Population Genetic Differentiation and Hybridization in The

Home , European herring gull

POPULATION GENETIC DIFFERENTIATION AND HYBRIDIZATION IN THE

GLAUCOUS GULL (LARUS HYPERBOREUS)

Emma Lachance Linklater

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

January 2021

ABSTRACT

Climate change poses a significant threat to the future of Arctic ecosystems. To effectively conserve Arctic species, genetically differentiated populations must be defined for adaptive and neutral genetic variation to be appropriately managed. This project examines population genetic differentiation in the Glaucous Gull (Larus hyperboreus) – a circumpolar

Arctic species – and assesses hybridization between Glaucous Gulls and three closely-related species. The Glaucous Gull is a valuable species, biologically and culturally. As apex predators,

Glaucous Gulls develop high levels of toxins in their tissue and are, therefore, excellent bioindicators of the long-range transport of contaminants in the Arctic. Glaucous Gulls also hybridize with other white-headed gull species where breeding ranges overlap. Although the

International Union for Conservation of Nature currently lists Glaucous Gull as Least Concern, declines have been reported across their range in Arctic Canada. Currently no fine•scale population genetic information exists for this species, and management units have not been delineated. Double digest restriction-site associated DNA sequencing was conducted on 62

Glaucous Gulls, 18 American Herring Gulls (L. smithsonianus), 6 European Herring Gulls (L. argentatus), and 15 Glaucous-winged Gulls (L. glaucescens) sampled across the North American and European Arctic. Despite the geographic distance between sampling locations, molecular assignments and principal coordinates analysis suggest only weak population differentiation between sampled European and Canadian colonies of Glaucous Gull. Interspecific analyses using

2145 loci show that Glaucous Gull and Glaucous-winged Gull are genetically distinct species but that Glaucous Gull and the two species of Herring Gull are only weakly differentiated. Several sampled individuals may represent hybrids between Glaucous Gulls and American Herring

Gulls. Detailed information on population genetic structure and hybridization will help

iii conservation practitioners effectively manage the long-term persistence of Glaucous Gull populations. Proactive management strategies for this species will benefit both Glaucous Gull and the entire Arctic ecosystem.

ACKNOWLEDGEMENTS

I’d like to start by acknowledging that much of the sampling for this project was done in the

Inuit Nunangat (ᐃᓄᐃᑦ ᓄᓇᖓᑦ, the traditional homeland of the Inuit), particularly in Nunavut

(ᓄᓇᕗᑦ), and Nunavik (ᓄᓇᕕᒃ). The contemporary sampling for this project would not have been possible without the support and collaboration of Inuit communities. I’d also like to acknowledge that due to the advanced age of some of the samples used in this project, many samples were taken by and for the Government of Canada at a time when forced assimilation of indigenous people was the national policy. These older samples were taken from indigenous lands, rather than given. Finally, I am a settler scientist and much of the work on this thesis was done at Queen’s University which sits on the traditional lands of the Haudenosaunee and

Anishinaabe.

I would also like to acknowledge the environmental impact of this research. All my samples were extracted using a salt extraction protocol to reduce the amount of harmful chemical waste produced, however, chloroform was still required to remove fat from many of the samples. My project required plane and/or boat travel to access remote regions for sampling, shipments of samples over long distances, storage of samples in ultra-cold freezers, use of a high-performance computing cluster, and thousands of single-use plastics in the lab. I estimated that DNA extraction, purification, and quality assessment required 2-3 plastic tubes, and 20+ plastic pipette tips per sample under our current extraction protocol. Our lab has collaborated with the

Sustainability Action committee to work towards the diversion of recyclable plastics from landfills and will be implementing new protocols this year.

Thank you to my supervisor Vicki Friesen. Over the last three years, you have thought of me and reached out over and over to provide opportunities to do research both in the lab and in the field. Thank you for your guidance, support, and empathy through my Honours and Masters theses. Thank you for inspiring a love of Arctic seabirds.

A huge thanks to my incredible lab members Katie Birchard, Drew Sauve, Russell Turner,

Brody Crosby, Nadege Allan, Ryan Franckowiak, Bronwyn Harkness, Becky Taylor, Chris

Boccia, Alisa Samuelson, Heather Vanderlip, and Maleeka Thaker. Thank you for your companionship, encouragement, and assistance! In particular, I need to thank Lila Colston-

Nepali who has provided an unbelievable amount of advice, support, friendship, and guidance through my Masters. Even when you were on the other side of the world (literally 10 hours and

45 minutes ahead), you still made time for me and I’m so thankful for that! Thank you to my friends in the department and to the wonderful BioBabes. You are such an incredible group of smart, talented, hilarious, and amazing women. I am endlessly inspired by each of you!

Thank you to Tim Birt for your help with all my lab-related questions and to Evelyn Jensen,

Peiwen Li, Ying Chen, Lila Colston-Nepali, and the members of the Bioinformatics and R (BnR)

Club for providing bioinformatics help. Thank you to my committee members Dr. Jannice

Friedman and Prof. John Smol and to Strategic Grant collaborators Dr. Grant Gilchrist, Dr. Greg

Robertson, Dr. Jenn Provencher, and Dr. Tony Gaston from Environment and Climate Change

Canada (ECCC); Prof. Marie-Josée Fortin from the University of Toronto; Dr. Kyle Elliott from

McGill University; and Dr. George Divoky from Friends of Cooper Island. Thank you to my co- authors Dr. Sarah Sonsthagen, Dr. Freydís Vigfúsdóttir, and Dr. Greg Robertson.

I’d like to acknowledge the work of all the sample collectors, this project would not have been possible without their contributions! For providing and helping to coordinate samples for

vi this project, I’d like to thank Sarah Sonsthagen and the United States Geological Survey

(USGS); Oliver Haddrath and the Royal Ontario Museum (ROM); Sharon Birks and the Burke

Museum; Þorvaldur Björnsson, Freydís Vigfúsdóttir, and the Icelandic Institute of Natural

History; and Lana Dolgova, Guy Savard, Kelsey Ha, Bronwyn Harkness and the National

Wildlife Research Centre (NWRC). Thank you to Summer Work Experience Program (SWEP) students Alex Tsiofas and Arjun Augustine who helped with DNA extractions, and quality assessment of samples in the summer of 2019. Thank you to Zhengxin Sun for your expertise preparing my library for sequencing, and for your kindness, guidance, and incredible foresight regarding the pandemic. I would not be finishing my Masters right now if you hadn’t insisted on shipping in February. Thank you to Karen Ho and the Centre for Applied Genomics (TCAG) in

Toronto for sequencing and to Joanne Surette for all your hard work behind the scenes at

Queen’s.

Thank you to the Natural Sciences and Engineering Research Council (NSERC Strategic

Projects Grant 493789-16), the Northern Scientific Training Program (NSTP), the School of

Graduate Studies, and the Queen’s University Biology Department for funding.

I couldn’t have done this without my incredibly supportive family and friends. Thank you to my parents for inspiring a love of nature, and always encouraging my curiosity; and to my grandparents, Gerry and Grandpa, who saved the songbirds that hit the window so we could look at them when we came to visit. Finally, thank you to my partner, Aidan. You have been an absolute rock through all the highs and lows of my thesis. Thank you for listening to me talk about gulls for two straight years, thank you for always believing in me and encouraging me, thank you for drying my tears. I love you so much.

vii

TABLE OF CONTENTS

Abstract………………………………………………………………………………….. ii

Acknowledgements………………………………………………………………. iv

Table of Contents………………………………………………………………… vii

List of Tables…………………………………………………………………….. x

List of Figures……………………………………………………………………. xi

List of Abbreviations…………………………………………………………….. xiii

A Note on Systematics…..………………………………………………………. xv

CHAPTER 1: Introduction……………………………...………………………... 1

Conservation of genetic variation………………………………………... 3

Hybridization……………………………………………………………... 4

Glaucous Gull…………………………………………………………….. 6

Ecology…………………………………………………………… 6

Population differentiation………………………………………... 8

Hybridization…………………………………………………….. 9

Conservation concerns…………………………………………… 11

Methodology……………………………………………………………... 13

Research Aims and Predictions…………………………………………... 13

Significance………………………………………………………………. 15

CHAPTER 2: Methods……..……………………………………………………. 16

Sample collection and DNA extraction…………………………………... 16

Library preparation and sequencing……………………………………… 16

Identifying loci and filtering data………………………………………… 17

viii

Hybridization……………………………………………………………... 19

Differentiation between the four species………………...……….. 19

Migration rate……………………………………………………. 22

Population differentiation within the Glaucous Gull..…………………… 22

Population genetic structure…………...………………………… 22

Genetic diversity…………………………………………………. 23

Mantel test………………………………………………………... 24

Loci potentially under selection………………………………….. 24

Migration rate………………………...………………………….. 25

CHAPTER 3: Results………...…………………………………………………... 29

Hybridization……………………………………………………………... 29

Differentiation between the four species…………………………. 29

Migration rate…………………...……………………………….. 31

Population differentiation within the Glaucous Gull..…………………… 31

Population genetic structure……………...……………………… 31

Loci potentially under selection…………………………………. 32

Genetic diversity……………………..…………………………... 32

Migration rate………………………………...………………….. 32

Mantel test…………………...……..…………………………….. 33

CHAPTER 4: Discussion………...………………………………………………. 58

Hybridization…………………………………………………………….. 58

Differentiation between four gull species………………………... 58

Glaucous Gull hybrids……………...……………………………. 60

Other hybrids………………...…………………………………... 61

Population differentiation within the Glaucous Gull……..……………… 63

Historical association and contemporary gene flow………...…... 63

Selection pressures……………………………………………….. 65

Subspecies designations………………………………………….. 66

Conservation Implications……………………………………………….. 68

Hybridization……………………………………………………... 68

Population genetic structure………………………...…………… 70

Future Directions…………………………………………………………. 71

Conclusions………………………………………………………………. 74

Literature Cited…………………………………………………………………... 75

LIST OF TABLES

Table 2.1. Sampling location, number of samples, source of samples, and sampling year…………………………………………………….. 26 Table 3.1. Most likely K selected by three methods for three and four species of gulls………………………………………………………... 50 Table 3.2. Pairwise FST and probability of significance for four species of gulls………………………………………………………... 51 Table 3.3. Migration rate, residency rate, and standard deviation for four species of gulls…………………………………………………... 52 Table 3.4. Most likely K selected by three methods for Glaucous Gulls…………………………………………………………………… 53 Table 3.5. Pairwise FST and probability of significance for five colonies of Glaucous Gulls……………………………………………. 54 Table 3.6. Rarefied allelic richness and expected heterozygosity for six colonies of Glaucous Gulls………………………………………... 55 Table 3.7. Frequency of private alleles from three colonies of Glaucous Gulls………………………………………………………... 56 Table 3.8. Migration rate, residency rate, and standard deviation for two regions of Glaucous Gulls………………………………………... 57

LIST OF FIGURES

Figure 2.1. Map of the sampling locations, number of samples, and breeding range for each species………………………………………………………… 27 Figure 2.2. Map of the sampling locations, number of samples, and breeding range of Glaucous Gull after hybrids were removed…………………………. 28 Figure 3.1. Probability of assignment to K genetic clusters from STRUCTURE for all four species; a) admixture model, b) admixture and priors model…………………………………………………………………... 34 Figure 3.2. Probability of assignment to K genetic clusters from STRUCTURE for all four species using the priors model…………………… 35 Figure 3.3. Probability of assignment to K genetic clusters from STRUCTURE for three species; a) admixture model, b) admixture and priors model, and c) priors model…………………………………………………… 36 Figure 3.4. Principal component analysis of all four species………………... 37 Figure 3.5. Principal coordinates analysis of all four species………………... 38 Figure 3.6. Principal component analysis (a) and principal coordinates analysis (b) of Glaucous Gulls and American Herring Gulls………………… 39 Figure 3.7. Principal component analysis (a) and principal coordinates analysis (b) of Glaucous Gulls and European Herring Gulls…………………. 40 Figure 3.8. Principal component analysis (a) and principal coordinates analysis (b) of Glaucous Gulls and Glaucous-winged Gulls…………………. 41 Figure 3.9. Probability of assignment to hybrid categories with NEWHYBRIDS for Glaucous Gulls and Glaucous-winged Gulls…………… 42 Figure 3.10. Probability of assignment to K genetic clusters from STRUCTURE for Glaucous Gulls using priors model……………………….. 43 Figure 3.11. Probability of assignment to K genetic clusters from STRUCTURE for Glaucous Gull using admixture and priors model………... 44 Figure 3.12. Probability of assignment to K genetic clusters from STRUCTURE for Glaucous Gulls using admixture model…………………... 45

xii

Figure 3.13. Principal component analysis for Glaucous Gulls without hybrids………………………………………………………………………… 46 Figure 3.14. Principal coordinates analysis for Glaucous Gulls without hybrids………………………………………………………………………… 47 Figure 3.15. Histogram of frequency of private alleles in three colonies of Glaucous Gulls………………………………………………………………... 48

Figure 3.16. Slatkin’s linearized FST vs. log-transformed marine distance for Glaucous Gull………………………………………………………………… 49

xiii

LIST OF ABBREVIATIONS

AFLP Amplified fragment length polymorphisms AUs Adaptive units bp Base pairs °C Degrees Celsius DBRs Degenerate base regions ddRADseq Double digest restriction site-associated DNA sequencing DNA Deoxyribonucleic acid ECCC Environment and Climate Change Canada ESUs Evolutionarily significant units

FST Wright’s fixation index HCl Hydrochloric acid IBD Isolation by distance IINH Icelandic Institute of Natural History IOC International Ornithologist’s Committee K Number of genetic clusters km Kilometres MCMC Markov Chain Monte Carlo µl Microlitre mg Milligram ml Millilitre mtDNA Mitochondrial DNA MUs Management units NaCl Sodium chloride NWRC National Wildlife Research Centre p Statistical probability value PCA Principal component analysis PCoA Principal coordinates analysis PCR Polymerase chain reaction r Correlation coefficient

xiv

RAD Restriction site-associated DNA RADseq Restriction site-associated DNA sequencing ROM Royal Ontario Museum SNPs Single nucleotide polymorphisms SPRI Solid phase reversible immobilization SD Standard deviation TRIS Tris(hydroxymethyl)aminomethane USGS United States Geological Survey

A Note on Systematics

The phylogeny of the large white-headed gulls (order Charadriiformes; family Laridae) is controversial and there is no consensus between the International Ornithologist’s Committee

(IOC) World Bird List and the American Ornithological Society. Throughout this thesis, species naming conventions will follow the suggestion of the IOC World Bird List v. 10.2 (Gill et al.

