AN ASSESSMENT OF POPULATION GENETIC STRUCTURE

IN GUILLEMOTS (CEPPHUS)

by

BRONWYN ANNECY SEABROOK HARKNESS

A thesis submitted to the Department of Biology

in conformity with the requirements for the

degree of Master of Science

Queen’s University

Kingston, Ontario, Canada

October 2017

Copyright © Bronwyn Annecy Seabrook Harkness, 2017

ii

Abstract

Understanding how evolutionary forces shape population genetic structure is a principal goal of

evolutionary geneticists. Guillemots (Aves: Charadriiformes: Alcidae: Cepphus) provide useful

systems for investigating population genetic differentiation in a genus with a wide geographic

range that disperses according to a one-dimensional stepping stone model. Guillemots may also

be highly vulnerable to climate change in parts of their range. For this reason, identifying

population genetic structure is important for their successful management and conservation. I

used various genetic markers and methods to assess population genetic variation in pigeon

guillemots (C. columba) and black guillemots (C. grylle). In both species, I found evidence of

limited dispersal and restricted gene flow between distant populations. Genetic differentiation

appeared to follow a pattern of isolation by distance in both species. In pigeon guillemots,

evidence suggested that population genetic structure arose through a combination of both

historical and contemporary processes. Based on my results, I recommend that pigeon guillemot

subspecies be revised. I also suggest that pigeon guillemot populations in the Aleutian Islands be

considered for designation as an evolutionarily significant unit (ESU) as their differentiation has

been strongly influenced by historical processes. Population genetic structure in black guillemots

did agree with current subspecies delineations. Based on the level of variation detected within the

subspecies however, I would recommend that black guillemots be treated as management units.

The results from this study hold promising insight into how patterns of gene flow can lead to

different patterns of population genetic structure.

iii

Co-authorship

Chapter 2

Bronwyn A. S. Harkness*, Veronica F. Poland, Gabriela Ibarguchi, John F. Piatt, and Vicki L.

Friesen

Bronwyn A. S. Harkness

Generated nuclear data for Southeast Alaska population,

performed all statistical analyses, and wrote final manuscript.

Veronica F. Poland

Generated mitochondrial data for pigeon guillemots.

Gabriela Ibarguchi

Generated microsatellite and intron data for pigeon guillemots

(with the exception of the Southeast Alaska population).

John F. Piatt

Helped with initial project design, sample collection, and funding.

Vicki L. Friesen

Helped with initial project design, provided funding, advised on

project progression, and edited drafts of manuscript.

Chapter 3

Bronwyn A. S. Harkness*, Gregory J. Robertson, and Vicki L. Friesen

Bronwyn A. S. Harkness

Helped with initial project design, generated ddRADseq data,

performed all statistical analyses, and wrote final manuscript.

Gregory J. Robertson

Helped with initial project design, sample collection, and funding.

Vicki L. Friesen Helped with initial project design, provided funding, collected

samples, advised on project progression, and edited drafts of

manuscript.

iv

Acknowledgements

Firstly, thank you to my supervisor, Vicki Friesen, for giving me the opportunity to work on this

project. Thank you for your patience, understanding, and continued support – being a graduate

student in your lab has been an incredible learning experience! Thank you to Laurene Ratcliffe

and Rob Colautti for sharing their bountiful knowledge and offering insightful critiques as

members of my advisory committee. Thank you to Tim Birt for helping me to develop my lab

skills and for providing guidance and wisdom when I was faced with challenges in the lab. To

Greg Robertson, thank you for introducing me to the wonderful world of fieldwork and allowing

me to join your team each summer, I will always cherish my time in Newfoundland. Thank you

to all the members of the Friesen lab for their continued support. Thank you especially to Becky

Taylor, Drew Sauve, and Anna Tigano for their friendship and guidance throughout this degree.

Each chapter of this thesis could not have been completed without the help of a large

number of people and organizations. Thanks to Chapter 2 co-authors V.L. Friesen, G. Ibarguchi,

V.F Poland, and J.F. Piatt for their individual contributions to this study. Thanks to T.Birt, L.

Hayes, J. Holder, M. Litzow, A. Marr, J. Moy, J. Pitocchelli, A. Pritchard, S. Speckman, B.

Sydeman, A. Burger, and T. Van Pelt, for collection of samples, and M. Kidd and J. Moy for

DNA extractions. T. Birt, Z. Sun, and R. Taylor provided technical advice during lab work. S.

Wallace assisted with analyses. V.L. Friesen and R. Colautti provided comments / suggestions

for improving the manuscript. I am grateful to D. Jelinski, V. Byrd, the research vessel Tiglax

and crew, AMNWR, USFWS and USGS for logistical help. Funding for this chapter was

provided by the Exxon Valdez Oil Spill Restoration Office, a Natural Science and Engineering

Research Council (NSERC) Undergraduate Student Research Award, and an NSERC Discovery

Grant to Vicki Friesen.

v

Thanks to Chapter 3 co-authors G.J. Robertson and V.L. Friesen for their individual

contributions to this study. Thanks to G. Divoky, R. Ronconi, T. Gaston, M. Kassera, and M.

Kidd for collection of samples. Funding for this chapter was provided by the STAGE program of

Environment and Climate Change Canada and an NSERC Discovery Grant to Vicki Friesen.

Thank you to the American Ornithological Society, Canadian Society of Ornithologists, and

Queen’s University for funding travel to meetings.

Finally, thank you to my family for their continued love and support, I couldn’t have done

it without you!

vi

Table of Contents

Abstract..................................................................................................................................... ii

Co-Authorship........................................................................................................................... iii

Acknowledgements................................................................................................................... iv

Table of Contents...................................................................................................................... vi

List of Tables............................................................................................................................ ix

List of Figures........................................................................................................................... x

List of Abbreviations................................................................................................................ Xii

Chapter 1 : General Introduction.............................................................................................. 1

Study Species.................................................................................................................... 3

Methodology..................................................................................................................... 4

Purpose.............................................................................................................................. 6

Importance........................................................................................................................ 6

Chapter 2 : Both historical and contemporary processes lead to population genetic differentiation

in pigeon guillemots (Cepphus columba) ................................................................................ 8

Abstract............................................................................................................................. 8

Introduction....................................................................................................................... 9

Methods............................................................................................................................. 11

Sampling and DNA Screening.................................................................................. 11

Mitochondrial Control Region.......................................................................... 12

Microsatellites................................................................................................... 12

Introns............................................................................................................... 13

Data Analyses........................................................................................................... 13

Tests of Neutrality and Variability................................................................... 13

Population Differentiation................................................................................ 14

Phylogenetic Relationships............................................................................... 15

Population History............................................................................................ 16

Results............................................................................................................................... 17

Tests of Neutrality and Variability................................................................... 17

Population Differentiation................................................................................ 19

vii

Phylogenetic Relationships............................................................................... 20

Population History............................................................................................ 21

Discussion......................................................................................................................... 21

Population Genetic Structure.................................................................................... 21

Evolutionary History................................................................................................. 22

Evidence for Historical Isolation...................................................................... 22

Evidence for Contemporary Differentiation..................................................... 23

Subspecies................................................................................................................. 24

Future Directions...................................................................................................... 25

Chapter 3 : An assessment of population genomic structure in black guillemots (Cepphus grylle)

in North America...................................................................................................................... 40

Abstract............................................................................................................................. 40

Introduction....................................................................................................................... 41

Methods............................................................................................................................. 45

Sample Collection and DNA Extraction................................................................... 45

Library Preparation and Quality Filtering................................................................. 46

Alignment with Reference Genome.................................................................. 46

Data Analyses........................................................................................................... 47

Population Genetic Analysis............................................................................. 47

Outlier Analysis................................................................................................ 48

Results............................................................................................................................... 49

Population Genetic Analysis............................................................................. 49

Outlier Analysis................................................................................................ 50

Discussion......................................................................................................................... 51

Population Genetic Structure............................................................................ 51

Isolation by Distance......................................................................................... 53

Outlier Analysis : Possible Adaptation to Climatic Regions............................ 54

Subspecies and Management............................................................................ 55

Future Directions.............................................................................................. 55

Chapter 4 : General Discussion................................................................................................. 65

Population Genetic Structure............................................................................................ 65

viii

Isolation by Distance......................................................................................................... 65

Evolutionary History......................................................................................................... 66

Outlier Analysis................................................................................................................ 67

Subspecies and Management............................................................................................ 67

Importance........................................................................................................................ 68

Future Directions.............................................................................................................. 68

Summary : Chapter 2................................................................................................................ 70

Summary : Chapter 3................................................................................................................ 71

Literature Cited......................................................................................................................... 72

Appendix A : Supplementary Material for Chapter 2............................................................... 85

Appendix B : Supplementary Material for Chapter 3............................................................... 101

ix

List of Tables

Table 2.1 Subspecies, sampling regions, abbreviations, sites, and numbers (n) for pigeon

guillemots.................................................................................................................................... 26

Table 2.2 Number of haplotypes, haplotype diversity, nucleotide diversity, observed and

expected F values for Ewens-Watterson test of neutrality and probabilities of neutrality for

Chakraborty’s tests...................................................................................................................... 27

Table 2.3 Observed/expected heterozygosity estimates for pigeon guillemot nuclear loci....... 28

Table 2.4 Estimates of ΦST / FST for pairwise comparisons of pigeon guillemot regional

samples........................................................................................................................................ 29

Table 3.1 Subspecies, sampling regions, region abbreviations, sites, and sample numbers (n)

for black guillemots.................................................................................................................... 57

Table 3.2 Summary statistics for all populations across variant nucleotide sites and all

sites.............................................................................................................................................. 58

Table 3.3 Estimates of FST for pairwise comparisons of black guillemot regional samples...... 59

x

List of Figures

Figure 2.1 Breeding distribution of pigeon guillemots, and sampling sites.............................. 32

Figure 2.2 A statistical parsimony haplotype network derived for control region sequences

of pigeon guillemots................................................................................................................... 33

Figure 2.3 Regression plot from Mantel test testing for correlation between genetic distance

(Slatkin’s linearized FST) and geographic distance (log-transformed shoreline distance)

using both mitochondrial and nuclear data................................................................................. 34

Figure 2.4 Probabilities of assignment of individuals to each of two genetic populations

(K=2) derived from the program STRUCTURE, based on genetic variation in nuclear loci of

pigeon guillemots....................................................................................................................... 35

Figure 2.5A Results of principal component analysis of nuclear variation in pigeon

guillemots (grouped by population)........................................................................................... 36

Figure 2.5B Results of principal component analysis of nuclear variation in pigeon

guillemots (grouped by subspecies)........................................................................................... 37

Figure 2.5C Results of principal component analysis of nuclear variation in pigeon

guillemots (grouped by membership to Aleutian Islands)......................................................... 38

Figure 2.6 Phylogenic relationships among mitochondrial control region sequences of

pigeon guillemots from different regions, using BEAST........................................................... 39

Figure 3.1 Distribution map of black guillemots and sampling sites........................................ 61

Figure 3.2A Probabilities of assignment of individuals to different genetic populations

derived from the program STRUCTURE, based on genetic variation in nuclear loci of black

guillemots (K = 2)...................................................................................................................... 62

Figure 3.2B Probabilities of assignment of individuals to different genetic populations

derived from the program STRUCTURE, based on genetic variation in nuclear loci of black

guillemots (K = 3)...................................................................................................................... 62

Figure 3.3A Results of principal component analysis of nuclear variation in black

guillemots (grouped by population)........................................................................................... 63

Figure 3.3B Results of principal component analysis of nuclear variation in black

guillemots (grouped by subspecies)........................................................................................... 63

Figure 3.4 Regression plot from Mantel test testing for correlation between genetic distance

xi

(Slatkin’s linearized FST) and geographic distance (log-transformed shoreline

distance)......................................................................................................................................

64

xii

List of Abbreviations

α Significance level of the test

ABC Approximate Bayesian computation

AFLP Amplified fragment length polymorphism

AMOVA Analysis of molecular variance

ANOVA Analysis of variance

B-Y Benjamini-Yekutieli

BIC Bayesian information criterion

bp Base pair

D Slatkin’s linearized FST

ddRADseq Double-digest restriction-site associated DNA sequencing

ESU Evolutionarily significant unit

FIS Inbreeding coefficient

FST Fixation index, measure of population differentiation (nuclear DNA)

h Haplotype diversity

HKY+G Hasegawa-Kishino-Yano substitution model with a gamma distribution

Ho Observed heterozygosity

IBIS l’Institut de biologie integrative et des systèmes

IUCN International Union for the Conservation of Nature

K Number of genetic populations

MCMC Markov chain Monte Carlo

mtDNA Mitochondrial DNA

MU Management unit

MYA Million years ago

n Sample size

π Nucleotide diversity

p Statistical probability

P Average major allele frequency

PCA Principal components analysis

PCR Polymerase chain reaction

xiii

r Correlation coefficient

ra Rate of admixture

SNP Single nucleotide polymorphism

tn Time of divergence event

ΔK Second order change in posterior probability

ΦST Measure of population differentiation (mitochondrial DNA)

1

Chapter 1

General Introduction

Gene flow is one of the major processes that shape the population genetic structure of a species.

It is often thought of as a constraining force to evolution, but it can also be a creative force

(Slatkin 1987; Ellstrand and Rieseberg 2016). The role of gene flow in genetic differentiation

and ultimately speciation depends on a number of historical and contemporary factors, one of

which is the pattern of gene flow. If migrants follow an n-island model of dispersal (i.e. are

exchanged at random among populations), then in theory one migrant per generation will

homogenize allele frequencies (Wright 1969). However if migrants disperse according to a

stepping stone model (i.e. primarily between neighbouring colonies), especially if populations are

dispersed linearly, then we would expect gene flow to be less effective at counteracting genetic

drift (Kimura and Weiss 1964). As a result, we would expect population genetic structure to be

stronger in species that follow a one-dimensional stepping stone model of dispersal than in those

following an n-island model. As well, in the case of the stepping stone model, dispersal is often

constrained by distance and this can result in isolation by distance (Mantel 1967). Consequently,

populations situated more closely geographically tend to be more genetically similar (Hutchison

and Templeton 1999). With the island model of dispersal, genetic differentiation would be

independent of geographic distribution (Allendorf et al. 2013).

Isolation by distance through restricted dispersal has been demonstrated in species across a

number of taxa. Balanophyllia elegans, a solitary stony coral, possesses crawling larvae, which

are capable of only limited dispersal (Hellberg 1995). At a spatial scale of 1 to 50 km, B. elegans

appeared to follow the one-dimensional stepping stone model of dispersal. Gene flow and

genetic drift had equilibrated and the author suggested that isolation by distance could be

facilitating local adaptation. (At a broader spatial scale this pattern was not observed). The

2

Australian barramundi (Lates calcarifer), a catadromous estuarine fish, disperses linearly along

the coastline and gene flow is limited to the parental catchment area and neighbouring rivers

(Russell and Garrett 1988; Chenoweth et al. 1998). Geographic distance explained 23% -79% of

mitochondrial gene flow in these fish, suggesting that genetic differentiation was a result of

isolation by distance. The red drum (Sciaenops ocellatus) breeds in coastal waters of the Atlantic

Ocean, close to the mouths of bays and estuaries (Matlock 1987). Adults are capable of large-

scale dispersal, but movement of young is relatively limited. Genetic variation in red drum was

spatially distributed according to an isolation by distance pattern. Dispersal occurred according

to a modified stepping stone model, where movement is not limited strictly to adjacent water

bodies but dispersal to these areas occurs with a higher probability (Gold et al. 2001). The

Arizona tree frog (Hyla wrightorum), a dryland anuran that breeds in ponds, showed significant

genetic differentiation between populations (Mims et al. 2016). Geographic distance appeared to

be the primary driver of population genetic structure (with influences by slope and canopy cover

as well). This was in contrast to other pond breeding anurans that exhibited little / no genetic

structure but it was likely due to differences in life history traits (e.g. desiccation tolerance,

mobility, fecundity, population sizes, and duration of larval phase). White-throated dipper

(Cinclus cinclus) populations in the Iberian Peninsula were genetically differentiated, with

genetic distances significantly correlating with geographical distances (Hernández et al. 2016).

The white-throated dipper is a riverine bird that is highly dependent on fast flowing streams with

shallow, clear water. Genetic variation between populations could possibly be due to a lack of

suitable habitat, limiting dispersal and colonization. The Eurasian badger (Meles meles)

experienced limited dispersal and exhibited strong philopatry (Pope et al. 2006). The amount of

genetic diversity in badgers is thought to have been influenced by Pleistocene glaciations, but

these effects have contemporarily been replaced by a strong pattern of isolation by distance. This

3

pattern of genetic differentiation is attributed to restricted dispersal among individuals. From

these examples, it is clear that life history traits and especially patterns of gene flow can lead to

different patterns of population genetic structure.

Specific patterns of gene flow are rarely known for seabirds. A review by Friesen et al.

(2007) assessed studies of geographic variation in mtDNA (mitochondrial DNA) in seabirds to

test the importance of various factors in generating population genetic and phylogeographic

structure and found that colony distribution (nesting in small colonies on coastal cliffs / islands

vs. nesting primarily in large colonies on offshore islands) had a weak effect on population

genetic structure. Genetic distance also correlated with geographic distance in many of the

studies reviewed by Friesen et al. (2007), which is what we would expect if dispersal occurred

primarily between neighbouring colonies. Birds in the family Alcidae (order Charadriiformes)

are varied in their colony distribution and population genetic structure.

Study Species

Guillemots (Charadriiformes : Alcidae : Cepphus) are arctic seabirds that breed in polar regions

and partway down the Atlantic and Pacific coasts of North America, Europe, and Asia.

Guillemots will avoid flying over large expanses of land and open water / ice (Udvardy 1963),

and are found in higher densities along boundaries of land or ice and marine water (Bradstreet

1979). Philopatry is thought to be relatively strong in guillemots, with breeding dispersal

occurring much less often than natal dispersal and over much shorter distances (Austin 1929;

Frederiksen and Petersen 1999, 2000). Guillemots possess many characteristics that set them

apart from other seabirds and even other alcids. Unlike other alcids, guillemots are essentially

resident at colonies year-round and overwinter at the ice edge (Ewins 1985). Guillemots typically

forage on small demersal fish in shallow coastal waters closely surrounding colonies (Golet et al.

2000; Divoky 2011). Guillemots are also one of the few arctic seabirds with geographic variation

4

in morphology, with clines in body size, wing and tarsal length, culmen length, and quantity of

white on the wing depending on the species (Storer 1952). In a comparison between guillemots

and murres (Charadriiformes : Alcidae : Uria) Storer (1952) suggested that the greater

geographical variation in morphology in guillemots compared to murres could be due to the

difference in colony distribution: guillemots nest primarily in small, dispersed colonies on coastal

cliffs and nearshore islands (Cairns 1980), whereas murres nest in fewer, much larger colonies on

offshore islands. A previous study on guillemots using the mitochondrial control region found

pronounced population genetic structure which could have arisen as a result of historical isolation

in glacial refugia during the Pleistocene (Kidd and Friesen 1998). This is in contrast to many

other alcids that show little to no genetic differentiation between populations (reviewed in

Friesen 2015). The combination of these factors points towards limited dispersal in guillemots

and the potential for population genetic structure.

Methodology

Double digest restriction-site associated DNA sequencing (ddRADseq) (Peterson et al.

