Journal of the Linnean Society 27: 269-285. Hoffmann, A. A. & Sgrò, C. M. 2011. Climate change and evolutionary adaptation. Nature 470: 479-485. Holder, K., Montgomerie, R. & Friesen, V. L. 1999. A test of the glacial refugium hypothesis using patterns of mitochondrial and nuclear DNA sequence variation in rock ptarmigan (Lagopus mutus). Evolution: 1936-1950. Holenweg, A. K. & Reyer, H. U. 2000. Hibernation behavior of Rana lessonae and R. esculenta in their natural habitat. Oecologia 123: 41-47. Holt, R. D. 1985. Population dynamics in two-patch environments: some anomalous consequences of an optimal habitat distribution. Theoretical Population Biology 28: 181-208. Holt, R. D. & Gomulkiewicz, R. 1997. How does immigration influence local adaptation? A reexamination of a familiar paradigm. The American Naturalist 149: 563-572. Holt, R. D. & McPeek, M. A. 1996. Chaotic population dynamics favors the evolution of dispersal. The American Naturalist 148: 709-718. Howard, W. E. 1960. Innate and environmental dispersal of individual vertebrates. American Midland Naturalist 63: 152-161. Hubbell, S. P. 2001. The Unified Neutral Theory of Biodiversity and Biogeography. Princeton Univ. Press, Princeton, NJ. Hubisz, M. J., Falush, D., Stephens, M. & Pritchard, J. K. 2009. Inferring weak population structure with the assistance of sample group information. Molecular Ecology Resources 9: 1322-1332. Hutchings, J. A. & Fraser, D. J. 2008. The nature of fisheries-and farming-induced evolution. Molecular Ecology 17: 294-313. Hutchings, J. A. & Gerber, L. 2002. Sex-biased dispersal in a salmonid fish. Proceedings of the Royal Society B: Biological Sciences 269: 2487-2493. Hutchings, J. A. & Myers, R. A. 1994. The evolution of mating strategies in variable  151  environments. Evolutionary Ecology 8: 256-268. Ims, R. A. & Hjermann, D. (2001) Condition-dependent dispersal. In: Dispersal (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.). pp. 203-216. Oxford University Press,, Oxford. IPCC 2007. Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC. Jakobsson, M. & Rosenberg, N. A. 2007. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23: 1801-1806. Johnson, L. 1980. The Arctic charr. In: Charrs: salmonid fishes of the genus Salvelinus, (Balon, E. K., ed.). pp. 15-98. Dr. W. Junk bv Publishers, The Hague. Johnson, M. L. & Gaines, M. S. 1990. Evolution of dispersal: theoretical models and empirical tests using birds and mammals. Annual Review of Ecology, Evolution and Systematics 21: 449-480. Jonsson, B. & Hindar, K. 1982. Reproductive strategy of dwarf and normal Arctic charr (Salvelinus alpinus) from Vangsvatnet Lake, western Norway. Canadian Journal of Fisheries and Aquatic Sciences 39: 1404-1413. Jonsson, B. & Jonsson, N. 2001. Polymorphism and speciation in Arctic charr. Journal of Fish Biology 58: 605-638. Jordan, D. S. & Evermann, B. W. 1896. The Fishes of North and Middle America. U.S. National Museum Bulletin 47, part 1. Kawecki, T. J. 2003. Sex-Biased Dispersal and Adaptation to Marginal Habitats. The American Naturalist 162: 415-426. Kawecki, T. J. & Ebert, D. 2004. Conceptual issues in local adaptation. Ecology Letters 7: 1225-1241. Kawecki, T. J. & Holt, R. D. 2002. Evolutionary consequences of asymmetric dispersal rates. The American Naturalist 160: 333-347. Keller, L. F. & Waller, D. M. 2002. Inbreeding effects in wild populations. Trends in Ecology & Evolution 17: 230-241. Kelly, B. P., Whiteley, A. & Tallmon, D. The Arctic melting pot. Nature 468: 891-891.  152  Kennedy, P. L. & Ward, J. M. 2003. Effects of experimental food supplementation on movements of juvenile Northern Goshawks (Accipiter gentilis atricapillus). Oecologia 134: 284-291. King, R. & Lawson, R. 1995. Color-pattern variation in Lake Erie water snakes: the role of gene flow. Evolution 49: 885-896. Kirkpatrick, M. & Barton, N. H. 1997. Evolution of a species' range. Am. Nat. 150: 1-23. Klemetsen, A., Amundsen, P. A., Dempson, J. B., Jonsson, B., Jonsson, N., O'Connell, M. F. & Mortensen, E. 2003. Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus L.: a review of aspects of their life histories. Ecology of Freshwater Fish 12: 1-59. Knowles, L. L. & Maddison, W. P. 2002. Statistical phylogeography. Molecular Ecology 11: 2623-2635. Knowles, L. L. 2004. The burgeoning field of statistical phylogeography. Journal of Evolutionary Biology 17: 1-10. Knudsen, R., Klemetsen, A., Amundsen, P.-A. & Hermansen, B. 2006. Incipient speciation through niche expansion: an example from the Arctic charr in a subarctic lake. Proceedings of the Royal Society B: Biological Sciences 273: 2291-2298. Koskinen, M. T., Nilsson, J., Veselov, A. J., Potutkin, A. G., Ranta, E. & Primmer, C. R. 2002. Microsatellite data resolve phylogeographic patterns in European grayling, Thymallus thymallus, Salmonidae. Heredity 88: 391-401. Kristoffersen, K. 1994. The influence of physical watercourse parameters on the degree of anadromy in different lake populations of Arctic charr (Salvelinus alpinus (L.)) in northern Norway. Ecology of Freshwater Fish 3: 80-91. Kurta, A. & Murray, S. W. 2002. Philopatry and migration of banded Indiana bats (Myotis sodalis) and effects of radio transmitters. Journal of Mammalogy 83: 585589. Kyle, C. J. & Boulding, E. G. 2000. Comparative population genetic structure of marine gastropods (Littorina spp.) with and without pelagic larval dispersal. Marine Biology 137: 835-845. Lamaze, F. C., Sauvage, C., Marie, A., Garant, D. & Bernatchez, L. 2012. Dynamics of  153  introgressive hybridization assessed by SNP population genomics of coding genes in stocked brook charr (Salvelinus fontinalis). Molecular Ecology 21: 2877-2895. Langeland, A. & L'Abée-Lund, J. H. 1998. An experimental test of the genetic component of the ontogenetic habitat shift in Arctic charr (Salvelinus alpinus). Ecology of Freshwater Fish 7: 200-207. Larsen, P., Nielsen, E., Koed, A., Thomsen, D., Olsvik, P. & Loeschcke, V. 2008. Interpopulation differences in expression of candidate genes for salinity tolerance in winter migrating anadromous brown trout (Salmo trutta L.). BMC genetics 9: 12. Larsson, S., Forseth, T., Berglund, I., Jensen, A. J., Naslund, I., Elliott, J. M. & Jonsson, B. 2005. Thermal adaptation of Arctic charr: experimental studies of growth in eleven charr populations from Sweden, Norway and Britain. Freshwater Biology 50: 353-368. Lavergne, S., Mouquet, N., Thuiller, W. & Ronce, O. 2010. Biodiversity and climate change: integrating evolutionary and ecological responses of species and communities. Annual Review of Ecology, Evolution, and Systematics 41: 321-350. Leggett, W. C. & Carscadden, J. E. 1978. Latitudinal variation in reproductive characteristics of American shad (Alosa sapidissima): evidence for population specific life history strategies in fish. Journal of the Fisheries Research Board of Canada 35: 1469-1478. Leibold, M. A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., Hoopes, M. F., Holt, R. D., Shurin, J. B., Law, R. & Tilman, D. 2004. The metacommunity concept: a framework for multi-scale community ecology. Ecology letters 7: 601613. Lenormand, T. 2002. Gene flow and the limits to natural selection. Trends in Ecology and Evolution 17: 183-189. Levin, S. A., Muller-Landau, H. C., Nathan, R. & Chave, J. 2003. The ecology and evolution of seed dispersal: a theoretical perspective. Annual Review of Ecology, Evolution, and Systematics: 575-604. Levins, R. 1969. Some demographic and genetic consequences of environmental heterogeneity for biological control. Bulletin of the Entomological Society of  154  America 15: 237-240. Librado, P. & Rozas, J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: 1451-1452. Loehr, J., Worley, K., Grapputo, A., Carey, J., Veitch, A. & Coltman, D. W. 2006. Evidence for cryptic glacial refugia from North American mountain sheep mitochondrial DNA. Journal of Evolutionary Biology 19: 419-430. Loewen, T. (2008) Life history, morphometric and habitat use variation in Arctic charr (Salvelinus alpinus) populations of Southern Baffin Island, Nunavut, Canada. In: Department of Biological Sciences, Vol. MSc. pp. 205. University of Manitoba, Winnipeg, Manitoba. Loewen, T. N., Gillis, D. & Tallman, R. F. 2009. Ecological niche specialization inferred from morphological variation and otolith strontium of Arctic charr Salvelinus alpinus L. found within open lake systems of southern Baffin Island, Nunavut, Canada. Journal of Fish Biology 75: 1473-1495. Loewen, T. N., Gillis, D. & Tallman, R. F. 2010. Maturation, growth and fecundity of Arctic charr, Salvelinus alpinus (L.), life-history variants co-existing in lake systems of Southern Baffin Island, Nunavut, Canada. Hydrobiologia 650: 193202. Mäkinen, H. S., Cano, J. M. & Merilä, J. 2006. Genetic relationships among marine and freshwater populations of the European three-spined stickleback (Gasterosteus aculeatus) revealed by microsatellites. Molecular Ecology 15: 1519-1534. Malécot, G. 1950. Quelques schémas probabilistes sur la variabilité des populations naturelles. Annales de l'Université de Lyon 13: 37-60. Mallet, J. 1986. Dispersal and gene flow in a butterfly with home range behavior: Heliconius erato (Lepidoptera: Nymphalidae). Oecologia 68: 210-217. Manel, S., Gaggiotti, O. E. & Waples, R. S. 2005. Assignment methods: matching biological questions with appropriate techniques. Trends Ecol. Evol. 20: 136-142. Manel, S., Joost, S., Epperson, B. K., Holderegger, R., Storfer, A., Rosenberg, M. S., Scribner, K. T., Bonin, A. & Fortin, M. J. 2010. Perspectives on the use of landscape genetics to detect genetic adaptive variation in the field. Molecular Ecology 19: 3760-3772.  