2020). I will be referring to the American Herring Gull (Larus smithsonianus) and European

Herring Gull (L. argentatus) as valid and distinct species rather than treating the American and

European Herring Gulls as subspecies of the Herring Gull (L. a. smithsonianus and L. argentatus, respectively). Singular names (e.g. Glaucous Gull) will be used to refer to the taxa, and plural names (e.g. Glaucous Gulls) will be used to refer to individual organisms.

CHAPTER 1: INTRODUCTION

Arctic species are facing a variety of threats due to anthropogenic climate change, and the conservation of Arctic species will require proactive and informed management strategies in response to these threats. Anthropogenic climate change is having and will continue to have severe impacts on the Arctic. The annual mean extent of Arctic sea ice has decreased by 3-4% per decade since 1979 (IPCC 2014) and current climate projections suggest that the most northern Arctic communities may experience the greatest declines in shore-fast sea ice (Cooley et al. 2020). The Arctic is expected to experience greater increases in average sea surface temperature and average precipitation than the global mean (IPCC 2014). Rising average temperatures and the reduction of sea ice in Arctic environments have diverse and wide-reaching impacts for the persistence and health of Arctic organisms and ecosystems. A study on the impact of climate change on Arctic and subarctic mammals found that indirect impacts of climate change such as restrictions to dispersal ability and changes to community composition may be just as threatening as climate change itself (Hof et al. 2012). Rapid changes to the Arctic climate can lead to the range expansion of southern species into Arctic environments. These invasive southern species could include pathogens, predators and competitors for the already limited resources of Arctic ecosystems. A study on pathogens in Polar Bears (Ursus maritimus) found that bears in the Southern Beaufort Sea were being exposed to a new array of pathogens and that climate change-induced behavioural changes (such as time spent on land) had a significant impact on their exposure risk (Atwood et al. 2017). The range expansions of southern species into Arctic territories may also lead to increased predation (e.g., invasive Red King Crab

(Paralithodes camtschaticus): Christiansen et al. 2015), competition for habitat and resources

(e.g., Red Fox (Vulpes vulpes) vs. Arctic Fox (V. lagopus): Elmhagen et al. 2017), and

2 hybridization (e.g., Grizzly Bear (U. arctos horribilis) x Polar Bear: Pongracz et al. 2017;

Harbour Porpoise (Phocoena phocoena) x Dall’s Porpoise (Phocoenoides dalli): Crossman et al.

2014).

The balance of Arctic ecosystems may also be threatened by continued human interference. The reduction of sea ice represents a unique economic opportunity and many industries are already taking advantage of this unprecedented Arctic access. For many shipping companies in the Northern Hemisphere, trade routes through the Northwest Passage have the potential to be shorter and cheaper than shipping routes through the Panama Canal (Somanathan et al. 2007). The reduction of sea ice in the last decade has led the total number of kilometres travelled by ships through the Arctic to triple (Dawson et al. 2017). Arctic tourism is another avenue by which industry has an opportunity to exploit an untapped market. Nunavut has experienced a 70% increase in expedition cruise tourism, and a 400% increase in pleasure craft tourism in the past decade (Johnston et al. 2019). Although the Government of Canada has collected information from a range of sources and is developing Low Impact Shipping Corridors, the suggestions of people in Arctic communities across the Inuit Nunangat (Inuit homeland) have not yet been taken into account (Dawson et al. 2020). Increased shipping activity poses many threats to Arctic ecosystems including the introduction of invasive species through ballast water

(Ware et al. 2016), ship strikes on marine mammals, accidental entanglement in equipment by birds and mammals, anthropogenic noise causing disorientation and migratory disruption, and most significantly, the accidental or illegal release of oil pollution (Arctic Marine Shipping

Assessment 2009).

To protect vulnerable Arctic ecosystems, proactive and informed management plans are critical. The development of protected areas is an example of a popular management strategy to

3 conserve a region or ecosystem. For instance, at the end of 2019, Canada had 12% of terrestrial areas and inland water, and 13% of marine and coastal areas federally protected (Environment and Climate Change Canada 2020). Land management by Indigenous peoples is also an effective form of ecosystem conservation; Schuster et al. (2019) found that Indigenous-managed land in

Australia, Brazil and Canada was more vertebrate-rich and hosted more threatened vertebrate species than federally protected areas. Individual species such as the Polar Bear may also have detailed management plans both federally and internationally (Durner et al. 2018).

Conservation of genetic variation

Effective conservation must work to protect three levels of diversity: ecosystems, species and genes (McNeely et al. 2010). The conservation of ecosystems and individual species gets considerable attention, but the conservation of genetic variation is less commonly discussed

(Coates et al. 2018). Genetic diversity has long been recognized as an essential component of a species’ evolutionary potential as it is the foundation for adaptation and long-term population viability (Frankel 1970; Avise 2008). Population size has an impact on genetic diversity since small populations tend to have less standing genetic diversity and experience more genetic drift

(i.e. are more likely to have beneficial alleles reduce in frequency and maladaptive alleles increase in frequency due to chance) (Frankham 1996). Small population sizes can also lead to a greater risk of inbreeding depression, and the formation of an extinction vortex which could be further exacerbated by environmental pressures (Gilpin and Soulé 1986). The maintenance of genetic diversity, therefore, may be critical to a species’ ability to respond to climate change.

Several conservation strategies involve the division of species into population units to appropriately conserve genetic variation. Funk et al. (2012) suggested a hierarchical method of determining conservation units based on different genetic markers. They proposed that all loci should be used to assess population genetic differentiation, and that genetically distinct populations should be considered evolutionarily significant units (ESUs). Genetic variation that is putatively neutral to selection versus putatively adaptive (functional) is identified through methods such as outlier analysis (Foll and Gaggiotti 2008). Funk et al. (2012) suggested that neutral loci be used to designate management units (MUs), and that putatively adaptive loci be used to designate adaptive units (AUs) within the pre-defined ESUs. Barbosa et al. (2018) demonstrated this method as a strategy for conservation and species management of the Cabrera

Vole (Microtus cabrerae).

Hybridization

Appropriate species management can be complicated by hybridization between species.

Conservation policies need to be flexible when setting guidelines for hybridizing species since the frequency of hybridization, extent of introgressed genetic variation, evolutionary impacts of hybridization, and causes of hybridization are likely to be unique to each species (Allendorf et al.

2001). Furthermore, hybridization between species can have both positive and negative effects.

Climate change can facilitate range shifts that bring closely related species into secondary contact. Depending on the reproductive barriers between these species, range shifts due to climate change can lead to hybridization between previously separate species (Chunco 2014).

Hybrid offspring may exist anywhere on the spectrum from complete sterility to greater fertility

5 than either parental species. Fertile first-generation (F1) hybrids may experience a phenomenon known as hybrid vigour where the F1 individual has higher fitness than either of its parental species, but later generations of hybrid crosses often experience a reduction in fitness known as outbreeding depression. Hybridization between species can be a part of the natural formation of new species (hybrid speciation) such as in the Golden-crowned Manakin (Lepidothrix vilasboasi)

(Barrero et al. 2018) and the ‘Big Bird’ lineage of Darwin’s finch on Daphne Major (Geospiza fortis x G. conirostris) (Lamichhaney et al. 2018). Hybridization can also lead to the genetic rescue of declining species (Becker et al. 2013). For instance, research by Wells et al. (2019) found that hybridization between genetically highly divergent Brook Trout (Salvelinus fontinalis) populations had mostly neutral or positive effects on offspring fitness.

However, hybridization can also eliminate evolutionary adaptations to the local environment through genetic admixture (Verhoeven et al. 2011), and lead to the extinction of populations or even species (Rhymer and Simberloff 1996). Furthermore, when hybridization is occurring between a threatened species and an introduced or invasive species, the imbalance in population size and stability of the species can lead to the threatened species being overwhelmed both demographically and genetically. For example, in the 1860s Mallard Ducks (Anas platyrhynchos) were introduced to New Zealand. Since their introduction, Mallard Ducks have hybridized with the native Grey Ducks (A. superciliosa) which, coupled with the pressures of overhunting and habitat reduction, led to the near extinction of pure representatives of the Grey

Duck. By the 1980s, pure Grey Ducks represented only 4.5% of the population in Otago, New

Zealand (Gillespie 1985). A hybrid swarm can occur when hybrid offspring have a similar level of genetic fitness to their parental species and can backcross with their parental species.

Anthropogenic hybridization is the hybridization of two species caused by human activity (either

6 intentionally or unintentionally). Anthropogenic hybridization can occur between species that never could have bred in natural conditions (hybridization between wild and domesticated species: Randi 2008; hybridization between native and introduced species: Hohenlohe et al.

2011). Identifying whether hybridization is due to anthropogenic effects or natural processes can be challenging but is extremely important for determining appropriate management strategies

(Allendorf et al. 2001).

Glaucous Gull

This study will explore population genetic differentiation and hybridization in an Arctic seabird, the Glaucous Gull (Larus hyperboreus), and may inform the division of conservation units in collaboration with Environment and Climate Change Canada (ECCC).

Ecology

The Glaucous Gull is a large species of Arctic seabird and the second-largest species of gull in the world. Glaucous Gulls are large-bodied, deep-chested, and thick-necked with males averaging larger in weight and longer in wing length than females (Weiser and Gilchrist 2020).

They take four years to mature and have distinct plumage morphs each year. Juvenile Glaucous

Gulls have pale feathers with intricate greyish-brown streaking, dark eyes, and pink bills with black tips (Weiser and Gilchrist 2020). As adults, both sexes have light grey mantles and upperwings, white underwings and wingtips, thick yellow beaks with a red subterminal spot, and pink legs (Weiser and Gilchrist 2020).

Glaucous Gulls have a circumpolar breeding distribution. They nest on the ground or on cliff ledges, often in mixed-species colonies with Thick-billed Murres (Uria lomvia), Common

Murres (U. aalge), Black-legged Kittiwakes (Rissa tridactyla), geese (Anser and Branta sp.),

Eiders (Somateria sp.), Northern Fulmars (Fulmarus glacialis), and other species (Gaston and

Nettleship 1981; Smith et al. 1994; Gilchrist and Robertson 1999; Mallory and Gilchrist 2005).

Glaucous Gulls nest at low densities (average <100 pairs per colony), and nest density tends to decrease as distance from coastal areas increases (Strang 1976; Barry and Barry 1990). Mating pairs are philopatric to a nest site if it has been successful in the past (Portenko 1972). Juvenile

Glaucous Gulls are also natally philopatric, with 40% of chicks that survived to breed returning to their natal colony (Gaston et al. 2009). Established breeding pairs frequently remain together for multiple years; the divorce rate at Coats Island, Nunavut, over 22 years was 9% (Gaston et al.

2009).

Across the breeding range, four subspecies of Glaucous Gulls are recognized based on plumage colour and body size (Banks 1986). Larus h. pallidissimus is the largest and breeds across northern Siberia; L. h. hyperboreus breeds in northern Scandinavia and into northwestern

Siberia; L. h. leuceretes breeds in Iceland, Greenland, and in Arctic Canada east of the

Mackenzie River; and L. h. barrovianus is the smallest and breeds in Alaska and Arctic Canada west of the Mackenzie River (Banks 1986).

Due to their wide geographic range and remote breeding locations, estimates of a global population size are challenging. More than 2768 breeding colonies may exist in the circumpolar

Arctic, with between 138,600 and 218,600 breeding pairs (Petersen et al. 2015). Population estimates in Canada vary, with Gilchrist (2001) estimating 69,000 individuals in 1000 colonies and Gaston et al. (2012) estimating a minimum of 25,000 individuals.

Population differentiation

Previous work on population genetic differentiation in Glaucous Gulls revealed some evidence of structure. Research using mitochondrial DNA (mtDNA) showed that Glaucous Gulls were biphyletic: Glaucous Gulls sampled in Europe shared a clade of haplotypes that differed from those of the Glaucous Gulls sampled in North America (Liebers et al. 2004; Sternkopf et al.

2010). However, this structure was not evident in amplified fragment length polymorphism

(AFLP) data, which could not distinguish between North American and European Glaucous

Gulls (Sternkopf et al. 2010). Research using eleven microsatellite loci indicated that Glaucous

Gulls were weakly structured by sampling location with clusters loosely representing south- western Alaska; northern Alaska, and Canada; Iceland; and Greenland, and Svalbard

(Sonsthagen et al. 2012).

The mtDNA haplotypes of North American Glaucous Gull samples were shared by

American Herring Gulls (Larus smithsonianus) (Liebers et al. 2004; Sternkopf et al. 2010).

Sonsthagen et al. (2012) also found that microsatellite and nuclear intron data grouped Glaucous

Gulls and Herring Gulls together. Sternkopf et al. (2010) hypothesized that some Glaucous Gulls radiated into Northern Europe from a North American glacial refugium and hybridized with

European Herring Gulls (L. argentatus), undergoing a complete mtDNA replacement since the last Glacial Maximum. A study using allozymes found that the genetic distance between

Glaucous Gulls and Iceland Gulls (L. glaucoides) sampled on Baffin Island, Canada was less than between Glaucous Gulls sampled on Baffin Island and Glaucous Gulls sampled from

Bjarnarhafnarfjall, Iceland (Snell 1991). Another study that used mtDNA and incorporated the allozyme data from Snell’s (1991) work found that when the gull species tested had two haplotypes, the less frequent haplotype was identical to a common haplotype in another species

(Crochet et al. 2003). Crochet et al. (2003) found that the less frequent of the two Glaucous Gull haplotypes was shared by European Herring Gulls, suggesting that much of the intraspecific diversity of Larus gulls is due to hybridization between species, rather than genetic processes within species.

Hybridization

Hybridization between Glaucous Gulls and other members of the Herring Gull species complex is relatively common and has occurred across the breeding range of Glaucous Gull

(Sonsthagen et al. 2016). The frequency of hybridization within this species complex may be due to the small number of morphological features that enforce reproductive isolation including iris colour, fleshy eye-ring colour, leg colour, and bill colour (Smith 1966; Pierotti 1987). Some hybrid combinations such as Glaucous Gulls and Great Black-backed Gulls (Larus marinus) have very few sightings, and reports rely solely on unusual or intermediate plumage colouration and body size as evidence of hybridization (e.g. Glaucous Gull x Great Black-backed Gull sighting in Limerick, Ireland in 1948: Wilson 1951). Other crosses are so common that they have their own names within “gulling” communities. Glaucous Gull x (European or American)

Herring Gull hybrids are sometimes called “Viking Gulls” or “Nelson’s Gulls”, and Glaucous

Gull x Glaucous-winged Gull (L. glaucescens) hybrids are sometimes called “Seward Gulls”.

These hybrid crosses have been identified visually through intermediate characteristics

(Glaucous x Herring: Ingólfsson 1970; Spear 1987; Glaucous x Glaucous-winged: Swarth et al.