2012) is a tool that has become increasingly popular for answering questions about population

differentiation within species. It offers a number of advantages over other genotyping techniques,

for example, it does not require the design of species-specific primers and so no prior knowledge

about the organism’s genomic sequence is required (Dierickx et al. 2015). This makes

ddRADseq well suited for studies involving endangered or non-model species. In addition,

because of the large numbers of markers produced, fewer individuals are needed to resolve

accurate population genetic parameters and structure, which makes ddRADseq particularly useful

for species that might be rare or difficult to sample in the field (Dierickx et al. 2015). Finally, the

cost of ddRADseq is relatively low and it often provides a more robust analysis of the species’

population genetics. For example, in cases where low genetic differentiation is found using low

5

numbers of markers, larger numbers of markers from next-generation sequencing are usually able

to resolve some structure within the species (Toews et al. 2016). Ruegg et al. (2014)

demonstrated this when they revisited population structure in Wilson’s warblers (Wilsonia

pusilla). Irwin et al. (2011) had originally assayed Wilson’s warblers with amplified fragment

length polymorphisms (AFLPs) and found that the species was split into well-defined western

and eastern groups and possible species level differences were detected. Ruegg et al. (2014)

found the same broad patterns when they compared populations using ddRADseq but they also

found evidence for structuring within the western subspecies that had been missed by AFLPs,

although some structuring had previously been detected by mtDNA (Kimura et al. 2002, Paxton

et al. 2013). Results from ddRADseq studies are now being used to determine implications of

genetic differentiation for conservation purposes. For example, Dierickx et al. (2015) compared

variation of Hawaiian and Japanese black-footed albatrosses (Phoebastria nigripes). Previous

studies on this species performed with mtDNA and microsatellites reported contrasting

conclusions, with some studies finding significant differentiation between populations (Walsh

and Edwards 2005; Ando et al. 2014) and others suggesting that they do not differ significantly

(Eda et al. 2008). With ddRADseq, Dierickx et al. (2015) found that genetic variation was low,

but differed between Hawaiian and Japanese populations. While this might not have a great

impact from an evolutionary perspective, even a small amount of genetic differentiation can be

important for species conservation and population management (Benestan et al. 2015). This is

especially true for populations that are demographically independent and for populations that are

likely to be affected differently by climate change, as are Hawaiian and Japanese albatrosses. For

this reason, Dierickx et al. (2015) recommended that the populations be considered separate

management units. This example highlights the benefit of using a genome-wide approach to

address population genetics and management questions.

6

Purpose

In the present study I used various genetic markers and methods (mtDNA, microsatellites,

introns, single nucleotide polymorphisms (SNPs) from reduced-representation sequencing) to

assess population genetic variation in pigeon guillemots and black guillemots. In Chapter 2, my

objective was to investigate (1) the extent to which pigeon guillemots exhibit population

differentiation at neutral markers, and (2) whether differentiation is a result of historical isolation

or contemporary processes. I predicted that regional populations would differ at neutral molecular

markers. I also predicted that genetic differentiation in pigeon guillemots could have arisen

historically, through isolation in glacial refugia during the Pleistocene similar to results from

Kidd and Friesen (1998). Alternatively, they could have become differentiated in situ through

genetic drift and / or selection due to restricted gene flow and therefore follow a pattern of

isolation by distance. In Chapter 3, I conducted a genome-wide survey of SNPs identified by

ddRADseq with the objective of determining the extent to which black guillemot colonies differ

at neutral and putatively adaptive markers. I predicted that regional populations of black

guillemots are genetically distinct and that this differentiation might follow a pattern of isolation

by distance.

Importance

Guillemots provide a useful system for testing population differentiation in a species with

a wide range that disperses according to a one-dimensional stepping stone model. Previous

studies on systems that disperse according to a stepping stone model have largely been on

organisms with limited mobility or a lack of suitable habitat due to fragmentation. Seabirds in

general are strong fliers and as such are unlikely to encounter many physical barriers to dispersal

however, as outlined previously, factors other than physical barriers could lead to population

genetic differentiation in guillemots. Using genetics to understand gene flow in this system will

7

allow us to understand better how different patterns of gene flow affect population genetic

structure. From an applied perspective, the results from this study will be useful for management

and conservation of guillemots in the future. We currently have a poor understanding of

population genetic structure of Cepphus. The same characteristics that make them likely to have

pronounced population structure also makes them more vulnerable to disturbances such as oil

spills. Identifying population genetic structure in guillemots will give us baseline measurements

of genetic differentiation and potentially allow us to gauge whether guillemots are well suited to

adapt as they continue to face environmental changes due to climate change.

8

Chapter 2

Both historical and contemporary processes lead to population genetic differentiation in

pigeon guillemots (Cepphus columba)

Abstract

Understanding the forces that shape population genetic structure is fundamental to the study of

evolution. Pigeon guillemots (Cepphus columba) are Pacific seabirds with few obvious barriers

to gene flow. As seabirds, they are strong fliers and have the ability to disperse long distances.

They experience many potential non-geographic barriers to gene flow however, which could

result in strong population genetic structure. Variation in the mitochondrial control region, four

microsatellite loci, and two introns were analysed for 202 pigeon guillemots representing three of

five subspecies. The mitochondrial sequences showed strong population structure. Nuclear loci

also exhibited significant population structure, however structure was weaker than for the

mitochondrial control region. Marked differences were found between the Aleutian Island

populations and mainland populations and strong evidence was found for differentiation in

pigeon guillemots arising as a result of both historical isolation and contemporary processes.

Genetic variation did not support current subspecies designations.

9

Introduction

Because most birds have strong dispersal abilities, gene flow in most avian species is potentially

high and population genetic structure potentially weak. However, recent studies based on

molecular markers indicate that population genetic structure in birds is varied. In some species

population genetic structure is weak or absent (e.g. mallard, Anas platyrhynchos, Kraus et al.

2012; cerulean warbler, Setophaga cerulea, Deane et al. 2013; little auk, Alle alle, Wojczulanis-

Jakubas et al. 2014; regent honeyeater, Anthochaera phrygia, Kvistad et al. 2015, crowned

solitary eagle, Buteogallus coronatus, Canal et al. 2017). In others, population genetic structure is

strong suggesting that gene flow is restricted (e.g. Cory’s shearwater, Calonectris borealis,

Gomez-Diaz et al. 2009; whiskered auklets, Aethia pygmaea, Pshenichnikova et al. 2015; red-

billed chough, Pyrrhocorax pyrrhocorax, Morinha et al. 2017; Grauer's swamp warbler,

Bradypterus graueri, Kahindo et al. 2017). Gene flow may be restricted for many reasons.

Friesen et al. (2007; 2015) reviewed studies of seabirds to test the importance of various factors

in generating population genetic structure. The role of geographic barriers in reducing gene flow

is well supported. However, many species show evidence of population genetic structure in the

absence of geographic barriers to dispersal. In these species the extent of population genetic

structure is best explained by nonbreeding distribution. Seabirds that are resident at colonies

year-round or that migrate to population-specific nonbreeding areas may have less opportunity

for gene flow than those with a single, species-specific migratory destination or that simply

disperse during the nonbreeding season. Foraging range during the breeding season has a weak

influence on population genetic structure. Species that forage inshore (within <8 km of land)

would have less opportunity for gene flow among colonies, and accordingly inshore foraging is

associated with population genetic structure. The geographic distribution of colonies can also

affect gene flow: species that nest in small colonies scattered along coastlines likely follow a one-

10

dimensional stepping stone model of dispersal where individuals disperse primarily to

neighbouring colonies rather than at random (Kimura and Weiss 1964). In the one-dimensional

stepping stone model, gene flow is less effective at countering genetic drift. Philopatry alone

may also reduce gene flow, but it usually acts in combination with other barriers to gene flow

(Friesen et al. 2007; Friesen 2015).

Guillemots (Cepphus spp.; Charadriiformes: Alcidae) provide a potentially interesting

system for investigating population genetic structure in seabirds. Guillemots will not fly over

land or ice, so large expanses of land or ice can act as geographic barriers to dispersal (Udvardy

1963). They generally breed in small colonies on islands and headlands (Gaston and Jones 1998)

and forage on small demersal fish in shallow coastal waters closely surrounding colonies (Golet

et al. 2000). Pigeon guillemots (C. columba) breed along the northern coastlands of the Pacific

Ocean (Figure 2.1), and have an estimated population size of less than 69,000 breeding birds

(Kushlan et al. 2002). They winter in pack ice relatively close to their breeding colonies, with the

northern limit of their distribution defined by the availability of open water, similar to their close

relative the black guillemot (C. grylle, Ainley et al. 1990; Divoky et al. 2016). Previously

guillemots were believed to have strong natal philopatry (Gaston and Jones 1998), but a recent

review suggests that past estimates of philopatry in seabirds may be exaggerated (Coulson 2016).

This is corroborated by a banding study that found that natal dispersal in black guillemots may be

relatively high (Frederiksen and Petersen 2000). Pigeon guillemots are morphologically variable

throughout their range and five subspecies have been designated based on clines in wing length,

tarsus length, culmen length, and quantity of white on the wing (Storer 1952; Figure 2.1).

According to Storer (1952) wing length, tarsus length, and culmen length decrease from south to

north, while the quantity of white on the wing increases.

Previously, Kidd and Friesen (1998) compared variation in the conserved central domain

11

of the mitochondrial control region among populations of all three species of guillemots.

Assuming a substitution rate of 2% per million years, they estimated the date of divergence of the

mitochondrial lineages of populations of pigeon guillemots to be 0.1 and 0.07 million years ago

(MYA). This places the divergence time during the Pleistocene glaciations. They suggested that

genetic differences may exist among local populations of pigeon guillemots, but their sampling

was too limited for rigorous analyses of population genetic structure. Furthermore, their study did

not include any nuclear genes. Nuclear data can sometimes be incongruous with mitochondrial

data, making it important to assess both (e.g. Walsh and Edwards 2005; Eda et al. 2008; Ando et

al. 2014).

In the present study, I analysed variation in the more variable sections of the

mitochondrial control region and six nuclear genes among pigeon guillemots sampled throughout

the north-eastern Pacific. My objectives were to investigate (1) the extent to which pigeon

guillemots exhibit population differentiation at neutral markers, and (2) whether differentiation

has arisen as a result of historical isolation or contemporary processes. Given that pigeon

guillemots breed in small colonies scattered along coastlines, forage and overwinter near

breeding colonies, and have extensive geographic variation in morphology, I predicted that

regional populations would differ at neutral molecular markers according to a pattern of isolation

by distance. I also predicted that genetic differentiation in pigeon guillemots could have arisen

historically, through isolation in glacial refugia during the Pleistocene similar to results from

Kidd and Friesen (1998). Alternatively, they could have become differentiated in situ through

genetic drift and / or selection due to restricted gene flow.

Methods

Sampling and DNA screening

12

Solid tissue, blood or feather samples were collected from 202 pigeon guillemots from 11 sites

throughout the breeding range (Table 2.1, Figure 2.1). Most Alaskan samples consisted of solid

tissue (heart, liver and / or striated muscle) from adults in breeding condition collected for dietary

analyses in close proximity to colonies during the breeding season. Samples from elsewhere

generally consisted of blood from adults or chicks caught at nests. Samples were archived at

Queen's University, the Royal Ontario Museum, the Burke Museum, the American Museum of

Natural History and/or the University of Alaska Museum at Fairbanks. DNA was extracted using

a standard protease K phenol/chloroform technique followed by isopropanol precipitation

(Friesen et al. 1997).

Mitochondrial control region

The polymerase chain reaction (PCR) was used to amplify and sequence two overlapping

fragments of the mitochondrial control region with guillemot-specific primers (CGL56 and

CGH549, and CGL486 and CGH1006) following protocols detailed in Kidd and Friesen (1998).

Sequences were obtained for 723 bp (base pairs) of the control region, excluding 141 bp 5' to

primer CgL56 and a 92 bp region of ambiguity between primers CgL486 and CgH589, and

including parts of Domain I, II, and III. Sequences were aligned using the program BioEdit (Hall

1999), and haplotypes were identified using the program TCS v 1.13 (Clement et al. 2000).

Microsatellites

A genomic library was developed and screened for dinucleotide (CA) repeats following

previously published protocols (Ibarguchi et al. 2000). PCR primers were developed for two loci

(Supplementary Table 2.1). A panel of six samples was tested for amplification using these

primers as well as primers developed previously for marbled murrelets (Brachyramphus

marmoratus; Friesen et al. 2005), thick-billed and common murres (Uria lomvia and Uria aalge,

respectively; Ibarguchi et al. 2000) and yellow warblers (Dendroica petechia; Dawson et al.

13

1997). MgCl2 concentrations and annealing temperatures were optimized, and the presence of

length variation was determined in four loci (Supplementary Table 2.1) by electrophoresis

through polyacrylamide gels (Ibarguchi et al. 2000). Microsatellites were genotyped using the

GenomeLab GeXP Genetic Analysis System (Beckman Coulter Inc., USA) and scored using the

associated software v 10.2.3.

Introns

Amplifications were attempted on four to six samples with 30 pairs of PCR primers previously

designed to amplify nuclear introns from vertebrates (Friesen et al. 1997; Friesen 1999; Friesen

unpubl.) using previously published protocols with the addition of 62.5 µg/mL BSA and 0.01

mg/mL gelatin. Various annealing temperatures and concentrations of MgCl2 and DMSO were

tested to optimize amplifications. Loci for which clean amplification products could be derived

consistently were then chosen for population screening using a combination of single-stranded

conformational polymorphisms and direct sequencing (Friesen et al. 1997). Individuals from the

Alaska Peninsula were sequenced directly (Genome Quebec, McGill University, Montreal,

Quebec). Sequences were aligned using Geneious (Kearse et al. 2012), and variable sites were

confirmed from chromatograms. Two loci generated repeatable results (Supplementary Table

2.1).

Data Analyses

Tests of neutrality and variability

The program ARLEQUIN v 3.5 (Excoffier and Lischer 2010) was used to test control region

sequences for departures from neutral expectations using Ewens-Watterson and Chakraborty tests

(Ewens 1972; Watterson 1978; Chakraborty 1990), and to test nuclear markers for deviations

from Hardy-Weinberg proportions and linkage equilibrium. The program MICROCHECKER

(van Oosterhout et al. 2004) was used to check microsatellites for null alleles. Haplotype

14

frequencies, haplotype diversity (h, Nei 1987) and nucleotide diversity (π, Tajima 1983) were

calculated for the control region sequences, and regional allele frequencies were calculated for

nuclear loci.

Population differentiation

Analysis of molecular variance (AMOVA) was used to calculate the proportion of genetic

variation distributed among populations (ΦST for control region sequences, FST for nuclear genes)

in ARLEQUIN. ΦST and FST were calculated for both the entire sample and for pairwise

comparisons of populations. Statistical significance was tested by randomization using 10,000

permutations of the data with a rejection level (α) of 0.05. Kimura’s two-parameter model of

sequence evolution (Kimura 1980) with a gamma value of 0.45 was used for the control region

sequences. Potential Type I statistical errors were addressed by applying Benjamini-Yekutieli (B-

Y) corrections to pairwise comparisons (Benjamini and Yekutieli 2001; Narum 2006).

Mantel tests (Mantel 1967) were used to determine whether genetic differentiation

increases with geographic distance between populations. Pairwise estimates of Slatkin’s

linearized ΦST or FST were tested for correlation with log-transformed geographic distance

between populations using ARLEQUIN, with 10,000 permutations of the data. The shortest

geographic distance between regions was calculated using an online distance calculator (Michels

1997). Additionally, because pigeon guillemots are unlikely to fly over large expanses of open

water, distances along shorelines were estimated using Google Maps (1:500 km; Google Maps

2017) and log transformed. When more than one sampling site was included in a region (e.g. for

comparisons involving the Aleutian Islands), the geographic midpoint of all sites was used.

The program STRUCTURE v 2.3.4 (Pritchard et al. 2000; Pritchard et al. 2010) was used

to infer population structure among populations using the nuclear data. The program was run

under the admixture model with correlated allele frequencies, without sampling location as prior

15

information, a burn-in of 10000 iterations and 100000 iterations after the burn-in. Each value of

K (number of populations) from 1 to 5 was run 10 times and delta K (ΔK, the most probable

number of genetic populations) was calculated using STRUCTURE HARVESTER (Evanno et al.

2005; Earl and vonHoldt 2012). Posterior probabilities were calculated using the equation

provided in the STRUCTURE manual (Pritchard 2010). Membership probabilities were averaged

among runs using CLUMPP v 1.1.2. (Jakobsson and Rosenberg 2007) and results were displayed

using DISTRUCT v.1.1 (Rosenberg 2004). A principal component analysis (PCA) was

performed on the nuclear data in R v 3.3.1 (R Core Team 2017) using the packages adegenet

(Jombart 2008) and ade4 (Dray and Dufour 2007). Missing data were replaced with global mean

allele frequencies.

Phylogenetic relationships

A statistical parsimony network was derived for control region sequences using PopART

(Clement et al. 2002; http://popart.otago.ac.nz). Sites with undefined states were masked. The

program BEAST v 1.8.4 was used to construct a phylogenetic tree for the mitochondrial control

region data (Drummond et al. 2012). Individuals were grouped by region (Table 2.1) and five

sequences from C. grylle and one sequence from C. carbo were used as outgroups. The

Hasegawa-Kishino-Yano substitution model with a gamma distribution (HKY+G) was used

(Hasegawa et al. 1985), which was selected as the best model using the Bayesian information

criterion (BIC) in jModelTest2 (Darriba et al. 2012). A strict molecular clock with a mutation

rate of 7.4% (Wenink et al. 1996) and the coalescent tree prior that assumes a constant population

size with Jeffrey’s prior (Jeffreys 1946; Kingman 1982; Drummond et al. 2005) were selected.

The BEAST analysis was run for 107 MCMC (Markov chain Monte Carlo) runs with a burn-in of

106 runs. TRACER v1.6 (Rambaut et al. 2014) was used to analyse the trace files generated by

BEAST and to assess convergence. The analysis was run three times and output files were

16

combined using LogCombiner v 1.8.4, with a burn-in of 10% of trees for each original file. A

consensus tree was generated in TreeAnnotator v 1.8.4 (Drummond et al. 2012) and the tree was

displayed graphically using FigTree v 1.4.3 (Rambaut 2007).

Population history

Three scenarios for population history were compared using coalescent-based approximate

Bayesian computation analyses (ABC) in DIYABC (Cornuet et al. 2008). This approach makes

it possible to model different population scenarios and generate simulated datasets to fit each

scenario. These simulated datasets are then compared to determine which scenario best fits the

observed dataset. Mitochondrial control region samples were grouped into three population

groups based on results from population genetics analyses: North (Andreanof Is. and Fox Is.),

Alaska (Shumigan Is., Eastern Alaska Peninsula, Cook Inlet, Prince William Sound, and

Southeast Alaska), and South (Vancouver Is., Central Oregon, and Central California). Scenario

1 (Supplementary Figure 2.1) simulated the existence of two glacial refugia during the last

glaciation, with divergence between North and South populations occurring pre-glaciation and no

gene flow post-glaciation. Scenario 2 simulated isolation of pigeon guillemots in a single glacial

refugium, followed by expansion and colonization of their current range post-glaciation.