155  Matthysen, E. 2005. Density-dependent dispersal in birds and mammals. Ecography 28: 403-416. MacArthur, R. H. & Wilson, E. O. 1967. The theory of island biogeography. Princeton Univ Pr. Mayr, E. 1942. Systematics and the origin of species. Columbia Univ Press, New York. Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge, MA. McCairns, R. J. & Bernatchez, L. 2008. Landscape genetic analyses reveal cryptic population structure and putative selection gradients in a large-scale estuarine environment. Molecular Ecology 17: 3901-3916. McDowall, R. M. 1987. Evolution and the importance of diadromy: the occurence and distribution of diadromy in fishes. American Fisheries Symposium 1: 1-13. McDowall, R. M. 1997. The evolution of diadromy in fishes (revisited) and its place in phylogenetic analysis. Reviews in Fish Biology and Fisheries 7: 443-462. McPeek, M. A. & Holt, R. D. 1992. The evolution of dispersal in spatially and temporally varying environments. The American Naturalist 140: 1010-1027. McPhail, J. D. 1961. A systematic study of the Salvelinus alpinus complex in North America. Journal of the Fisheries Research Board of Canada 18: 793-816. Molofsky, J. & Ferdy, J. B. 2005. Extinction dynamics in experimental metapopulations. Proceedings of the National Academy of Sciences of the United States of America 102: 3726. Moore, J.-S., Gow, J. L., Taylor, E. B. & Hendry, A. P. 2007. Quantifying the constraining influence of gene flow on adaptive divergence in the lake-stream threespine stickleback system. Evolution 61: 2015-2026. Moore, J. S. & Hendry, A. P. 2009. Can gene flow have negative demographic consequences? Mixed evidence from stream threespine stickleback. Philosophical Transactions of the Royal Society B: Biological Sciences 364: 1533-1542. Moore, J. W. 1975. Distribution, movements, and mortality of anadromous arctic char, Salvelinus alpinus L., in the Cumberland Sound area of Baffin Island. Journal of Fish Biology 7: 339-348. Morinville, G. R. & Rasmussen, J. B. 2003. Early juvenile bioenergetic differences between anadromous and resident brook trout (Salvelinus fontinalis). Canadian  156  Journal of Fisheries and Aquatic Sciences 60: 401-410. Morjan, C. L. & Rieseberg, L. H. 2004. How species evolve collectively: implications of gene flow and selection for the spread of advantageous alleles. Molecular Ecology 13: 1341-1356. Morton, W. 1955. Charr or char: history of a common name for Salvelinus. Science 121: 874. Myers, R. A. 1984. Demographic consequences of precocious maturation of Atlantic salmon (Salmo salar). Canadian Journal of Fisheries and Aquatic Sciences 41: 1349-1353. Myers, R. A. & Hutchings, J. A. 1987. Mating of anadromous Atlantic salmon, Salmo salar L., with mature male parr. Journal of Fish Biology 31: 143-146. Nathan, R., Perry, G., Cronin, J. T., Strand, A. E. & Cain, M. L. 2003. Methods for estimating long-distance dispersal. Oikos 103: 261-273. Nei, M. 1987. Molecular evolutionary genetics. Columbia Univ Pr. Neville, H. M., Isaak, D. J., Dunham, J. B., Thurow, R. F. & Rieman, B. E. 2006. Finescale natal homing and localized movement as shaped by sex and spawning habitat in Chinook salmon: insights from spatial autocorrelation analysis of individual genotypes. Molecular Ecology 15: 4589-4602. Nichols, R. A. & Hewitt, G. M. 1994. The genetic consequences of long distance dispersal during colonization. Heredity 72: 312-317. Nordeng, H. 1983. Solution to the «char problem» based on Arctic char (Salvelinus alpinus) in Norway. Canadian Journal of Fisheries and Aquatic Sciences 40: 1372-1387. Nordeng, H. 2009. Char ecology. Natal homing in sympatric populations of anadromous Arctic char Salvelinus alpinus (L.): roles of pheromone recognition. Ecology of Freshwater Fish 18. Nordeng, H. & Bratland, P. 2006. Homing experiments with parr, smolt and residents of anadromous Arctic char Salvelinus alpinus and brown trout Salmo trutta: transplantation between neighbouring river systems. Ecology of Freshwater Fish 15: 488-499. Nosil, P. 2004. Reproductive isolation caused by visual predation on migrants between  157  divergent environments. Proceedings of the Royal Society of London B 271: 15211528. Nosil, P. & Crespi, B. J. 2004. Does gene flow constrain adaptive divergence or vice versa? A test using ecomorphology and sexual isolation in Timema cristinae walking-sticks. Evolution 58: 102-112. Nosil, P., Funk, D. J. & Ortiz-Barrientos, D. 2009. Divergent selection and heterogeneous genomic divergence. Molecular Ecology 18: 375-402. Nosil, P., Vines, T. H. & Funk, D. J. 2005. Reproductive isolation caused by natural selection against immigrants from divergent habitats. Evolution 59: 705-719. Nunavut Fisheries Strategy 2005. Government of Nunavut and Nunavut Tunngavik Incorporated pp. 50.. Nunes, S., Zugger, P. A., Engh, A. L., Reinhart, K. O. & Holekamp, K. E. 1997. Why do female Belding's ground squirrels disperse away from food resources? Behavioral Ecology and Sociobiology 40: 199-207. Paetkau, D., Calvert, W., Stirling, I. & Strobeck, C. 1995. Microsatellite analysis of population structure in Canadian polar bears. Molecular Ecology 4: 347-354. Paetkau, D., Slade, R., Burdens, M. & Estoup, A. 2004. Genetic assignment methods for the direct, real-time estimation of migration rate: a simulation-based exploration of accuracy and power. Molecular Ecology 13: 55-65. Papst, M. H. (1994) Variation in Growth of Hatchery Reared Arctic Charr, Salvelinus Alpinus (L.). Vol. PhD. pp. 402. University of Manitoba. Park, S. 2001. Microsatellite toolkit. Department of Genetics, Trinity College, Dublin, Ireland. Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual Reviews in Ecology, Evolution and Systematics 37: 637-69. Parmesan, C. & Yohe, G. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421: 37-42. Perrin, N. & Goudet, J. (2001) Inbreeding, kinship, and the evolution of natal dispersal. In: Dispersal (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.). pp. 123-142. Oxford University Press, Oxford, UK. Perrin, N. & Mazalov, V. V. 2000. Local competition, inbreeding, and the evolution of  158  sex-biased dispersal. The American Naturalist 155: 116. Perron, G. G., Gonzalez, A. & Buckling, A. 2007. Source–sink dynamics shape the evolution of antibiotic resistance and its pleiotropic fitness cost. Proceedings of the Royal Society London, Series B 274: 2351-2356. Petit, R. J. 2008. The coup de grâce for the nested clade phylogeographic analysis? Molecular Ecology 17: 516-518. Petit, R. J., Aguinagalde, I., de Beaulieu, J.-L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, R., Fineshi, S., Grivet, D., Lascoux, M., Mohanty, A., Müller-Stark, G., Demesure-Musch, B., Palmé, A., Martín, J. P., Rendell, S. & Vendramin, G. G. 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 1563-1565. Petit, R. J. & Excoffier, L. 2009. Gene flow and species delimitation. Trends in Ecology and Evolution 24: 386-393. Piry, S., Alapetite, A., Cornuet, J. M., Paetkau, D., Baudoin, L. & Estoup, A. 2004. GENECLASS2: a software for genetic assignment and first-generation migrant detection. Journal of. Heredity 95: 536-539. Porter, J. H. & Dooley, J., J.L. 1993. Animal dispersal patterns: a reassessment of simple mathematical models. Ecology 74: 2436-2443. Posada, D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution 25: 1253-1256. Power, G. (1969) The salmon of Ungava Bay. pp. 22. Arctic Institute of North America. Power, G. 2002. Charrs, glaciations and seasonal ice. Environmental biology of fishes 64: 17-35. Power, M., Dempson, J. B., Reist, J. D., Schwarz, C. J. & Power, G. 2005. Latitudinal variation in fecundity among Arctic charr populations in eastern North America. Journal of Fish Biology 67: 255-273. Power, G. & Power, M. 1995. Ecotones and fluvial regimes in arctic lotic environments. Hydrobiologia 303: 111-124. Power, M., Power, G., Reist, J. D. & Bajno, R. 2009. Ecological and genetic differentiation among the Arctic charr of Lake Aigneau, Northern Québec. Ecology of Freshwater Fish 18: 445-460.  159  Pritchard, J. K., Stephens, M. & Donnelly, P. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945-959. Priest, H. & Usher, P. J. (2004) Final Report: Nunavut Wildlife Harvest Study 1996-2001, prepared for the Nunavut Wildlife Management Board. pp. 816. Prouzet, P. 1981. Observation d'une femelle de tacon de saumon atlantique (Salmo salar L.) parvenue à maturité sexuelle en rivière. Bulletin Français de Pisciculture: 1619. Provan, J. & Bennett, K. D. 2008. Phylogeographic insights into cryptic glacial refugia. Trends in Ecology & Evolution 23: 564-571. Pulliam, H. R. 1988. Sources, sinks, and population regulation. The American Naturalist 132: 652-661. Quinn, T. P. 1993. A review of homing and straying of wild and hatchery-produced salmon. Fisheries Research 18: 29-44. Quinn, T. P. 2005. The behavior and ecology of Pacific salmon and trout. UBC Press. Quinn, T. P., Hodgson, S., Flynn, L., Hilborn, R. & Rogers, D. E. 2007. Directional selection by fisheries and the timing of sockeye salmon (Oncorhynchus nerka) migrations. Ecological Applications 17: 731-739. Quinn, T. P., Kinnison, M. T. & Unwin, M. J. 2001. Evolution of chinook salmon (Oncorhynchus tshawytscha) populations in New Zealand: pattern, rate, and process. Genetica 112: 493-513. R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing Vienna Austria. Rambaut, A. 2010. FigTree. Version 1.3. 1. Institute of Evolutionary Biology, University of Edinburgh, Edinburgh. Available online: http://tree. bio. ed. ac. uk/software/figtree. Read, C. J. 2000. Information from Arctic charr fisheries in the Baffin region, Nunavut, 1995 to 1999. Canadian Data Report in Fisheries and Aquatic Sciences 1067: x + 176 p. Réale, D., McAdam, A. G., Boutin, S. & Berteaux, D. 2003. Genetic and plastic responses of a northern mammal to climate change. Proceedings of the Royal Society B: Biological Sciences 270: 591-596.  160  Redenbach, Z. & Taylor, E. B. 2002. Evidence for historical introgression along a contact zone between two species of char (Pisces: Salmonidae) in northwestern North America. Evolution 56: 1021-1035. Reist, J. D. 1989. Genetic structuring of allopatric populations and sympatric life history types of charr, Salvelinus alipuns/malma, in the western Arctic, Canada. Physiology and Ecology Japan, Special Volume 1: 405-420. Reist, J. D., Gyselman, E., Babaluk, J. A., Johnson, J. D. & Wissink, R. 1995. Evidence for two morphotypes of Arctic char (Salvelinus alpinus L.) from Lake Hazen, Ellesmere Island, Northwest Territories, Canada. Nordic Journal of Freshwater Research. Reist, J. D., Johnson, J. D. & Carmichael, T. J. 1997. Variation and specific identity of char from northwestern arctic Canada and Alaska. American Fisheries Society Symposium 19: 250-261. Reist, J. D. & Sawatzky, C. D. (2010) Arctic char: indicator #6. In: Arctic biodiversity trends 2010 - selected indicators of change. pp. 41-44. Akureyi, Iceland: CAFF International Secretariat. Reusch, T. B. H. & Wood, T. E. 2007. Molecular ecology of global change. Molecular Ecology 16: 3973-3992. Rexroad III, C. E., Coleman, R. L., Martin, A. M., Hershberger, W. K. & Killefer, J. 2001. Thirty-five polymorphic microsatellite markers for rainbow trout (Oncorhynchus mykiss). Animal Genetics 32: 317. Rideout, R. M. & Tomkiewicz, J. 2011. Skipped spawning in fishes: more common than you might think. Marine and Coastal Fisheries 3: 176-189. Riechert, S. 1993. Investigation of potential gene flow limitation of behavioral adaptation in an aridland spider. Behavioural Ecology and Sociobiology 32: 355-363. Rikardsen, A. H. & Elliott, J. M. 2000. Variations in juvenile growth, energy allocation and life-history strategies of two populations of Arctic charr in North Norway. Journal of Fish Biology 56: 328-346. Rikardsen, A. H., Diserud, O. L. A. H., Elliott, J., Dempson, J. B., Sturlaugsson, J. & Jensen, A. J. 2007. The marine temperature and depth preferences of Arctic charr (Salvelinus alpinus) and sea trout (Salmo trutta), as recorded by data storage tags.  161  Fisheries Oceanography 16: 436-447. Rikardsen, A. H., Thorpe, J. E. & Dempson, J. B. 2004. Modelling the life-history variation of Arctic charr. Ecology Of Freshwater Fish 13: 305-311. Roff, D. A. 1996. The evolution of threshold traits in animals. The Quarterly Review of Biology 71: 3. Roff, D. A. & Fairbairn, D. J. 2001. The genetic basis of dispersal and migration, and its consequences for the evolution of correlated traits. In: Dispersal (Clobert, J., Danchin, E., Dhondt, A. A. & Nichols, J. D., eds.). pp. 191-202. Oxford University Press, Oxford. Rogers, A. R. & Harpending, H. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9: 552-569. Ronce, O. 2007. How does it feel to be like a rolling stone? Ten questions about dispersal evolution. Annual Reviews in Ecology, Evolution and Systematics. 38: 231-253. Ronce, O. & Kirkpatrick, M. 2001. When sources become sinks: migrational meltdown in heterogeneous habitats. Evolution 55: 1520-1531. Rosenberg, N. A. 2004. DISTRUCT: a program for the graphical display of population structure. Molecular Ecology Notes 4: 137-138. Rousset, F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 1219-1228. Roux, M. J., Tallman, R. F. & Lewis, C. W. 2011. Small-scale Arctic charr Salvelinus alpinus fisheries in Canada's Nunavut: management challenges and options. Journal of Fish Biology 79: 1625-1647. Rull, V. 2009. Microrefugia. Journal of Biogeography 36: 481-484. Savvaitova, K. A. 1995. Patterns of diversity and processes of speciation in Arctic charr. Nordic Journal of Freshwater Research 71: 81-91. Schenekar, T. & Weiss, S. 2011. High rate of calculation errors in mismatch distribution analysis results in numerous false inferences of biological importance. Heredity 107: 511-512. Schindler, D. E., Hilborn, R., Chasco, B., Boatright, C. P., Quinn, T. P., Rogers, L. A. & Webster, M. S. 2010. Population diversity and the portfolio effect in an exploited  162  species. Nature 465: 609-612. Schluter, D. 1996. Ecological speciation in postglacial fishes. Philosophical Transactions of the Royal Society of London B 351: 807-814. Schluter, D. 2002. The Ecology of Adaptive Radiation. Oxford University Press. Schluter, D. & Ricklefs, R. E. (1993) Species diversity: an introduction to the problem. In: Species diversity in ecological communities, (Ricklefs, R. E. & Schluter, D., eds.). pp. 1-10. Schneider, S. & Excoffier, L. 1999. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. Genetics 152: 1079. Scott, W. B. & Crossman, E. J. 1973. Freshwater fishes of Canada. Fisheries Research Board of Canada Bulletin 184. Shafer, A., Côté, S. D. & Coltman, D. W. 2010. Hot spots of genetic diversity descended from multiple Pleistocene refugia in an alpine ungulate. Evolution 65: 125-138. Sinervo, B. & Clobert, J. 2003. Morphs, dispersal behavior, genetic similarity, and the evolution of cooperation. Science 300: 1949-1951. Skúlason, S., Sorrason, S. S. & Jonsson, B. (1999) Sympatric morphs, populations and speciation in freshwater fish with emphasis on arctic charr. In: Evolution of biological diversity, (Magurran, A. E. & May, R. M., eds.). pp. 70-92. Oxford University Press, Oxford, UK. Slatkin, M. 1973. Gene flow and selection in a cline. Genetics 75: 733-756. Slatkin, M. 1976. The rate of spread of an advantageous allele in a subdivided population. In: Population genetics and ecology, (Karlin, S. & Nevo, E., eds.). pp. 767-780. Academic Press, Inc., New York. Slatkin, M. 1985. Gene flow in natural populations. Annual Reviews in Ecology, Evolution and Systematics 16: 393-430. Slatkin, M. 1987. Gene flow and the geographic structure of natural populations. Science 236: 787-792. Slatkin, M. 1993. Isolation by distance in equilibrium and non-equilibrium populations. Evolution 47: 264-279. Slatkin, M. & Barton, N. H. 1989. A comparison of three indirect methods for estimating  163  average levels of gene flow. Evolution 43: 1349-1368. Slatkin, M. & Hudson, R. R. 1991. Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics 129: 555562. Slatkin, M. & Maddison, W. P. 1989. A cladistic measure of gene flow inferred from the phylogenies of alleles. Genetics 123: 603-613. Slettan, A., Olsaker, I. & Lie, Ø. 1995. Atlantic salmon, Salmo salar, microsatellites at the SSOSL 25, SSOSL 85, SSOSL 311, SSOSL 417 loci. Animal genetics(Print) 26: 281-282. Sloggett, J. J. & Weisser, W. W. 2002. Parasitoids induce production of the dispersal morph of the pea aphid, Acyrthosiphon pisum. Oikos 98: 323-333. Smouse, P. E., Long, J. C. & Sokal, R. R. 1986. Multiple regression and correlation extensions of the Mantel test of matrix correspondence. Systematic zoology 35: 627-632. Snorrason, S. S., Skúlason, S., Jonsson, B., Malmquist, H. J., Jonasson, P. M., Sandlund, O. T. & Lindem, T. 1994. Trophic specialization in Arctic charr Salvelinus alpinus (Pisces; Salmonidae): morphological divergence and ontogenetic niche shifts. Biological Journal of the Linnean Society 52: 1-18. Soltis, D. E., Morris, A. B., McLachlan, J. S., Manos, P. S. & Soltis, P. S. 2006. Comparative phylogeography of unglaciated eastern North America. Molecular Ecology 15: 4261-4293. Spares, A. D., Stokesbury, M. J. W., O'Dor, R. K. & Dick, T. A. 2012. Temperature, salinity and prey availability shape the marine migration of Arctic char, Salvelinus alpinus, in a macrotidal estuary. Marine Biology 159: 1633-1646. Stamford, M. D. & Taylor, E. B. 2004. Phylogeographical lineages of Arctic grayling (Thymallus arcticus) in North America: divergence, origins and affinities with Eurasian Thymallus. Molecular Ecology 13: 1533-1549. Stearns, S. & Sage, R. 1980. Maladaptation in a marginal population of the