1934), but they have also been uncovered through genetic analysis (Vigfúsdóttir et al. 2008;

Pálsson et al. 2009; Sonsthagen et al. 2012; Sonsthagen et al. 2016). Hybridization has also been reported between Glaucous-winged Gulls and Herring Gulls (called “Cook Inlet Gulls”)

(Williamson and Peyton 1963; Smith 1966); as well as between Herring Gulls and Lesser Black-

10 backed Gulls (L. fuscus) (“Appledore Gulls”) (Smith 1966), Great Black-backed Gulls (“Great

Lakes Gulls”) (Smith 1966), and Kelp Gulls (L. dominicanus) (“Chandeleur Gulls”) (Dittmann and Cardiff 2005); and between Glaucous-winged Gulls and Western Gulls (L. occidentalis)

(“Olympic Gulls” or “Puget Sound Gulls”) (Heinl and Piston 2009).

The central and eastern coast of Iceland represents a well-researched zone of secondary contact between Glaucous Gulls and European Herring Gulls. The earliest unquestionable record of a European Herring Gull in Iceland was an immature bird shot in Husavik in 1909 (Ingólfsson

1970). European Herring Gulls were observed breeding in Iceland by 1925 and have since become established on the eastern coast of Iceland (Ingólfsson 1970; Vigfúsdóttir et al. 2008) and are encroaching on key Glaucous Gull breeding territory in central Iceland (Petersen 1998).

European Herring Gulls may be competing with Glaucous Gulls for nesting sites and resources.

First generation hybrids and fertile backcrosses have been identified through both morphometric and genetic analyses (Ingólfsson 1970; Vigfúsdóttir et al. 2008; Pálsson et al. 2009). Visual identification of Glaucous Gull x European Herring Gull hybrids is challenging and is typically based on the identification of intermediate traits. Due to the significant overlap of appearance between ‘pure’ individuals of Glaucous Gull and European Herring Gull, a bird with intermediate traits could suggest an admixed individual or could represent the natural range of traits within one species or the other.

Hybrid gulls show little or no indication of reduced fitness (Neubauer et al. 2009). A study on the hybrid zone between Herring Gulls and Caspian Gulls (Larus cachinnans) found that hybrid females have a slightly shorter lifespan than females or males in either parental species (Neubauer et al. 2014). This result agrees with Haldane’s Rule, which predicts that the heterogametic sex (females in birds) will experience a greater reduction in hybrid fitness

(Haldane 1922). When species are recently diverged, genetic incompatibilities may not have had time to develop and hybridization can lead to hybrid superiority (Grant and Grant 1992) such as in the hybrid zone of Glaucous-winged Gulls and Western Gulls (Good et al. 2000).

Conservation concerns

Glaucous Gulls are currently listed as Least Concern by the International Union for

Conservation of Nature (IUCN) (Bird Life International 2018); however, estimates of the stability of Glaucous Gull populations vary greatly with geographic location. Glaucous Gull populations are stable or increasing in Alaska, Greenland and Russia but declining in Canada,

Iceland and Svalbard (Petersen et al. 2015). These populations face a wide array of threats, which also differ by geographic location. In Canada, population declines may be due to low reproductive success and high mortality. Breeding success at the declining colony at Prince

Leopold Island, Nunavut was significantly lower than at the stable Coats Island, Nunavut colony

(Gaston et al. 2005; 2009). Studies in two Canadian colonies (St. Helena Island, Nunavut and

Coats Island, Nunavut) estimated adult survival of Glaucous Gull to be about 0.85 (Gaston et al.

2009; Allard et al. 2010), which is lower than the estimated survival rate of 0.89 for more temperate Larus species including Herring Gulls, Lesser Black-backed Gulls and California

Gulls (L. californicus) (Gaston et al. 2009). The lower survival rates at Canadian colonies may be attributable to several interacting threats: contaminant loads, which are higher in Glaucous

Gulls than in other Canadian seabirds (Buckman et al. 2004); outbreaks of avian cholera due to scavenging off infected carcasses (Allard et al. 2010); severe or irregular weather; and other unknown causes (e.g., some Glaucous Gulls have been found dead despite good physical condition) (Mallory et al. 2009).

Glaucous Gull declines in Iceland are not well understood. Adult birds and eggs are harvested in Iceland, but harvest rates have declined by more than half from an annual average harvest of 3847 Glaucous Gulls between 1995 and 2002, to an annual average of 1722 between

2004 and 2011 (Petersen et al. 2015). Nest predation by Arctic Foxes has grown as fox populations have increased in Iceland in recent decades, but this predation pressure may be resulting in relocation of nest sites rather than direct population declines (Hersteinsson 2004).

Reduction of food availability may also contribute to the population decline of Glaucous Gulls, as Iceland is more tightly managing garbage dumps and the disposal of fish refuse from processing plants (Petersen et al. 2015). Hybridization between Glaucous Gulls and European

Herring Gulls in the past century may also play a part in population declines for Icelandic

Glaucous Gulls (Ingólfsson 1970; Vigfúsdóttir et al. 2008; Pálsson et al. 2009).

Populations of Glaucous Gulls in Bjørnøya, Norway are likely declining due to increased competition for resources and the bioaccumulation and biomagnification of contaminants. Great

Skua (Stercorarius skua) became established on Bjørnøya in 1970 and have since increased to more than 300 breeding pairs; Glaucous Gulls have experienced a 65% reduction to 600 breeding pairs since 1986 which may, in part, be due to increased competition for resources with and predation by Great Skuas (Strøm 2007). Large numbers of dead and dying Glaucous Gulls are found annually on Bjørnøya and autopsies suggest that these birds have high levels of organochlorine pesticides (OCP), polychlorinated biphenyls (PCB), and polybrominated diphenyl ethers (PBDE) in their tissue (Sagerup et al. 2009). A wide array of negative consequences have been linked to high levels of environmental contaminants in Glaucous Gulls

(Verreault et al. 2010), including reduced reproductive success (Bustnes et al. 2003), reduced adult survival (Bustnes et al. 2003), weakened immune function (Sagerup et al. 2000), altered

13 hormone production (Verreault et al. 2004), and altered breeding and nesting behaviour (Bustnes et al. 2001). Finally, hybridization between Glaucous Gulls and European Herring Gulls may not be a major cause of population declines on Bjørnøya but some mixed-species pairs have been observed (Bertram and Lack 1933).

Methodology

High-throughput sequencing technology has allowed large quantities of detailed genetic data to be produced for a relatively low cost. Double digest restriction site-associated DNA (ddRAD) sequencing is a high-throughput sequencing method in which DNA is digested by two restriction enzymes, ligated with barcoded adaptors, and amplified to produce many copies of the same

DNA fragments (Peterson et al. 2012). ddRAD sequencing depends on high quality samples but can produce millions of fragments of DNA for analysis. High-throughput sequencing has allowed researchers to detect more detailed genetic structure and evolutionary history (Emerson et al. 2010), observe subtle gene flow (Zarza et al. 2016), resolve phylogenies (McCormack et al. 2013), identify novel viruses (Rwahnih et al. 2018), and detect ancient hybridization events

(Lecaudey et al. 2018) among other things.

Research Aims and Predictions

The aim of this research is to use high-throughput sequencing to determine 1) if hybrid individuals or introgressed genetic material from closely-related members of the Herring Gull species complex are present in the sampled Glaucous Gull colonies, and 2) if populations of

Glaucous Gulls are genetically differentiated across their range.

I predict that first generation (F1) hybrids and backcrosses will be present in the dataset due both to the apparent frequency of hybridization between Glaucous Gulls and closely-related species, and to the challenges of identifying hybrid gulls visually, leading to their unintentional sampling. I predict that due to the natal philopatry of Glaucous Gulls, hybrids will be present in colonies where Glaucous Gulls and other members of the Herring Gull species complex breed sympatrically, and that hybrids and introgression with other gulls will be less frequent or non- existent where Glaucous Gulls breed allopatrically.

I predict that population genetic structure will be present in Glaucous Gulls but will be relatively weak despite the wide geographic range of the samples and the large number of genetic markers included in this study. Glaucous Gulls breed at low densities, are natally philopatric and form monogamous breeding partnerships, which could contribute to strong population genetic differentiation; however, Glaucous Gulls are recently diverged from other

Larus gulls, hybridize with other Larus species, and are highly mobile, which could weaken or eliminate population genetic differentiation. With consideration for these divergent forces and previous research that suggests little population genetic differentiation, I predict that population genetic structure will be weak but that colonies may show a pattern of isolation by distance, and that North American and European colonies may be differentiated.

Due to the wide circumpolar range of Glaucous Gull, breeding colonies likely experience different climates and predation pressures. Under different selective pressures, populations of

Glaucous Gulls may begin to develop local adaptations. For this reason, I predict that I will find some putatively adaptive single nucleotide polymorphisms (SNPs) in outlier analyses.

Significance

This work is part of a greater collaborative project between Queen’s University, McGill

University, the University of Toronto and ECCC to determine the potential of Arctic seabird species to adapt to climate change, including Glaucous Gull, Northern Fulmar, Common Murre,

Black-legged Kittiwake, Common Eider (Somateria mollissima), Ivory Gull (Pagophila eburnea) and Black Guillemot (Cepphus grylle). This work will inform ECCC’s management decisions for the long-term persistence of Glaucous Gulls in a changing Arctic. Furthermore, this work aims to uncover the extent of hybridization and introgression between Glaucous Gulls and closely related members of the Herring Gull species complex.

CHAPTER 2: METHODS

Sample collection and DNA extraction

Tissue samples were collected from 133 Glaucous Gulls, 15 Glaucous-winged Gulls, 20

American Herring Gulls, and 24 European Herring Gulls from 25 sampling locations (Table 2.1).

Samples are archived at -80°C at Queen’s University, Kingston. DNA was extracted using a

NaCl precipitation protocol (Aljanabi and Martinez 1997). Samples were digested with 20μl proteinase K (20mg/ml) at 56°C in 200μl of salt-homogenizing buffer (400mM NaCl; 10mM

TRIS-HCl pH 8.0; 2mM EDTA pH 8.0) and 40μl of 10% SDS (Aljanabi and Martinez 1997).

After denaturing, 2.5μl of RNAse A/T1 (2mg/ml) was added and samples were incubated at

37°C for 30 minutes (Aljanabi and Martinez 1997). Many samples had large concentrations of fat and were further purified using a chloroform treatment (Friesen et al. 1997). DNA was further purified using ethanol precipitation and samples were resuspended in 30μl of 10mM

TRIS buffer (Sambrook et al. 1989). To assess sample quality visually, 2μl of each sample was subjected to electrophoresis through a 2% agarose gel with a 100 base pair (bp) ladder in TRIS- acetate buffer. To assess contamination and DNA concentration, 1μl of each sample was analyzed in a Denovix QFX Fluorometer (Denovix, Wilmington, DE, USA). DNA quality was highly variable due to the preservation method and advanced age of some samples.

Library preparation and sequencing

In silico digestion of an annotated American Herring Gull reference genome (B10K-DU-

002-28) from the Bird 10,000 Genome Project was performed using the program ddRADseqTools (Mora-Márquez et al. 2017) to optimize the selection of enzymes and fragment

17 sizes. The enzymes SbfI and MluCl, and fragment sizes between 200 and 400 bp (325 – 525 bp including the adaptors) were selected to maximize the average coverage per sample and number of fragments for further analyses. Preparation of a ddRAD library was done in-house at Queen’s

University by Zhengxin Sun based on a protocol by Peterson et al. (2012). DNA samples were digested using enzymes MluCl and SbfI and cleaned using SPRI beads (Beckman Coulter,

Indianapolis, IN, USA). Unique combinations of adaptors with degenerate base regions were ligated to each individual sample from a selection of 48 barcodes and 4 indexes. Adaptors had degenerate base regions (DBR) to facilitate the removal of PCR duplicates during bioinformatic processing (Schweyen et al. 2014; Vendrami et al. 2017). Samples were pooled and cleaned again with SPRI beads. Samples were shipped to the Centre for Applied Genomics in Toronto for 2x125 bp paired-end sequencing in a single lane on an Illumina HiSeq 2500 automated sequencer (Illumina, San Diego, CA, USA).

Identifying loci and filtering data

Quality of raw sequences was explored using the program FastQC v 0.11.9 (Andrews

2010). Then PCR duplicates were removed using the Python script ParseDBR_ddRAD.py

(github.com/Eljensen/ParseDBR_ddRAD) which was designed specifically to complement adaptors with DBRs (see above). Demultiplexing of samples was done using process_radtags in the program STACKS v 2.3e (Catchen et al. 2013). Reads with uncalled bases were removed, reads with low quality scores were discarded, barcodes and RADtags were rescued, and reads were trimmed to 104 bp (-c -q -r -t 104) using STACKS. Reads were aligned to the American

Herring Gull reference genome (B10K-DU-002-28) using the Burrows-Wheeler Aligner v 0.7.17

BWA-MEM (Li and Durbin 2010). Alignment success was assessed using SAMtools v 1.10 (Li et al. 2009) and unmapped reads and secondary alignments were removed.

Aligned reads were assembled in STACKS using the ref_map.pl pipeline and populations was used to select only the first SNP from every locus to reduce linkage disequilibrium.

VCFtools v 0.1.16 (Danecek et al. 2011) was used to filter to the highest quality SNPs. The order and value of filters used was informed by O’Leary et al. (2018) and was conducted iteratively to retain the greatest number of individuals and SNPs while also ensuring the highest possible quality data. In VCFtools, the function --minQ was used to retain only sites with a quality value of 20 or greater, and the function --minDP was used to retain only individuals with an average depth of 3 or greater. The function --min-meanDP was used to retain only sites with an average depth of 10 or greater. Discarding sites and individuals with low average depth decreases the chances of making incorrect site calls because sites with low depth can be inadvertently labelled homozygous due to the lack of an alternative allele among the reads. The function --mac was used to set the minor allele count to 3 meaning that an allele must appear in at least 3 individuals to be retained. The function --max-missing was used to remove sites that had 50% or more missing data, and then --missing-indv was used to assess the missing data per individual. The function --remove was used to remove individuals with 90% or more missing data. Next, the functions --max-missing, --missing-indv, and --remove were used cyclically to gradually increase the severity of the filters up to removing sites that had 10% or more missing data and removing individuals with 30% or more missing data.