Scenario 3 tested that pigeon guillemots were historically isolated in two refugia with secondary

contact occurring post-glaciation. In all scenarios the prior distribution for all historical

parameters was set to uniform. Due to the absence of available data for population size, broad

priors were used for Ne (10 – 100 000). Mean generation time of pigeon guillemots was

estimated to be 8.8 years (Southern et al. 1965, Hudson 1985; Moum and Árnason 2001). Time

of divergence of the ancestral population (t2 and tc) were set to 1250 – 100 000 generations ago

(0.011 – 0.88 MYA), before the last glacial period (Wisconsin glaciation). Admixture events and

more recent divergences (t1, ta, and tb) were set to 0 – 1250 generations ago (present day – 0.011

17

MYA), after the Wisconsin glaciation. Rate of admixture (ra) was left as the default (0.001 –

0.999). The Hasegawa-Kishino-Yano substitution model with a gamma distribution (HKY+G)

was selected, as determined by JModelTest2 (Hasegawa et al. 1985, Darriba et al. 2012). Mean

mutation rate was set to 6.5X10-7 per site / per generation based on pairwise divergence rates of

the avian mitochondrial control region suggested by Wenink et al. (1996). Eight single-sample

summary statistics were used for each population (number of distinct haplotypes, number of

segregating sites, mean pairwise difference, variance of the number of pairwise differences,

Tajima’s D statistics (Tajima 1989), number of private segregating sites, mean of the numbers of

the rarest nucleotide at segregating sites, variance of the numbers of the rarest nucleotide at

segregating sites) and five two-sample summary statistics were used (number of distinct

haplotypes in the pooled sample, number of segregating sites in the pooled sample, mean of

within sample pairwise differences, mean of between sample pairwise differences, FST between

two samples (Hudson et al. 1992)). I ran 3 million simulations and compared the scenarios by

estimating posterior probabilities using logistic regression (Cornuet et al. 2010, Inoue et al. 2014,

Perktas et al. 2015, Pérez-Alvarez et al. 2016). Type I and Type II error rates were estimated

using the method described in Cornuet et al. 2010 and Inoue et al. 2014.

Results

Tests of neutrality and variability

Mitochondrial control region sequences were similar to those published previously for guillemots

(Kidd and Friesen 1998), and contained the conserved sequence blocks typical of other species of

birds (F, D, and C Boxes and CSB-1; Marshall and Baker 1997). Eighty-five control region

haplotypes defined by seventy-eight variable sites were identified (Supplementary Table 2.2;

Supplementary Table 2.3; Figure 2.2). Fifty-nine variable sites occurred in Domain I, seventeen

18

occurred in Domain II, and two occurred in Domain III. Ewens-Watterson and Chakraborty tests

did not reveal any deviations from neutrality except at Mandarte Island. In the Chakraborty test,

Mandarte Island had significantly more haplotypes than expected (14 observed vs. 8.6 expected,

P = 0.02; Table 2.2), however it did not show a significant deviation from neutrality using the

Ewens-Watterson test.

All nuclear loci were variable, although only three alleles were detected for Dpu16

(Supplementary Table 2.4). Of these three alleles two were rare, meaning that Dpu16 was

monomorphic for the same allele in most populations. No null alleles were indicated by

MICROCHECKER for microsatellite data. None of the nuclear loci showed significant

deviations from Hardy-Weinberg proportions in any of the populations, with the exception of

ribosomal protein 40 intron V in Mandarte Island (Table 2.3). Cco5-9 did show a significant

departure from Hardy-Weinberg proportions at the species level due to an excess of homozygotes

(Table 2.3). Six of the eleven populations showed evidence of linkage disequilibrium at different

loci (Supplementary Table 2.5). Analyses were rerun excluding each of these loci and results

remained consistent, so results from analyses with all loci were included in the following

sections.

Six alleles, defined by variation at four (cytochrome c intron I) and 15 sites (ribosomal

protein 40 intron V), were found within each of the two introns (Supplementary Table 2.4;

Supplementary Figure 2.2; Supplementary Table 2.6). Alleles differed by one to twelve

substitutions. Numbers of alleles at microsatellite loci ranged from three to 11 (Supplementary

Table 2.4). For all seven nuclear loci, several alleles were present at high frequency at most sites,

and the remaining alleles occurred in only one or two individuals each (Supplementary Table

2.4).

19

Population differentiation

Of the 85 control region haplotypes, 67 were private (found in only one region). With four

exceptions, shared haplotypes were always found in geographically adjacent regions

(Supplementary Table 2.3). Several haplotypes occurred within individual populations at high

frequencies (AA01, 25%; AD01, 22%; AF01, 19%; AG01, 19%; AL01, 41%). Results from

AMOVA indicated strong geographic structure in sequence variation (global Φst = 0.37, p <

0.001). Pairwise estimates of Φst were statistically significant for all comparisons except

Shumigan Island versus Semidi Island, Semidi Island versus Cook Inlet and Prince William

Sound, and Cleland Island versus Mandarte Island (Table 2.4).

Estimates of global population genetic structure were weak but statistically significant

both for introns (Fst = 0.05, p < 0.001; Supplementary Table 2.7) and for all loci combined (Fst =

0.09, p < 0.001; Table 2.4). Estimates for microsatellites were slightly higher, and also significant

(Fst = 0.11, p < 0.001; Supplementary Table 2.7). Estimates of FST were significant for the

majority of pairwise population comparisons for microsatellites, and some pairwise comparisons

for introns (Supplementary Table 2.7).

Mantel tests for correlations between genetic differentiation and geographic distance were

significant for all molecular markers (mitochondrial control region, r = 0.56, p < 0.05; all nuclear

loci, r = 0.59, p < 0.05; Figure 2.3). Mantel tests were also performed after removing populations

from the Aleutian Islands (Andreanof Island and Fox Island), as these populations appeared

genetically similar to each other but quite differentiated from other populations (Table 2.4; see

below). The tests remained significant, however the relationship between genetic differentiation

and geographic distance became slightly weaker (mitochondrial control region, r = 0.52, p <

0.05; all nuclear loci, r = 0.52, p < 0.05).

20

Results from STRUCTURE and STRUCTURE HARVESTER (ΔK and posterior

probabilities) indicated that pigeon guillemots comprise two genetic groups (K=2; Figure 2.4).

Assignment probabilities showed that genetic groups were not completely geographically

segregated, but appeared to follow an isolation by distance pattern consistent with the Mantel test

results. Results from the PCA also showed weak differentiation among populations with the first

three principal components explaining 17.5% of the variance (Figure 2.5A). When populations

were grouped by subspecies in the PCA, C. kaiurka and C. eureka were separated from each

other, whereas C. adianta overlapped with both other subspecies, with the first three principal

components explaining 17.5% of the variance (Figure 2.5B). Populations were also grouped as

Aleutian Islands versus others to test whether nuclear loci displayed as marked a difference

between these groups as did the mitochondrial control region (Figure 2.5C). The PCA separated

the two groups, but the relationship was weak (Table 2.4) with the first three principal

components explaining 17.5% of the variance.

Phylogenetic relationships

The phylogeny produced by BEAST showed mtDNA sequences of most birds from the

Andreanof and Fox Islands (Aleutian Islands) as a monophyletic group, divergent from other

pigeon guillemot populations (Figure 2.6). There were two exceptions: one individual from

Shumigan Island was in the monophyletic clade, and one individual from Fox Island was mixed

with the other populations. Support for the split between the Aleutian Islands and other pigeon

guillemots was strong, with a posterior probability of 1. Relationships between other populations

were not as clear, and posterior probabilities were low for the more recent divergences.

Divergence dates between mtDNA sequences of the black guillemot outgroup and pacific birds

was estimated to be between 0.63 – 1.2 MYA (with a posterior probability of 1). The divergence

between the spectacled guillemots and pigeon guillemots was estimated to have occurred

21

between 0.43 – 0.87 MYA (with a posterior probability of 1). My dates were in agreement with

those estimated by Kidd and Friesen (1998). The split between the Aleutian Islands and other

pigeon guillemot populations was estimated to have occurred between 0.16 - 0.32 MYA.

Population history

Results from analysis with the mitochondrial control region in DIYABC identified Scenario 1 as

the most highly supported, with strong posterior probability (0.8282; Supplementary Table 2.8).

The other two scenarios had much lower support. The Type I and Type II error rate estimates for

all 3 scenarios were low (Supplementary Table 2.8).

Discussion

Population Genetic Structure

My first objective was to investigate the extent to which regional populations of pigeon

guillemots exhibit population differentiation at neutral markers. Mitochondrial control region

sequence variation was strongly structured geographically. The majority of haplotypes were

private (Supplementary Table 2.2; Supplementary Table 2.3) and all estimates of Φst (global and

pairwise) were statistically significant, with the exception of a few populations that are within

200 - 650 km of each other. Marked population structuring was also found in nuclear DNA,

although it was weaker than for mtDNA. The estimate of global FST was statistically significant

and the majority of pairwise FST estimates were statistically significant for microsatellite loci

(Table 2.4; Supplementary Table 2.7). Pairwise FST estimates from introns were similar to those

from microsatellite loci, although fewer were significant (Supplementary Table 2.7). The strong

population genetic structure in pigeon guillemots is similar to a few species such as Kittlitz’s

murrelet (Brachyramphus brevirostris, Birt et al. 2011) but contrasts with weak or absent genetic

22

variation in most other high-latitude seabird species (e.g. razorbill, Alca torda, Moum and

Árnason 2001; reviewed in Friesen 2015).

My results clearly indicate a major genetic disjunction between Aleutian Islands and

mainland populations. Pairwise ΦST estimates showed that guillemots from the Andreanof and

Fox islands (which are both part of the Aleutian islands) were very similar to one another but

highly differentiated from the other populations (Table 2.4). The Aleutian Islands did not share

haplotypes with any other populations (with the exception of two) and their haplotypes were

highly differentiated from others within the species (Figure 2.2, Supplementary Table 2.3).

Although Aleutian island populations did not segregate as strongly at the nuclear level, they were

still significantly differentiated from more southern populations (Table 2.4; Figure 2.4; Figure

2.5C). The BEAST tree agreed with the mitochondrial and nuclear population genetic results:

Andreanof Island and Fox Island mtDNA sequences were divergent from continental

populations. Genetic studies of other bird species in the Aleutian Islands have also found marked

differences from populations elsewhere (e.g., Canada goose, Pierson et al. 2000; rock ptarmigan,

Lagopus mutus, Holder et al. 2000; marbled murrelet, Congdon et al. 2000; Kittlitz’s murrelet,

Birt et al. 2011).

Evolutionary History

Given that I found significant genetic differentiation between regional populations of pigeon

guillemots, my second objective was to investigate whether differentiation had arisen as a result

of historical isolation or in situ following recession of the glaciers through genetic drift and / or

selection due to restricted gene flow.

Evidence for historical isolation

MtDNA haplotypes from Aleutian Island guillemots were highly divergent from continental

North American haplotypes (except for one shared haplotype) (Figure 2.2; Supplementary Table

23

2.3). The distinctiveness of these haplotypes suggests that guillemots from the Aleutian Islands

were historically isolated from other populations. Estimates of Φst and Fst were also greatest

between Aleutian Island and continental samples (Table 2.4), suggesting that they diverged a

long time ago and have experienced limited / restricted gene flow in more recent years. The

Aleutian Islands also represented a monophyletic split within the population tree (Figure 2.6).

Assuming a strict molecular clock and a mutation rate of 7.4% for guillemot control regions

(Wenink et al. 1996), the split between the mtDNA lineages of the Aleutian Islands and other

populations was estimated to have occurred 0.16 – 0.36 MYA, i.e. prior to recession of the

Wisconsin glaciation. Results from DIYABC using mtDNA also suggested that the most likely

scenario was a historical split between the Aleutian Island and continental populations prior to

the Wisconsin glaciation, with continental populations diverging from each other post-glaciation.

This evidence suggests that a cryptic barrier to gene flow occurs within the Aleutian Islands

(Friesen et al. 2007). Other high arctic species that exhibit population genetic structure between

the Aleutian Islands and the Alaska Peninsula often occur in areas that are putative Pleistocene

refugia (Dyke and Prest 1987). These divergent populations are often associated with evidence

of historical fragmentation, suggesting long-term isolation.

Thus, differentiation of Aleutian Island versus continental North American pigeon

guillemots may be explained at least in part by historical fragmentation, probably by extensive

Pleistocene ice fields that would have separated tracts of rocky coastline from each other. These

findings agree with Udvardy’s (1963) hypothesis that geographic variation in pigeon guillemots

is due to isolation in multiple glacial refugia.

Evidence for contemporary differentiation

Pairwise estimates of ΦST and FST were strongly correlated with geographic distance between

populations, suggesting that populations of pigeon guillemots might experience isolation by

24

distance (Figure 2.3). Mitochondrial haplotypes were mostly private and shared haplotypes were

only found between adjacent populations (Supplementary Table 2.3). Results from STRUCTURE

and PCA also suggest increasing genetic differences with geographic distance (Figure 2.4; Figure

2.5). A pattern of isolation by distance exists when dispersal of individuals is constrained by

distance, so that gene flow occurs primarily between neighbouring populations (Hutchison and

Templeton 1999). As a result, populations that are closer geographically tend to be more

genetically similar. Thus, if pigeon guillemots are experiencing isolation by distance, this

differentiation would have occurred post-glaciation – either due to secondary contact between

two populations isolated in separate refugia, or due to recolonization after isolation in a single

refugium. This does not discount the possibility of historical isolation, but does provide strong

evidence for contemporary differentiation.

The combination of results from all analyses provides strong evidence for differentiation

in pigeon guillemots arising from both historical isolation and contemporary processes.

Population structure resulting from both contemporary and historical processes has been reported

across several species (e.g. marbled murrelets, Congdon et al. 2000; Eurasian badger, Meles

meles, Pope et al. 2006; wolverine, Gulo gulo, Zigouris et al. 2013; spectacle case pearly mussel,

Cumberlandia monodonta, Inoue et al. 2014; Chilean dolphin, Cephalorhynchus eutropia, Pérez-

Alvarez et al. 2016).

Subspecies

According to Storer (1952), subspecies of pigeon guillemots show marked morphological

differences that vary along a latitudinal cline. While pigeon guillemots did exhibit strong genetic

structure and genetic variation appears to follow a cline similar to morphological differentiation,

differences at neutral molecular markers did not agree with current subspecies delineations.

Although samples of C. c. kaiurka appeared highly differentiated from other populations and

25

subspecies based on the mitochondrial data, they were similar to guillemots from Fox Island,

which belong to C. c. adianta (Figure 2.2). C. c. eureka did not seem to differ greatly from C. c.

adianta, although the two diverged recently suggesting they might be exhibiting incomplete

lineage sorting (Table 2.4, Figure 2.6). Nuclear variation did not differ between samples of C. c.

kaiurka and most other Alaskan samples (Figure 2.5B). The topic of avian subspecies is highly

controversial and there is currently no standard way to designate subspecies, although an

integrative approach is often recommended (Torstrom et al. 2014; Patten 2015). My results

highlight a discordance between morphological and genetic groupings, and I recommend that

pigeon guillemot subspecies be reviewed. I also suggest that pigeon guillemot populations in the

Aleutian Islands be considered for designation as an evolutionarily significant unit (ESU) as their

differentiation has been strongly influenced by historical processes.

Future Directions

While using both mitochondrial and nuclear loci gives strong insight into the population structure

of pigeon guillemots, using only a handful of loci does give limited resolution of genetic

structure. Reduced-representation and genome-wide techniques that produce thousands of

markers would not only allow us to examine population structure at a higher resolution, but

would also give us the opportunity to explore adaptive differences and the extent of introgression

at sites of secondary contact in pigeon guillemots. My study included three of the five current

subspecies. Including samples from Cepphus columba columba and Cepphus columba snowi

would give us a broader understanding of the population genetics of pigeon guillemots. Other

sampling gaps exist in the Gulf of Alaska and southern California. Increasing sample sizes for

populations with fewer individuals would also make analyses more robust.

26

Table 2.1. Subspecies, sampling regions, abbreviations, sites, and numbers (n) for pigeon

guillemots.

Subspecies Region Abbrev-

iation Site n

kaiurka Andreanof Is. Andr Tanaklak I. 4 Great Sitkin I. 2 Adak I. 4 adianta Fox Is. Fox Anangula I. 1 Unalaska I. 2 Aiktak I. 5 Shumigan Is. Shum Belkofski Bay 3 Yukon Hr. 4 Eastern Alaska

Peninsula Semi Semidi Is. 4

Flat I. 3 Cook Inlet Cook Shuyak I. 2 Kachemak Bay 32 Prince William Sound PWS Jackpot I. 12 Naked I. 18 Southeast Alaska SEAK Midway I. 6 Couverden I. 3 Vancouver I. Clel Cleland I. 7 Mand Mandarte I. 29 eureka Central Oregon cOre Coos Bay 24 Central California cCal Point Reyes National Park 1 Southeast Farallon I. 34 Ano Nuevo I. 2

Total 202

27

Table 2.2. Number of haplotypes, haplotype diversity (h), nucleotide diversity (π) as a percent,

observed and expected F values for Ewens-Watterson test of neutrality (none significant) and

probabilities of neutrality for Chakraborty’s tests.

Sampling site Number of

haplotypes

h π Ewens-

Watterson

Chakraborty

Andr 10 0.98±0.05 0.73±0.44 0.12/0.12 0.76

Fox 6 0.89±0.11 1.25±0.73 0.22/0.20 0.56

Shum 7 1.00±0.08 1.19±0.72 NA NA

Semi 7 1.00±0.08 0.79±0.49 NA NA

Cook 21 0.96±0.02 1.04±0.55 0.07/0.07 0.65

PWS 16 0.95±0.02 0.68±0.38 0.08/0.10 0.87

SEAK 8 0.97±0.06 0.47±0.30 0.14/0.14 0.77

Clel 3 0.76±0.11 0.29±0.21 0.35/0.45 0.86

Mand 14 0.83±0.07 0.51±0.29 0.20/0.12 0.02

cOre 10 0.90±0.04 0.58±0.33 0.14/0.17 0.78

cCal 12 0.88±0.03 0.47±0.27 0.14/0.16 0.66

28

Table 2.3. Observed/expected heterozygosity estimates for pigeon guillemot nuclear loci. Region

abbreviations as in Table 2.1.

Sampling Site Dpu16 Uaa5-8 Cco5-9 Cco5-21 Cytochrome

c Intron I

Ribosomal Protein 40 Intron V

Andr NA 0.89/0.75 0.67/0.68 0.86/0.86 0.89/0.74 0.44/0.63 Fox NA 0.43/0.49 1.00/0.85 0.85/0.97 0.86/0.63 0.29/0.26

Shum NA 0.57/0.65 0.86/0.85 0.86/0.78 0.71/0.54 0.14/0.14 Semi 0.2/0.19 0.40/0.57 0.9/0.82 0.80/0.79 0.40/0.57 0.40/0.35 Cook 0.03/0.03 0.64/0.59 0.85/0.85 0.79/0.78 0.67/0.62 0.30/0.40 PWS NA 0.59/0.59 0.86/0.84 0.79/0.80 0.59/0.58 0.38/0.31

SEAK NA 0.67/0.50 0.78/0.71 0.56/0.55 0.44/0.57 0.22/0.40 Clel 0.13/0.13 0.33/0.57 0.88/0.74 0.67/0.56 0.63/0.58 0.50/0.50

Mand NA 0.49/0.56 0.72/0.68 0.45/0.46 0.62/0.60 0.17/0.36* cOre 0.26/0.23 0.45/0.63 0.52/0.50 0.74/0.71 0.61/0.49 0.43/0.39 cCal 0.12/0.10 0.69/0.65 NA 0.66/0.61 0.21/0.25 0.17/0.23

*Significantly different from 0 at α = 0.05 both before and after B-Y correction

29

Table 2.4. Estimates of ΦST / FST for pairwise comparisons of guillemot regional samples based on sequence variation in the

mitochondrial control region (below diagonal) and seven nuclear markers (above diagonal). Region abbreviations as in Table 2.1.