mosquito fish, Gambusia affinis. Evolution 34: 65-75. Steele, C. A. & Storfer, A. 2006. Coalescent-based hypothesis testing supports multiple Pleistocene refugia in the Pacific Northwest for the Pacific giant salamander  164  (Dicamptodon tenebrosus). Molecular Ecology 15: 2477-2487. Steinberg, E. K., Lindner, K. R., Gallea, J., Maxwell, A., Meng, J. & Allendorf, F. W. 2002. Rates and patterns of microsatellite mutations in pink salmon. Molecular Biology and Evolution 19: 1198. Stenseth, N. C. & Lidicker, W. Z. 1992. Animal dispersal: small mammals as a model. Springer. Stewart, J. R., Lister, A. M., Barnes, I. & Dalén, L. 2010. Refugia revisited: individualistic responses of species in space and time. Proceedings of the Royal Society of London B 277: 661-671. Stokes, K. & Law, R. 2000. Fishing as an evolutionary force. Marine Ecology Progress Series 208. Storfer, A. & Sih, A. 1998. Gene flow and ineffective antipredator behavior in a streambreeding salamander. Evolution 52: 558-565. Strasburg, J. L. & Rieseberg, L. H. 2008. Molecular demographic history of the annual sunflowers Helianthus annuus and H. petiolaris - large effective population sizes and rates of long-term gene flow. Evolution 62: 1936-1950. Strasburg, J. L. & Rieseberg, L. H. 2011. How robust are "Isolation with Migration" analyses to violations of the IM model? A simulation study. Molecular Biology and Evolution 27: 297-310. Svenning, M.-A. & Gullestad, N. 2002. Adaptations to stochastic environmental variations: the effects of seasonal temperatures on the migratory window of Svalbard Arctic charr. Environmental Biology of Fishes 64: 165-174. Svenning, M., Smith-Nilsen, N. & Jobling, M. 1992. Sea water migration of Arctic charr (Salvelinus alpinus L.): correlation between freshwater growth and seaward migration, based on back-calculation from otoliths. Nordic Journal of Freshwater Research 67: 18-26. Swanson, H. K., Kidd, K. A., Babaluk, J. A., Wastle, R. J., Yang, P. P., Halden, N. M. & Reist, J. D. 2010. Anadromy in Arctic populations of lake trout (Salvelinus namaycush): otolith microchemistry, stable isotopes, and comparisons with Arctic char (Salvelinus alpinus). Canadian Journal of Fisheries and Aquatic Sciences 67: 842-853.  165  Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585. Tallman, R. F. & Healey, M. C. 1994. Homing, straying, and gene flow among seasonally seperated populations of Chum salmon (Oncorhynchus keta). Canadian Journal of Fisheries and Aquatic Sciences 51: 577-588. Tallman, R. F., Saurette, F. & Thera, T. 1996. Migration and life history variation in Arctic charr, Salvelinus alpinus. Écoscience 3: 33-41. Tallmon, D. A., Luikart, G. & Waples, R. S. 2004. The alluring simplicity and complex reality of genetic rescue. Trends in Ecology & Evolution 19: 489-496. Taylor, E. B. 1991. A review of local adaptation in Salmonidae, with particular reference to Pacific and Atlantic salmon. Aquaculture 98: 185–208. Taylor, E. B., Foote, C. J. & Wood, C. C. 1996. Molecular genetic evidence for parallel life-history evolution within a pacific salmon (sockeye salmon and kokanee, Oncorhynchus nerka). Evolution 50: 401-416. Taylor, E. B., Lowery, E., Lilliestråle, A., Elz, A. & Quinn, T. P. 2008. Genetic analysis of sympatric char populations in western Alaska: Arctic char (Salvelinus alpinus) and Dolly Varden (Salvelinus malma) are not two sides of the same coin. Journal of Evolutionary Biology 21: 1609-1625. Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7: 381397. Tessier, N. & Bernatchez, L. 2000. A genetic assessment of single versus double origin of landlocked Atlantic salmon (Salmo salar) from Lake Saint-Jean, Québec, Canada. Canadian Journal of Fisheries and Aquatic Sciences 57: 797-804. Thériault, V., Bernatchez, L. & Dodson, J. J. 2007a. Mating system and individual reproductive success of sympatric anadromous and resident brook charr, Salvelinus fontinalis, under natural conditions. Behavioral Ecology and Sociobiology 62: 51-65. Thériault, V., Dunlop, E. S., Dieckmann, U., Bernatchez, L. & Dodson, J. J. 2008. The impact of fishing-induced mortality on the evolution of alternative life-history tactics in brook charr. Evolutionary Applications 1: 409-423.  166  Thériault, V., Garant, D., Bernatchez, L. & Dodson, J. 2007b. Heritability of life-history tactics and genetic correlation with body size in a natural population of brook charr (Salvelinus fontinalis). Journal of Evolutionary Biology 20: 2266-2277. Thomaz, D., Beall, E. & Burke, T. 1997. Alternative reproductive tactics in Atlantic salmon: factors affecting mature parr success. Proceedings of the Royal Society of London. Series B: Biological Sciences 264: 219-226. Thorpe, J. E., Mangel, M., Metcalfe, N. B. & Huntingford, F. A. 1998. Modelling the proximate basis of salmonid life-history variation, with application to Atlantic salmon, Salmo salar L. Evolutionary Ecology 12: 581-599. Toews, D. P. L. & Brelsford, A. 2012. The biogeography of mitochondrial and nuclear discordance in animals. Molecular Ecology 21: 3907-3930. Tonteri, A., Veselov, A. J., Titov, S., Lumme, J. & Primmer, C. R. 2007. The effect of migratory behaviour on genetic diversity and population divergence: a comparison of anadromous and freshwater Atlantic salmon Salmo salar. Journal of Fish Biology 70: 381-398. Trakhtenbrot, A., Nathan, R., Perry, G. & Richardson, D. M. 2005. The importance of long-distance dispersal in biodiversity conservation. Diversity and Distributions 11: 173-181. Travis, J. M. J. & Dytham, C. 1999. Habitat persistence, habitat availability and the evolution of dispersal. Proceedings of the Royal Society of London. Series B: Biological Sciences 266: 723-728. Travis, J. M. J., Mustin, K., Barto!, K. A., Benton, T. G., Clobert, J., Delgado, M. M., Dytham, C., Hovestadt, T., Palmer, S. C. F. & Van Dyck, H. 2012. Modelling dispersal: an eco-evolutionary framework incorporating emigration, movement, settlement behaviour and the multiple costs involved. Methods in Ecology and Evolution: doi: 10.1111/j.2041-210X.2012.00193.x Tremblay, N. O. & Schoen, D. J. 1999. Molecular phylogeography of Dryas integrifolia: glacial refugia and postglacial recolonization. Molecular Ecology 8: 1187-1198. Turgeon, J. & Bernatchez, L. 2001. Clinal variation at microsatellite loci reveals historical secondary intergradation between glacial races of Coregonus artedi (Teleostei: Caregoninae). Evolution 55: 2274-2286.  167  Valone, T. J. & Templeton, J. J. 2002. Public information for the assessment of quality: a widespread social phenomenon. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 357: 1549-1557. Van Oosterhout, C., Hutchinson, W. F., Wills, D. P. M. & Shipley, P. 2004. MICROCHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4: 535-538. Van Valen, L. 1971. Group selection and the evolution of dispersal. Evolution 25: 591598. Venable, D. L. & Brown, J. S. 1988. The selective interactions of dispersal, dormancy, and seed size as adaptations for reducing risk in variable environments. American Naturalist: 360-384. Verspoor, E. & Cole, L. J. 1989. Genetically distinct sympatric populations of resident and anadromous Atlantic salmon, Salmo salar. Canadian Journal of Zoology 67: 1453–1461. Via, S. 1999. Reproductive isolation between sympatric races of pea aphids. I. Gene flow restriction and habitat choice. Evolution: 1446-1457. Vitalis, R. 2002. Sex-specific genetic differentiation and coalescence times: estimating sex-biased dispersal rates. Molecular Ecology 11: 125-138. Walter, R. P., Aykanat, T., Kelly, D. W., Shrimpton, J. M. & Heath, D. D. 2009. Gene flow increases temporal stability of Chinook salmon (Oncorhynchus tshawytscha) populations in the Upper Fraser River, British Columbia, Canada. Canadian Journal of Fisheries and Aquatic Sciences 66: 167-176. Waples, R. S. & Gaggiotti, O. 2006. What is a population? An empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Molecular Ecology 15: 1419-1439. Waples, R. S. 1998. Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. Journal of Heredity 89: 438-450. Watkinson, A. R. & Sutherland, W. J. 1995. Source, sinks and pseudo-sinks. Journal of Animal Ecology 64: 126-130. Weir, B. S. & Cockerham, C. C. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370.  