The above filters were run separately on two datasets: 1) a dataset of 133 Glaucous Gull samples; and 2) a combined dataset of 133 Glaucous Gulls, 15 Glaucous-winged Gulls, 20

American Herring Gulls, and 24 European Herring Gulls. After filtering, the “Glaucous Gull

19 dataset” retained 62 samples, nine populations (KARR_NU, AKPA_NU, BEAR_NO,

DEVI_NU, BJAR_IS, BAFF_NU, KUUJ_QC, INUK_QC, and ARCO_AK; Table 2.1), and 621 loci; the “combined species dataset” retained 59 Glaucous Gulls (nine populations), 15

Glaucous-winged Gulls (3 populations), 6 European Herring Gulls (one population), 18

American Herring Gulls (two populations), and 2145 loci.

Finally, the program PGDSpider v 2.1.1.5 (Lischer and Excoffier 2012) was used to convert the Glaucous Gull dataset and the combined species dataset from .vcf format to the appropriate input formats for further analyses.

Hybridization

Hybridization between Glaucous Gull and Glaucous-winged Gull, American Herring

Gull and European Herring Gull was examined using several methods: molecular assignment using STRUCTURE v 2.3.4 (Pritchard et al. 2000), principal components analysis, principal coordinates analysis, Wright’s indices of population differentiation (pairwise FST), and

NewHybrids v 1.0 (Anderson and Thompson 2001). The program BayesAss v 3.0.4 (Wilson and

Rannala 2003) was used to estimate migration rates between the four species.

Differentiation between the four species

STRUCTURE v 2.3.4 is a Bayesian clustering program which assigns sampled individuals to genetic populations by minimizing deviations from Hardy-Weinberg and linkage equilibrium. STRUCTURE was run on the combined species dataset with three models: a) the admixture model without prior population information, b) the admixture model with the species designation as prior information, and c) the species designation as prior information without the

20 admixture model. Each run used 1,000,000 Markov chain Monte Carlo (MCMC) repetitions after a burn-in of 100,000 repetitions. This dataset was tested for a best K (assumed number of genetic populations) from 1 to 4, and was run 5 times for each value of K.

STRUCTURE was also run on the combined species dataset without the Glaucous- winged Gull samples because the strength of the genetic differentiation between Glaucous- winged Gulls and the other three species may mask weaker differentiation between Glaucous

Gull and the two Herring Gull species. This reduced dataset was also run with the three models described above, as well as 1,000,000 MCMC repetitions after 100,000 burn-in repetitions. This dataset was tested for a K from 1 to 3 and was run 5 times for each value of K.

To determine the best value of K for each of the STRUCTURE analyses described above, three methods were implemented in KFinder (Wang 2019): the Evanno method (Evanno et al.

2005), Bayes’ Rule (Pritchard et al. 2000), and the parsimony method (Wang 2019).

STRUCTURE plots were visualized using the package pophelper v 2.3.0 (Francis 2017) in R v

4.0.2 (R Core Team 2020). The best value of K was selected through a combination of visual analysis of STRUCTURE plots and the K value selected by the above methods.

Principal component analyses (PCAs) and principal coordinates analyses (PCoAs) are a priori dimensionality-reduction methods that preserve the variability of the original dataset while finding linear variables to explain aspects of the data. A PCA generates clusters based on correlations among samples, while a PCoA generates clusters based on differences between samples. PCoAs provide a better fit than PCAs when data are missing (Rohlf 1972). PCAs and

PCoAs were conducted on the combined species dataset using the ade4 package (Dray and

Dufour 2007) and visualized using the ggplot2 package (Wickham 2016) in R. PCAs and PCoAs were also conducted on the Glaucous Gull and American Herring Gull samples (77 individuals),

Glaucous Gull and European Herring Gull samples (65 individuals), and Glaucous Gull and

Glaucous-winged Gull samples (74 individuals).

Pairwise FST compares the variation in genotype frequencies within and between sampled populations to index population structure. Pairwise comparisons of FST between the four species were estimated on the “combined species dataset” using Arlequin v 3.5.2.2 (Weir and

Cockerham 1984; Excoffier and Lischer 2010). Values of FST were considered significant when p < 0.05.

NewHybrids v 1.0 is a program designed to estimate the probability that a sampled individual represents any one of several defined hybrid categories. This program was run on paired species datasets of Glaucous Gull and Glaucous-winged Gull (74 individuals); Glaucous

Gull and American Herring Gull (77 individuals); Glaucous Gull and European Herring Gull (65 individuals); and Glaucous Gull and Herring Gull (American and European Herring Gull pooled;

83 individuals) using 100,000 MCMC repetitions and a burn-in of 10,000 repetitions, and

200,000 MCMC repetitions and a burn-in of 20,000 repetitions. NewHybrids is unable to handle large numbers of loci so 10 datasets of 500 randomly sampled loci were generated for each species pairing. A reduced set of unlinked loci with the highest FST between the species pairings was also used to try to increase confidence in differentiating between species. These high-FST loci were selected by the function genepop_toploci in genepopedit (Stanley et al. 2016) in R. The

Glaucous Gull and American Herring Gull dataset had 154 unlinked high-FST loci; the Glaucous

Gull and European Herring Gull dataset had 128 loci; the Glaucous Gull and Glaucous-winged

Gull dataset had 184 loci; and the Glaucous Gull and pooled Herring Gull dataset had 146 loci.

These reduced loci datasets were run with 100,000 MCMC repetitions and 10,000 burn-in

22 repetitions. To increase efficiency, this program was run through the GUI platform EasyParallel

(Zhao et al. 2020) and graphed in hybriddetective (Wringe et al. 2017) in R.

Migration rate

The program BayesAss v 3.0.4 is a Bayesian method of estimating contemporary migration between populations. This program was run using the combined species dataset with four populations designated by species. BayesAss was run three times with different seed values using 3,000,000 MCMC iterations, a burn-in of 300,000 repetitions, and sampling every 2000 iterations. The mixing parameter for allele frequencies (-a) was adjusted from the default (0.1) to

0.3 to optimize mixing. To examine convergence between runs, the program Tracer v 1.7.1

(Rambaut et al. 2018) was used to graph the trace files and deviance was calculated using an R script developed by Meirmans (2014).

Population differentiation within the Glaucous Gull

Hybrids were removed and population genetic differentiation within Glaucous Gulls was assessed using several methods: molecular assignment using STRUCTURE, PCA, PCoA,

Wright’s indices of population differentiation (pairwise FST) and a Mantel test. BayeScan v 2.1

(Foll and Gaggiotti 2008) was used to identify loci potentially under selection, and BayesAss was used to estimate migration rate between populations.

Population genetic structure

STRUCTURE was run on the Glaucous Gull dataset without three identified hybrids (see

Results), and without the sole representatives of three populations (KUUJ_QC, INUK_QC, and

ARCO_AK) because STRUCTURE can be biased by uneven sample size (Puechmaille 2016).

Three models were run: a) the admixture model without prior population information, b) the admixture model with the sampling location as prior information, and c) the sampling location as prior information without the admixture model. STRUCTURE was run for K from 1 to 8, and was run 5 times for each value of K. To determine the best value of K, the three methods described above were implemented in KFinder. STRUCTURE plots were visualized using the package pophelper in R. The best value of K was selected through a combination of visual analysis of STRUCTURE plots and the K value selected by the above methods.

PCAs and PCoAs were conducted on the Glaucous Gull dataset without the three identified hybrids as described above.

Pairwise comparisons of FST were estimated for the Glaucous Gull dataset without the three identified hybrids, and without the individuals from ARCO_AK, KUUJ_QC, INUK_QC, and BAFF_NU using Arlequin (Weir and Cockerham 1984). These individuals were removed because pairwise FST estimates can be biased by small or unequal sample sizes. Values of FST were considered significant when p < 0.05.

Genetic diversity

Three indices of genetic diversity were calculated to estimate the genetic distinctiveness and diversity of the populations: rarefied allelic richness, expected heterozygosity, and private alleles. Rarefied allelic richness was calculated for five populations (KARR_NU, AKPA_NU,

BEAR_NO, DEVI_NU, and BJAR_IS) using the package hierfstat (Goudet 2005; Kalinowski

2004) in R. Expected heterozygosity was estimated from allelic variation (Smith and Grassle

1977) for the same five populations. Private alleles were computed for three populations of

24 similar sample size (KARR_NU, BEAR_NO, and DEVI_NU) in the package poppr v 2.81

(Kamvar et al. 2015) in R.

Mantel test

A Mantel test was conducted to determine if genetic distance between sampled colonies is correlated with geographic distance. Slatkin’s linearized FST (FST/(1-FST)) (Slatkin 1991) was calculated in Arlequin. Marine distance (the shortest distance between two sites that does not cross over land) was approximated to the nearest 250 km in Google Maps (Google Maps 2020) and was used instead of Euclidean distance. Marine distance may be a better estimate of flying distance between colonies because seabirds tend not to fly over large expanses of land or ice.

Marine distances were log-transformed (base = 10) based on the recommendations for analysis of isolation by distance in Rousset (1997). A Mantel test was conducted on Slatkin’s linearized

FST and log-transformed marine distance between five colonies of Glaucous Gull (AKPA_NU,

BEAR_NO, DEVI_NU, KARR_NU, and BJAR_IS) using the package ade4 in R with 99 repetitions. The relationship between Slatkin’s linearized FST and log-transformed marine distance was graphed using ggplot2 in R.

Loci potentially under selection

BayeScan v 2.1 was used to search for SNPs potentially under selection. BayeScan was run on the Glaucous Gull dataset twice: once using population designations (KARR_NU,

AKPA_NU, BEAR_NO, DEVI_NU, BJAR_IS, BAFF_NU, KUUJ_QC, INUK_QC, and

ARCO_AK), and once using regional designations (either Europe or North America). BayeScan was run using 100,000 repetitions following a burn-in of 50,000 repetitions, and prior odds of both 100 and 1000.

Migration rate

The program BayesAss was run on the Glaucous Gull dataset without three identified hybrids using the two genetic populations (North America and Europe), as described above.

Table 2.1. Sampling location, year, species designation, and source for sampled individuals. NWRC = National Wildlife Research Centre, Canada; ROM = Royal Ontario Museum, Canada; IINH = Icelandic Institute of Natural History, Iceland; BURKE = Burke Museum, USA; FRIESEN = Queen’s University, Canada

SPECIES SAMPLING SHORT COUNTRY COORDINATES SEQUENCED SAMPLES AFTER SAMPLE YEAR LOCATION FORM SAMPLES FILTERING SOURCE GLAUCOUS GULL Bjørnøya, Svalbard BEAR_NO Norway 74.46, 19.12 16 16 NWRC 1996 Skrúður Iceland 64.9, -13.62 3 0 IINH 1965 Bjarnarhafnarfjall BJAR_IS Iceland 64.99, -23.01 16 4 ROM/IINH 1986/1971-72 Ittoqqortoormiit Greenland 70.5, -22.0 12 0 Devil Island, Nunavut DEVI_NU Canada 76.51, -90.46 18 14 NWRC 2005-06 Akpatok Island, Nunavut AKPA_NU Canada 60.59, -68.11 4 4 NWRC 1983 Karrak Lake, Nunavut KARR_NU Canada 67.26, -100.27 21 17 NWRC 2004, 2007 Baffin Island, Nunavut BAFF_NU Canada 68.98, -68.17 7 4 ROM 1985 Kuujjuarapik, Quebec KUUJ_QC Canada 55.27, -77.76 2 1 NWRC 1992 Inukjuak, Quebec INUK_QC Canada 58.45, -78.11 3 1 NWRC 1991 Simpson Lake, Nunavut Canada 68.5, -91.4 10 0 NWRC 2007 Arctic Ocean, Alaska ARCO_AK USA 71.2, -163.85 10 1 BURKE 1993 Kigigak Island, Alaska USA 60.9, -165.0 9 0 ROM 1994 North Slope Borough, USA 71.3, -156.5 2 0 BURKE 1995 Alaska EUROPEAN Skrúður SKRU_IS Iceland 64.9, -13.62 19 6 ROM 1986 HERRING GULL Bjarnarhafnarfjall Iceland 64.9, -22.9 5 0 IINH 1965-1966 AMERICAN Southampton Island, SOUT_NU Canada 64.0, -81.92 9 8 FRIESEN 2013 HERRING GULL Nunavut Witless Bay, WITB_NL Canada 47.28, -52.83 11 10 FRIESEN 2012 Newfoundland GLAUCOUS- Bare Island, British BARE_BC Canada 48.63, -123.29 3 3 ROM 1996 WINGED GULL Columbia Buldir Island, Alaska BULD_AK USA 52.37, 175.92 4 4 ROM 1994 Amchitka Island, Alaska AMCH_AK USA 51.53, 179.03 8 8 ROM 1994

Figure 2.1. A map of sampling locations of gulls retained after data filtering. Species breeding ranges (Bird Life International 2020; Cornell Lab of Ornithology 2019; Hayward and Verbeek

2020; Weiser and Gilchrist 2020) are designated by colour, and the number of samples from each location after data filtering is contained in the coloured points and in Table 2.1.

Figure 2.2. Sampling map of Glaucous Gull colonies after hybrid individuals were removed. The breeding range (Weiser and Gilchrist 2020) of Glaucous Gull is shown in blue and the number of samples from each location is contained in the coloured points.

CHAPTER 3: RESULTS

Prior to data filtering, mean effective per sample coverage was 13.1 (24.8 SD). The minimum coverage was 1.2 and the maximum coverage was 156.8. The “Glaucous Gull dataset” had 133 individuals, 14 populations, and 60,464 loci; and the “combined species dataset” had

192 individuals, 21 populations across four species, and 77,657 loci.

After data filtering, the Glaucous Gull dataset retained 62 Glaucous Gulls from nine populations (KARR_NU, AKPA_NU, BEAR_NO, DEVI_NU, BJAR_IS, BAFF_NU,

KUUJ_QC, INUK_QC, and ARCO_AK), and 621 loci. After data filtering, the combined species dataset retained 59 Glaucous Gulls from nine populations (KARR_NU, AKPA_NU,

BEAR_NO, DEVI_NU, BJAR_IS, BAFF_NU, KUUJ_QC, INUK_QC, and ARCO_AK), 15

Glaucous-winged Gulls from three populations (BARE_BC, AMCH_AK, and BULD_AK), 18

American Herring Gulls from two populations (SOUT_NU and WITB_NL), 6 European Herring

Gulls from one population (SKRU_IS), and 2145 loci.

Hybridization

Differentiation between the four species

Analyses by the program STRUCTURE showed that Glaucous-winged Gulls are genetically distinct from the other three species of gull, but showed no clear distinctions between Glaucous

Gulls and European Herring Gulls and only weak distinctions between American Herring Gulls and Glaucous Gulls, and between American Herring Gulls and European Herring Gulls (Figures

3.1-3.3, Table 3.1). Regardless of the model, one American Herring Gull (from Witless Bay,

Newfoundland) grouped closely with the Glaucous-winged Gulls (Figures 3.1-3.3). Furthermore,

30 three Glaucous Gulls and one European Herring Gull grouped with the American Herring Gulls

(Figure 3.2-3.3).