Andr Fox Shum Semi Cook PWS SEAK Clel Mand cOre cCal

Andr 0.01 0.04 0.04 0.02 0.03* 0.11** 0.08* 0.14** 0.11** 0.26** Fox 0.11* 0.01 0.01 0.02 0.00 0.07** 0.07* 0.13** 0.10** 0.26**

Shum 0.48** 0.39** 0.00 0.03* 0.01 0.11** 0.08* 0.08** 0.08** 0.21** Semi 0.57** 0.47** -0.01 0.01 -0.01 0.11** 0.01 0.05* 0.05** 0.16** Cook 0.50** 0.49** 0.14* 0.01 0.00 0.11** 0.04* 0.10** 0.09** 0.18** PWS 0.60** 0.55** 0.17** -0.01 0.04* 0.11** 0.05** 0.10** 0.09** 0.18**

SEAK 0.66** 0.58** 0.14** 0.13* 0.15** 0.21** 0.06* 0.12** 0.12** 0.26** Clel 0.67** 0.58** 0.12* 0.13* 0.13* 0.12* 0.27** 0.01 0.03* 0.18**

Mand 0.66** 0.62** 0.15** 0.11* 0.17** 0.15** 0.13** 0.00 0.07** 0.16** cOre 0.66** 0.61** 0.21** 0.24** 0.28** 0.29** 0.27** 0.19* 0.21** 0.02* cCal 0.73** 0.70** 0.38** 0.44** 0.42** 0.47** 0.46** 0.45** 0.45** 0.11**

*Significantly different from 0 at α = 0.05 before B-Y correction **Significantly different from 0 at α = 0.05 after B-Y correction

30

Figure Captions

Figure 2.1. Breeding distribution (heavy lines) of pigeon guillemots, and sampling sites (white

dots). Thin lines represent approximate ranges of subspecies. Sampling site codes from northwest

to southeast: Ad = Adak I., Tn = Tanaklak I., GS = Great Sitka Is., Aa = Anangula Is., Un =

Unalaska Is., Ai = Aiktak Is., Be = Belkofski Bay, Yu = Yukon Hr., Se = Semidi Is., Fl = Flat Is.,

Sh = Shuyak Is., Ka = Kachemak Bay, Ja = Jackpot Is., Nk = Naked Is., Cv = Couverden Is., Mi

= Midway Is., Cl = Cleland Is., Ma = Mandarte Is., Co = Coos Bay, PR = Point Reyes National

Park, SF = Southeast Farallon Is., AN = Ano Nuevo Is.

Figure 2.2. A statistical parsimony haplotype network derived for control region sequences of

pigeon guillemots using PopART. Sizes of circles are proportional to haplotype frequencies.

Black dots represent inferred intermediate haplotypes that were not found in the present

sampling. Haplotype names have been removed for clarity. Region abbreviations as in Table 2.1.

Figure 2.3 Regression plot from Mantel test testing for correlation between genetic distance

(Slatkin’s linearized FST) and geographic distance (log-transformed shoreline distance). White

circles represent comparisons made using mitochondrial data and black circles represent

comparison made using nuclear data.

Figure 2.4. Probabilities of assignment of individuals to each of two genetic populations (K=2)

derived from the program STRUCTURE, based on genetic variation in nuclear loci of pigeon

guillemots. The program was run under the admixture models with correlated allele frequencies,

31

without sampling location as a prior, a burn-in of 10000 iterations and 100000 iterations after the

burn-in. Each value of K (number of populations) from 1 to 5 was run 10 times. Populations are

ordered as they appear geographically, from northwest to southeast (left to right). Region

abbreviations as in Table 2.1.

Figure 2.5. Results of principal component analysis of nuclear variation in pigeon guillemots

grouped by A) population, B) subspecies, and C) membership to Aleutian Islands.

A) 1 = Andreanof Is., 2 = Fox Is., 3 = Shumigan Is., 4 = Eastern Alaska Peninsula, 5 = Cook

Inlet, 6 = Prince William Sound, 7 = South East Alaska, 8 = Cleland Is., 9 = Mandarte Is., 10 =

Central Oregon, 11 = Central California.

B) 1 = kaiurka (Andreanof Is.), 2 = adianta (Shumigan Is., Eastern Alaska Peninsula, Cook Inlet,

Prince William Sound, South East Alaska, Cleland Is., Mandarte Is.), 3 = eureka (Central

Oregon, Central California).

C) 1 = Andreanof Is. and Fox Is., 2 = Shumigan Is., Eastern Alaska Peninsula, Cook Inlet, Prince

William Sound, South East Alaska, Cleland Is., Mandarte Is., Central Oregon, Central California.

Figure 2.6. Phylogenic relationships among mitochondrial control region sequences of pigeon

guillemots from different regions, using BEAST. C. grylle and C. carbo are used as an outgroup.

Posterior probabilities are shown at major nodes. Nodes with low posterior probabilities are not

labelled.

32

Fig. 2.1

Nk Ja

Yu

Cv Mi

GS Tn Ad a d

i a n

t a

33

Fig. 2.2

34

Fig. 2.3

35

Fig. 2.4

36

Fig. 2.5A

1

10 11

2 3

4

5

6

7

8

9

Eigenvalues

37

Fig. 2.5B

1

2

3

Eigenvalues

38

Fig. 2.5C

1 2

Eigenvalues

39

Fig. 2.6

0.99

0.99

40

Chapter 3

An assessment of population genomic structure in black guillemots (Cepphus grylle) in

North America

Abstract

Identifying genetically differentiated populations can be important for successful species

conservation. If local populations become genetically differentiated then the loss of a population

can result in partial loss of a species’ genetic diversity, which can affect a species’ ability to adapt

to stressors. Black Guillemots (Cepphus grylle) are seabirds that may be highly vulnerable to

climate change, but we have little knowledge of their population genetics and demographics. I

conducted a genome-wide survey of genetic variation using double-digest restriction-site

associated DNA sequencing (ddRADseq) to determine the extent to which regional populations

of black guillemots differ at neutral and putatively adaptive markers. I tested the hypothesis that

regional populations of black guillemots are genetically distinct. Results from my study showed

that black guillemots exhibit clear population genetic structure, suggesting that gene flow among

black guillemot populations is restricted. These results confirm results of previous studies of

population genetic structure in black guillemots using mtDNA. Possible reasons for this marked

genetic structure include strong philopatry, physical barriers to gene flow, tendency to remain

close to breeding colonies year-round, and historical isolation in multiple refugia during the

Pleistocene. Outlier analyses detected two putatively adaptive loci. Genotypes at these loci were

sorted according to latitude. My research indicates that regional populations of black guillemots

should be managed as separate units.

41

Introduction

Identifying genetically differentiated populations has important implications for successful

species conservation. If local populations are genetically differentiated, then extirpation of a

population results in partial loss of a species’ genetic diversity. A species that is losing genetic

diversity is increasingly at risk of extinction, as it has decreasing standing genetic variation for

adaptation to new environmental pressures, such as climate change (Allendorf et al. 2013).

Population genetic techniques can allow us to determine the level of genetic diversity within a

population and assess the potential impact of natural and anthropogenic stress on the success of

the species, which can be used to gauge a population's vulnerability. To prioritize conservation

efforts, biologists can use information from population genetics studies to recommend that

populations be considered as evolutionarily significant units (ESUs) or management units (MUs).

An ESU is defined as a group of organisms that has been isolated from other conspecific groups

for long enough to have become genetically and ecologically distinctive (Paetkau 1999). An MU

is a group of organisms that have become demographically independent and exhibit low levels of

connectivity (Paetkau 1999). Classification as either an ESU or an MU determines how that

group should be managed. Population genetics have been used to successfully estimate

population genetic structure to aid conservation in arctic seabirds such as marbled murrelets

(Brachyramphus marmoratus; Friesen et al. 2005) and thick-billed murres (Uria lomvia; Tigano

et al. 2015).

Black guillemots (Cepphus grylle) are colonial seabirds that breed in the Arctic and North

Atlantic oceans and Europe. Arctic populations of black guillemots are considered pack ice

obligates, meaning that many aspects of their life history are dependent on the presence of pack

ice (Divoky 2011). Their close association with snow and ice habitats makes them potentially

vulnerable to climate change. Rising temperatures and decreasing sea ice cover are negatively

42

affecting the breeding success of black guillemots that breed in the arctic, specifically on Cooper

Island, AK. Increases in atmospheric temperature and decreases in sea ice have indirectly led to

decreased breeding success in black guillemots through the alteration of prey distribution and

abundance, and the introduction of new competitors and predators (Divoky 2011). Historically,

the majority of the guillemots’ diet consisted of Arctic cod (Boreogadus saida), which is the

primary prey fish in the ice-associated ecosystem (Bradstreet 1980, Cairns 1987, Divoky et al.

2015). In recent years, however, the diet composition of black guillemot nestlings on Cooper

Island, Alaska has shifted. Arctic cod, which used to account for 95% of the birds’ diet, now

comprises less than 5% of their diet. Near-shore demersal fish, sculpin (Cottidae) in particular,

now comprise the majority of their diet (Divoky et al. 2015). This shift away from Arctic cod has

been associated with a five-fold increase in the rate of nestling starvation, and reductions in

nestling growth and fledgling mass (Divoky et al. 2015). Melting sea ice has also created the

opportunity for subarctic species to move farther north where they may have a negative impact on

true arctic species. One example is recent colonization of black guillemot breeding grounds by

horned puffins (Fratercula corniculata). These subarctic seabirds have moved north in response

to the warming climate and now compete aggressively with black guillemots for nesting cavities.

Horned puffins now kill up to half of the black guillemot nestlings on Cooper Island, Alaska

(Divoky 2011). Predation on black guillemot nestlings and eggs from polar bears (Ursus

maritimus) that arrive on the island after the summer ice retreats has also had a negative impact

on black guillemot breeding success, resulting in an almost complete nestling failure in some

years (Divoky 2011). Black guillemots are currently listed as a species of ‘Least Concern’ by the

International Union for the Conservation of Nature (IUCN), however they are clearly vulnerable

to climate change and may become more at risk as climate change continues to affect their

surrounding environment (BirdLife International 2016). Gaining a better understanding of

43

population genetic structure in black guillemots will allow us to make informed decisions about

management and conservation. If high arctic populations represent MUs or ESUs, they are

especially vulnerable and should be managed separately.

We might expect black guillemots to experience restrictions to gene flow for a number of

reasons. Black guillemots have a widespread breeding range (Friesen et al. 2007; 2015). As

black guillemots will not typically fly over land or ice, these can act as potential barriers to gene

flow (Udvardy 1963). Guillemots also nest in small, isolated colonies on coastal cliffs and islands

and forage inshore, close to their breeding sites. This limits their opportunity to encounter

individuals from other populations, thereby minimizing gene flow. Black guillemots are thought

to follow a one-dimensional stepping stone model of dispersal, where individuals disperse

primarily to neighbouring colonies rather than dispersing at random (Kidd and Friesen 1998).

Gene flow is less effective at counteracting genetic drift within the one-dimensional stepping

stone model. Black Guillemots have high seasonal connectivity meaning that they migrate to

nearby population specific non-breeding areas as opposed to dispersing at random during the

non-breeding season or migrating to a species-specific non-breeding area (Divoky et al 2016).

Again, this lowers the probability that guillemots will encounter individuals from other

populations and return to a different breeding site. Guillemots from different latitudes also inhabit

different climate regions. A number of climatic variables vary with latitude including temperature

(land and sea surface), photoperiod, sea ice extent, etc. (Global Observing Systems Information

Center; Diamond and Lief 2009). The two latitude groups occupy different climate regions:

higher latitude populations experience a tundra climate, whereas lower latitude populations

experience a warm summer continental climate (Natural Resources Canada 1957). As well, high

latitude habitats experience close pack ice or solid ice cover during the year, while low latitude

habitats are either completely ice free or experience only open pack ice (Natural Resources

44

Canada 1971). Finally, black guillemots are morphologically variable throughout their range and

six subspecies have been designated based on variation in wing length, tarsus length, and culmen

length (Storer 1952; Figure 3.1).

Some genetic evidence exists for restricted gene flow in black guillemots. Kidd and

Friesen (1998) conducted an analysis of variation in mtDNA (using the mitochondrial control

region) in guillemots from seven colonies in the Northern Atlantic and Arctic Oceans. They

found that all populations differed genetically, although genetic divergence was not correlated

with geographic distance. Evidence from breeding distributions, timing of divergence events,

phylogeography, gene flow and genetic variation within populations led Kidd and Friesen (1998)

to conclude that divergence of peripheral populations of black guillemots likely resulted from

Pleistocene glaciations. Pleistocene glaciers would have acted as substantial barriers to gene flow

between populations. The correlation between breeding distributions of guillemots and putative

glacial refugia also supports glacial vicariance. While several populations of black guillemots

apparently remained isolated since retreat of the glaciers (e.g. those from Maine and Finland),

contemporary gene flow appears to occur between populations in the Chukchi Sea, Iceland,

Norway and Nunavut (Kidd and Friesen 1998). This study had several limitations: Analysis of a

single mitochondrial locus gives limited resolution of genetic structure compared to studies that

employ hundreds to thousands of markers. As well, results from mtDNA are often incongruous

with data from nuclear DNA. Assessing both mtDNA and nuclear DNA is therefore important to

gain a more complete understanding of the population genetics within the system. Finally, only

three populations from North America were included in this study, meaning that we still have a

very limited understanding of the population genetics of black guillemots in North America.

Further studies including nuclear genes, more markers, and more populations will provide more

45

robust inferences regarding genetic differentiation and dispersal among populations in Canada

and the U.S.A.

I conducted a genome-wide survey of single nucleotide polymorphisms (SNPs) identified

by double-digest restriction-site associated DNA sequencing (ddRADseq) to determine the extent

to which black guillemot colonies differ at neutral and putatively adaptive markers (Peterson et

al. 2012). I tested the hypothesis that regional populations of black guillemots are genetically

distinct and that this differentiation might follow a pattern of isolation by distance. Alternatively,

black guillemots might exist as one large population that mixes genetically and demographically.

In this case, black guillemot populations will show low genetic differentiation between regional

populations at neutral markers and exhibit evidence of high levels of gene flow. I also used

outlier analyses to detect loci putatively under selection and to investigate whether adaptive

variation was associated with latitude. I predicted that higher latitude populations would differ

from low latitude populations at adaptive markers.

Methods

Sample Collection and DNA Extraction

Blood, feather, or solid tissue samples were collected from 55 breeding adults or chicks from 9

black guillemot colonies along the Canadian east and arctic coasts as well as Alaska (Table 3.1;

Figure 3.1). Samples are archived at -80° C in ethanol at Queen’s University, Kingston. DNA

was purified from the samples using a standard proteinase-K phenol/chloroform extraction and

ethanol precipitation (Sambrook et al. 1989). DNA concentrations were standardized to 20 ng/µl

using a Qubit 3.0 fluorometer and a Qubit dsDNA Broad Sensitivity Assay Kit (Invitrogen,

Carlsbad, California, USA).

46

Library Preparation and Quality Filtering

Purified DNA samples were sent to l’Institut de biologie integrative et des systèmes (IBIS) de

L’Université Laval, Quebec for ddRADseq library construction. The restriction enzymes SbfI and

MspI were used to cut the DNA into fragments. Fragments were then ligated to two adapters, one

of which is a DNA barcode unique to each individual and one that is a common adapter (Davey

and Blaxter 2010). Fragments were size-selected using a BluePippin bioanalyzer (Sage Science,

U.S.A) and the resulting RADseq library was sequenced on a HiSeq 2000, using single read 100

bp sequencing at the Genome Quebec Innovation Center (McGill University, Montreal, Quebec).

Upon receiving the sequenced reads, the quality of the raw data was checked using FASTQC

(Andrews 2010).

Alignment with reference genome

Sequences were demultiplexed by barcode length using process_radtags in STACKS v 1.45

(Catchen et al. 2013). Reads were truncated to 90 bp and bases with a Phred score less than 30

were discarded. Individuals with less than 100 000 reads were then removed. BOWTIE2

(Langmead et al. 2012) was used to align reads to the pre-assembled reference genome of the

thick-billed murre (Uria lomvia) (Tigano et al. 2017a). Alignments were quality filtered,

retaining only those alignments with a mapping quality of 10. Ref_map (STACKS) was used to

align the reads to the thick-billed murre genome. The minimum number of identical, raw reads

required to create a stack was set to 10 and a population map with the sampling location of each

individual was supplied. Populations (STACKS) was then used to generate output files for

various programs to calculate population genetics statistics. For a locus to be included in the

analyses it had to be present in four of the six populations and possessed by 75% of the

individuals in a population. (Alternative values for these parameters were tested in order to select

those that allowed for the highest level of stringency while still retaining the majority of

47

informative markers.) A minimum minor allele frequency of 0.05 was required to process a

nucleotide site at a locus, and data analysis was restricted to one random SNP per locus.

VCFTOOLS (Danecek 2011) was used to determine the amount of missing data of each

individual. Individuals with more than 11% missing data were removed. A total of 32 individuals

and 5424 SNPs remained for further analyses.

Data Analyses

Population genetic analysis

Global FST and pairwise FST between population pairs were calculated with an analysis of

molecular variance (AMOVA) in GENODIVE and tested for significance using 10000

permutations (values were corroborated using the R-package hierfstat (Goudet 2004)) (Weir and

Cockerham 1984; Meirmans and Tienderen 2004). Global and population diversity indices and F-

statistics for each SNP were tested in STACKS and GENODIVE. Major allele frequencies,

percentage of polymorphic sites, and π indicate the general level of genetic diversity in a

population. FIS is the inbreeding coefficient, and positive values indicate nonrandom mating or

subpopulation structure (Allendorf et al. 2013). An analysis of variance (ANOVA) was used to

test for significant differences between diversity indices of populations in R v 3.3.1 (R Core

Team 2017).

The program STRUCTURE v 2.3.4 (Pritchard et al. 2000, Pritchard et al. 2010) was used

to infer population structure. Each run used a burn-in of 50,000 followed by 100,000 MCMC

iterations, using the admixture model with correlated allele frequencies and both with and

without sampling location as prior information. Each value of K from 1 to 10 was run 10 times

and delta K (ΔK) was used to test for the most probable number of genetic populations using

STRUCTURE HARVESTER (Evanno et al. 2005, Earl and vonHoldt 2012). Posterior

48

probabilities were calculated using the equation provided in the STRUCTURE manual (Pritchard

2010). Membership probabilities were averaged among runs using CLUMPP v 1.1.2. (Jakobsson

and Rosenberg 2007) and results were displayed using DISTRUCT v.1.1 (Rosenberg 2004). A

principal component analysis (PCA) was performed using the R-packages adegenet (Jombart

2008) and ade4 (Dray and Dufour 2007).

A correlation between geographic distance and genetic distance was tested using a Mantel

test (Mantel 1976) executed with the R-packages ecodist (Goslee and Urban 2007) and ade4 with

500000 replicates. Geographic distances were estimated using a measuring tool in Google Maps

(scale of 1:500 km) and were log-transformed (Google Map 2017). Shoreline distance was tested

as well as linear distance, because black guillemots will avoid flying over land or ice and so the

shortest functional distance between two populations would be along the coastline. Slatkin’s

linearized FST (D) was used as a measure of genetic distance. FST values were calculated in

GENODIVE and transformed to Slatkin’s FST (D = FST/(1-FST)). The Mantel tests were run with

the null hypothesis that r <= 0 and the alternative hypothesis that r > 0.