168  Whitlock, M. C. 2011. G'st and D do not replace Fst. Molecular Ecology 20: 1083-1091. Whitlock, M. C. & Barton, N. H. 1997. The effective size of a subdivided population. Genetics 146: 427. Whitlock, M. C. & McCauley, D. E. 1999. Indirect measures of gene flow and migration: FST " 1/(4Nm+ 1). Heredity 82: 117-125. Whitlock, M. C. & Phillips, P. C. 2000. The exquisite corpse: a shifting view of the shifting balance. Trends in ecology and evolution 15: 347. Williams, S. E., Shoo, L. P., Isaac, J. L., Hoffmann, A. A. & Langham, G. 2008. Towards an integrated framework for assessing the vulnerability of species to climate change. PLoS Biology 6: 2621-2626. Williamson, K. S., Cordes, J. F. & May, B. 2002. Characterization of microsatellite loci in chinook salmon (Oncorhynchus tshawytscha) and cross species amplification in other salmonids. Molecular Ecology Notes 2: 17-19. Wilson, C. C. & Bernatchez, L. 1998. The ghost of hybrid past: fixation of arctic charr (Salvelinus alpinus) mitochondrial DNA in an introgressed population of lake trout (S. namaycush). Molecular Ecology 7: 127-132. Wilson, C. C., Hebert, P. D. N., Reist, J. D. & Dempson, J. B. 1996. Phylogeography and postglacial dispersal of arctic charr Salvelinus alpinus in North America. Molecular Ecology 5: 187-197. Wirtz, P. 1999. Mother species-father species: unidirectional hybridization in animals with female choice. Animal behaviour 58: 1-12. Wolff, J. O. 1992. Parents suppress reproduction and stimulate dispersal in opposite-sex juvenile white-footed mice. Nature 359: 409-410. Wood, C. C. & Foote, C. J. 1996. Evidence for sympatric genetic divergence of anadromous and nonanadromous morphs of sockeye salmon(Oncorhynchus nerka). Evolution 50: 1265-1279. Wright, S. 1931. Evolution in Mendelian populations. Genetics 16: 97-159. Wright, S. 1932. The roles of mutation, inbreeding, crossbreeding and selection in evolution. In: Proceedings of the VIth International Congress of Genetics, Vol. 1. pp. 356-366. Wright, S. 1943. Isolation by distance. Genetics 28: 114.  169  Wu, C. 2001. The genic view of the process of speciation. Journal of Evolutionary Biology 14: 851-865. Yeaman, S. & Otto, S. P. 2011. Establishment and maintenance of adaptive genetic divergence under migration, selection, and drift. Evolution 65: 2123-2129. Zwickl, D. J. (2006) GARLI: genetic algorithm for rapid likelihood inference. In: See http://www. bio. utexas. edu/faculty/antisense/garli/Garli. html, Vol. PhD Dissertation. pp. The University of Texas at Austin, Austin.  170  >//5$6)?*>@*A'//&5.5$104=*.0154)0&(*B#4*"<0/154*2** Table A.1. Name and geographic coordinates of sampling locations where specimens of Arctic Char (ARCH, Salvelinus alpinus) and Dolly Varden (DVCH, Salvelinus malma) used for mtDNA sequencing were collected. ‘N’ indicates the sample size from each sampling location. The samples are approximately arranged from west to east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Table S1. Continued A)15*$0.5%  C58)#$%  A/5%)5(%  D01)1'65%  D#$8)1'65%  E%  Y'IP(.<%1,C./%  >"<)(%U(#"<L%  01&2%  36549%  A**95:=%  8%  B"?#I/$E")%1,C./%  R/,<;.%0#-./@%+IE<L%  01&2%  34564%  A***584%  *4%  B"?#E)%1,C./%  R/,<;.%0#-./@%+IE<L%  01&2%  345::%  A**:5F3%  :4%  BEEZZ"E%1,C./%  R/,<;.%0#-./@%+IE<L%  01&2%  3*5:6%  A**75=6%  :*%  BEE)%1,C./%  R/,<;.%0#-./@%+IE<L%  01&2%  345=3%  A**:576%  *4%  64%!,#.%T").%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F5:9%  A*4954:%  7%  >$/I<%>"$%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F543%  A*4F5:6%  *=%  &"P-/,L?.%>"$%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F5*6%  A*4859:%  :6%  S./?E(I<%T").%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F563%  A*4=54=%  *4%  B,@,?"%T").%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F5*8%  A*4=563%  *4%  G,('-I<.%T").%  +5W5%O,;@I/,"%U(#"<L%  01&2%  7F568%  A*485*4%  *4%  GI##"(@I<%R.<,<(E#"%  O,;@I/,"%U(#"<L%  01&2%  7F5:3%  A**85F3%  8%  0P,@E)%  &I/<Q"##,(%U(#"<L%  01&2%  3=544%  AF8544%  *=%  1.(I#E@.%  &I/<Q"##,(%U(#"<L%  01&2%  3857F%  AF85F8%  ::%  &'"/%  &I/<Q"##,(%U(#"<L%  01&2%  7F57:%  A*465=9%  9%  HI/@'%T").%  &I/<Q"##,(%U(#"<L%  01&2%  38533%  AF=5*4%  8%  >./<"/L%2"/-IE/%  &I/I<"@,I<%DE#J%0/."%  01&2%  79599%  A**85=4%  :=%  1$P./%RI,<@%  &I/I<"@,I<%DE#J%0/."%  01&2%  7F54:%  A**65:=%  64%  T5%H"E$E)%  B.<@%R.<,<(E#"%  01&2%  79568%  A*4353=%  *4%  T,@@#.%H"E$E)%T").%  B.<@%R.<,<(E#"%  01&2%  79568%  A*4353=%  36%  &I<@QI$@I%T").%  HGY%."(@%IJ%D/."@%>."/%T5%  01&2%  7=59:%  A***543%  F%  >/I;)%T").%  R"E#"@E)%0/."%  01&2%  7F589%  A*:656=%  =%  2I/<"L"$%1,C./%  R"E#"@E)%0/."%  01&2%  7F566%  A*:653F%  89%  0//IQ(P,@'%1,C./%  R.##$%>"$%  01&2%  79563%  AF4566%  **%  >.;'./%1,C./%  R.##$%>"$%  01&2%  7957:%  AF45==%  *:%  B.,@'%>"$%1,C./%  R.##$%>"$%  01&2%  7956:%  A995:3%  =%  B.##.@%1,C./%  R.##$%>"$%  01&2%  79566%  AF4549%  *:%  YIE/,(@%1,C./%  R.##$%>"$%  01&2%  795F4%  AF4576%  86%  >.C./#$%U<#.@%  !.#C,##.%U(#"<L%  01&2%  3=543%  A*43596%  7%  NE<L"(%2"/-IE/%  N.CI<%U(#"<L%  01&2%  38574%  A9:58:%  9%  !E//"$%  W##.(P./.%U(#"<L%  01&2%  9*566%  A7F5=3%  *4%  T").%>%  W##.(P./.%U(#"<L%  01&2%  9:5*6%  A79589%  3%  0#.["<L/"%  W##.(P./.%U(#"<L%  01&2%  9*533%  A7=5=6%  9%  0#.["<L/"%  W##.(P./.%U(#"<L%  01&2%  9*533%  A7=5=6%  6*%  &/",?%  W##.(P./.%U(#"<L%  01&2%  9*59=%  A795F:%  *6%  2"K.<%T").%  W##.(P./.%U(#"<L%  01&2%  9*536%  A3:583%  =:%  2.,<@K.#P"<%  W##.(P./.%U(#"<L%  01&2%  9*534%  A775F6%  :3%  T").%YE-I/?%  W##.(P./.%U(#"<L%  01&2%  945F=%  A3=5=9%  8%  !E()I[%S,I/L%  W##.(P./.%U(#"<L%  01&2%  37584%  A935*6%  :3%  &"V.%1I-./@%>/IQ<%  !.#C,##.%R.<,<(E#"%  01&2%  735=9%  A9*536%  =%  172  Table S1. Continued A)15*$0.5%  C58)#$%  A/5%)5(%  D01)1'65%  D#$8)1'65%  E%  D/,<.##%T").%  !.#C,##.%R.<,<(E#"%  01&2%  7F5=3%  A965F:%  F%  2"##%T").%  !.#C,##.%R.<,<(E#"%  01&2%  79573%  A9:5=4%  F%  U;.%T").%  !.#C,##.%R.<,<(E#"%  01&2%  7F5:3%  A98573%  F%  R/,<;.%T").%  !.#C,##.%R.<,<(E#"%  01&2%  7F5=3%  A985*9%  F%  O"<%1""$%T").%  !.#C,##.%R.<,<(E#"%  01&2%  7F5:3%  A96576%  F%  S"#(.%2IV.%T").%  !.#C,##.%+IE<L%  01&2%  795*9%  A*475=9%  7%  0<?,P"ZE\%1,C./%  !.#C,##.%+IE<L%  01&2%  795*3%  A*475:9%  7%  &"V.%2IIV./%  >"JJ,<%U(#"<L%  01&2%  79586%  A7759=%  =%  YE?""@%  >"JJ,<%U(#"<L%  01&2%  3:543%  A9456=%  :F%  B,<?<",@%  >"JJ,<%U(#"<L%  01&2%  7756:%  A78563%  =%  H"#E(,"\%T").%  >"JJ,<%U(#"<L%  01&2%  77579%  A7=548%  :=%  ]"(,?,"@%  >"JJ,<%U(#"<L%  01&2%  7857:%  A7756*%  8%  D,JJI/L%  >"JJ,<%U(#"<L%  01&2%  3456:%  A9654=%  7%  1"C<%  >"JJ,<%U(#"<L%  01&2%  34583%  A3F5=4%  ::%  B"@E\"M%R/,@K#./%2"/-IE/%  >"JJ,<%U(#"<L%  01&2%  7F5*8%  A*4=563%  6%  N,"<"%1,C./%  1"<),<%U<#.@%0/."%  01&2%  7:596%  AF:563%  9%  !.#,"L,<.%1,C./%  1"<),<%U<#.@%0/."%  01&2%  7:599%  AF:5*6%  9%  2I/(.('I.%T").%  ]E^-.;%  01&2%  7:569%  A3653:%  F%  T";%T"S#"PP.%  ]E^-.;%  01&2%  7*566%  A3653:%  3%  T").%0,?<."E%  ]E^-.;%  01&2%  =35:6%  A345*:%  8:%  R,<?"#E,@%  ]E^-.;%  01&2%  7*5:9%  A36579%  :4%  U))"/E"#,)%T").%%  >.#;'./%U(#"<L(%  01&2%  ==5F=%  A3F5=4%  *4%  UP,@"C,)%  >.#;'./%U(#"<L(%  01&2%  ==5F=%  A3F5=4%  **%  UV"/"E@"%T").%_TI<?%T").`%  >.#;'./%U(#"<L(%  01&2%  ==5F=%  A3F5=4%  =%  B"(.?"#,)%T").%  >.#;'./%U(#"<L(%  01&2%  ==5F=%  A3F5=4%  3%  Y"(./((E,@%T").a1,C./%  D/..<#"<L%  01&2%  33534%  A7F584%  38%  *  %2"V#I@$V.(%@").<%J/IP%D.<-"<)5%b/,?,<"##$%VE-#,('.L%,<%0#.)(.$.C%.@%"#5%:44F% %2"V#I@$V.(%@").<%J/IP%D.<>"<)%J/IP%0$./(%_:4*4`5%Y'I(.%("PV#.(%Q./.%<I@%,<;#EL.L%,<%@'.%',(@I/,;"#% L.PI?/"V'$%"<"#$(,(%-.;"E(.%U%L,L%<I@%'"C.%";;.((%@I%J/.\E.<;$%L"@"% :  173  Table A.2. Sampling location information for the samples used in the microsatellite analysis of Arctic Char (Salvelinus alpinus) A)15*%#65*  A)15*$0.5*  C58)#$*  D01*  D#$*  E*  D)B5F<)(1#4=*  &T%  &"/,-IE%T").%  +IE@'."(@%0#"()"%  7458=%  A*=8544%  :8%  T"<L#I;).L%  +T%  +EPP,@%T").%  +IE@'."(@%0#"()"%  =F534%  A*=6579%  =7%  T"<L#I;).L%  YT%  YI?,")%T").%  +IE@'."(@%0#"()"%  =F57*%  A*=F576%  ::%  T"<L#I;).L%  T)*46%  T").%*46%  XE)I<%HI/@'%(#IV.%  7F586%  A*6F5=3%  84%  T"<L#I;).L%  T)*48%  T").%*48%  XE)I<%HI/@'%(#IV.%  7F586%  A*6F574%  84%  T"<L#I;).L%  1"L%  1"LL,%T").%  >"<)(%U(#"<L%  3*573%  A*:653=%  *:%  T"<L#I;).L%  !,L%  !,LL#.%T").%  >"<)(%U(#"<L%  3*596%  A*:=566%  :F%  T"<L#I;).L%  &"V%  &"V/I<%T").%  >"<)(%U(#"<L%  3*596%  A*:85::%  :F%  T"<L#I;).L%  +";%  +";'(%1,C./%  >"<)(%U(#"<L%  3*5F3%  A*:=5*4%  *3% 0<"L/IPIE(%  Y'I%  Y'IP(.<%1,C./%  >"<)(%U(#"<L%  365::%  A**F5=8%  *:% 0<"L/IPIE(%  BEE%  BEE)%1,C./%  G5%O,;@I/,"%U(#"<L%  345=3%  A**:576%  :4% 0<"L/IPIE(%  B"?%  B"?#E)%1,C./%  G5%O,;@I/,"%U(#"<L%  345::%  A**:5F3%  *=% 0<"L/IPIE(%  H"#%  H"#I"?$E)%1,C./%  G5%O,;@I/,"%U(#"<L%  345::%  A**:5::%  *3% 0<"L/IPIE(%  BEEZ%  BEZZE"%1,C./%  G5%O,;@I/,"%U(#"<L%  3*5:6%  A**75=6%  :4% 0<"L/IPIE(%  H"E$%  T,@@#.%H"E$E)%T").%  &"P-/,L?.%>"$%0/."%  79568%  A*4353=%  :7%  T"<L#I;).L%  1.(%  1.(I#E@.%T").%  &I/<Q"##,(%U(#"<L%  38579%  AF8599%  84%  T"<L#I;).L%  0P,%  0P,@E)%T").%  &I/<Q"##,(%U(#"<L%  3=544%  AF8544%  *4%  T"<L#I;).L%  TI/%  TI/L%T,<L(.$%T").%  >II@',"%R.<,<(E#"%  345*4%  AF:5:=%  :9% 0<"L/IPIE(%  YIE%  YIE/,(@%1,C./%  >II@',"%R.<,<(E#"%  795F4%  AF4576%  :F% 0<"L/IPIE(%  0//%  0//IQ(P,@'%1,C.%  >II@',"%R.<,<(E#"%  79563%  AF4566%  *6% 0<"L/IPIE(%  >.;%  >.;'./%1,C./%  >II@',"%R.<,<(E#"%  7957:%  AF45==%  3%  0<"L/IPIE(%  R/,%  R/,<;.%T").%  !.#C,##.%R.<,<(E#"%  7F5=3%  A985*9%  F%  0<"L/IPIE(%  D/,%  D/,<<.##%T").%  !.#C,##.%R.<,<(E#"%  7F5=3%  A965F:%  F%  0<"L/IPIE(%  2"#%  2"##%T").%  !.#C,##.%R.<,<(E#"%  79573%  A9:5=4%  *7% 0<"L/IPIE(%  D,J%  D,JJI/L%1,C./%  G.(@./<%>"JJ,<%U(#"<L%  34583%  A3F5=4%  :F% 0<"L/IPIE(%  1"C%  1"C<%1,C./%  G.(@./<%>"JJ,<%U(#"<L%  3456:%  A9654=%  64% 0<"L/IPIE(%  0#.[%  0#.["<L/"%T").%  W##.(P./.%U(#"<L%  9*533%  A7=5=6%  64%  T"<L#I;).L%  T").>%  T").%>%  W##.(P./.%U(#"<L%  9:5*6%  A79589%  :3%  T"<L#I;).L%  !E/%  !E//"$%T").%  W##.(P./.%U(#"<L%  9*566%  A7F5=3%  64%  T"<L#I;).L%  !E(%  !E()I[%S,I/L%  W##.(P./.%U(#"<L%  37583%  A9759=%  :F% 0<"L/IPIE(%  B,<%  B,<?