In the STRUCTURE analysis without Glaucous-winged Gulls, there was almost complete consensus across the three models and the three methods that the most likely number of genetic populations was 2 (Table 3.1). Visual analysis of the STRUCTURE plot suggested a weak divide between American Herring Gulls and European Herring Gulls and between

American Herring Gulls and Glaucous Gulls (Figure 3.3).

Principal component analysis (PCA) showed clear differentiation along species boundaries for Glaucous-winged Gulls and American Herring Gulls but no clear differentiation between Glaucous Gulls and European Herring Gulls (Figure 3.4). Principal coordinates analysis

(PCoA) showed a similar pattern with slightly more of the variation explained by each axis

(Figure 3.5). Both the PCA and PCoA showed five individuals that did not group with their species designations: three Glaucous Gulls grouped with the American Herring Gulls; an

American Herring Gull grouped with the Glaucous-winged Gulls; and a European Herring Gull grouped with the American Herring Gulls (Figures 3.4-3.5). The same five individuals also appeared as potential hybrids in the paired species PCAs and PCoAs (Figures 3.6-3.8).

Pairwise estimates of FST were significantly greater than zero between all sampled species (Table 3.2), however, estimates of population differentiation between European Herring

Gulls and the other three species should be interpreted cautiously due to the low number of samples representing European Herring Gull.

NewHybrids analyses between Glaucous Gull and Glaucous-winged Gull samples with

500 randomly sampled loci consistently identified several putative hybrids (Figure 3.9),

however, analyses with the reduced panel of high FST loci were unable to identify Glaucous- winged Gull as a separate species and instead identified them as F2 hybrids. This failure to identify Glaucous-winged Gull as a distinct species suggests that NewHybrids run parameters did not converge and, therefore, putative hybrids cannot be identified by this analysis. All analyses between Glaucous Gull and American Herring Gull, European Herring Gull, and

Herring Gull (pooled American and European Herring Gull) also failed to converge.

Migration rate

BayesAss run parameters converged, estimates of migration and residency rate were consistent between runs, and the run with the lowest deviance was selected as the best (Table

3.3). However, the estimate of residency rate for the European Herring Gull samples was too low to be able to confidently estimate migration between this species and the other three. Otherwise, only one estimate of migration (between Glaucous Gulls and American Herring Gulls) was significantly different from zero (Table 3.3).

Population differentiation within the Glaucous Gull

Population genetic structure

Estimates of the most likely number of genetic populations were inconsistent across the three models in STRUCTURE (Table 3.4). Visual analysis of STRUCTURE plots showed clear division between European colonies (Bjørnøya and Bjarnarhafnarfjall) and North American colonies of Glaucous Gull when sampling location was used as prior information (Figure 3.10).

Population differentiation between North America and Europe was also visible in the K=7 plots using the other two models (Figures 3.11-3.12).

The PCA and PCoA showed weak differentiation between North American and European colonies of Glaucous Gull, with colonies separating along the secondary axis of the PCA (Figure

3.13a) and along the primary axis of the PCoA (Figure 3.14a).

Pairwise estimates of FST were significantly greater than zero between Bjørnøya and Karrak

Lake or Devil Island, Nunavut (Table 3.5). Pairwise estimates of FST were also significantly greater than zero between Karrak Lake and Bjarnarhafnarfjall, Iceland, however, this may be a statistical artifact due to the low sample size from Bjarnarhafnarfjall (Table 3.5).

Loci potentially under selection

BayeScan did not identify any putatively adaptive loci.

Genetic diversity

The six Glaucous Gull colonies had similar levels of allelic richness and expected heterozygosity (Table 3.6). The three colonies with sufficient sample size (KARR_NU,

DEVI_NU, and BEAR_NO) had 166 private alleles in total (Table 3.7). Colonies at KARR_NU and BEAR_NO had a similar number of private alleles, and DEVI_NU had the least (Table 3.7).

Over 45% of the private alleles were found in just one individual per population, however, nearly

40% were found in two individuals and more than 15% were found in three or more individuals per population (Figure 3.15).

Migration rate

BayesAss run parameters converged, estimates of migration and residency rate were consistent between runs, and the run with the lowest deviance was selected as the best (Table

3.8). However, the residency rate for North American Glaucous Gulls is too low to be able to confidently estimate migration between North America and Europe.

Mantel test

The Mantel test did not find a significant correlation between Slatkin’s linearized FST and log-transformed marine distance (r = 0.21, p = 0.29) (Figure 3.16).

Figure 3.1. Results of analyses using the program STRUCTURE on the combined species dataset showing probability of assignment to two, three, and four genetic populations using a) the admixture model, and b) the admixture model and species as prior population information.

Figure 3.2. Results of analyses using the program STRUCTURE on the combined species dataset showing probability of assignment to two, three, and four genetic populations without admixture using species as prior population information.

Figure 3.3. Results of analyses using the program STRUCTURE on the combined species dataset without Glaucous-winged Gulls showing probability of assignment to two genetic populations using a) the admixture model, b) the admixture model and species as prior population information, and c) species as prior population information.

Figure 3.4. Results of principal component analysis of genomic variation in Glaucous Gull, European Herring Gull, American Herring Gull, and Glaucous-winged Gull samples; a) depicts principal components one and two and bar chart shows eigenvalues, b) depicts principal components two and three. Putative hybrids are labelled with their sampling colony. See Table 2.1 for colony abbreviations.

Figure 3.5. Results of principal coordinates analysis of genomic variation in Glaucous Gull, European Herring Gull, American Herring Gull, and Glaucous-winged Gull samples; a) depicts principal coordinates one and two and bar chart shows eigenvalues, b) depicts principal coordinates two and three. Putative hybrids are labelled with their sampling colony. See Table 2.1 for colony abbreviations.

Figure 3.6. Results of principal component analysis (a) and principal coordinates analysis (b) of genomic variation in Glaucous Gull and American Herring Gull samples. Bar charts show eigenvalues. Putative hybrids are labelled with their sampling colony. See Table 2.1 for colony abbreviations.

Figure 3.7. Results of principal component analysis (a) and principal coordinates analysis (b) of genomic variation in Glaucous Gull and European Herring Gull samples. Bar charts show eigenvalues. Putative hybrids are labelled with their sampling colony. See Table 2.1 for colony abbreviations.

Figure 3.8. Results of principal component analysis (a) and principal coordinates analysis (b) of genomic variation in Glaucous Gull and Glaucous-winged Gull samples. Bar charts show eigenvalues. Putative hybrids are labelled with their sampling colony. See Table 2.1 for colony abbreviations.

Figure 3.9. Results of analyses using the program NewHybrids on Glaucous Gull and Glaucous- winged Gull samples showing probability of assignment to each of six hybrid categories using 500 randomly sampled loci.

Figure 3.10. Results of analyses using the program STRUCTURE on the Glaucous Gull dataset showing probability of assignment to two, three and four genetic populations using sampling location as prior population information. See Table 2.1 for colony abbreviations.

Figure 3.11. Results of analyses using the program STRUCTURE on the Glaucous Gull dataset showing probability of assignment to two, four and seven genetic populations using admixture model and sampling location as prior population information. See Table 2.1 for colony abbreviations.

Figure 3.12. Results of analyses using the program STRUCTURE on the Glaucous Gull dataset showing probability of assignment to two, three and seven genetic populations using the admixture model. See Table 2.1 for colony abbreviations.

Figure 3.13. Results of principal component analysis of genomic variation in Glaucous Gull samples from 6 colonies; a) depicts principal components one and two and bar chart shows eigenvalues, b) depicts principal components two and three.

Figure 3.14. Results of principal coordinates analysis of genomic variation in Glaucous Gull samples from 6 colonies; a) depicts principal coordinates one and two and bar chart shows eigenvalues, b) depicts principal coordinates two and three.

80 75

70 65

Number of Private Alleles ofPrivate Number 20 16

10 5 2 2 1 0 1 2 3 4 5 6 7 Frequency of Private Allele

Figure 3.15. Histogram of the combined frequency (number of occurrences) of private alleles in three colonies of Glaucous Gull (BEAR_NO, KARR_NU, and DEVI_NU). See Table 2.1 for colony abbreviations.

Figure 3.16. Correlation between Slatkin’s linearized FST and log-transformed marine distance between five colonies of Glaucous Gull (AKPA_NU, BEAR_NO, KARR_NU, DEVI_NU, and BJAR_IS) with linear regression line. See Table 2.1 for colony abbreviations.

Table 3.1. Most likely number of genetic populations (K) from STRUCTURE analyses of four species of gull, selected by three methods: Evanno (Evanno et al. 2005), Bayes’ Rule (Pritchard et al. 2000), and parsimony (Wang 2019). See page 19 for a description of each STRUCTURE model.

Model Evanno Bayes’ Rule Parsimony Combined species Admixture 2 4 2 Priors 2 4 3

Admixture + Priors 2 3 2

Combined species Admixture 2 2 2 without Glaucous- winged Gull Priors 2 2 2

Admixture + Priors 2 2 1

Table 3.2. Estimates of pairwise FST between the four sampled species of gull. FST values are below the diagonal, p-values are above the diagonal. Estimates that are significantly different from 0 (p < 0.05) are bold.

Glaucous Gull European American Glaucous- Herring Gull Herring Gull winged Gull Glaucous Gull 0.01 0.00 0.00 European Herring Gull 0.03 0.01 0.00 American Herring Gull 0.09 0.03 0.00 Glaucous-winged Gull 0.29 0.18 0.18

Table 3.3. Estimates of migration rate for Glaucous Gull, American Herring Gull, European Herring Gull and Glaucous-winged Gull are given as the proportion of the recipient population which originated from the source population. Residency rate appears along the diagonal and standard deviations are in brackets. Values displayed are from the run with the lowest deviance. Estimates that are significantly different from zero are in bold with an asterisk.

Source Glaucous Gull European American Glaucous- Populations Herring Gull Herring Gull winged Gull Recipient Glaucous 0.962 (0.013) 0.011 (0.007) 0.022* (0.01) 0.006 (0.005) Populations Gull European 0.200 (0.049) 0.700 (0.03) 0.066 (0.04) 0.035 (0.03) Herring Gull American 0.016 (0.015) 0.015 (0.015) 0.938 (0.027) 0.031 (0.021) Herring Gull Glaucous- 0.018 (0.017) 0.018 (0.017) 0.035 (0.023) 0.929 (0.031) winged Gull

Table 3.4. Most likely number of genetic populations (K) from STRUCTURE analysis of Glaucous Gull samples selected by three methods: Evanno (Evanno et al. 2005), Bayes’ Rule (Pritchard et al. 2000), and parsimony (Wang 2019). See page 23 for a description of each STRUCTURE model.

Model Evanno Bayes’ Rule Parsimony Admixture 2 7 3

Priors 2 4 2

Admixture + Priors 4 7 2

Table 3.5. Estimates of pairwise FST between five sampled colonies of Glaucous Gull displayed east to west. FST values are below the diagonal, p-values are above the diagonal. Estimates that are significantly different from 0 (p < 0.05) are bold and have an asterisk. See Table 2.1 for colony abbreviations.

BEAR_NO BJAR_IS DEVI_NU AKPA_NU KARR_NU BEAR_NO 0.06 0.00* 0.26 0.00* BJAR_IS 0.02 0.22 0.16 0.03* DEVI_NU 0.03* 0.01 0.25 0.12 AKPA_NU 0.00 0.03 0.01 0.36 KARR_NU 0.01* 0.02* 0.01 0.00

Table 3.6. Rarefied allelic richness and expected heterozygosity from five colonies of Glaucous Gull. See Table 2.1 for colony abbreviations.

Allelic Expected Richness Heterozygosity KARR_NU 1.13 0.13 AKPA_NU 1.14 0.14 BEAR_NO 1.13 0.13 DEVI_NU 1.13 0.13 BJAR_IS 1.14 0.14

Table 3.7. Frequency of private alleles from three Glaucous Gull colonies. See Table 2.1 for colony abbreviations.

Frequency KARR_NU BEAR_NO DEVI_NU 1 27 22 26 2 26 10 29 3 3 7 6 4 4 0 1 5 1 1 0 6 1 1 0 7 1 0 0 Total: 63 62 41

Table 3.8. Estimates of migration rate for Glaucous Gulls from two genetically differentiated regions: North America (4 populations, 36 individuals), and Europe (2 populations, 20 individuals) are given as the proportion of the recipient population which originated from the source population. Residency rate appears along the diagonal, and standard deviations are in brackets. Values displayed are from the run with the lowest deviance.

Source Populations North America Europe Recipient North America 0.676 (0.009) 0.324 (0.009) Populations Europe 0.025 (0.019) 0.975 (0.019)

CHAPTER 4: DISCUSSION

As predicted, several individuals included in this analysis were identified as potential hybrids between Glaucous Gulls and American Herring Gulls. Furthermore, one individual was identified as an American Herring Gull x European Herring Gull hybrid, and one individual was identified as a Glaucous-winged Gull x American Herring Gull hybrid. This research also provided evidence for weak population genetic structure between North American and European colonies of Glaucous Gulls, as predicted.

Hybridization

Identification of hybrid individuals was challenging due to the low differentiation between

Glaucous Gulls and American Herring Gulls, European Herring Gulls and Glaucous-winged

Gulls. However, as predicted, several Glaucous Gulls sampled in areas of breeding range overlap were identified as hybrids.

Differentiation between four gull species

The Glaucous Gull and Glaucous-winged Gull were the most strongly differentiated of the species pairs. This result agrees with previous work using nuclear intron and microsatellite data that showed that Glaucous Gulls and Glaucous-winged Gulls group in separate clusters within the Herring Gull species complex (Sonsthagen et al. 2016). Glaucous Gulls and American

Herring Gulls were confidently differentiated by most methods in the present study, although

NewHybrids was unable to identify the American Herring Gull samples as distinct from the

Glaucous Gull samples. This is likely because NewHybrids is limited in the number of loci it can analyze, and 500 loci was not enough for NewHybrids to reach convergence between such

59 closely related species. Previous research on Glaucous Gulls and American Herring Gulls could not confidently differentiate the species using microsatellite loci and nuclear introns (Sonsthagen et al. 2016) or mtDNA (Liebers et al. 2004) and the phylogenetic relationship between these species is still controversial. Glaucous Gulls and European Herring Gulls were significantly differentiated using pairwise FST but the value was low and these species could not be differentiated by any other methods. This may be due to historical association between the species as well as contemporary gene flow and introgression in Iceland (Ingólfsson et al. 1970;

Vigfúsdóttir et al. 2008). In future work, a greater number of nuclear loci may be necessary to distinguish these two species.