Outlier analysis

Outlier analysis was performed using BAYESCAN v 2.1 (Foll and Gaggiotti 2008),

which uses differences in allele frequencies between populations to identify putative candidate

SNPs under selection. Input files were converted to BAYESCAN format using PGDSPIDER v

2.1.1.2 (Lischer and Excoffier 2011). BAYESCAN was used to search for the signal of selection

among all SNPs in several pairwise comparisons, allowing for a false discovery rate of 5%.

Comparisons were made between a) those populations at higher latitudes (Alaska and Nunavut)

vs. those at lower latitudes (St. Lawrence, Nova Scotia, New Brunswick, and Maine), and b)

populations in the Arctic (Alaska) vs. populations in the Atlantic (Nunavut, St. Lawrence, Nova

Scotia, New Brunswick, and Maine). Prior odds for the neutral model were set to 1:10 and 20

49

pilot runs consisting of 5000 iterations each were run, followed by 10,000 iterations with a burn-

in length of 50,000 iterations as suggested per the manual (Foll 2012). Convergence was

assessed using the R package CODA (Plummer et al. 2006). BAYESCAN output was visualized

in R v 3.3.1. Posterior odds were used to indicate how much more likely the model with

selection was compared to the neutral model according to Jeffrey’s scale of evidence. Positive

alpha values were then used to distinguish SNPs under diversifying selection, while negative

alpha values were used to indicate balancing selection (Foll 2012). After detection of outliers,

the VCF file output from STACKS and VCFTOOLS were used to determine the frequency of

alleles and genotypes at these particular loci.

Results

Population genetic analysis

Most values for genetic diversity indices were similar among populations, with the exception of

nucleotide diversity and FIS (Table 3.2). The average frequency of polymorphic sites across all

sites was 0.47%, the average frequency of private alleles was 0.003% for all sites and 0.5% for

variant sites, and the average major allele frequency (P) among variant positions was 81.8%. An

estimate of global FIS was significantly greater than zero (FIS = 0.047, p < 0.05). Estimates of FIS

for individual populations were low but significantly greater than zero for all populations except

Nova Scotia (Table 3.2). The estimate of global population genetic structure was weak but

statistically significant (FST = 0.042, p <0.05). Estimates of FST between pairs of regional

samples were weak but significant for the majority of comparisons (Table 3.3). When location

was not used as prior information, results from STRUCTURE and STRUCTURE HARVESTER

indicated that black guillemots comprise two genetic groups (K=2; Supplementary Table 3.1).

However when location was used as prior information, results from STRUCTURE and

50

STRUCTURE HARVESTER were mixed. Delta K (ΔK) most strongly supported two genetic

groups (K=2; Figure 3.2A; Supplementary Table 3.1), while posterior probabilities supported

three genetic groups (K=3; Figure 3.2B; Supplementary Table 3.1). Assignment probabilities

suggested that genetic groups were not completely geographically segregated, but genetic

differentiation appeared to increase with geographic distance. Results from the PCA showed

marked differentiation among populations (Figure 3.3A). The first two principal components

explained 11.6% of the variance. When populations were grouped by subspecies in the PCA, C.

c. mandtii, C. c. ultimus, and C. c. arcticus all clustered as unique groups (Figure 3.3B). The first

two principal components explained 11.6% of the variance.

The Mantel test using shortest distance as the measurement of geographic distance

between regional samples was not significant (r = 0.35, p = 0.13). A Mantel test using shoreline

distance was significant however, and a positive correlation was found between geographic and

genetic distance (r = 0.63, p = 0.03; Figure 3.4). Results were the same from both ecodist and

ade4.

Outlier analysis

Different pairwise comparison in BAYESCAN produced a different number of outlier

SNPs: the comparison between high and low latitudes detected three outliers (SNP 3478, SNP

6283, and SNP 15077), and the comparison between the Arctic and Atlantic ocean populations

detected one outlier (SNP 13821). Log10 posterior odds for each outlier locus ranged from 1.2 to

3.4 (Supplementary Table 3.2), which according to Jeffrey’s scale of evidence (Foll 2012)

strongly supports the model with selection over the neutral model in all cases. All of these loci

produced positive alpha values (Supplementary Table 3.2), which suggests that they all

experienced diversifying selection. FST estimates averaged over all populations ranged from 0.19

to 0.24 for all loci (Supplementary Table 3.2). For each locus one allele was typically more

51

frequent across populations than the other (Supplementary Table 3.2). There did appear to be

some sorting of genotypes according to which biological group the population belonged to (i.e.

high latitude vs. low latitude). High latitude populations were exclusively homozygous for C/C

at SNP 3478, whereas low latitude populations contained a mix of homozygous C/C (27.3%),

heterozygous C/A (27.3%), and homozygous A/A (45.4%). At SNP 15077, high latitude

populations contained a mix of homozygous A/A (23.8%), heterozygous A/G (47.6%), and

homozygous G/G (28.6%), whereas low latitude populations were exclusively homozygous for

A/A. Upon closer inspection one of the SNPs (SNP 6283) from the latitudinal comparison and

the single SNP from the Pacific/Atlantic comparison were missing ~34% of their data making it

difficult to draw any further conclusions for these SNPs.

Discussion

Population genetic structure

Black guillemots exhibit weak but significant population genetic structure among regional

populations in North America at neutral nuclear loci. The estimate of global FST was statistically

significant and the majority of pairwise FST estimates were statistically significant (Table 3.3).

Defined population structure was also visible from STRUCTURE (Figure 3.2) and PCA analyses

(Figure 3.3A). These results suggest that gene flow might be restricted among black guillemot

populations to some degree. Populations from Maine and New Brunswick showed high

similarity, with birds from both populations identifying as almost 100% from the same genetic

cluster in STRUCTURE and overlapping in the PCA. This is supported by a low FST estimate for

pairwise population comparisons (Table 3.3). Similarity between these two populations is not

surprising, as New Brunswick and Maine are the closest sampling locations geographically. The

regions they inhabit (Gulf of Maine and Bay of Fundy) also share similar biological and physical

52

characteristics, which differ from those of the next nearest population in Nova Scotia, located

along the Scotian Shelf (Greenlaw et al. 2013; Ward-Paige and Bundy 2016). For this reason, it

makes sense that birds from Maine and New Brunswick might be more similar to each other than

to those from Nova Scotia as they are potentially well adapted to that region and potentially

overwinter in proximity to one another. Similar differences have been shown in Atlantic herring

(Clupea harengus), with fish sampled from the Scotian Shelf showing weak but significant

differentiation from those sampled in the Gulf of Maine and Bay of Fundy (McPherson et al.

2004). Birds from Nova Scotia, St. Lawrence, and Nunavut showed intermediate assignment to

the two genetic clusters identified by STRUCTURE at K=2 (Figure 3.2A). The distinction

becomes a bit clearer in the STRUCTURE plot from K=3 (Figure 3.2B): a third genetic group is

formed comprising Nunavut, St. Lawrence and Nova Scotia samples. Birds from Nunavut and St.

Lawrence then showed some probability of membership to the same genetic cluster as Alaskan

birds, while birds from Nova Scotia showed partial membership to the same genetic cluster as

birds from New Brunswick and Maine. STRUCTURE does not work well under isolation by

distance, which could explain the discrepancy (Pritchard 2010). Populations appeared to group

together somewhat according to their geographic distribution. More broadly, these differences

could also be associated with latitude. Black guillemots from Alaska are quite genetically

distinct from southeastern populations of black guillemots in North America. These populations

are the furthest from one another geographically, which could be leading to this marked

differentiation. The pattern of differentiation visible in black guillemots also corresponds to

timing of ice-retreat following the last glacial maximum during the Pleistocene (Dyke 2004).

The ice retreated first from the east coast of North America between Maine, New Brunswick and

Nova Scotia and these populations are more similar to each other than to the other populations.

The ice then retreated from the St. Lawrence, Nunavut, and Alaska at later dates.

53

The strong population genetic structure in black guillemots is similar to that seen in

pigeon guillemots (Chapter 2) and as with pigeon guillemots, this contrasts with weak or absent

genetic population differentiation in most other high-latitude seabird species (reviewed in Friesen

2015). Overall my results were in agreement with the findings of Kidd and Friesen (1998).

Although I did not survey all of the same populations, both studies detected marked

differentiation between North American populations. Kidd and Friesen found higher values of

FST in their analyses using the mitochondrial control region, however FST is a relative measure of

genetic differentiation and this could have been affected by the use of different sampling

locations in each project as well as the use of nuclear DNA as opposed to mtDNA.

Estimates of FIS were significantly greater than zero for all populations except Nova

Scotia. Positive values of FIS indicate a deficit of heterozygotes relative to Hardy-Weinberg

proportions and are often caused by nonrandom mating within populations or population

subdivision (Allendorf et al. 2013). Inbreeding is one consequence of small population sizes.

Although population sizes of black guillemots in North America are largely undocumented,

populations are typically small (Butler and Buckley 2002). Alternatively, significant estimates of

FIS could be due to the presence of null alleles.

Isolation by distance

Differentiation among black guillemot populations showed a strong positive correlation between

geographic distance (measured along the shoreline) and genetic distance. This often results from

a species’ dispersal being constrained by distance so that gene flow is most likely to occur

between neighbouring colonies (Hutchison and Templeton 1999), corresponding to a stepping

stone model (Kimura and Weiss 1964). As a result, populations that are closer together

geographically tend to be more genetically similar as suggested by the current results. This

supports our current understanding of dispersal among guillemots, as they are unlikely to fly over

54

land and water and will typically stay close to their breeding colonies throughout the year

minimizing the opportunity for gene flow. Isolation by distance can also result from secondary

contact between historically isolated populations (Bertl and Blum 2016). Further testing on the

evolutionary history of black guillemots is necessary. Kidd and Friesen (1998) found no

evidence of a relationship between genetic and geographic distance (measured both linearly and

along the shoreline between colonies) in their study, which was conducted at a larger scale.

Future analyses involving more widely distributed populations of black guillemots will be

necessary to determine whether the pattern of isolation by distance (at nuclear markers) holds up

on a global scale. The prevalence of isolation by distance in seabirds is not well known. In many

cases distance is not a barrier to dispersal in seabirds and geographic distance has been found to

only weakly explain the extent of population genetic structure (e.g. black-legged kittiwake, Rissa

tridactyla; reviewed in Friesen et al. 2007). While distance is clearly a factor in the

differentiation among guillemots, other factors are likely involved as well.

Outlier analysis: possible adaptation to climatic regions

The comparison between high and low latitude populations in BAYESCAN produced two viable

outliers that showed signatures of diversifying selection. Population genotypes varied according

to their membership in the high or low latitude group at both markers. Individuals from high

latitude populations were all C/C homozygotes at the first marker, and a combination of A/A,

A/G, and G/G genotypes at the second. Conversely, individuals from low latitudes were a

combination of C/C, C/A, and A/A genotypes at the first marker, and all A/A homozygotes at the

second. Populations at different latitudes are exposed to different environmental pressures and

evidence from population genetics analyses suggests that black guillemots do not migrate to

distant populations. As such, we might expect black guillemots to be locally adapted to the

specific environmental conditions of their region (such as climate and exposure to pack ice). A

55

study on thick-billed murres found that outlier loci grouped individuals from different

populations according to their non-breeding distribution suggesting that outliers may be

informative about adaptation and connectivity (Tigano et al. 2017b). Aside from this, molecular

methods have not been widely used to test for local adaptation in seabirds (some examples

include Li et al 2014; Vásquez-Carillo et al. 2014). Although the results from the present study

are not enough to conclude that black guillemots are locally adapted, it highlights the need for

further research in this area.

Subspecies and management

Subspecies of black guillemots were previously designated based on variation in morphological

differences (Storer 1952). Differentiation at neutral markers agrees with the current subspecies

delineations. Populations from different subspecies clustered together in genetic analyses,

suggesting that they are more similar to each other than to populations from other subspecies

(Figure 3.2; Figure 3.3). However, population structure also exists within C. c. arcticus. As

such, regional populations should be managed as separate units (with the exception of New

Brunswick and Maine).

Future directions

The use of ddRADseq to sequence thousands of markers allowed me to examine population

structure at a higher resolution than has previously been done. The use of a reference genome

created by whole-genome sequencing can be used in combination with RADseq methods to

improve the quality of population genetic analyses (Andres and Luikart 2014). Whole-genome

sequencing can also be used to further explore patterns of adaptive molecular evolution and to

link genes to phenotypes (Toews et al. 2016). This approach would be useful for furthering our

understanding of how black guillemots might be genetically adapted to local climates. As well,

this study made use of only a small number of individuals per population. Increasing sample

56

sizes will improve power of analyses. My sample distribution spanned most of North America,

however some important sampling gaps still exist, such as Newfoundland and Labrador and areas

of the Chukchi and Beaufort Seas. This study also focused solely on populations in North

America, however black guillemots are distributed throughout the Arctic and North Atlantic.

Future analyses including samples from other parts of the distribution and all of the subspecies

will allow us to make more general conclusions about population genetic structure in black

guillemots. As well, future work on the evolutionary history of black guillemots is necessary to

determine whether genetic differentiation in black guillemots arose as a result of historical

isolation or contemporary processes.

57

Table 3.1. Subspecies, sampling regions, region abbreviations, sites, and sample numbers (n) for

black guillemots.

Subspecies Region

Site n Abbrev- iation

mandtii Alaska AK Cooper Is. 8

ultimus Nunavut NU Prince Leopold Is. 4

Digges Sound 9

Coats Island 3

East Bay 3

arcticus St. Lawrence STL Long Pèlerin 5

Nova Scotia NS Country Is. 10

New Brunswick NB Kent Is. 5

Maine ME Great Duck Is. 8

Total 55

58

Table 3.2. Summary statistics for all populations across variant nucleotide sites and all sites.

Statistics include the number of private alleles in each population (private), the number of

polymorphic (top) or total (bottom) nucleotide sites across the dataset (sites), percentage of

polymorphic loci (% Poly), the average frequency of the major allele (P), the average observed

heterozygosity per locus (Ho), the average nucleotide diversity (π), and the average Wright’s

inbreeding coefficient (FIS). FIS values significantly greater than zero (p < 0.05) are indicated with

‘*’. Region abbreviations as in Table 3.1.

Pop ID Private Sites % Poly P (%) Ho π FIS Variant positions

AK 64 5265 80.36 81 0.267 0.279 0.0331* NU 26 5355 69.49 82.1 0.268 0.269 0.0083* STL 6 5311 47.49 84.27 0.252 0.263 0.0169* NS 26 4866 86.79 80.57 0.274 0.287 0.0374 NB 6 5236 65.15 81.99 0.276 0.283 0.0124* ME 27 5216 82.27 81.06 0.265 0.281 0.0405*

All positions AK 64 810913 0.52 99.88 0.002 0.002 0.0002 NU 26 820550 0.45 99.88 0.002 0.002 0.0001 STL 6 808934 0.31 99.9 0.002 0.002 0.0001 NS 26 747877 0.56 99.87 0.002 0.002 0.0002 NB 6 808377 0.42 99.88 0.002 0.002 0.0001 ME 27 804210 0.53 99.88 0.002 0.002 0.0003

59

Table 3.3. Estimates of FST for pairwise comparisons of black guillemot regional samples based

on nuclear loci. Region abbreviations as in Table 3.1.

AK NU STL NS NB ME n 8 3 2 8 3 8 AK

NU 0.03* STL 0.07* 0.05

NS 0.05* 0.03* 0.05* NB 0.07* 0.05 0.07 0.02*

ME 0.07* 0.05* 0.08* 0.03* 0.01* *Significantly different from 0 at α = 0.05

60

Figure 3.1 Distribution map of black guillemots (modified from Storer 1952; Svanberg and

Ægisson 2006). Thin lines represent approximate ranges of subspecies. White dots represent

sampling sites. Sampling site codes from northwest to southeast: CI = Cooper Is., PLI = Prince

Leopold Is., DS = Digges Sound, Cts = Coats Island, EB = East Bay, LP = Long Pèlerin, Ctr =

Country Is., KI = Kent Is., GDI = Great Duck Is.

Figure 3.2 Probabilities of assignment of individuals to different genetic populations derived

from the program STRUCTURE, based on genetic variation in nuclear loci of black guillemots

with sampling location used as prior information. Populations are ordered as they appear

geographically, from northwest to southeast (left to right). See Table 3.1 for abbreviations.

A) Probabilities of assignment of individuals to each of two genetic populations (K=2).

B) Probabilities of assignment of individuals to each of three genetic populations (K=3).

Figure 3.3 Results of principal component analysis of nuclear variation in black guillemots

grouped by A) population and B) subspecies.

A) 1 = Alaska 2 = Nunavut 3 = St. Lawrence 4 = Nova Scotia, 5 = New Brunswick, 6 =

Maine.

B) 1 = mandtii (Alaska), 2 = ultimus (Nunavut), 3 = arcticus (St. Lawrence, Nova Scotia,

New Brunswick, Maine).

Figure 3.4. Regression plot from Mantel test testing for correlation between genetic distance

(Slatkin’s linearized FST) and geographic distance (log-transformed shoreline distance). Points

that represent comparisons between subspecies are circled and labeled.

61

Fig. 3.1

62

A

B

Fig. 3.2

63

A

B

Fig. 3.3

d = 5

1

2

3

4

5

6

Eigenvalues

d = 5

1

2

3

Eigenvalues

64

Fig. 3.3

65

Chapter 4

General Discussion

Population Genetic Structure

My first objective was to investigate the extent to which regional populations of both pigeon

guillemots (Cepphus columba) and black guillemots (C. grylle) exhibited differentiation at

neutral markers. Genetic variation was strongly structured geographically in both species of

guillemot. In pigeon guillemots, mitochondrial control region sequences showed high variability

and the majority of haplotypes were private. Weak population structure was also found using

nuclear DNA (four microsatellites and two introns). Global estimates of ΦST and FST were

statistically significant and the majority of pairwise estimates for genetic differentiation between

populations were significant. Analyses using STRUCTURE and PCA also provided evidence for

geographically structured genetic variation. In black guillemots, clear population genetic

structure was evident from analyses with nuclear DNA (5424 single nucleotide polymorphisms

(SNPs) from double-digest restriction-site associated DNA sequencing (ddRADseq)). A global

estimate of FST was significant and estimates of FST between population pairs were weak but

significant for the majority of comparisons. Analyses using STRUCTURE and principal

components analysis (PCA) also provided evidence for geographically structured genetic

variation. This marked genetic structure suggests that gene flow among guillemot populations

within a species might be restricted. I was therefore able to reject the null hypothesis that

guillemots are not genetically differentiated at neutral markers.

Isolation by distance

As part of my first objective I also wanted to determine whether genetic differentiation

was affected by distance between populations in both pigeon and black guillemots. I predicted

66

that genetic differentiation in guillemots would follow a pattern of isolation by distance as a

result of dispersal following a stepping stone model. In both species, genetic distance was

strongly correlated with geographic distance (measured as the shortest distance along the

shoreline between populations). As well, in pigeon guillemots, shared haplotypes from the

mitochondrial control region occurred primarily between neighbouring colonies. Results from

STRUCTURE and PCA in both species also suggested increasing genetic differences with

geographic distance. This pattern of genetic differentiation in which populations that are closer

geographically are more similar genetically is typical of isolation by distance (Mantel 1967;

Hutchison and Templeton 1999). I was able to reject the null hypothesis that distance between

populations had no effect on genetic differentiation. These results contradict those from Kidd

and Friesen (1998), who found no association between geographic and genetic distance in black

guillemots. Isolation by distance has not been well studied in seabirds. Often, distance is not a

barrier to dispersal among seabirds as they are typically strong fliers and individuals from some

species have been known to travel long distances for foraging trips and migration (Shealer 2002;

Egevang et al. 2010). In seabird species that do exhibit genetic differentiation, geographic

distance only weakly explained the extent of population genetic structure (reviewed in Friesen et

al. 2007). As such, guillemots provide a useful system for studying how genetic differentiation is

affected by limited dispersal in a highly mobile organism.