<",@%  +IE@'."(@%>"JJ,<%U(#"<L%  7856:%  A77569%  8*% 0<"L/IPIE(%  U\"%  U\"#E,@%  +IE@'."(@%>"JJ,<%U(#"<L%  735*:%  A7=54:%  84% 0<"L/IPIE(%  U(E%  U(E,@E\%  +IE@'."(@%>"JJ,<%U(#"<L%  795:4%  A77596%  89% 0<"L/IPIE(%  H"#E%  H"#E,(,"\%T").%  +IE@'."(@%>"JJ,<%U(#"<L%  7359=%  A775:6%  *=% 0<"L/IPIE(%  B"(%  B"(.?"#,)%T").%  >.#;'./%U(#"<L(%  ==5F=%  A3F5=4%  :=% 0<"L/IPIE(%  R.@%  R.@./%T").%  G.(@./<%2EL(I<%>"$%  765*6%  AF:594%  :6% 0<"L/IPIE(%  U>%  U)"L#,C,)%>/II)%  T"-/"LI/%  =956%  A7:5*3%  :8% 0<"L/IPIE(%  174  *  !"#"$"%&'"()" %  ;'<=>+%  ;'<=>?%  9:+,%  '8.%  67*,2%  67*,0%  67*,1%  67*,.%  67*,,%  67*,/%  )*+0,%  )*+0/%  )*+.-%  )*+.5%  )*+.4%  )*+.3%  )*+.2%  )*+.1%  )*+.0%  )*+..%  )*+.,%  !"#$%&'($%  )*+./%  )*+,-%  Table A.3. Frequency of haplotypes of Arctic Char (Salvelinus alpinus) in each sampling locations.  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  **%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  0%  +%  +%  +%  +%  +%  +%  +%  +%  1"2"/$33"45$6,# %  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  7%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  ;%  +%  +%  +%  +%  +%  +%  +%  +%  +%  19#)9-)" %  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  <%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *  59-"/")' %  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  =%  +%  +%  +%  +%  +%  +%  +%  +%  +%  >9#:)%?:6,-%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  @%  +%  +%  +%  +%  @%  @%  +%  ABC,%?:6,-%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  <%  +%  +%  +%  7%  7%  +%  +%  D-"(,-%E"),%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  0%  +%  +%  +%  +%  +%  +%  +%  5"-#9)%E"),%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *%  +%  @%  +%  +%  +%  *%  *%  &"-:FBB%E"),%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *0%  +%  +%  +%  +%  +%  +%  +%  G9CC:3%E"),%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  7%  +%  *%  +%  +%  +%  +%  E"),%*+<%  @@%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *  &',-,.',/ % *  89(:/)" % *  E"),%*+H%  **%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  &"/B,%?:6,- %  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  0%  +%  +%  +%  +%  +%  +%  +%  D:-3'%?:6,-%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  @+%  +%  +%  +%  I%  *%  +%  +%  ?"JJ:%  *7%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  !:JJ#,%  =%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  &"K-B/%  =%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  1,-/"-J%?:6,-%  ;%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  59K3"/%E"),%  7%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  L"-),-%?:6,-%  0%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  G".'(%?:6,-%  ;%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  M'BC(,/%?:6,-%  <%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *  175  )*+,-%  )*+./%  )*+.,%  )*+..%  )*+.0%  )*+.1%  )*+.2%  )*+.3%  )*+.4%  )*+.5%  )*+.-%  )*+0/%  )*+0,%  67*,/%  67*,,%  67*,.%  67*,1%  67*,0%  67*,2%  '8.%  9:+,%  ;'<=>?%  ;'<=>+%  Table A3. Continued  5"2#B-$9")%?:6,-%  ;%  +%  +%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  5"2#9)%?:6,-%  @+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  599NN"9%?:6,-%  @+%  +%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  599)%?:6,-%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  <+%!:#,%E"),%  7%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  1$-B/%1"$%  *0%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  &"CF-:J2,%1"$%  @<%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  D,-29(B/%E"),%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  5:3:2"%E"),%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  >:('FB/,%E"),%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  >B##"(3B/%L,/O%  H%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  PC:39)%  ;%  7%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  ?,(B#93,%  *I%  0%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  &'"-%  I%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  AB-3'%E"),%  H%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  1,-/"-J%Q"-FB9-%  @<%  +%  +%  +%  +%  @%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  ?$C,-%LB:/3%  <+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  EO%A"9$9)%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  EO%A"9$9)%E"),%  I@%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  &B/3RB$3B%E"),%  ;%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  1-B.)%E"),%  0%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  QB-/"J"$%?:6,-%  H=%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  P--BR(C:3'%?:6,-%  *+%  +%  +%  +%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  1,.',-%?:6,-%  **%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  5,:3'%1"$%?:6,-%  H%  +%  +%  +%  +%  +%  +%  +%  +%  +%  *%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  +%  !"#$%&'($%  176  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able A3. Continued  &"K,%?BF,-3%  177  )*+,-%  )*+./%  )*+.,%  )*+..%  )*+.0%  )*+.1%  )*+.2%  )*+.3%  )*+.4%  )*+.5%  )*+.-%  )*+0/%  )*+0,%  67*,/%  67*,,%  67*,.%  67*,1%  67*,0%  67*,2%  '8.%  9:+,%  ;'<=>?%  ;'<=>+%  Table A3. Continued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Table A.4. Summary population statistics for the microsatellite analysis across nine loci in Arctic Char (Salvelinus alpinus). ‘HO’ is observed heterozygosity, ‘HE’ is Nei’s (1987) unbiased expected heterozygosity, ‘SD’ is standard deviation, and ‘AR’ is allelic richness calculated in FSTAT based on a minimum sample size of 1 diploid individual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Figure A.1. Details of the STRUCTURE analysis for Arctic Char (Salvelinus alpinus) assayed at nine microsatellite DNA loci are presented in the text. (a) lnProbability of the data calculated for each K (Pritchard et al. 2000). Note that for unknown reasons, two of the 20 independent runs at K=3 did not converge. This led to high variance which biased the deltaK calculations. I eliminated these two runs to make the following graphs in STRUCTURE harvester and from all subsequent analyses. (b) Delta K (Evanno et al. 2005) values for each K.  (a)  (b)  180  260  250  300  350  400  450 270  260  140  IB  IB  CL SL TL Lk103 Lk104 Rad Mid Cap Sac Tho Kuu Kag Nal Kuuj Nauy Res Ami Alex LakeB Mur Mus  260  IB  220  Lor Tou Arr Bec Pri Gri Hal Gif Rav Kin Iqa Isu Nalu Kas Pet  180  Lor Tou Arr Bec Pri Gri Hal Gif Rav Kin Iqa Isu Nalu Kas Pet  Ssosl456  Lor Tou Arr Bec Pri Gri Hal Gif Rav Kin Iqa Isu Nalu Kas Pet  Smm24  CL SL TL Lk103 Lk104 Rad Mid Cap Sac Tho Kuu Kag Nal Kuuj Nauy Res Ami Alex LakeB Mur Mus  220  Smm22  180  280  290  300 150  200  250  Smm24 | Allele lenght (bp)  80  100  140  OtsG253b  181  IB  Lor Tou Arr Bec Pri Gri Hal Gif Rav Kin Iqa Isu Nalu Kas Pet  CL SL TL Lk103 Lk104 Rad Mid Cap Sac Tho Kuu Kag Nal Kuuj Nauy Res Ami Alex LakeB Mur Mus  100  140  180  OtsG253b | Allele lenght (bp) Omm1105 | Allele lenght (bp)  Sco215  140  Smm24 | Allele lenght (bp)  220  Sco220  Ssosl456 | Allele lenght (bp)  180  Sco215 Sco220 | Allele lenght (bp) Allele length (bp)| Allele lenght (bp)  Sco200  CL SL TL Lk103 Lk104 Rad Mid Cap Sac Tho Kuu Kag Nal Kuuj Nauy Res Ami Alex LakeB Mur Mus  IB  Lor Tou Arr Bec Pri Gri Hal Gif Rav Kin Iqa Isu Nalu Kas Pet  CL SL TL Lk103 Lk104 Rad Mid Cap Sac Tho Kuu Kag Nal Kuuj Nauy Res Ami Alex LakeB Mur Mus  140  Smm22 | Allele lenght (bp)  Figure A.2. Bubbles plots of microsatellite allele frequencies showing introgression of  Labrador Arctic Char (Salvelinus alpinus) alleles (the site at the extreme right) in other  population of the Arctic Archipelago. Sampling locations are roughly arranged from west  (left) to east (right) and locations where most of the Labrador alleles were observed are  between the dashed lines. Marker Omm1128 is not shown because it was nearly  monomorphic.  Omm1105  ,00$12"3%45%!600-$7$1#89:%78#$9"8-.%;+9%<=80#$9%>% Gonadosomatic index I compared the frequency distributions of gonadosomatic index (GSI) of male and female Arctic Char (Salvelinus alpinus) to determine the reproductive status of individual fish. GSI was calculated by dividing the gonad weight (g) by the total round weight (g) of each fish. Males and females were analyzed separately because of different gonad size (a breeding male can have gonads that are smaller than that of a non-breeding female). Figure B.1 Frequency distributions of gonadosomatic index of males (top) and females (bottom) Arctic Char (Salvelinus alpinus). The dashed lines denote the cutoffs between breeding and nonbreeding individuals discussed in the text. M 50 40 30  Frequency (%)  20 10 0  F 50 40 30 20 10 0 0.00  0.05  0.10  0.15  0.20  0.25  GSI  I found bimodal distributions of GSI for both males and females with almost no overlap between the two distributions (Figure S1). Based on the distributions, I assigned all males with a Figure S1. Frequency distributions of gonadosomatic index (GSI) for male (top) and female (bottom) Arctic char.  