Residency rates of Glaucous Gulls, American Herring Gulls, and Glaucous-winged Gulls estimated using BayesAss were sufficiently high to assess the migration rate between these species. Unfortunately, the estimated residency rate of European Herring Gulls was too low to accurately measure the migration rate between European Herring Gulls and the other three species which may suggest that gene flow is (or has recently been) high between them.

Furthermore, results seem to suggest significant migration occurs from Glaucous Gull into

European Herring Gull which is consistent with previous research (Vigfúsdóttir et al. 2008).

Most estimates of migration were not significantly different from zero, providing no evidence for gene flow between these species. In contrast, the estimated migration rate from American

Herring Gull into Glaucous Gull was 2.2% (1.0% SD). This estimate of migration is higher than some estimates of migration between populations within structured species (e.g. New Zealand

Blue Duck (Hymenolaimus malacorhynchos): Grosser et al. 2016), which further highlights the limited barriers to gene flow between members of the Herring Gull species complex. Since filtering did not retain samples from regions of breeding range overlap between Glaucous Gulls

60 and Glaucous-winged Gulls, the rate of migration estimated by this analysis may not accurately represent the extent of gene flow between these species. Overall, results from BayesAss suggest significant migration from American Herring Gull into Glaucous Gull, and from Glaucous Gull into European Herring Gull.

Glaucous Gull hybrids

Through a combination of methods, three individuals were identified as hybrids between

Glaucous Gulls and American Herring Gulls. All three individuals were originally identified as

Glaucous Gulls and were sampled near where the southern edge of the Glaucous Gull breeding range overlaps with the northern edge of the American Herring Gull breeding range (Figure 2.1).

Only four individuals from the Baffin Island colony were retained through data filtering and two individuals were identified as potential Glaucous Gull x American Herring Gull hybrids (hybrid frequency of 50%). Seventeen individuals from the Karrak Lake colony were retained through data filtering and one was identified as a Glaucous Gull x American Herring Gull hybrid (hybrid frequency of 6%). With so few samples retained from Baffin Island, this value is probably not indicative of the true hybrid frequency at this colony. However, these results may suggest that hybrids are more frequent at Baffin Island than at Karrak Lake. Interestingly, the colonies at

Akpatok Island and Southampton Island had no potential hybrids despite the colonies’ placement within the overlap between the American Herring Gull and Glaucous Gull breeding ranges

(Figure 2.1). Overall, three hybrid individuals were identified from 59 Glaucous Gull samples, which is a hybrid frequency of 5%. Previous research on the frequency of Glaucous Gull x

American Herring Gull hybrids near the Mackenzie Delta used intermediate morphology to assign 4.4% of the birds assessed as hybrids (Spear 1987).

Despite clear genetic differentiation between Glaucous Gulls and Glaucous-winged Gulls, no

Glaucous Gull x Glaucous-winged Gull hybrids were identified. Hybridization between

Glaucous Gulls and Glaucous-winged Gulls is frequent in some regions of Alaska where breeding ranges overlap (Swarth 1934; Strang 1977). Therefore, the lack of Glaucous Gull x

Glaucous-winged Gull hybrids is likely because no Alaskan Glaucous Gull colonies were retained through data filtering.

No individuals were identified as Glaucous Gull x European Herring Gull hybrids despite the frequency of hybridization between these species. A study of gull morphology at

Bjarnarhafnarfjall and Skrúður, Iceland in the 1960s showed that 22% of the gulls sampled at

Bjarnarhafnarfjall and 51% of the gulls sampled at Skrúður were Glaucous Gull x European

Herring Gull hybrids (Ingólfsson 1970). Hybridization between Glaucous Gulls and European

Herring Gulls is likely much less frequent in Bjørnøya, Norway, although mixed species pairs and F1 hybrids have been recorded (Bertram and Lack 1933). Due to the lack of strong genetic differentiation between Glaucous Gulls and European Herring Gulls, if hybrid individuals were present among the samples from Iceland (Bjarnarhafnarfjall and Skrúður) or Norway (Bjørnøya), our methods may not have been able to identify them. A greater number of loci as well as morphological data for each sample may be necessary to accurately differentiate Glaucous Gulls and European Herring Gulls.

Other hybrids

Analyses also uncovered one European Herring Gull x American Herring Gull hybrid in

Iceland (Skrúður), and one American Herring Gull x Glaucous-winged Gull hybrid in

Newfoundland, Canada (Witless Bay). American Herring Gull sightings in Iceland are uncommon but not rare, and hybridization between these two species is frequent. Not all

62 organizations consider the American Herring Gull and European Herring Gull to be separate species but treat the American Herring Gull as a subspecies of the European Herring Gull (see

Introduction).

Some Glaucous-winged Gulls have been reported in Newfoundland, but they are extremely rare (Mactavish 2005; Clarke 2006) and no Glaucous-winged Gulls have been recorded breeding in Newfoundland. A Glaucous-winged Gull x American Herring Gull hybrid hatched on the east coast of Canada would be an extraordinarily rare individual. It may instead be a hybrid between an American Herring Gull and another gull species. Due to the close genetic relationships between members of the Herring Gull species complex, the Witless Bay individual may actually be a hybrid between American Herring Gull and another gull species that Glaucous-winged

Gulls are genetically similar to. Previous research by Sonsthagen et al. (2016) revealed that

Glaucous-winged Gulls form a genetic clade with Mew Gulls (Larus canus), Ring-billed Gulls

(L. delawarensis), Heermann’s Gulls (L. heermanni), Yellow-footed Gulls (L. livens) and

Western Gulls (L. occidentalis), while Glaucous Gulls and American Herring Gulls are part of a separate genetic clade. However, most of the species that form a genetic clade with Glaucous- winged Gulls are also not common breeders in Newfoundland, and Ring-billed Gulls (which do breed in Newfoundland) have no recorded instances of hybridization with American Herring

Gulls. Another possibility is that the Witless Bay individual hatched on the west coast (where hybridization between Glaucous-winged Gulls and American Herring Gulls is relatively common) and subsequently migrated to Newfoundland and found an American Herring Gull breeding partner.

Population differentiation within the Glaucous Gull

As predicted, this study revealed weak population genetic differentiation between North

American and European colonies of Glaucous Gulls using several different methods. Pairwise

FST indicated significant differentation between Bjørnøya and Devil Island, and between

Bjørnøya and Karrak Lake, which are the most geographically distant pairs of colonies (Figure

2.2). Using the most sensitive model (using sampling sites as prior information, without genetic admixture), STRUCTURE was able to uncover clear differentiation between North American and European colonies of Glaucous Gull. Principal component analysis and principal coordinates analysis each displayed a single, diffuse cluster of samples, with North American populations on one side and European populations on the other and significant overlap between them, suggesting weak differentiation between North American and European populations. Finally, an isolation by distance analysis showed a positive but nonsignificant relationship between geographic and genetic distance between sampled colonies.

Historical association and contemporary gene flow

Previous researchers found differentiation between North American and European colonies of Glaucous Gulls in mitochondrial but not nuclear (AFLP) data (Liebers et al. 2004;

Sternkopf et al. 2010). Sternkopf et al. (2010) proposed that the observed mitochondrial divide

(biphyly) between North American and European Glaucous Gulls may be due to historical associations between different species of gulls in shared glacial refugia. They hypothesized that

Glaucous Gulls shared a glacial refugium with other gulls with Clade II haplotypes (North

American species). After the Last Glacial Maximum, Glaucous Gulls with Clade II haplotypes may have expanded into Europe and interbred with European Herring Gulls (Clade I haplotypes), eventually experiencing a near complete mitochondrial replacement with Clade I haplotypes at

European colonies. According to this hypothesis, the mitochondrial biphyly is a more recent development after the Last Glacial Maximum, and that Clade II is the ancestral mitochondrial lineage of Glaucous Gulls. Unlike Sternkopf et al. (2010), my research revealed a North

American/European divergence in nuclear loci which may warrant reconsideration of the theory proposed by Sternkopf et al. (2010).

If the North American/European divergence in nuclear loci reflects historical separation of Glaucous Gull colonies, North American Glaucous Gulls may have shared a glacial refugium with gulls with Clade II haplotypes, while European Glaucous Gulls shared a glacial refugium with gulls with Clade I haplotypes. However, if the North American/European differentiation in nuclear loci instead reflects a more contemporary lack of gene flow between North American and European Glaucous Gull colonies, population genetic structure at nuclear loci may have arisen after Clade II Glaucous Gulls expanded into Europe and underwent a mitochondrial replacement with Clade I haplotypes. The program BayesAss can be used to estimate contemporary migration rates and may have helped uncover whether this North

American/European divergence in Glaucous Gulls reflects historical or contemporary population genetic structure. Unfortunately, BayesAss is unable to accurately estimate migration rates if the proportion of migrants in the population is greater than 1/3 (Meirmans 2013). Therefore, despite convergence in BayesAss runs, and similar estimates of migration and residency rate between runs of different seed values, BayesAss was unable to accurately estimate the migration rate between North American and European colonies of Glaucous Gulls.

Without an accurate method to estimate contemporary migration between North

American and European colonies of Glaucous Gull, my study is unable to resolve whether the observed differentiation between North America and Europe reflects contemporary population

65 dynamics or historical associations between Glaucous Gulls in shared glacial refugia. Future work using Approximate Bayesian Computation (ABC) modelling may be able to distinguish between contemporary and historical causes of differentiation.

Private alleles can be an indicator of the uniqueness of a population (Barbosa et al. 2018).

Private alleles largely reflect mutations arising in a population and, therefore, large numbers of private alleles occurring at a high frequency in a population may suggest that the population has been genetically isolated for an extended time. Karrak Lake, Nunavut and Bjørnøya, Norway had similar numbers of private alleles, and Devil Island, Nunavut had the fewest. The greatest number of private alleles in Karrak Lake and Bjørnøya were found in one individual per population while at Devil Island, the greatest number of private alleles was found in two individuals. These results may suggest some restrictions to gene flow between the colonies but barriers to migration may not be very strong. An analysis of isolation by distance was not significant, however, there was a trend towards greater genetic distance between more distant colonies. Genetic and geographic distance may be positively correlated but small numbers of sampling sites may reduce the power to detect this relationship. Glaucous Gulls are mid-range migrants and so distance alone between the sampled colonies may reduce gene flow sufficiently to produce genetic differentiation (Weiser and Gilchrist 2020). Glaucous Gulls are also philopatric both to their birth location and to nesting locations where they have bred successfully in previous years (Portenko 1972; Gaston et al. 2009). Philopatry can reduce gene flow because adult Glaucous Gulls do not frequently switch to new breeding locations.

Selection pressures

Because Glaucous Gulls have such a large range, they likely do not experience the same selective pressures across their range in either the breeding or non-breeding season. Glaucous

Gulls in western Iceland and south-western Greenland are year-long residents, however, most

Glaucous Gulls breed in the Arctic and migrate to lower latitudes during the winter (Weiser and

Gilchrist 2020). Migratory Glaucous Gulls may breed as far north as Ellesmere Island in North

America and the north coast of Scandinavia in Europe, but winter as far south as North Carolina in North America and western France in Europe (Weiser and Gilchrist 2020). The temperature, precipitation, predators, prey availability, prey type, hunting pressure, and proximity to human settlements likely differ throughout the breeding and nonbreeding season across the sampled colonies. Despite the wide range of potential selective pressures across the sampled Glaucous

Gull colonies, no loci were identified as being putatively under either divergent or balancing selection by BayeScan. Since ddRAD sequencing only subsamples the genome, this result may simply reflect the low coverage of the genome that ddRAD sequencing provides.

Subspecies designations

Population genetic differentiation between North American and European colonies does not completely align with current subspecies designations. Glaucous Gulls breeding in Norway are designated as L. h. hyperboreus while Glaucous Gulls breeding in North America and Iceland are L. h. leuceretes. Subspecies designations are based on regional morphological differences between populations; for instance, L. h. leuceretes are typically larger and have a paler mantle than L. h. hyperboreus (Banks 1986). In this study, Icelandic Glaucous Gulls (Bjarnarhafnarfjall) genetically group with the breeding birds in Norway (Bjørnøya) rather than with the colonies in

North America. This result may suggest that Glaucous Gulls in Iceland are more genetically similar to L. h. hyperboreus individuals despite their morphological similarities to L. h. leuceretes individuals. However, since only four Glaucous Gulls from Iceland

(Bjarnarhafnarfjall) were retained through data filtering, further research may be necessary to

67 determine to which subspecies Icelandic Glaucous Gulls are more genetically similar at nuclear loci. Previous work by Vigfúsdóttir et al. (2008) using mtDNA and microsatellites showed that

Icelandic Glaucous Gulls group genetically with Europe rather than with Canada. They also found that populations sampled in eastern Greenland were more genetically similar to Canadian rather than Icelandic Glaucous Gull colonies (Vigfúsdóttir et al. 2008). This may suggest that the location of the geographic divide between the North American and European genetic populations is between Greenland and Iceland. Sonsthagen et al. (2012), however, found that Glaucous Gulls in Greenland grouped genetically with those in Svalbard, and that Icelandic Glaucous Gulls formed their own genetic cluster.

Population genetic differentiation between North American and European colonies has been found in other species within the Herring Gull species complex as well as other seabird species.

Sonsthagen et al. (2012) observed a divide in microsatellite loci between the eastern and western

Atlantic in Herring Gulls. However, this result may simply reflect the taxonomy used in their sampling. Sonsthagen et al. (2012) pooled Vega Gulls (Larus vegae), American Herring Gulls and European Herring Gulls under the Herring Gull species designation. Genetic differentiation between North America and Europe has also been found in other North Atlantic seabird species such as Common Murres (Morris-Pocock et al. 2008) and Razorbills (Alca torda) (Moum and

Árnason 2001). This shared pattern of population genetic differentiation between European and

North American colonies of seabirds is often attributed to glacial refugia during the Pleistocene.

Conservation implications

Hybridization

This research aligns with current species boundaries between Glaucous Gulls, American

Herring Gulls, and Glaucous-winged Gulls. However, STRUCTURE, principal component analysis, principal coordinates analysis and NewHybrids were unable to accurately identify

European Herring Gull as a species distinct from Glaucous Gull, and pairwise FST was significant but low between Glaucous Gull and European Herring Gull. This result may suggest that European Herring Gulls are more closely related to Glaucous Gulls than to American

Herring Gulls. Unfortunately, only six European Herring Gull samples were retained through data filtering, and all samples were from Iceland where European Herring Gulls have only recently become established breeders. European Herring Gulls breeding in Iceland may be less genetically differentiated from Glaucous Gulls due to frequent hybridization between these two species and introgression of Glaucous Gull alleles into European Herring Gulls in Iceland

(Vigfúsdóttir et al. 2008). Further work should be done to determine the genetic relationship between Glaucous Gulls and European Herring Gulls using a greater number of samples from a wider geographic range.