Evolutionary history

In Chapter 2, my second objective was to determine whether differentiation in pigeon

guillemots was a result of historical isolation or contemporary processes. I predicted that genetic

differentiation in pigeon guillemots could have arisen historically, through isolation in glacial

refugia or alternatively, they could have become differentiated in situ through genetic drift and /

67

or selection due to restricted gene flow. I found evidence for differentiation through both

historical and contemporary processes. MtDNA haplotypes and estimates of ΦST ad FST

supported historical isolation of pigeon guillemots in the Aleutian Islands versus continental

North America. The Aleutian Islands populations represented a monophyletic split in the

phylogenetic tree and divergence was estimated to have occurred prior to the Wisconsin

glaciation. Results from DIYABC also supported pre-glacial timing of this divergence.

Contemporary processes were supported by results from STRUCTURE, PCA, and a Mantel test,

all of which suggested that populations were more genetically similar to populations situated

nearby than to those further away. As well, shared haplotypes from the mitochondrial control

region occurred primarily between neighbouring colonies. The combination of these results

provides strong evidence that both historical and contemporary processes led to differentiation in

pigeon guillemots.

Outlier analysis

In Chapter 3, I also performed exploratory outlier analyses to identify loci putatively

under selection in black guillemots. I identified two viable loci from a pairwise comparison

between high and low latitude populations, and both loci showed signatures of diversifying

selection. At each locus, genotypes sorted according to latitude. One possible explanation is that

high and low latitude populations are found in different climate regions. Further study is required

to make definitive conclusions about loci under selection in black guillemots.

Subspecies and management

Subspecies have been described for both pigeon guillemots and black guillemots (Storer

1952). Three out of five pigeon guillemot subspecies and three out of six black guillemot

subspecies were represented in the present studies. Population genetic structure in pigeon

68

guillemots did not reflect the current subspecies divisions and it is my recommendation that these

be revised. I would also suggest that pigeon guillemot populations in the Aleutian Islands be

considered for designation as an Evolutionarily Significant Unit (ESU) as their differentiation has

been strongly influenced by historical processes (Casacci et al. 2014). Population genetic

structure in black guillemots did agree with current subspecies delineations. Based on the level of

variation detected within the subspecies however, I would recommend that black guillemots be

treated as management units (sensu Moritz 1994).

Importance

Understanding how evolutionary forces (i.e. mutation, genetic drift, gene flow, and

natural selection) influence genetic variation within and among populations and shape patterns of

biodiversity is fundamental to the study of evolutionary genetics. Guillemots provide a useful

system for investigating how these evolutionary forces shape population genetic differentiation in

a genus with a wide range that disperses according to a one-dimensional stepping stone model.

Despite their capacity for high mobility, I found evidence of limited dispersal and restricted gene

flow between distant populations. Genetic differentiation appeared to follow a pattern of

isolation by distance. As well, population genetic structure appeared to be a combination of both

historical and contemporary processes. The results from this study hold promising insight into

how certain patterns of gene flow can lead to different patterns of population genetic structure.

Future Directions

Future studies should employ high-throughput sequencing techniques to increase the

number of markers and to produce a reference genome. The use of a reference genome will

increase the quality of the filtered data (Andrews and Luikart 2014). Many populations are

represented by a small number of individuals. Increasing sample sizes and using roughly the

69

same number of samples for each population will improve the accuracy and power of analyses

(Quinn and Keough 2002). Several important sampling gaps existed in each study and gathering

samples from these areas would improve our understanding of population genetic structure in

both species.

70

Summary

Chapter 2

Both historical and contemporary processes lead to population genetic differentiation in

pigeon guillemots (Cepphus columba)

1. Understanding the forces that shape population genetic structure is fundamental to the study of

evolution.

2. Pigeon guillemots (Cepphus columba) experience many potential non-geographic barriers to

gene flow, which could result in strong population genetic structure.

3. I analysed variation in the mitochondrial control region, four microsatellite loci, and two

introns to assess genetic differentiation at neutral markers in pigeon guillemots and to determine

whether differentiation was a result of historical isolation or contemporary processes.

4. I found significant population genetic structure at both mitochondrial and nuclear markers.

Genetic differentiation followed a pattern of isolation by distance. Marked differences were

found between the Aleutian Island populations and mainland populations and strong evidence

was found for differentiation in pigeon guillemots arising as a result of both historical isolation

and contemporary processes.

5. I recommend that the current pigeon guillemot subspecies designations be reviewed and that

pigeon guillemot populations in the Aleutian Islands be considered for designation as an

evolutionarily significant unit (ESU) as their differentiation has been strongly influenced by

historical processes.

71

Summary

Chapter 3

An assessment of population genomic structure in black guillemots (Cepphus grylle) in

North America

1. If local populations are genetically differentiated, then extirpation of a population results in

partial loss of a species’ genetic diversity and local adaptations, which can affect a species’

ability to adapt to stressors.

2. Black Guillemots (Cepphus grylle) may be highly vulnerable to climate change in parts of their

range, but little research has been done to assess genetic variation in this species.

3. I conducted a genome-wide survey of genetic variation using double-digest restriction-site

associated DNA sequencing (ddRADseq) to determine the extent to which regional populations

of black guillemots differ at neutral and putatively adaptive markers. I tested the hypothesis that

regional populations of black guillemots are genetically distinct and I predicted that

differentiation in black guillemot would follow a pattern of isolation by distance.

4. I found that black guillemots exhibit clear population genetic structure, suggesting that gene

flow among black guillemot populations is restricted. Genetic differentiation followed a pattern

of isolation by distance.

5. Outlier analyses detected two putatively adaptive loci. Genotypes at these loci were sorted by

latitude.

6. My research indicates that regional populations of black guillemots should be managed as

separate units.

72

Literature Cited

Ainley, D.G., Boekelheide, R.J., Morrell, S.H., and C.S. Strong. 1990. Pigeon guillemot. Pages 276-305 in D.G. Ainley, and R.J. Boekelheide, editors. Seabirds of the Farallon Islands: Ecology, Dynamics, and Structure of an Upwelling-System Community. Stanford University Press, Stanford CA.

Allendorf, F.W., Luikart, G., and S.N. Aitken. 2013. Conservation and the Genetics of

Populations, 2nd edition. Blackwell Publishing, UK. Ando, H., L. Young, M. Naughton, H. Suzuki, T. Deguchi, and Y. Isagi 2014. Predominance of

unbalanced gene flow from the Western to Central North Pacific colonies of the Black-footed Albatross (Phoebastria nigripes). Pacific Science 38:309–318.

Andrews S. 2010. FastQC: a quality control tool for high throughput sequence data.

(http://www.bioinformatics.babraham.ac.uk/projects/fastqc) Andrews, K.R. and G. Luikart. 2014. Recent novel approaches for population genomics data

analysis. Molecular Ecology 23:1661–1667. Austin, O.L. 1929. “The Races of Cepphus grylle (Linn.)”. Bulletin of the Northeastern Bird-

Banding Association 5.1: 1–6. Benestan, L., Gosselin, T., Perrier, C., Sainte-Marie, B., Rochette, R., and L. Bernatchez. 2015.

RAD genotyping reveals fine-scale genetic structuring and provides powerful population assignment in a widely distributed marine species, the American lobster (Homarus americanus). Molecular Ecology 24:3299–3315.

Benjamini, Y. and D. Yekutieli. 2001. The control of false discovery rate under dependency.

Annals of Statistics 29:1165–1188. Bertl, J., and M.G.B Blum. 2016. Can secondary contact following range expansion be

distinguished from barriers to gene flow? bioRxiv (https://doi.org/10.1101/043398). BirdLife International. 2016. Cepphus grylle. The IUCN Red List of Threatened Species 2016.

(http://dx.doi.org/10.2305/IUCN.UK.2016-3.RLTS.T22694861A86848264.en). Birt, T.P., Mackinnon, D., Piatt, J.F., and V.L. Friesen. 2011. Genetic differentiation of the

Kittlitz’s Murrelet Brachyramphus brevirostris in the Aleutian Islands and Gulf of Alaska. Marine Ornithology 39:39-45.

Bradstreet, M.S.W. 1979. Thick-billed Murres and Black Guillemots in the Barrow Strait area,

N.W.T., during spring: distribution and habitat use. Canadian Journal of Zoology 57:1787-1802.

73

Bradstreet, M.S.W. 1980. Thick-billed murres and black guillemots in the Barrow Strait area, N.W.T., during spring: diets and food availability along ice edges. Canadian Journal of Zoology 58: 2120-2140.

Butler, R.G. and D.E. Buckley. 2002. Black Guillemot (Cepphus grylle). in A. Poole, editor. The

Birds of North America Online. Cornell Lab of Ornithology, Ithaca, New York. Cairns, D. K. 1980. Nesting density, habitat structure and human disturbance as factors in Black

Guillemot reproduction. Wilson Bulletin 92:352-361. Cairns, D.K. 1987. Diet and foraging ecology of Black Guillemots in northeastern Hudson Bay.

Canadian Journal of Zoology 65: 1257-1263. Canal, D., Roques, S., Negro, J.J., and J.H. Sarasola. 2017. Population genetics of the endangered

Crowned Solitary Eagle (Buteogallus coronatus) in South America. Conservation Genetics 18: 235-240.

Casacci, L.P., Barbero, F., and E. Balletto. 2014. The “Evolutionarily Significant Unit” concept

and its applicability in biological conservation. Italian Journal of Zoology 81:182-193. Catchen, J., Hohenlohe, P., Bassham, S., Amores, A., and W. Cresko. 2013 Stacks: an analysis

tool set for population genomics. Molecular Ecology. 22:3124-4140. Chakraborty, R. 1990. Mitochondrial DNA polymorphism reveals hidden heterogeneity within

some Asian populations. American Journal of Human Genetics 47:87–94. Chenoweth, S.F., Hughes, J.M., Keenan, C.P., and S. Lavery. 1998. Concordance between

dispersal and mitochondrial gene flow: isolation by distance in a tropical teleost, Lates calcarifer (Australian barramundi). Heredity 80:187-197.

Clement, M., Posada, D., and K. Crandall. 2000. TCS: a computer program to estimate gene

genealogies. Molecular Ecology 9: 1657-1660 Clement, M., Snell, Q., Walke, P., Posada, D., K. Crandall. 2002. TCS: estimating gene

genealogies. Proc. of the16th International Parallel and Distributed Processing Symposium 2:184.

Congdon, B.C., Piatt, J.F., Martin, K., and V.L. Friesen. 2000. Mechanisms of population

differentiation in marbled murrelets: historical versus contemporary processes. Evolution 54:974-986.

Cornuet, J.M., Ravigné, V., and A. Estoup. 2010. Inference on population history and model

checking using DNA sequence and microsatellite data with the software DIYABC (v1.0). BMC Bioinformatics 11:1-11.

74

Cornuet, J.M., Santos, F., Beaumont, M.A., Robert, C.P., Marin, J., Balding, D.J., Guillemaud, T., and A. Estoup. 2008. Inferring population history with DIY ABC: a user-friendly approach to approximate Bayesian computation. Bioinformatics 24:2713-2719.

Coulson, J.C. 2016. A review of philopatry in seabirds and comparisons with other waterbird

species. Waterbirds, 39:229-240. 2016 Danecek, P., Auton, A., Abecasis, G., Albers, C.A., Banks, E., DePristo, M.A., Handsaker, R.,

Lunter, G., Marth, G., Sherry, S.T. McVean, G., Durbin, R., and 1000 Genomes Project Analysis Group. 2011. The Variant Call Format and VCFtools, Bioinformatics, 27: 2156-2158.

Darriba, D., Taboada, G. L., Doallo, R., and D. Posada. 2012. jModelTest 2: more models, new

heuristics and parallel computing. Nature Methods 9:772. Davey, J.W. and M.L. Blaxter. 2010. RADSeq: next-generation population genetics. Briefings in

Functional Genomics 9: 416-423. Dawson, J. G., Gibbs, H. L., Hobson, K. A., and S. M. Yezerinac. 1997. Isolation of

microsatellite DNA markers from a passerine bird, Dendroica petechia (the yellow warbler), and their use in population studies. Heredity 79:506-514.

Deane, P.E., Mccoy, K.D., Robertson, R.J., Birt, T.P., and V.L. Friesen. 2013. Minimal genetic

structure in the cerulean warbler despite evidence for ecological differentiation among populations. Condor 115:178-185.

Diamond, H.J., and C.J. Lief. 2009. A Comprehensive Data Portal for Global Climate

Information. EOS, Transactions, American Geophysical Union 90:341-348. Dierickx, E.G., Shultz, A.J., Sato, F., Hiraoka, T., and S.V. Edwards. 2015. Morphological and

genomic comparisons of Hawaiian and Japanese Black-footed Albatrosses (Phoebastria nigripes) using double digest RADseq: implications for conservation. Evolutionary Applications 8: 662-678.

Divoky, G. 2011. Black Guillemots in a melting Arctic: Responding to shifts in prey,

competitors, and predators. Pages 125–130 in R. T. Watson, T. J. Cade, M. Fuller, G. Hunt, and E. Potapov, editors. Gyrfalcons and Ptarmigan in a Changing World, Volume I. The Peregrine Fund, Boise, Idaho, USA.

Divoky, G.J., Douglas, D.C., and I.J. Stenhouse. 2016. Arctic sea ice a major determinant in

Mandt's black guillemot movement and distribution during non-breeding season. Biology Letters 12: 1-5.

Divoky, G.J., Lukacs, P.M., and M.L. Druckenmiller. 2015. Effects of recent decreases in arctic

sea ice on an ice-associated marine bird. Progress in Oceanography. 136: 151-161.

75

Dray, S., and A.B. Dufour. 2007. The ade4 package: implementing the duality diagram for ecologists. Journal of Statistical Software. 22: 1-20.

Drummond, A.J., Rambaut, A., Shapiro, B., and O.G. Pybus. 2005. Bayesian coalescent

inference of past population dynamics from molecular sequences. Molecular Biology and Evolution 22:1185–1192.

Drummond, A.J., Suchard, M.A., Xie, D., and A. Rambaut. 2012. Bayesian phylogenetics with

BEAUti and the BEAST 1.7. Molecular Biology And Evolution 29: 1969-1973 Dyke, A.S. 2004. An outline of North American deglaciation with emphasis on central and

northern Canada. Developments in Quaternary Sciences 4:373-424.

Dyke, A.S. and V.K. Prest. 1987. Paleogeography of northern North America, 18 000–5000 years ago. Geological Survey of Canada, Map 1703A.

Earl, D.A. and B.M. vonHoldt. 2012. STRUCTURE HARVESTER: a website and program for

visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4:359-361.

Eda, M., Kawakami, K., Chiba, H., Suzuki, H., Horikoshi, K., and H. Koike. 2008. Genetic

characteristics of the Black-footed albatross Diomedea nigripes on the Bonin Islands and their implications for the species’ demographics history and population structure. Ornithological Science 7:109–116.

Egevang, C., Stenhouse, I.J., Phillips, R.A., Petersen, A., Fox, J.W., and J.R.D. Silk. 2010.

Tracking of Arctic terns Sterna paradisaea reveals longest animal migration. PNAS 197:2078-2081.

Ellstrand, N.C., and L.H. Rieseberg. 2016. When gene flow really matters: gene flow in applied

evolutionary biology. Evolutionary Applications 9:833-836. Evanno, G., Regnaut, S., and J. Goudet. 2005. Detecting the number of clusters of individuals

using the software STRUCTURE: a simulation study. Molecular Ecology 14:2611–2620. Ewens, W.J. 1972. The sampling theory of selectively neutral alleles. Theoretical Population

Biology. 3:87–112. Ewins, P.J. 1985. Colony attendance and censusing of Black Guillemots Cepphus grylle in

Shetland. Bird Study 32:176-185. Excoffier, L. and H.E.L. Lischer. 2010. Arlequin suite ver 3.5: A new series of programs to

perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: 564-567.

Foll, M. 2012. BayeScan v 2.1 User Manual. Ecology 20:1450-1462.

76

Foll, M. and O. Gaggiotti. 2008. A Genome-Scan Method to Identify Selected Loci Appropriate

for Both Dominant and Codominant Markers: A Bayesian Perspective. GENETICS 180:977-993.

Frederiksen, M. and A. Petersen. 1999. Philopatry and dispersal within a Black Guillemot colony.

Waterbirds: The International Journal of Waterbird Biology 22: 274-281. Frederiksen, M. and A. Petersen. 2000. The importance of natal dispersal in a colonial seabird,

the Black Guillemot Cepphus grylle. Ibis 142: 48-57. Friesen, V.L. 2015. Speciation in seabirds: why are there so many species...and why aren’t there

more? Journal of Ornithology 156: 27-39. Friesen, V.L., Birt, T.P., Piatt, J.F., Golightly, R.T., Newman, S.H., Hebert, P.N., Congdon, B.C.

and G. Gissing. 2005. Population genetic structure and conservation of marbled murrelets (Brachyramphus marmoratus). Conservation Genetics 6:607–614.

Friesen, V.L., Burg, T. and K.D. McCoy. 2007. Mechanisms of population differentiation in

seabirds. Molecular Ecology 16:1765-1785. Friesen, V.L., Congdon, B.C., Kidd, M.G., and T.P. Birt. 1999. PCR primers for the

amplification of five nuclear introns in vertebrates. Molecular Ecology 8:2147-2149. Friesen, V.L., Congdon, B.C., Walsh, H.E., and T.P. Birt. 1997. Intron variation in Marbled

Murrelets detected using analyses of single-stranded conformational polymorphisms. Molecular Ecology 6:1047-1058.

Gaston, A.J. and I.L. Jones. 1998. The Auks: Alcidae. Oxford University Press, New York. Gold, J.R., Burridge, C.P., and T.F. Turner. 2001. A modified stepping-stone model of population

structure in red drum, Sciaenops ocellatus (Sciaenidae), from the northern Gulf of Mexico. Genetica 111:305-317.

Golet, G.H., Kuletz, K.J., Roby, D.D., and D.B. Irons. 2000. Adult prey choice affects chick

growth and reproductive success in pigeon guillemots. The Auk, 117: 82-91. Gómez-Díaz, E., González-Solís, J., and M. A. Peinado. 2009. Population structure in a highly

pelagic seabird, the Cory’s shearwater Calonectris diomedea: an examination of genetics, morphology and ecology. Marine Ecology Progress Series 382: 197-209.

Google Maps. 2017. International Cartographic Association.

(https://www.google.ca/maps/@50.7990075,-150.9767058,3.87z) Goslee, S.C. and D.L. Urban. 2007. The ecodist package for dissimilarity-based analysis of

ecological data. Journal of Statistical Software 22(7):1-19.

77

Goudet, J. 2004. Hierfstat, a package for r to compute and test hierarchical F-statistics. Molecular

Ecology Resources 5:185-186. Greenlaw, M.E., Gromack, A.G., Basquill, S.P., MacKinnon, D.S., Lynds, J.A., Taylor, R.B.,

Utting, D.J., Hackett, J.R., Grant, J., Forbes, D.L., Savoie, F., Bérubé, D., Connor, K.J., Johnson, S.C., Coombs, K.A., and R. Henry. 2013. A Physiographic Coastline Classification of the Scotian Shelf Bioregion and Environs: The Nova Scotia Coastline and the New Brunswick Fundy Shore. Fisheries and Oceans Canada 1-49.