GSI greater than 0.007 as breeders, while males with a GSI lower than 0.007 were assigned as  182  non-breeders. For females, individuals with a GSI greater than 0.02 were assigned as breeder and individuals with GSI smaller than 0.02 were assigned as non-breeders. Microsatellite loci scoring Individual Arctic Char (Salvelinus alpinus) genotypes were obtained at 18 microsatellite loci combined in 4 multiplexes (Table S1). For each locus, the forward primer was labeled with a fluorescent dye, and the reverse primer was PIG-tailed to reduce stutter and facilitate genotyping (Brownsteain et al. 1996). PCR amplifications were carried out in 10µL volume reactions (see Table S1 for details). The PCR cycles were as follows: an initial denaturation step of 10 minutes at 95°C, 35 cycles of denaturation (45 seconds at 94°C), annealing (45 seconds at 55°C) and extension (45 seconds at 72°C), and a final extension cycle of 30 minutes at 72°C. Table B.1. Details of the primers and PCR reactions used for each of the 4 multiplexes. Multiplex  mpAC1  mpAC2a  mpAC2b  mpAC3  mpAC4  b  b  [Primer]  [MgCl2]  [dNTPs]  (!M)  (mM)  (!M)  1.50  200  1.00  2.00  200  0.50  1.50  200  0.50  2.00  200  0.50  2.00  200  1.00  Primer  Dye  Reference  Sco200  VIC  TDL33U#3U6#FH6HDU#$&&(#  0.40  Smm22  NED  !H3UD#DS#3A'#$&&%#  0.40  Sco220  6-FAM  TDL33U#3U6#FH6HDU#$&&(#  0.50  Sco215  PET  TDL33U#3U6#FH6HDU#$&&(#  0.30  Sco212  6-FAM  TDL33U#3U6#FH6HDU#$&&(#  0.50  Sco218  VQ!#  TDL33U#3U6#FH6HDU#$&&(#  0.50  Sfo18  @WT#  FU?DHE#DS#3A'#+**(#  0.20  Sco202  PET  TDL33U#3U6#FH6HDU#$&&(#  0.20  Smm21  VIC  !H3UD#DS#3A'#$&&%#  0.16  OMM1128  VQ!#  5DOH=36#DS#3A'#$&&+#  0.16  Smm24  @WT#  !H3UD#DS#3A'#$&&(#  0.20  OtsG253b  VQ!#  X8AA83GE=U#DS#3A'#$&&$#  0.12  SSOSL456  6-FAM  -ADSS3U#DS#3A'#+**)#  0.50  OMM1105  JW/#  5DOH=36#DS#3A'#$&&$#  0.20  OtsG83b  6-FAM  X8AA83GE=U#DS#3A'#$&&$#  0.50  Smm17  @WT#  !H3UD#DS#3A'#$&&%#  0.16  Sco109  VQ!#  Shaklee 2003  0.50  Sco216  JW/#  TDL33U#3U6#FH6HDU#$&&(#  0.40  a  AmpliTaq Gold® DNA Polymerase with Gold Buffer and MgCl2 solution from Applied Biosystems  b  mpAC2a and mpAC2b were combined post-PCR for the genotyping  Taq  a  183  Assessment of scoring error In order to assess the error rate associated with the genotyping procedure, I randomly selected 96 individuals for which I re-ran the PCR and re-genotyped. The scoring error rate was calculated for each locus by dividing the number of alleles that were mis-called (i.e., alleles that differed between the original genotypes and the 96 randomly selected individuals that were re-genotyped) by the total number of alleles called for that locus on the replicate plate. Because some loci did not amplify (either in the original genotyping or on the replicated plate), the total number of alleles used for the calculation of the error rate was often less than 184 (96 x 2 alleles). The 96 randomly selected individuals which I re-amplified and re-genotyped lead to the following locus-specific estimates of error rates: Sco200: 1.9%, Sco202: 2.5%, Sco212: 8.4%, Sco215: 0%, Sco216: 4.1%, Sco218: 3.3%, Sco220: 6.3%, Smm17: 0%, Smm21: 0%, Smm22: 0%, Smm24: 0.6%, Sco109: 2.7%, OtsG253b: 2.5%, OtsG83b: 1.3% Omm1105: 6.3%, Sfo18: 0%, SSOSL456: 1.9% for a total average scoring error rate of 2.5%. Locus Omm1128 failed on the replicated plate, and I therefore do not have an estimate of the error rate for this locus. The three loci identified by MICROCHECKER as problematic (Sco109, Sco212 and Sco218) also had higher than average scoring error rates, and once I eliminated them, the average scoring error rate was reduced to 2.0%.  184  Table B.2. Heterozygosity (observed and expected) and FIS (Weir & Cockerham 1984) per locus in samples of Arctic Char (Salvelinus alpinus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Table B2 cont’d !"#$%&'(&)*+,-. !"#$% '//(0  &'()*+,  &45)*+,  -/.3)+,  -&6)*+,  /5.3)+,  458:6:  456:97  456(;:  45669;  45678  4564((  1*#<  456878  4560(;  4567(=  456:(=  45:7(9  4566=:  45670:  4567:;  457=((  4540;  45486  454(=  454;6  E45;96  E4544=  E454:8  E45448  E454=(  1%23  45:6=0  45:6(:  456;(7  456=(9  45:7;6  456;88  456:80  45687=  1*#<  45::46  45:=86  45:9=6  45:6:(  456(87  456  457(=;  454(:  4548:  4546  45489  E454=9  454(8  E4540  &703)+,  &-6)*+,  86&)*+,  &'/)*+,  '/4)*+,  -&.)9:;  456  456800  456887  45684:  4568:6  456700  456===  456  45::7:  45608(  456970  45688:  45:649  E454(:  45;;;  45;48  4540=  45447  454;  45;=7  456878  456;;9  45:8  45:=98  456(;:  45:==0  45:86:  4564=:  4569(7  4569  45:=6;  456  45:096  4564::  456;(9  456  458476  454(8  454=;  45;4(  45497  E45449  454=:  E45;  E454=(  45(9=  1%23  457=69  4570:  4570;0  4579(9  457467  4570(:  457=:9  4567;(  457(8  457(9  4566  4570;=  4570  4570(0  4579:6  4579:7  1*#<  45747;  456667  456:=  457:6:  4576;;  4579(0  457:=  4570(7  45676=  4574(0  ;  457899  4566  456667  456667  45747;  4540=  454:;  4546;  E454;:  E454:  E45440  E454(0  E4549;  454=6  454=:  E454(8  E454;:  45460  45489  4546  4548(  1%23  45:987  45:(78  45:60=  45:0((  45:407  456(0  4564(=  45:807  458=68  458::0  45:8  45:6:;  4587=6  45:69:  45:89:  45:=48  1*#<  458:00  45:7=;  45:898  45::46  45:;:  456:=8  456860  45::;0  4589  45840:  ;  456=49  458;90  45608(  45608(  45:(:=  45;(;  E454:6  454=(  E454(6  E45446  E45490  E45487  E4544;  E45447  45;;7  E45(;(  E4540:  45;=(  E45487  E4547:  454;8  1%23  45:9:9  456640  45:6:7  4564==  456=76  45:::(  45:060  456446  456:79  4564(8  456  4568;9  456;8  4560=6  45::=  456=48  1*#<  456  456:7=  4569:;  456:(=  457(09  456;0  456=:6  456:;0  457=((  45:74:  ;  456(:8  456  4574=(  45:(=;  45:799  >?<&@ABCD  E45409  454;  E4546  E454:9  E4547;  E4540;  E45;48  E4546;  E4549;  454(:  E45;0=  45406  4540  E4548(  454:(  45490  1*#<  458:00  45:7=;  45:898  45::46  45:;:  456:=8  456860  45::;0  4589  45840:  ;  456=49  458;90  45608(  45608(  45:(:=  45;(;  E454:6  454=(  E454(6  E45446  E45490  E45487  E4544;  E45447  45;;7  E45(;(  E4540:  45;=(  E45487  E4547:  454;8  1%23  45:9:9  456640  45:6:7  4564==  456=76  45:::(  45:060  456446  456:79  4564(8  456  4568;9  456;8  4560=6  45::=  456=48  1*#<  456  456:7=  4569:;  456:(=  457(09  456;0  456=:6  456:;0  457=((  45:74:  ;  456(:8  456  4574=(  45:(=;  45:799  E45409  454;  E4546  E454:9  E4547;  E4540;  E45;48  E4546;  E4549;  454(:  E45;0=  45406  4540  E4548(  454:(  45490  >?<&@ABCD F,<H6=#  &'3)*+,  456:6:  >?<&@ABCD F,<H6=#  -2-3)+,  457496  >?<&@ABCD '//;:  &/1)*+,  456799  >?<&@ABCD ')*(;8  /0/)*+,  456789  >?<&@ABCD ''F'G098  -&.)*+,  1%23  >?<&@ABCD  186  Table B2 cont’d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Table B2 cont’d !"#$%&'(&)*+,-. !"#$% '//(0  &'()*+,  1&4)-+,  121)0+,  &'2)5+,  '2.)6+,  1&3()+,78932/()+,789 423)6+,  4567:0  4566;  45:4(<  45666  45679:  456;;6  456888  45:408  456;48  4566<9  45:480  4567::  456:7<  456<0  45:(76  45640:  4568<:  456999  45;7:(  4568<<  456096  45:9;8  456999  456;8  456:<:  45666:  456:0;  454;<  E45480  45<48  45478  454;0  45<0:  4540<  4549  E454(7  45488  454<6  454(:  E454<<  454(:  4569(:  456798  456089  45668<  456(69  45679:  456;4<  456860  4564<:  45;7(6  45;98  456((<  45;4:7  45;;:6  45;8  45:4(0  456748  456;40  457:;;  45;7:(  456468  456;8  45;8  45777;  45;(:(  45;696  45;74:  45;6:8  45<<<  E45499  E45447  454(7  45<7:  45<0:  4546<  E45449  454;8  45<96  454<6  4547  E4547<  454<8  1%23  45:840  45:(47  45:080  45:0(  45:046  45:(4<  45:090  45:96:  45:08:  45:8(6  45:080  45:90  45:8(<  45:94;  1*#=  4568;<  45:8  45:06;  <  457970  <  45:4:<  45:77;  45:906  45:;8  45:;69  45;(0<  456748  456:0;  >?=&@ABCD  >?=&@ABCD  45<<  E454<:  454<  E4548(  4599;  E4540;  45406  E454<9  454(9  E454<<  E454(0  45(0<  45<46  45477  1%23  45;:;8  45697(  45;;<  45;;99  45;809  45;;8<  458680  45;(99  45;798  45;89<  45;;((  457:;7  45787:  45;66<  1*#=  456(((  45;8  45;96<  45;04;  45677;  45;7:(  458<<<  45;  45:<0:  4568  45;<;0  457;8;  457;9:  45;976  E454(  45<<7  45488  4548<  E45<96  45406  45<96  4540:  E45<66  E45<<7  4546(  45408  E454<8  454:(  1%23  45;;:8  45;787  4564;8  45;:77  456498  4564;;  456:<<  45696:  45608:  4568<:  45;868  45;::6  45;6:;  456098  1*#=  45;999  45;9<;  45;<09  45;;;6  45;8  45607(  45:<<<  45:999  456;(9  456;8  456(:6  45:<7;  45;9:<  4579<7  454;  4548;  45<(;  45499  454;6  E45446  E454<<  E454:7  E454(<  E454<0  E45469  E45<9(  454;8  45(;7  456(((  45;8  45;96<  45;04;  45677;  45;7:(  458<<<  45;  45:<0:  4568  45;<;0  457;8;  457;9:  45;976  E454(  45<<7  45488  4548<  E45<96  45406  45<96  4540:  E45<66  E45<<7  4546(  45408  E454<8  454:(  1%23  45;;:8  45;787  4564;8  45;:77  456498  4564;;  456:<<  45696:  45608:  4568<:  45;868  45;::6  45;6:;  456098  1*#=  45;999  45;9<;  45;<09  45;;;6  45;8  45607(  45:<<<  45:999  456;(9  456;8  456(:6  45:<7;  45;9:<  4579<7  454;  4548;  45<(;  45499  454;6  E45446  E454<<  E454:7  E454(<  E454<0  E45469  E45<9(  454;8  45(;7  >?=&@ABCD  >?=&@ABCD 1*#= >?=&@ABCD F,=H69#  123)0+, 123()+, 1&4)0+,  456789  1*#=  F,=H69#  &./)-+,  1*#= >?