Although hybridization may not currently pose a conservation concern to colonies of

Glaucous Gulls north of the American Herring Gull breeding range, American Herring Gulls may expand their range northward as climate change impacts the Arctic (Brommer et al. 2012).

Increasing air and water temperatures through the summer may allow American Herring Gulls to breed in progressively higher latitudes and increase their breeding range overlap with Glaucous

Gulls. For instance, climate change may be responsible for the northward expansion and population growth of Lesser Black-backed Gulls in Greenland (Boertmann and Frederiksen

2016). However, Boertmann and Frederiksen (2016) also studied European Herring Gull populations and discovered that they have not significantly expanded their range in Greenland since the first breeding pair was reported in 1986 (Boertmann and Frederiksen 2016). A northward range shift could bring more American Herring Gulls into contact with colonies of

Glaucous Gulls and other high Arctic gulls and lead to more hybridization between these species.

Hybridization can be beneficial through the introduction of novel genetic diversity but it can also lead to introgression of maladaptive variation, loss of local adaptation, and outbreeding depression (Allendorf et al. 2001). Previous research has shown that introgression between

Glaucous Gulls and European Herring Gulls in Iceland is asymmetrical: a greater proportion of

Glaucous Gull mtDNA haplotypes introgressed into European Herring Gulls than vice versa

(Vigfúsdóttir et al. 2008). This is likely due to Glaucous Gulls being well-established in Iceland while European Herring Gulls have only recently begun breeding there. A similar pattern of asymmetric introgression could emerge if American Herring Gulls expand northward and hybridize with Glaucous Gulls more extensively in the Canadian Arctic.

When managing a species that hybridizes as readily as Glaucous Gulls do, it is important to consider that hybridization can be natural or anthropogenically induced. Allendorf et al.

(2001) suggested a framework for hybrid management with six hybrid categories that distinguish between natural and anthropogenic hybridization, and between hybridization without introgression, occasional hybridization leading to some introgression, widespread introgression, and complete admixture. Across their range, populations of Glaucous Gulls may fall into Type 2

(natural introgression), Type 3 (natural hybrid zone) and Type 5 (widespread introgression due to human interference), depending on which species they hybridize with, how frequently hybridization occurs, and if these hybrid crosses are anthropogenically induced or naturally

70 occurring. Genetic and demographic associations between Glaucous Gulls and other members of the Herring Gull species complex are complicated and management of Glaucous Gulls should reflect that.

My research strengthens the suggestion that hybrid gulls are challenging to identify morphologically. All the samples used in this analysis were identified as ‘pure’ members of their respective species based on their morphology. Due to the frequency of hybridization between

Glaucous Gulls and American Herring Gulls in Canada, future genomic work on either species should attempt to identify and remove hybrids before assessing population differentiation, as accidental inclusion of hybrids could bias results.

Population genetic structure

Due to the lack of genetic differentiation between Canadian colonies at neutral loci, Glaucous

Gulls may be best organized under a single management unit (MU) in Canada. However, since this study did not identify any putatively adaptive SNPs, this research alone should not be used to inform adaptive units (AU) or evolutionarily significant units (ESU) under the framework proposed by Funk et al. (2012). Furthermore, ecological information should be considered when determining appropriate conservation units. Crandall et al. (2000) suggested using measures of both genetic and ecological exchangeability to determine if populations were truly demographically and genetically distinct. Without ecological data, my study cannot assess ecological exchangeability between Glaucous Gulls in North America and Europe using traits such as life history, morphology, and habitat as Crandall et al. (2000) suggested. Glaucous Gulls in Europe and North America are considered separate subspecies but this division is due mainly to differences in morphology and this may not be sufficient for Glaucous Gulls in North America and Europe to be considered ecologically non-exchangeable (Banks 1986).

Before designating all North American Glaucous Gull colonies under a single conservation unit, Glaucous Gulls sampled from Alaska and the western Canadian Arctic need to be analyzed. A previous study using microsatellite loci showed weak evidence for two genetic clusters in North America: one from south-western Alaska and one from northern Alaska, and

Baffin Island (Sonsthagen et al. 2012). Morphological differences between Alaskan Glaucous

Gulls and central Canadian Glaucous Gulls has led to their designation as separate subspecies (L. h. barrovianus, and L. h. leuceretes respectively). Larus h. barrovianus are smaller in size than

L. h. leuceretes and have the darkest mantle of all four subspecies (Banks 1986). Some seabird species show population genetic differentiation between eastern and western Nunavut centered around the Victoria Strait region (e.g., Common Eider, Turner 2019). All sampling in the present study was east of Victoria Strait, Nunavut, and so this research cannot inform whether colonies of Glaucous Gull in Alaska and the western Canadian Arctic form a separate genetic group. A future study may support the results of Sonsthagen et al. (2012), reveal population genetic differentiation between Glaucous Gull east and west of the Victoria Strait, or uncover structure between the subspecies, L. h. barrovianus and L. h. leuceretes.

Future Directions

My work has identified Glaucous Gull x American Herring Gull hybrids through genetic methods, and revealed that population genetic structure exists within the Glaucous Gull. Due to the genetic similarity between Glaucous Gulls and American and European Herring Gulls, not all hybrid individuals may have been uncovered. In the future, I propose to experiment with some other methods of analysis, including hzar (Derryberry et al. 2014), gghybrid (Bailey 2020),

SplitsTree (Huson and Bryant 2006), pcadapt (Luu et al. 2016), Linear Discriminant Analysis

(LDA), and ConStruct (Bradburd et al. 2018) to attempt to differentiate between the closely- related species. Other methods may also uncover other hybrid individuals in the data. I also propose to use ABC modelling in the program ABCtoolbox (Wegmann et al. 2010) to attempt to distinguish between historical and contemporary drivers of population genetic differentiation in

Glaucous Gulls.

An increase in contemporary sampling would significantly improve the ability to produce high quality genomic data. The oldest samples included in this study were collected in 1965, and the most recent were collected in 2013. Just 40% of the samples were collected since 2000.

Samples collected prior to the invention of high-throughput sequencing technology may not have been stored under the right conditions to preserve large amounts of high quality DNA. Increasing the sampling effort and ensuring that blood or tissue samples are frozen after collection and then transferred to an ultra-cold (-70°C) freezer will reduce the number of samples that drop out of analyses after sequencing.

This study could also be improved by increasing sampling in several key geographic areas. All the samples from Greenland (Ittoqqortoormiit region) and Alaska (Kigigak Island, and

North Slope Borough region) were eliminated during data filtering. Incorporating samples from the Yukon and the Northwest Territories as well as more samples from Alaska would improve our understanding of population genetic structure in North America. Sampling colonies in

Alaska would also increase the likelihood of sampling Glaucous Gull x Glaucous-winged Gull hybrids.

Greenland represents a key sampling area because of its geographic placement between

North America and Europe. Colonies of Glaucous Gulls in Greenland may group genetically with either the European or North American genetic populations, or may represent a mix of both.

Since seabirds tend not to fly across large expanses of land or ice, Greenland could represent the geographic barrier to gene flow between North America and Europe. Previous work by

Vigfúsdóttir et al. (2008) showed that Glaucous Gulls in eastern Greenland grouped with North

American colonies, but Sonsthagen et al. (2012) found that Glaucous Gulls from the same region in eastern Greenland grouped with samples from Svalbard. However, neither of these studies used high-throughput sequencing methods. Greenland could also have both Glaucous Gull x

American Herring Gull, and Glaucous Gull x European Herring Gull hybrids.

Finally, including Iceland Gulls from high Arctic breeding colonies would uncover the prevalence of hybridization between Glaucous Gulls and Iceland Gulls. Despite large differences in body size between Glaucous Gulls and Iceland Gulls, these species have been observed hybridizing under experimental (Smith 1966) and natural conditions (Swarth 1934). Iceland

Gulls are challenging to sample during the breeding season due to the extremely high latitudes of their breeding sites. For that reason, Iceland Gulls have often not been included in research on the relationships among members of the Herring Gull species complex, or have had a very small number of samples included in analyses (Crochet et al. 2003; Liebers et al. 2004). However, research by Snell (1991) using allozymes revealed that Glaucous Gulls and Iceland Gulls sampled from Home Bay, Baffin Island, were more similar to each other than to members of their own species sampled from other colonies. Genetic data produced with high-throughput sequencing may be able to differentiate between Glaucous Gulls and Iceland Gulls and identify hybrid individuals.

Conclusion

To conclude, Glaucous Gulls have a significant impact on Arctic ecosystems. As facultative scavengers, Glaucous Gulls contribute to pest and disease control (Moleón et al.

2014). As apex predators, they represent a significant source of mortality for sympatric breeding seabirds. For instance, researchers estimated that ~13,000 Glaucous Gulls consumed 21,000

Emperor Goose (Anser canagicus), 34,000 Canada Goose (Branta canadensis), and 16,000

White-fronted Goose (Anser albifrons) goslings on the Yukon-Kuskokwim Delta, Alaska in a single breeding season (Bowman et al. 2004). Glaucous Gull is also a culturally significant species. Because they bioaccumulate and biomagnify high levels of environmental contaminants, they are ideal bioindicators to track contamination in the Arctic (Newman et al. 2007; Verreault et al. 2010). They are also one of six species selected to be monitored by all circumpolar countries under the Circumpolar Seabird Group of the Conservation of Arctic Flora and Fauna

(CAFF) program. Finally, Glaucous Gulls are a traditional food source for some Inuit communities and are harvested for eggs and meat across the circumpolar Arctic (Weiser and

Gilchrist 2020). Glaucous Gulls deserve an informed and proactive management strategy to protect this biologically and culturally important species for generations to come.

Because populations of Glaucous Gulls show only weak population genetic differentiation between Europe and North America, Canadian Glaucous Gulls may be best organized under a single management unit. Furthermore, managers should be cognizant of the interconnectedness of this complex of hybridizing species. Management decisions (e.g. protections, hunt and harvest allowances, culling) directed towards one species in the Herring

Gull species complex may impact closely related species due to the frequency of hybridization and morphological similarity between Larus species.

Literature Cited

Aljanabi, S.M., and I. Martinez. 1997. Universal and rapid salt-extraction of high quality

genomics DNA for PCR-based techniques. Nucleic Acids Research 25: 4692-4693.

Allard, K.A., H.G. Gilchrist, A.R. Breton, C.D. Gilbert, and M.L. Mallory. 2010. Apparent

survival of adult Thayer’s and Glaucous Gulls nesting sympatrically in the Canadian

High Arctic. Ardea 98: 43-50.

Allendorf, F.W., R.F. Leary, P. Spruell, and J.K. Wenburg. 2001. The problems with hybrids:

Setting conservation guidelines. Trends in Ecology and Evolution 16: 613-622.

Allendorf, F.W., G. Luikart, and S.N. Aitken. 2013. Conservation and the genetics of

populations. Hoboken: John Wiley & Sons.

Anderson, E.C. and E.A. Thompson. 2001. A model-based method for identifying species

hybrids using multilocus genetic data. Genetics 160: 1217-1229.

Anderson, E.C., R. Waples, and S. Kalinowski. 2008. An improved method for predicting the

accuracy of genetic stock identification. Canadian Journal of Fisheries and Aquatic

Sciences 65: 1475-1486.

Andrews, S. 2010. FastQC: A quality control tool for high throughput sequence data. Available

online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.

Arctic Marine Shipping Assessment 2009 Report. Arctic Council. Protection of the Arctic

Marine Environment (PAME) working group.

Atwood, T.C., C. Duncan, K.A. Patyk, P. Nol, J. Rhyan, M. McCollum, M.A. McKinney, A.M.

Ramey, C.K. Cerqueira-Cézar, O.C.H. Kwok, J.P. Dubey, and S. Hennager. 2017.

Environmental and behavioral changes may influence the exposure of an Arctic apex

predator to pathogens and contaminants. Scientific Reports 7:13193.

Bailey, R.I. 2020. gghybrid: R package for evolutionary analysis of hybrids and hybrid zones.

Zenodo. v 1.0.0.

Banks, R. C. 1986. Subspecies of the Glaucous Gull, Larus hyperboreus (Aves:

Charadriiformes). Proceedings of the Biological Society of Washington 99: 149-159.

Barbosa, S., F. Maestre, T.A. White, J. Paupério, P.C. Alves, and J.B. Searle. 2018. Integrative

approaches to guide conservation decisions: Using genomics to define conservation units

and functional corridors. Molecular Evolution 27: 3452-3465.

Barrera-Guzmán, A.O., A. Aleixo, M.D. Shawkey, and J.T. Weir. 2018. Hybrid speciation leads

to novel male secondary sexual ornamentation of an Amazonian bird. Proceedings of the

National Academy of Sciences 115: E218-E225.

Barry S. J. and T. W. Barry. 1990. Food habits of Glaucous Gull in the Beaufort Sea. Arctic 43:

43-49.

Becker, M., N. Gruenheit, M. Steel, C. Voelckel, O. Deusch, P.B. Heenan, P.A. McLenachan, O.

Kardailsky, J.W. Leigh, and P.J. Lockhart. 2013. Hybridization may facilitate in situ

survival of endemic species through periods of climate change. Nature Climate Change 3:

1039-1043.

Bertram, G.C.L., and D. Lack. 1933. Notes on the birds of Bear Island. Ibis 3: 283-301.

Bird Life International. 2018. Larus hyperboreus. The IUCN Red List of Threatened Species

2018: e.T22694343A132544122.

Bird Life International. 2020. Larus argentatus.

http://datazone.birdlife.org/species/factsheet/european-herring-gull-larus-argentatus.

Boertmann, D., and M. Frederiksen. 2016. Status of Greenland populations of Great Black-

backed Gull (Larus marinus), Lesser Black-backed Gull (Larus fuscus) and Herring Gull

(Larus argentatus). Waterbirds 39: 29-35.

Bowman, T.D., R.A. Stehn, and K.T. Scribner. 2004. Glaucous Gull predation of goslings on the

Yukon-Kuskokwim Delta, Alaska. The Condor 106: 288-298.

Bradburd, G.S., G.M. Coop, and P.L. Ralph. 2018. Inferring continuous and discrete population

genetic structure across space. Genetics 210: 33-52.

Brommer, J.E., A. Lehikoinen, and J. Valkama. 2012. The breeding ranges of central European

and Arctic species move poleward. PLOS One 7: e43648.

Buckman, A.H., R.J. Norstrom, K.A. Hobson, N.J. Karnovsky, J. Duffe, and A.T. Fisk. 2004.