Hall, T.A. 1999. “BioEdit: a user-friendly biological sequence alignment editor and analysis

program for Windows 95/98/NT”. Nucleic Acids Symposium Series 41: 95-98. Hasegawa, M., Kishino, H., T. Yano. 1985. Dating of human-ape splitting by a molecular clock

of mitochondrial DNA. Journal of Molecular Evolution. 22:160–174. Hellberg, M. E. 1995. Stepping-stone gene flow in the solitary coral Balanophyllia elegans:

equilibrium and nonequilibrium at different spatial scales. Marine Biology 123: 573-581. Hernández, M. A., Campos, F., Santamaría, T., Rojo, M. A. and S. Dias. 2016. Is isolation by

distance the cause of the genetic structure of the Iberian white-throated dipper populations? Journal of Zoology 299:27-36.

Holder, K., Montgomerie, R., and V.L. Friesen. 2000. Genetic diversity and management of

Nearctic rock ptarmigan (Lagopus mutus). Canadian Journal of Zoology 82:564–575. Hudson, P.J. 1985. Population parameters for the Atlantic Alcidae. Pages 233 -261 in D.N.

Nettleship, T.R. Birkhead, editors. The Atlantic Alcidae, Academic Press, New York. Hudson, R., Slatkin, M., and W.P. Maddison. 1992. Estimation of levels of gene flow from DNA

sequence data. Genetics 132, 583-589. Hutchison, D.W. and A.R. Templeton. 1999. Correlation of pairwise genetic and geographic

distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53:1898-1914.

Ibarguchi, G., Birt, T.P., Warheit, K.I., Boag, P.T., and V.L. Friesen. 2000. Microsatellite loci

from common and thick-billed murres, Uria aalge and U. lomvia. Molecular Ecology 9:629-644.

Inoue, K., Monroe, E.M., Elderkin, C.L., and D.J. Berg. 2014. Phylogeographic and population

genetic analyses reveal Pleistocene isolation followed by high gene flow in a wide ranging, but endangered, freshwater mussel. Heredity 112:282-290.

78

Irwin, D. E., Irwin, J. H., and T. B. Smith. 2011. Genetic variation and seasonal migratory connectivity in Wilson’s Warblers (Wilsonia pusilla): Species-level differences in nuclear DNA between western and eastern populations. Molecular Ecology 20:3102–3115.

Jakobsson, M., and N.A, Rosenberg. 2007. CLUMPP: a cluster matching and permutation

program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23:1801–1806.

Jeffreys, H. 1946. An invariant form for the prior probability in estimation problems. Proceedings

of the Royal Society of London. Series A, Mathematical and Physical Sciences. 186: 453–461.

Jombart, T. 2008. adegenet: a R package for the multivariate analysis of genetic

markers. Bioinformatics 24:1403-1405. Kahindo, C. M., Bates, J. M. and R.C.K. Bowie. 2017. Population genetic structure of Grauer's

Swamp Warbler Bradypterus graueri, an Albertine Rift endemic. Ibis 159: 415–429. Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., Buxton, S.,

Cooper, A., Markowitz, S., Duran, C., Thierer, T., Ashton, B., Mentjies, P., and A. Drummond. 2012. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28:1647-1649.

Kidd, M.G. and V.L. Friesen. 1998. Analysis of the mechanisms of microevolutionary change in

Cepphus guillemots using patterns of control region variation. Evolution 52: 1158-1168. Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions

through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111– 120.

Kimura, M. and G.H. Weiss. 1964. The stepping stone model of population structure and the

decrease of genetic correlation with distance. Genetics 49:561-576. Kimura, M., Clegg, S.M., Lovette, I.J., Holder, K.R., Girman, D.J., Milá, B., Wade, P., and T. B.

Smith. 2002. Phylogeographical approaches to assessing demographic connectivity between breeding and overwintering regions in a Nearctic–Neotropical warbler (Wilsonia pusilla). Molecular Ecology 11:1605– 1616.

Kingman, J.F.C. 1982. The coalescent. Stochastic Processes and their Applications 13:235-248. Kraus, R.H.S, van Hooft, P., Megens, H.J., Tsvey, A., Fokin, S.Y., Ydenberg, R.C., and H.H.

Prins. 2012. Global lack of flyway structure in a cosmopolitan bird revealed by a genome wide survey of single nucleotide polymorphisms. Molecular Ecology 22:41-55.

79

Kushlan, J.A., Steinkamp, M.J., Parsons, K.C., Capp, J., Cruz, M.A., Coulter, M., Davidson, I., Dickson, L., Edelson, N., Elliot, R., Erwin, R.M., Hatch, S.A., Kress, S., Milko, R., Miller, S., Mills, K.L., Paul, R., Phillips, R., Saliva, J.E., Syderman, B., Trapp, J., Wheeler, J. and K.D. Wohl. 2002. Waterbird conservation for the Americas: the North American Waterbird Conservation Plan, version 1. Waterbird Conservation for the Americas. Washington, DC, U.S.A.

Kvistad, L., Ingwersen, D., Pavlova, A., Bull, J.K., and P. Sunnucks. 2015. Very low population

structure in a highly mobile and wide-ranging endangered bird species. PLoS ONE 10: e0143746.

Langmead B., and S. Salzberg. 2012 Fast gapped-read alignment with Bowtie 2. Nature Methods

9:357-359. Li, C., Zhang, Y., Li, J., Kong, L., Hu, H., Pan, H., Xu, L., Deng, Y., Li, Q., Jin, L., Yu, H.,

Chen, Y., Liu, B., Yang, L., Liu, S., Zhang. Y., Lang, Y., Xia, J., He, W., Shi, Q., Subramanian, S., Millar, C.D., Meader, S., Rands, C.M., Fujita, M.K.M, Greenwold, M.J., Castoe, T.A., Pollock, D.D, Gu, W., Nam, K., Ellegren, H., Ho, S.Y.W., Burt, D.W., Ponting, C.P., Jarvis, E.D., Gilbert, M.T.P., Yang, H., Wang, J., Lambert, D.M., Wang, J., and G. Zhang. 2014. Two Antarctic penguin genomes reveal insights into their evolutionary history and molecular changes related to the Antarctic environment. GigaScience 3:1-15.

Lischer, H.E.L. and L. Excoffier. 2012. PGDSpider: An automated data conversion tool for

connecting population genetics and genomics programs. Bioinformatics 28: 298-299. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach.

Cancer Research. 27: 209–220. Marshall, H.D., and A.J. Baker. 1997. Mitochondrial control region sequences as tools for

understanding evolution of avian taxa. Pages 49-80 in D. P. Mindell, editor. Avian molecular systematics and evolution. Academic Press, New York, U.S.A.

Matlock, G.C., 1987. The life history of the red drum, pp. 1–47 in Manual of Red Drum

Aquaculture, edited by G.W. Chamberlain, R.J. Miget & M.G. Haby. Texas Agricultural Extension Service and Sea Grant College Program, Texas A&M University, College Station, Texas.

McPherson A.A., O’Reilly, P.T., and C.T. Taggart. 2004. Genetic Differentiation, Temporal

Stability, and the Absence of Isolation by Distance among Atlantic Herring Populations. Transactions of the American Fisheries Society 133:434-446.

Meirmans, P., and P.H. van Tienderen. 2004. GENOTYPE and GENODIVE: two programs for

the analysis of genetic diversity of asexual organisms. Molecular Ecology Resources 4:792-794.

80

Michels, C. 1997. Latitude/Longitude Distance Calculator . Northern Arizona University Information Technology Services. (http://jan.ucc.nau.edu/~cvm/latlongdist.html)

Mims, M.C., Hauser, L., Goldberg, C.S., and J.D. Olden. 2016. Genetic differentiation, isolation-

by-distance, and metapopulation dynamics of the Arizona Treefrog (Hyla wrightorum) in an isolated portion of its range. PLoS ONE 11:1-23.

Morinha, F., Dávila, J. A., Bastos, E., Cabral, J. A., Frías, Ó., González, J. L., Travassos, P.,

Carvalho, D., Milá, B. and G. Blanco. 2017. Extreme genetic structure in a social bird species despite high dispersal capacity. Molecular Ecology 10:2812-2825.

Moritz, C., 1994. Applications of mitochondrial DNA analysis in conservation: a critical review.

Molecular Ecology 3:401-411. Morris-Pocock, J.A., Taylor, S.A., Birt, T.P., Damus, M., Piatt, J.F., Warheit, K.I., and V.L.

Friesen. 2008. Population genetic structure in Atlantic and Pacific Ocean common murres (Uria aalge): natural replicate tests of post-Pleistocene evolution. Molecular Ecology 17: 4859-4873.

Moum, T. and E. Árnason. 2001. Genetic diversity and population history of two related seabird

species based on mitochondrial DNA control region sequences. Molecular Ecology 10:2463–2478.

Narum, S.R. 2006. Beyond Bonferroni: Less conservative analyses for conservation genetics.

Conservation Genetics 7:783-787. Natural Resources Canada. Climatic Regions [map]. 1:10000000. Atlas of Canada, 3rd Edition

030. Canada: Natural Resources Canada, 1957. Natural Resources Canada. Ice [map]. 1:35000000. Atlas of Canada, 4th Edition 013_14. Canada:

Natural Resources Canada, 1971. Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York, NY,

U.S.A. Paetkau, D. 1999. Using genetics to identify intraspecific conservation units: a critique of current

methods. Conservation Biology. 13: 1507 – 1509. Patten, M.A. 2015. Subspecies and the philosophy of science. The Auk 132:481-485. Paxton, K.L., Yau, M., Moore, F.R., and D.E. Irwin. 2013. Differential migratory timing of

western populations of Wilson’s Warbler (Cardellina pusilla) revealed by mitochondrial DNA and stable isotopes. The Auk 130:689–698.

81

Pérez-Alvarez, M.J., Olavarría, C., Moraga, R., Baker, C.S., Hamner, R.M., E. Poulin. 2016. Historical dimensions of population structure in a continuously distributed marine species: The case of the endemic Chilean dolphin. Scientific Reports – Nature 6:1-11.

Perktas, U., Gür, H., and E. Ada. 2015. Historical demography of the Eurasian green

woodpecker: integrating phylogeography and ecological niche modelling to test glacial refugia hypothesis. Folia Zoologica 64:284-295.

Peterson, B.K., Weber, J.N., Kay, E.H., Fisher, H.S., and H.E. Hoekstra. 2012. Double Digest

RADseq: An inexpensive method for de novo SNP discovery and genotyping in model and non-model species. PLoS One 7: 1-11.

Peterson, B.K., Weber, J.N., Kay, E.H., Fisher, H.S., and H.E. Hoekstra. 2012. Double Digest

RADseq: An Inexpensive Method for De Novo SNP Discovery and Genotyping in Model and Non-Model Species. PLoS One 7: 1-11.

Pierson, B.J., Pearce, J.M., Talbot, S.L., Shields, G.F., And K.T. Scribner. 2000. Molecular

genetic status of Aleutian Canada Geese from Buldir and the Semidi Islands, Alaska. Condor 102:172-180.

Plummer, M., Best, N., Cowles, K., and K. Vines. 2006. CODA: Convergence Diagnosis and

Output Analysis for MCMC. R News 6:7-11. PopART. Allan Wilson Centre. (http://popart.otago.ac.nz) Pope, L.C., Domingo-Roura, X., Erven, K., and T. Burke. 2006. Isolation by distance and gene

flow in the Eurasian badger (Meles meles) at both a local and broad scale. Molecular Ecology 15:371-386.

Pritchard, J.K., Stephens, M., and P. Donnelly. 2000. Inference of population structure using

multilocus genotype data. Genetics 155:945–959. Pritchard, J.K., Wen, X., and D. Falush. 2010. Documentation for STRUCTURE software.

Version 2.3. Department of Human Genetics, University of Chicago, Chicago, Illinois. Pshenichnikova, O.S., Klenova, A.V., Sorokin, P.A., Zubakin, V.A., Konyukhov, N.B.,

Kharitonov, S.P., and Y.B. Artukhin. 2015. The Crested Auklet, Aethia cristatella (Alcidae, Charadriiformes), does not vary geographically in genetics, morphology or vocalizations. Marine Biology 162: 1329-1342.

Quinn, G.P., and M.J. Keough. 2002. Experimental Design and Data Analysis for Biologists.

Cambridge University Press, New York, U.S.A. R Core Team. 2017. R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria. (https://www.R-project.org/)

82

Rambaut, A. 2007. FigTree. Institute of Evolutionary Biology. (http://tree.bio.ed.ac.uk/software/figtree).

Rambaut, A., Suchard, M.A., Xie, D., and A.J. Drummond. 2014. Tracer v1.6.

(http://beast.bio.ed.ac.uk/Tracer) Rosenberg, N.A. 2004. DISTRUCT: a program for the graphical display of population structure.

Molecular Ecology Notes 4:137–138. Ruegg, K.C., Anderson, E.C., Paxton, K.L., Apkenas, V., Lao, S., Siegel, R.B, Desante, D.F.,

Moore, F., and T.B. Smith. 2014. Mapping migration in a songbird using high-resolution genetic markers. Molecular Ecology 23: 5726-5739.

Russell, D. J. and R. N. Garrett. 1988. Movements of juvenile barramundi, Lates calcarifer

(Bloch), in north-eastern Queensland. Australian Journal of Marine and Freshwater Research 39:117 – 123.

Sambrook J, Fritsch, E.F., and T. Maniatis. 1989. Molecular cloning: a laboratory manual 2nd ed.

Cold Spring Harbor Laboratory Press, Plainview, New York, U.S.A. Shealer, D.A. 2002. Foraging Behaviour and Food of Seabirds. Pages 137-170 in E.B. Schreiber,

and J. Burger, editors. Biology of Marine Birds. CRC Press LLC, Florida, U.S.A. Slatkin, M. 1987. Gene Flow in Natural Populations. Annual Review of Ecology and Systematics

16:393-430. Southern, H.N., Carrick R., and W.G. Potter. 1965. The natural history of a population of

guillemots Uria aalge. Journal of Animal Ecology 35:1–11. Storer, R.W. 1952. A comparison of variation, behaviour, and evolution in the seabird genera

Uria and Cepphus. University of California Publications in Zoology, 52, 121-222. Svanberg, I. and S. Ægisson. 2006. Black Guillemot (Cepphus grylle) in North European Folk

Ornithology. Scripta Islandica 56:101-114. Tajima, F. 1983. Evolutionary relationship of DNA sequences in finite populations. Genetics and

Breeding 105:437–460. Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA

polymorphism. Genetics 123:585-595. Tigano, A., Damus, M., Morris-Pocock, J.A., Artukhin, Y.B., and V.L. Friesen. 2015. The Arctic:

Glacial refugium or area of secondary contact? Inference from the population genetic structure of the thick-billed murre (Uria lomvia), with implications for management. Journal of Heredity 106: 238-246.

83

Tigano, A., Sackton, T.B., and V.L. Friesen. 2017a. Assembly and RNA-free annotation of highly heterozygous genomes: The case of the thick-billed murre (Uria lomvia). Molecular Ecology Resources doi: 10.1111/1755-0998.12712.

Tigano, A., Shultz, A.J., Edwards, S.V., Robertson, G.J., and V.L. Friesen. 2017b. Outlier

analyses to test for local adaptation to breeding grounds in a migratory arctic seabird. Ecology and Evolution 7:2370-2381.

Toews, D.P.L., Campagna, L., Taylor, S.A., Balakrishnan, C.N., Baldassarre, D.T., Deane-Coe,

P.E., Harvey, M.G., Hooper, D.M., Irwin, D.E., Judy, C.D., Mason, N.A., McCormack, J.E., McCracken, K.G., Oliveros, C.H., Safran, R.J., Scordato, E.S.C., Faust Stryjewski, K., Tigano, A., Uy, J.A.C., and B.M. Winger. 2016. Genomic approaches to understanding population divergence and speciation in birds. The Auk 133: 13-30.

Torstrom, S.M., Pangle, K.M., and B.J. Swanson. 2014. Shedding subspecies: The influence of

genetics on reptile subspecies taxonomy. Molecular Phylogenetics and Evolution 76:134:143.

Udvardy, M. D. F. 1963. Zoogeographical study of the Pacific Alcidae. Pages 85 – 111 in J.L.

Gressitt, editor. Bishop Museum, HI Pacific Basin Biogeography, Hawaii, U.S.A. van Oosterhout, C., Hutchinson, W. F., Wills, D. P., and P. Shipley. 2004. MICRO‐CHECKER:

software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes 4:535-538.

Vásquez-Carrillo, C., Friesen, V., Hall, L., and M.Z. Peery. 2014. Variation in MHC class II B

genes in marbled murrelets: implications for delineating conservation units. Animal Conservation 17: 244-255.

Walsh, H. E. and S. V. Edwards 2005. Conservation genetics and pacific fisheries bycatch:

mitochondrial differentiation and population assignment in the Black-footed Albatrosses (Phoebastria nigripes). Conservation Genetics 6:289–295.

Ward-Paige, C.A. and A. Bundy. 2016. Mapping Biodiversity on the Scotian Shelf and in the Bay

of Fundy. Fisheries and Oceans Canada 1-89. Watterson, G. 1978. The homozygosity test of neutrality. Genetics 88:405–417 Weir, B.S., and C.C. Cockerham. 1984. Estimating F-Statistics for the Analysis of Population

Structure. Evolution 38: 1358-1370 . Wenink, P.W., Baker, A.J., Rösner, H., and M.G.J. Tilanus. 1996. Global mitochondrial DNA

phylogeography of holarctic breeding dunlins. Evolution 50:318-330. Wojczulanis-Jakubas, K., Kilikowska, A., Harding, A.M.A., Jakubas, D., Karnovsky, N.J., Steen,

H., Strøm, H., Welcker, J., Gavrilo, M., Lifjeld, J.T., and A. Johnsen. 2014. Weak

84

population genetic differentiation in the most numerous Arctic seabird, the little auk. Polar Biology 37:621-630.

Wright, S. 1969. Evolution and the Genetics of Populations, Vol. 2, The Theory of Gene

Frequencies. University of Chicago Press, Chicago. Zigouris, J., Shaefer, J.A., Fortin, C., and C.J. Kyle. 2013. Phylogeography and post-glacial

recolonization in wolverines (Gulo gulo) from across their circumpolar distribution. PLoS ONE 8: e83837.

85

Appendix A

Supplementary Material for Chapter 2

Supplementary Table 2.1. PCR primers and annealing temperatures for nuclear loci screened in

pigeon guillemots.

Locus Primer Sequence Annealing (forward/reverse) Temperature (̊C) Cytochrome c GAAAAAGGAGGCAAGCACAAGACTGG/ 53 Intron I1 TGTTCCTGGGGATGTACTTCTTTGGATT

Ribosomal protein 40 GGGCCTGATGTGGTGGATGCTGGC/ 63 Intron V2 GCTTTCTCAGCAGCAGCCTGCTC Cco5-91 TTCCTACCAGTAAAAGAGAGGA/ 55

GTACCCCTTTCCTAATTCAAG Cco5-211 TCAAGATGATGAAGACCCTAAT/ 55 AGAGTTGCACAGGTTAAATACC Dpu162 ACAGCAAGGTCAGAATTAAA/ 61 AACTGTTGTGTCTGAGCCT Uaa5-83 CAGTTTCTTTAAGTCGTGCCAG/ 60 CACTTAGGTCCAAAACCTAACC 1Friesen et al., unpubl. data

2Dawson et al. 1997 3Ibarguchi et al. 2000

86

Supplementary Table 2.2. Variable sites within a 723 bp fragment of the mitochondrial control region among 202 pigeon guillemots.