=&@ABCD  '//<;  &./)0+,  1%23  ''F'G087 1%23  ')*(<7  &'()-+,  >?=&@ABCD  188  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able and Cockerham 1984) each! Arctic Char ! B.3. Pairwise ! ! FST (Weir ! ! ! ! between ! !  ! !  !  !  !  )!2$:67(  !  !  ! ! ! sample !used in ! the study. ! The shaded ! ! (Salvelinus alpinus) part of the! matrix!  )!2$967(  !  !  )+)$:67(  !  !  !"+$<67(  !  !  "+0$%67(  !  !  )!*/$67>?@(  !  *+1/$67>?@( 2+*$%67(  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  denotes comparisons between two juvenile samples. The italics denote  comparisons two different years! at the same site.! The bold ! ! involving ! ! ! sampling ! ! !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  !  values are statistically significant (after Bonferroni correction).  189  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able! S3. Cont’d !  190  Age-class of juveniles Figure B.2. Frequency histograms of fork length of juvenile Arctic Char (Salvelinus alpinus) showing that multiple cohorts (seen as different length-classes in the histograms) were collected for most (but not all) sampling locations.  OPI09juv  25 20 15 10 5 0  KIN09juv  25 20 15 10 5 0  KAN10juv  25 20 15 10 5 0  IRV10juv  25 20 15 10 5 0  IQ209juv  25 20 15 10 5 0  IKP09juv  25 20 15 10 5 0  25 20 15 10 5 0  KIP09juv  25 20 15 10 5 0  AUN10juv  QAS09juv  25 20 15 10 5 0  ISU09juv  25 20 15 10 5 0  IQA09juv  25 20 15 10 5 0  IQ109juv  25 20 15 10 5 0  25 20 15 10 5 0  AVA09juv  Frequency  KEK10juv  25 20 15 10 5 0  0  25  50  75  100  125  Length (mm)  191  COLONY analysis I used the program COLONY to test for the presence of full-sib families in my samples of juveniles. The program takes into account the error rate in microsatellite scoring, which I estimated from the replicated samples as discussed above. I ran each site separately, assuming male and female polygamy, with medium length runs and medium likelihood precision. It was impossible to run the program with long length runs and high likelihood precision because it was computationally too expensive. The COLONY analysis identified the presence of putative full-sibs in all but two of the juvenile samples (Table S1). All but one individual from each families were removed from the analyses presented in the main article. The COLONY analysis also identified many half-sibs pairs in each population. Table B.4. Summary of sibship reconstruction among samples of juvenile Arctic Char (Salvelinus alpinus) using the program COLONY. N is the sample size I used in the COLONY analysis and also corresponds to the sample sizes indicated in table 1. All but one individual from each full-sib family was removed for some of the analyses (see main text) and the remaining sample size after this was done is indicated here. Sampling  Number of full-  Size of full-sib  Number of half-  N (after sibs  location  N  sib families  families  sib dyads  removed)  IQ209jv  48  1  2  70  47  KIN09jv  58  1  2  84  57  AVA09jv  65  0  NA  99  65  IAT09jv  48  1  2  64  47  KEK10jv  54  8  2; 2; 3; 4; 5; 5; 9; 11  289  21  IQ109jv  56  3  2; 2; 2  170  33  ISU09jv  40  5  2; 2; 2; 3; 6  76  30  KAN10jv  70  3  2; 2; 2  131  67  KIP09jv  61  3  2; 2; 2  96  58  AUN10jv  43  6  2; 2; 2; 2; 3; 3  96  35  IRV09jv  5  0  NA  ?  5  IKP09jv  60  1  2  100  59  OPI09jv  26  2  2; 2  63  24  IQA09jv  65  2  2; 2  ?  63  QAS09jv  65  1  2  122  64  192  The high number of half-sib identified by COLONY led us to investigate the possibility that COLONY was making type-I errors. The following analyses were conducted. First, the program was run on two samples of adults, which should be less likely to be related than juveniles according to the Allendorf-Phelps effect. While no fullsib families were identified in adult samples, several pairs of half-sibs were identified: 55 pairs in KIP04ad, and 77 pairs in ISU04ad. I do not expect so many adults to be related, unless the population sizes are very small, which apparently sustainable commercial exploitation of those stocks make difficult to believe. Second, I ran 57 adult samples from sites Kingnait and Kangerk together (the two populations are geographically distant and should exchange relatively few dispersers) and the program identified 36 pairs of halfsibs comprised of one individual from each population. The program also identified a pair of full sibs and 144 pairs of half-sibs when I ran the 101 adult samples from Isuituq (2003 and 2004 together). These numbers are similar to that typically obtained for samples of juveniles, further suggesting that Allendorf-Phelps effects are not important for Arctic char. Also note that COLONY identified a pair of full-sib juveniles from Iqaluit lake that were collected at two opposite ends of the lake – a finding that suggests that the higher prevalence of putative sibs in the juvenile samples is in fact not due to Allendorf-Phelps effects. Together, these observations suggest that (1) COLONY likely is too liberal in the identification of full-sibs, and (2) that even some of the full-sibs identified may be the result of type-I errors. Because there appears to be sufficient evidence that samples of adults are equally likely as samples of juveniles to contain half-sibs, I chose to retain the half-sibs in all analysis. On the other hand, while there is some evidence of type-I errors in the identification of full-sibs, I still remove them from the analysis. This conservative approach should ensure that our conclusions are not the result of increased relatedness among juvenile samples.  193  MIGRATE analysis I ran MIGRATE-n ver. 3.2.15 (Beerli & Felsenstein 2001) under the Brownian motion model for microsatellites and using UPGMA-based start genealogy with a full migration matrix model. I used a Bayesian search strategy with 1 long chain, sampling increments of 100, number of recorded chains of 5000 and a 10,000 chain-long burn-in. I used slicesampling for the proposal distributions and the uniform prior distributions for both ! and the migration parameter M were specified to the following values: !: min = 0, max = 100, window = 1; ": min = 0, max = 100, window = 10. I started by running four independent replicate runs under these conditions, and then used the final parameter estimates from those four replicates as the start parameters in another run with three independent replicates. The two different runs gave parameter estimates that were highly correlated (r2 = 0.704), which indicates that the runs converged according to the guidelines provided in the software manual.  194  Other biological correlates of dispersal propensity There was no significant difference between the mean trait value of dispersing and philopatric individuals for all traits that were examined. Previous studies, however, found that individuals with intermediate characteristics were more likely to be philopatric (Gyselman 1994), a fact that would be missed from a simple examination of means. Examination of frequency distributions for trait values of dispersing and philopatric individuals show that there is no such effect in the present study.  Age  25  10  D  0 10  15  20  25  Fork length  20  600  Condition factor (g/mm)  10  Count  30  30 20 10  50  D  20 10  0  0  200  400  Weight  600  800  0.5  ForkLength  25  P  15 10 0  25  D  20  5 0 0  1000  1.0  1.5  2.0  2.5  condition  20  (g)  5  10  40  0  30  10  15  50  40  D  Count  400  Gonadweight  0  20  Frequency  200  30  (mm)  Count  0  Age  P  P  5  120 100 80 60 40 20 0  120 100 80 60 40 20 0  D  0  5  0  (g)  5  20 10  P  15  25 15  Gonad weight  20  (yr) Count  Count  P  (a)  2000  3000  weight  4000  5000  6000  Figure B.3. Biological correlates of dispersal propensity in Arctic Char (Salvelinus alpinus). Frequency histograms of five traits comparing trait values of philopatric (P) and dispersing (D) individuals confirm that none of those five traits seem associated with dispersal propensity. (a) Analysis conducted in GeneClass2 with all samples, (b) Analysis conducted in GeneClass2 only with samples with an assignment score greater than 95%, and (c) analysis conducted in STRUCTURE.  Figur com ind ***  195  Figu  (b)  80 60 40 20 0  0 30 20 10 0 0  5  10  15  20  P 80 60 40 20 0  D  10  Frequency (%)  P  20  D  Frequency (%)  30  25  0  200  400  600  800  400  600  P  25 20 15 10 5 0  25 20 15 10 5 0  D  Frequency (%)  20 15 10 5 0  P 20 15 10 5 0  200  Gonad weight (g)  D  Frequency (%)  Age  0.5  Fork length (mm)  1.0  1.5  2.0  Condition factor (g/mm)  P  10 5 0  15 10  D  Frequency (%)  15  5 0 0  1000 2000 3000 4000 5000 6000  Weight (g)  196  0  5  10  15  20  P  80 60 40 20 0  80 60 40 20 0  D  25 20 15 10 5 0  Frequency (%)  P  25 20 15 10 5 0  D  Frequency (%)  (c)  25  0  P  10 0 20  D  30  0 400  600  800  0.5  Fork length (mm)  1.0  1.5  2.0  Condition factor (g/mm)  P  20 15 10 5 0  D  Frequency (%)  20  10  200  20 15 10 5 0  600  30  Frequency (%)  25 20 15 10 5 0  P 25 20 15 10 5 0  400  Gonad weight (g)  D  Frequency (%)  Age  200  0  1000 2000 3000 4000 5000 6000  Weight (g)  197  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073414/manifest

Comment

Related Items