Roman numerals indicate Domain. Arabic numerals indicate location relative to the 5’ end of the control region. ? indicates an

uncertain base. Hyphens represent deletions.

________________________________________________________________________________________

I II III 11111111111111111111111111112222222222222222222233333333333 34456 666667777788 99 44455555566666677788888899990001113334555566678800222445555 92501 228880388934 45 57903578902346702923457814561574674561367923500119689043456 87542 072393034098 07 ________________________________________________________________________________________ AA01 TCCTCTAGCCATTCTTCACCTTTCAAAGGTATTGCCCAAGCTCCGGGGGGGGCTATTTA AACAT--TCCGTAGTCGTT TT AB01 ..............C....T.............A.......C................. ................... .. AC01 ..............C....T.............A.......C......A.......... ................... .. AD01 ..............C......................G..................... ................... .. AD06 ..............C......................G..................... ................... CC AE01 ..............C............................................ ................... .. AF01 ..A.?...?.....C......................G..................... ................... .. AG01 ..............C.................................A.......... ................... .. AH01 ..............................G............................ ................... .. AI01 ...?T...T?..C.C....T.............A.......C................. ................... .. AJ01 C.AAT..?T.??C.C?A..T.............A.......C................. ................... .. AK01 C.AAT..?T..C?.C?A..T.............A.......C......AA......... ................... .. AL01 ............C.C....TC............A.......C......A.......... ................... .. AL12 ............C.C....TC............A.......C......A.......... ................A.. .. AM01 ............C.C..G...............A.......CT.A...A.......... ................... .. AN01 ............C.C....TC............A.......C................. ................... .. AO01 ............C......T..CT.........A.............AA.......... ................... .. AO05 ............C......T..CT.........A.............AA.......... ...........C....... .. AP01 ....T.........C....T.............A.......C......A.......... ................... ..

87

AQ01 ...?T..T....?.C......................G..................... ................... .. AR01 ........?...C?.....T.............A.......C....A.A.......... ................... .. AS02 ............C.C....T.............A.......C................. .........T......... .. AT01 ............C.C..................A.......CT.....A.......... ................... .. AU01 ........?...C?C?A..T.............A.......C.?............... ................... .. AV01 ............C.C....T.............A.......C................. ................... .. AW01 ............C.C....T.............A.......C..........T...... ................... .. AX03 ....T.........C.................C................A......... ..........A........ .C AY01 ....?.GA....C.C....TC............A.......C......A.......... ................... .. AZ02 ............CTC....T.............A.......CT.A...A.......... ...T.....T......... .. BA01 ............C.C....T.C...........A.......C................. ................... .. BB01 .............TC....TA....G.......A.......CT.A...AA......... ................... .. BC01 ............C.C...TT.............A.......C......A.......... ................... .. BD01 ....?...?...C.C...TT.............A.......C......AA......... G.................. .. BE01 ....?.......C.C....T.............A.......C......A...T...... ................... .. BF01 ....T........TC....TA....G.......A.......CT.A...A.......... ................... .. BG01 ....?.......CTC....T.....G.......A.......CT.A...A.......... ................... .. BH01 .............TC....TA....G.......A.......CT.A...A.......... ................... .. BI01 ....?.......CTC....T....GG.......A.......CT.A...A.......... G.................. .. BJ01 ....?........TC....TAC...G.......A.......CT.A...A...T...... ................... .. BK01 ....T.......C.C...TT.............A.......C......A.......... G.................. .. BL01 ............C.C....T.............A.......C....A.A.......... ................... .. BM04 ..AAT..?T..C?TC.A..T.....G.......A.......CT.A...A.......... ..................C .. BN01 ............C.C..................A.......C......A....?..... ................... .. BO02 ............CTC.?..T.............A........T.A...A.......... ...T.....T......... .. BP01 ............CTC....T.....G.......A....G..CT.A.............. ................... .. BQ01 ..?AT..TTT?CCTC?A.TT.............A.......C......AA......... G.................. .. BR01 ............C.C....TC...........................A.......... ................... .. BS01 ....T..T...CC?C....T.............A.......C......A.......... ................... .. BT01 ....T..?....CTC....TA....G.......A.......CT.A...A....?..... ................... .. BU01 ..T....A....C......T.............A.......C......A.......... ................... .. BV01 ............C......T..CT.........A.......C......A.......... ................... .. BX01 .............TC....TA....G.......A.......CT.A...A....C..... ................... .. BZ02 ....T...?...CTC....T.............A.......CT.A...A.......... ...T.....T......... ..

88

CB01 ....T..?...CCTC.A.TTA....G.......A.......CT.A...A.......... ................... .. CC01 ............C.C....T............................AA....G.... ................... .. CD08 ...................T...T.G.AAC...A.......CT.....A.......... ....-...T.A........ .? CD11 ...................T...T.G.AAC...A.......CT.....A.......... ..........A..A.-.C. .. CE09 ...................T...T.G.A.C...A.......CT.....A.......... ....-...T.A..AC.... .? CH01 ............C.C....T.............A..............A.....G.... ...?............... .. CJ01 .....C......C.C....T.............A.......C......A.......... ...?............... .. CK01 ..........?.C.C....T............................A.....G.... ...?............... .. CL01 ............C.C....T......G.......G..........T..A.....G.... ...?............... .. CN08 ...................T...T.G.AAC...A..T....CTT.T..AACC..GACCT ...?-...T.A........ .? CP08 ....................T..CT.G.AAC...A.......CTT....AA......... ...?-...T.A........ .? CR01 .............C.C....T..CT.........A.......C................. ...?............... .. CS01 ...........?.CTC...?..............A.......CT.A...AA......... G..?............... .. CT01 .............C.C....T...........C.A..T....C......A.......... ...?............... .. CU15 ....................T...T.G.AAC...A.......CTT....AA......... ...?......A........ .. CV01 .............C.C...CT..CT...............CG....T............. ...?............... .. CW01 ......?....C.C?C?A..T.............A.......C...T..A.......... ...?............... .. DB01 .............C.C..G...............A.......CT.....AA......... ....?????.......... .. DC01 .............C.C..................A.......CT.....AA......... ....?????.......... .. DD01 .............C.C.....C............A.......C......AA......... .......?........... .. DD05 .............C.C.....C............A.......C......AA......... ...........C....... .. DE05 .............C.C....T.............A..............A.......... ...........C....... .. DF01 .............C.C..G...............A.......C......A.......... ................... .. DG17 .....????????????A..T...T.G.AAC...A.T....?CT.....A.......... .....AT.T.A.G...... .C DH18 .....?????????????..TC..T.G.AAC...A......?C.?....A.......... ..........A........ .. DG19 .....???????????AA..T...T.G.AAC...A.......CT.....A.......... .....AT.T.A..A..... ?? DI20 .....??TT????????A..T...T.G.AAC...T.......CTT....A.......... ........T.A........ ?? DK18 ....................T..?T.G.?AC...A.......CTT............... ..........A........ .. DL21 .T.....?T...?.?.....T..?T.G.?AC...A.......C......A.......... .C...AT.T.A........ .C DM18 ........T...........T...T.G.AAC...A.......CTT....AA......... ..........A........ .. DN18 ....................T...T.G.?AC...A.......CTT....AA......... ..G.......A........ .. DO22 ...........?..TC....TA....G.......A.......CT.A...A.......... .......G........... ??

89

Supplementary Table 2.3. Frequencies of control region haplotypes within regional samples of

pigeon guillemots. Region abbreviations as in Table 2.1.

Haplotype

Region

Andr Fox Shum Semi Cook PWS SEAK Clel Mand cOre cCal

AA01

6 3 AB01

2 3 2 3

AC01

1 1

1 1 1 AD01

3 8

AD06

1 AE01

2 2

AF01

7 AG01

7

AH01

2 AI01

1

AJ01

1 AK01

1

AL01

3 12 2 AL12

1

AM01

2 2 AN01

4

AO01

1 AO05

1

AP01

1 AQ01

1

AR01

1 AS02

4

AT01

1

1 AU01

2

AV01

1

2 AW01

1

AX03

1 AY01

1

AZ02

1 1 BA01

2

BB01

1 BC01

1 1 1 3

BD01

1 3 1 BE01

2 2

90

BF01

4 2 BG01

3

BH01

1

1

3 BI01

1

BJ01

1 BK01

2

BL01

2 BM04

1

BN01

1

1 BO02

1

BP01

1 BQ01

4

BR01

3 BS01

1

BT01

2 BU01

2

BV01

1 BX01

1

BZ02

1 CB01

1

CC01

1 CD08

1 1

CD11

3 CE09

1

CH01

1 CJ01

1 1

CK01

1 CL01

1

CN08

1 CP08

1

CR01

1 CS01

1

CT01

1 CU15 2

CV01

1 CW01

1

DB01

1 DC01

1

DD01

1 DD05

1

DE05

2 DF01

1

DG17 1

91

DG19 1 DH18 1 DI20 1 DK18 1 DL21 1 DM18 1 DN18 1 DO22

1

Total 10 8 7 7 34 30 9 7 29 24 37

92

Supplementary Table 2.4. Numbers of samples and regional allele frequencies (absolute

numbers) for two introns and four microsatellites for pigeon guillemots. Region abbreviations as

in Table 2.1.

Andr Fox Shum Semi Cook PWS SEAK Clel Mand cOre cCal Cytochrome c Intron I 1 6 4 7 12 36 32

9 32 30 41

2 7 8 7 6 19 20 11 6 16 14 7 3

1

1 7

5 1 10

4 2

2

2 5 3 1

1 1 6

2

6

1 N 18 14 14 20 66 58 18 16 58 46 48

Ribosomal Protein 40 Intron V 1 9 12 13 16 48 47 14 11 45 34 40 2 7 2 1 3 18 11 1 4 12 12 6 3 1

1

2

1

4 1 5

1

6

1 N 18 14 14 20 66 58 18 16 58 46 46

Cco5-9 112 2 2 1 6 7 10 6 25 3 120 1 1 122 3 4 4 5 12 14 9 7 20 32 55 124 1 1 2 2 11 9 3 1 1 126 10 4 4 3 15 13 4 1 4 5 128 1 2 1 1 130 1 1 4 11 8 134 1 1 1 1 136 1 1 1 4 1 8 1 138 1 4 140 2 2 N 18 14 13 20 66 58 18 17 58 46 55

93

Cco5-21 126 2 2 2 1 3 1 128 2 130 1 132 3 1 4 6 16 10 11 12 42 22 25 136 1 3 3 2 3 10 26 138 3 3 1 4 3 1 1 7 140 4 2 3 2 18 9 2 1 9 2 6 142 5 5 5 7 20 22 2 2 1 144 1 1 2 2 1 1 1 146 2 3 6 1 148 1 1 N 18 14 14 20 66 58 17 18 58 46 58

Dpu16 143 2 1 1 6 2 147 18 14 14 18 65 58 18 15 58 40 53 149 1 N 18 14 14 20 66 58 18 16 58 46 58

Uaa5-8 110 8 9 3 10 39 34 12 12 23 20 22 112 5 5 8 9 12 15 5 5 31 18 25 114 1 1 116 1 118 2 1 1 12 6 3 5 10 120 1 2 1 3 1 1 1 1 N 18 14 14 20 64 58 18 18 58 44 58

94

Supplementary Table 2.5. Table of significant linkage disequilibrium for nuclear loci

(significance level = 0.05). Significance is indicated by a +.

Andr Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC

Uaa5-8 Dpu16 -

Cco5-9 - - Cco-5-21 + - +

RP40 - - - - CYTC + - - - -

Cook

Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC Uaa5-8

Dpu16 - Cco5-9 - -

Cco-5-21 - - + RP40 - - - -

CYTC - - - - -

PWS Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC

Uaa5-8 Dpu16 -

Cco5-9 - - Cco-5-21 - - -

RP40 + - - - CYTC - - - - -

Clel

Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC Uaa5-8

Dpu16 - Cco5-9 - -

Cco-5-21 - - - RP40 + + - -

CYTC - + - - +

95

Mand Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC

Uaa5-8 Dpu16 -

Cco5-9 + - Cco-5-21 - - -

RP40 - - - - CYTC - - - - -

cCal

Locus Uaa5-8 Dpu16 Cco5-9 Cco5-21 RP40 CYTC Uaa5-8

Dpu16 - Cco5-9 - -

Cco-5-21 - - - RP40 - + - -

CYTC - + - - +

96

Supplementary Table 2.6. Sequence variation among alleles for two nuclear introns for pigeon

guillemots. Dots indicate identity with the first sequence. Numbers refer to nucleotide positions

in Supplementary Figure 2.1.

(a) Cytochrome c I (b) Ribosomal Protein 40 V ______________ _________________________ Allele 11 Allele 11111222222233 6915 111569033368923 1329 145414257955066 ______________ _________________________ 1 GCGC 1 TCAGCTATGCGCACC 2 .T.. 2 ...A........... 3 ...A 3 ..............T 4 .TA. 4 ...........T... 5 .T.T 5 .....C......... 6 AT.. 6 CTG.T.GGCTTTTT. ______________ _________________________

97

Supplementary Table 2.7. Estimates of FST for pairwise comparisons of regional samples of pigeon guillemots based on microsatellites

(below diagonal) and introns (above diagonal). Region abbreviations as in Table 2.1.

Andr Fox Shum Semi Cook PWS SEAK Clel Mand cOre cCal Andr

0.04 0.10 0.06 0.04 0.06* 0.13* -0.01 0.07* 0.06* 0.26** Fox -0.01

-0.03 0.04 0.05 0.03 0.03 0.03 0.05 0.08* 0.29**

Shum 0.01 0.03

-0.01 0.03 -0.01 0.14* 0.02 0.02 0.03 0.17* Semi 0.02 0.00 0.00

-0.02 -0.03 0.19* -0.03 -0.02 -0.02 0.04

Cook 0.01 0.01 0.03* 0.02

0.00 0.18** -0.03 -0.01 -0.01 0.08** PWS 0.01 -0.01 0.02 0.00 0.00

0.20** -0.01 0.01 -0.01 0.08*

SEAK 0.10* 0.09* 0.09* 0.07* 0.07* 0.07**

0.16* 0.17** 0.25** 0.46** Clel 0.12** 0.09* 0.10* 0.03 0.07** 0.07** 0.00

-0.02 -0.03 0.10*

Mand 0.17** 0.16** 0.11** 0.08** 0.14** 0.15** 0.09** 0.02

0.01 0.07* cOre 0.13** 0.11** 0.10** 0.08** 0.12** 0.12** 0.05* 0.05* 0.10**

0.05*

cCal 0.28** 0.26** 0.23** 0.20** 0.23** 0.23** 0.17** 0.20** 0.21** 0.04*

*Significantly different from 0 at α = 0.05 before B-Y correction **Significantly different from 0 at α = 0.05 after B-Y correction

98

Supplementary Table 2.8. Type I and Type II error rates and posterior probabilities for each

scenario calculated from DIYABC.

True Scenario Type I error rate

Type II error rate

Posterior probability

Scenario 1 0.02 0.026 0.8285 [0.8123,0.8446]

Scenario 2 0.006 0.022 0.0342

[0.0000,0.1101]

Scenario 3 0.04 0.018 0.1373 [0.0562,0.2184]

99

Supplementary Figure 2.1. Demographic scenarios tested using DIYABC. All three scenarios

assume that there are three populations groups (North: Andreanof Is. and Fox Is., Alaska:

Shumigan Is., Eastern Alaska Peninsula, Cook Inlet, Prince William Sound, and Southeast

Alaska, and South: Vancouver Is., Central Oregon, and Central California) at the present time

(t0) and these populations diverged in the past. The time scale is shown on the right.

100

Supplementary Figure 2.2. Sequences of the most common alleles for three introns for pigeon guillemots. Variable sites are in bold.

Asterisks represent the beginnings and ends of introns. Positions are relative to the 3’ end of the 5’ primer.

Cytochrome c Intron I Position ACCCAACCTG AATGGCCTCT TTGGACGCAA GACTGGACAG GCTGAGGGCT TCTCTTATAC 60 GGATGCTAAT AAGAATAAAG GTAAACTTCA AGCTGTTCTT ATGAGTTACT AGGGACAAAA 120 * ATACTCAGTT CTTTTCAGTA CTTTCCAGA- ACTACTGCCG TCTGCCCAAA AGTAAAATAA 180 TATGGAGGAA AAGAATATAA ATATGTCTGT TATTTTAA-A GTCACTGACC ATCTC--TTT 240 TCT-TCTTTC TAGGTATCAC CTGGGGTGAG GATACT 276 * Ribosomal Protein 40 Intron V Position TCGGGAGGTC TTGCGCATGC GTGGCACCAT CTCCCGTGAG CACCCATGGG AAGTCATGCC 60 TGACTTGTAC TTCTACAGGG ATCCCGAGGA GGTAGGAGAC CTCAGGAAGG AAGCATGGCT 120 * GTGTGAGATG TCCTCCTTGT GCTCCTGCCC CATGGTCCTG CGTGCCTGCT GTTGAGGACA 180 GGTTTTGCCA ATATAATGAT TACACTTCAC AGTTAGGGTG GTGGTTAATT AGTCTTGTC- 240 ---GTTGGTT GGTGCGTGTT GTCTGGTATG TGATAGTGTG AGGGCGGTGA GGTGTTTCTG 300 GTGGGAGACT TAGGCTACTC TCTTCCGTTG CAGATCGAAA AGGAG 345

101

Appendix B

Supplementary Material for Chapter 3

Supplementary Table 3.1. Mean estimated Ln probabilities and delta K (ΔK) calculated for values

of K from 1-10 using data from STRUCTURE in STRUCTURE HARVESTER. Analyses were

run with and without location as prior information.

No Location

Location K Mean LnP(K) ΔK K Mean LnP(K) ΔK

1 -150410.33 NA 1 -150405.74 NA 2 -148239.76 40.293001 2 -148201.28 76.70808 3 -148310.08 3.808468 3 -148095.81 7.029713 4 -148994.31 10.914477 4 -148687.88 5.054467 5 -151722.08 3.681842 5 -150057.23 0.631452 6 -164720.98 0.235461 6 -152030.55 3.956399 7 -170162.33 0.243076 7 -168302.57 1.144388 8 -170224.89 0.000197 8 -161488.37 0.064289 9 -170290.17 0.221333 9 -155450.53 3.02419 10 -166491.88 NA 10 -175851.48 NA

102

Supplementary Table 3.2. Outlier loci statistics output from BAYESCAN. Statistics for each

locus include the posterior probability for the model including selection (Prob), the logarithm of

Posterior Odds to base 10 (log10(PO)) for the model with selection, the q-value for the model

with selection (q-value), the estimated alpha coefficient indicating the strength and direction of

selection (alpha), the FST coefficient averaged over populations (FST), and the allele frequencies

of each allele at that locus (Allele Freq).

SNP

Prob

log10(PO)

q-value

alpha

FST

Allele Freq

3478

0.992

2.105

0.004

1.765

0.204

C: 0.77 A: 0.23

6283

1.000

3.398

0.000

1.997

0.238

T: 0.52 C: 0.48

15077

0.934

1.154

0.025

1.633

0.190

A: 0.64 G: 0.36

13821

0.954

1.312

0.047

1.578

0.190

G: 0.83 A: 0.17