Organochlorine contaminants in seven species of Arctic seabirds from northern Baffin

Bay. Environmental Pollution 128: 327-338.

Bustnes, J.O., V. Bakken, K.E. Erikstad, F. Mehlum, and J.U. Skaare. 2001. Patterns of

incubation and nest-site attentiveness in relation to organochlorine (PCB) contamination

in Glaucous Gulls. Journal of Applied Ecology 38: 791-801.

Bustnes, J.O., K.E. Erikstad, J.U. Skaare, V. Bakken, and F. Mehlum. 2003. Ecological effects of

organochlorine pollutants in the Arctic: A study of the Glaucous Gull. Ecological

Applications 13: 504-515.

Catchen, J., P. Hohenlohe, S. Bassham, A. Amores, and W. Cresko. 2013. Stacks: An analysis

tool set for population genomics. Molecular Ecology 22: 3124-3140.

Christiansen, J.S., M. Sparboe, B.-S. Sæther, and S.I. Siikavuopio. 2015. Thermal behaviour and

the prospect of an invasive benthic top predator onto the Euro-Arctic shelves. Diversity

and Distributions 21: 1004-1013.

Clarke, J. 2006. eBird Checklist: https://ebird.org/canada/checklist/S69679164. eBird: An online

database of bird distribution and abundance [web application]. eBird, Ithaca, New York.

Available: http://www.ebird.org.

Coates, D.J., M. Byrne, and C. Moritz. 2018. Genetic diversity and conservation units: Dealing

with the species-population continuum in the age of genomics. Frontiers in Ecology and

Evolution 6: 165.

Cooley, S.W., J.C. Ryan, L.C. Smith, C. Horvat, B. Pearson, B. Dale, and A.H. Lynch. 2020.

Coldest Canadian Arctic communities face greatest reductions in shorefast sea ice. Nature

Climate Change 10: 533-538.

Cornell Lab of Ornithology. 2019. All About Birds: Herring Gull. Cornell Lab of Ornithology,

Ithaca, New York. https://www.allaboutbirds.org/guide/Herring_Gull/maps-range.

Crandall, K.A., O.R.P. Bininda-Emonds, G.M. Mace, and R.K. Wayne. 2000. Considering

evolutionary processes in conservation biology. Trends in Ecology and Evolution 15:

290-295.

Crochet, P.-A., J.Z. Chen, J.-M. Pons, J.-D. Lebreton, P.D.N. Hebert, and F. Bonhomme. 2003.

Genetic differentiation at nuclear and mitochondrial loci among large white-headed gulls:

Sex-biased interspecific gene flow? Evolution 57: 2865-2878.

Crossman, C.A., L.G. Barrett-Lennard, and E.B. Taylor. 2014. Population structure and

intergeneric hybridization in harbour porpoises Phocoena phocoena in British Columbia,

Canada. Endangered Species Research 26: 1-12.

Danecek, P., A. Auton, G. Abecasis, C.A. Albers, E. Banks, M.A. DePristo, R. Handsaker, G.

Lunter, G. Marth, S.T. Sherry, G. McVean, R. Durbin, and 1000 Genomes Project

Analysis Group. 2011. The variant call format and VCFtools. Bioinformatics 27: 2156-

2158.

Dawson, J., L. Copland, M.E. Johnston, L. Pizzolatto, S.E. Howell, R. Pelot, L. Etienne, L.

Matthews, and J. Parsons. 2017. Climate change adaptation strategies and policy options

for Arctic shipping: A report prepared for Transport Canada. Ottawa, Canada.

Dawson, J., N. Carter, N. van Luijk, C. Parker, M. Weber, A. Cook, K. Grey, and J. Provencher.

2020. Infusing Inuit and local knowledge into the Low Impact Shipping Corridors: An

adaptation to increased shipping activity and climate change in Arctic Canada.

Environmental Science and Policy 105: 19-36.

Derryberry, E.P., G.E. Derryberry, J.M. Maley, and R.T. Brumfield. 2014. HZAR: Hybrid zone

analysis using an R software package. Molecular Ecology Resources 14: 652-663.

Dittmann, D.L., and S.W. Cardiff. 2005. Origins and identification of Kelp x Herring Gull

hybrids: The “Chandeleur” Gull. Birding 37: 266-276.

Dray, S., and A. Dufour. 2007. The ade4 package: Implementing the duality diagram for

ecologists. Journal of Statistical Software 22: 1-20.

Durner, G.M., K.L. Laidre, and G.S. York (eds.). 2018. Polar Bears: Proceedings of the 18th

working meeting of the IUCN/SSC Polar Bear Specialist Group, 7-11 June 2016,

Anchorage, Alaska. Gland, Switzerland and Cambridge, UK: IUCN.

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.

Elmhagen, B., D. Berteaux, R.M. Burgess, D. Ehrich, D. Gallant, H. Henttonen, R.A. Ims, S.T.

Killengreen, J. Niemimaa, K. Norén, T. Ollila, A. Rodnikova, A.A. Sokolov, N.A.

Sokolova, A.A. Stickney, and A. Angerbjörn. 2017. Homage to Hersteinsson and

Macdonald: Climate warming and resource subsidies cause red fox range expansion and

Arctic fox decline. Polar Research 36:sup1.

Emerson, K.J., C.R. Merz, J.M. Catchen, P.A. Hohenlohe, W.A. Cresko, W.E. Bradshaw, and

C.M. Holzapfel. 2010. Resolving postglacial phylogeography using high-throughput

sequencing. Proceedings of the National Academy of Sciences 107: 16196-16200.

Environment and Climate Change Canada. 2020. Canadian environmental sustainability

indicators: Canada’s conserved areas. Available at: www.canada.ca/en/environment-

climate-change/services/environmental-indicators/conservedareas.html.

Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of clusters of individuals

using the software STRUCTURE: A simulation study. Molecular Ecology 14: 2611-

2620.

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., and O.E. Gaggiotti. 2008. A genome scan method to identify selected loci appropriate

for both dominant and codominant markers: A Bayesian perspective. Genetics 180: 977-

993.

Francis, R.M. 2017. pophelper: An R package and web app to analyze and visualize population

structure. Molecular Ecology Resources 17: 27-32.

Frankel, O.H. 1970. Sir William Macleay memorial lecture 1970: Variation – The essence of

life. Proceedings of the Linnean Society of New South Wales 95: 158-169.

Frankham, R. 1996. Relationship of genetic variation to population size in wildlife. Conservation

Biology 10: 1500-1508.

Friesen, V.L., B.C. Congdon, H.E. Walsh, and T.P. Birt. 1997. Intron variation in marbled

murrelets detected using analyses of single-stranded conformational polymorphisms.

Molecular Ecology 6: 1047-1058.

Funk, W.C., J.K. McKay, P.A. Hohenlohe, and F.W. Allendorf. 2012. Harnessing genomics for

delineating conservation units. Trends in Ecology and Evolution 27: 489-496.

Gaston, A. J., S. Descamps, and H. G. Gilchrist. 2009. Reproduction and survival of Glaucous

Gulls breeding in an Arctic seabird colony. Journal of Field Ornithology 80: 135-145.

Gaston, A.J., M.L. Mallory, and H.G. Gilchrist. 2012. Populations and trends of Canadian Arctic

seabirds. Polar Biology 35: 1221-1232.

Gaston, A. J. and D. N. Nettleship. 1981. The Thick-billed Murres of Prince Leopold Island.

Ottawa: Canada Wildlife Service.

Gilchrist, H. G. and G. J. Robertson. 1999. Population trends of gulls and Arctic Terns nesting in

the Belcher Islands, Nunavut. Arctic 52: 325-331.

Gilchrist, H.G. 2001. Glaucous Gull (Larus hyperboreus). In: Poole, A., ed. The Birds of North

American online. Ithaca, New York: Cornell Lab of Ornithology.

Gill, F., D. Donsker and P. Rasmussen (eds.). 2020. IOC World Bird List v. 10.2. doi:

10.14344/IOC.ML.10.2.

Gillespie, G.D. 1985. Hybridization, introgression, and morphometric differentiation between

mallard (Anas platyhynchos) and grey duck (Anas superciliosa) in Otago, New Zealand.

The Auk 102: 459-469.

Gilpin, M.E., and M.E. Soulé. 1986. Minimum viable populations: The processes of population

extinction. Conservation Biology: The Science of Scarcity and Diversity (ed. M.E.

Soulé). Sinauer Associates, Sunderland, Massachusetts. Pg. 13-34.

Good, T.P., J.C. Ellis, C.A. Annett, and R. Pierotti. 2000. Bounded hybrid superiority in an avian

hybrid zone: Effects of mate, diet, and habitat choice. Evolution 54: 1774-1783.

Google Maps. 2020. International Cartographic Association.

Goudet, J. 2005. HIERFSTAT, a package for R to compute and test hierarchical F-statistics.

Molecular Ecology Resources 5: 184-186.

Grant, P.R. and B.R. Grant. 1992. Hybridization of bird species. Science 256: 193-197.

Grosser, S., J. Abdelkrim, J. Wing, B.C. Robertson, and N.J. Gemmell. 2017. Strong isolation by

distance argues for separate population management of endangered Blue Duck

(Hymenolaimus malacorhynchos). Conservation Genetics 18: 327-341.

Haldane, J.B.S. 1922. Sex ratio and unisexual sterility in hybrid animals. Journal of Genetics 12:

101-109.

Hayward, J. L. and N. A. Verbeek. 2020. Glaucous-winged Gull (Larus glaucescens), version

1.0. In: Birds of the World (S. M. Billerman, Editor). Cornell Lab of Ornithology, Ithaca,

NY, USA.

Heinl, S.C., and A.W. Piston. 2009. Birds of the Ketchikan Area, Southeast Alaska. Western

Birds 40: 54-144.

Hersteinsson, P. 2004. Tófa [Arctic Fox]. In: Íslensk spendýr [Icelandic mammals]. Vaka-

Helgafell, Reykjavík. Pg. 74-85.

Hof, A.R., R. Jansson, and C. Nilsson. 2012. Future climate change will favour non-specialist

mammals in the (sub)Arctics. PLOS One 7: e52574.

Hohenlohe, P.A., S.J. Amish, J.M. Catchen, F.W. Allendorf, and G. Luikart. 2011. Next-

generation RAD sequencing identifies thousands of SNPs for assessing hybridization

between rainbow and westslope cutthroat trout. Molecular Ecology Resources 11: 117-

122.

Huson, D.H. and D. Bryant. 2006. Application of phylogenetic networks in evolutionary studies.

Molecular Biology and Evolution 23: 254-267.

Ingólfsson, A. 1970. Hybridization of Glaucous Gulls Larus hyperboreus and Herring Gulls L.

argentatus in Iceland. Ibis 112: 340-362.

IPCC, 2014: Climate Change 2014: Synthesis Report. Contribution of working groups I, II, and

III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change.

[Core Writing Team, R.K. Pachauri, and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland.

Johnston, M., J. Dawson, and E. Stewart. 2019. Marine tourism in Nunavut: Issues and

opportunities for economic development in Arctic Canada. In: Koster R., D. Carson

(eds.) Perspectives on Rural Tourism Geographies. Geographies of Tourism and Global

Change. Springer, Cham. https://doi.org/10.1007/978-3-030-11950-8_7

Jombart, T. 2008. adegenet: A R package for the multivariate analysis of genetic markers.

Bioinformatics 24: 1403-1405.

Kalinowski, S.T. 2004. Counting alleles with rarefaction: Private alleles and hierarchical

sampling designs. Conservation Genetics 5: 539-543.

Kamvar, Z.N., J.C. Brooks, and N.J. Grünwald. 2015. Novel R tools for analysis of genome-

wide population genetic data with emphasis on clonality. Frontiers in Genetics 6: 208.

Lamichhaney, S., F. Han, M.T. Webster, L. Andersson, B.R. Grant, and P.R. Grant. 2018. Rapid

hybrid speciation in Darwin’s finches. Science 359: 224-228.

Lecaudey, L.A., U.K. Schliewen, A.G. Osinov, E.B. Taylor, L. Bernatchez, and S.J. Weiss.

2018. Inferring phylogenetic structure, hybridization and divergence times within

Salmonidae (Teleostei: Salmonidae) using RAD-sequencing. Molecular Phylogenetics

and Evolution 124: 82-99.

Li, H., and R. Durbin. 2010. Fast and accurate long-read alignment with Burrows-Wheeler

Transform. Bioinformatics, Epub. [PMID: 20080505]

Li, H., B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. Marth, G. Abecasis, R.

Durbin, and 1000 Genome Project Data Processing Subgroup. 2009. The sequence

alignment/map (SAM) format and SAMtools. Bioinformatics 25: 2078-2079.

Liebers, D., P. de Knijff, and A.J. Helbig. 2004. The herring gull complex is not a ring species.

Proceedings of the Royal Society B 271: 893-901.

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.

Luu, K., E. Bazin, and M.G.B. Blum. 2016. Pcadapt: An R package to perform genome scans for

selection based on principal component analysis. Molecular Ecology Resources 17: 67-

77.

Mactavish, B. 2005. eBird Checklist: https://ebird.org/checklist/S68352062. eBird: An online

database of bird distribution and abundance [web application]. eBird, Ithaca, New York.

Available: http://www.ebird.org.

Mallory, M. L. and H. G. Gilchrist. 2005. Marine birds of the Hell Gate Polynya, Nunavut,

Canada. Polar Research 24: 87-93.

McCormack, J.E., M.G. Harvey, B.C. Faircloth, N.G. Crawford, T.C. Glenn, and R.T.

Brumfield. 2013. A phylogeny of birds based on over 1500 loci collected by target

enrichment and high-throughput sequencing. PLOS One 8: e54848.

McNeely, J.A., K.R. Miller, W.V. Reid, R.A. Mittermeier, and T.B. Werner. 1990. Conserving

the world’s biological diversity. International Union for Conservation of Nature and

Natural Resources. Gland, Switzerland. Pg. 193.

Mehl. K.R., R.T. Alisauskas, K.A. Hobson, and D.K. Kellett. 2004. To winter east or west?

Heterogeneity in winter philopatry in a central-arctic population of King Eiders. The

Condor 106: 241-251.

Meirmans, P.G. 2014. Nonconvergence in Bayesian estimation of migration rates. Molecular

Ecology Resources 14: 726-733.

Moleón, M., J.A. Sánchez-Zapata, A. Margalida, M. Carrete, J.A. Donázar, and N. Owen-Smith.

2014. Humans and scavengers: The evolution of interactions and ecosystem services.

BioScience 64: 394-403.

Mora-Márquez, F., V. García-Olivares, B.C. Emerson, and U. López de Heredia. 2017.

ddRADseqTools: A software package for in silico simulation and testing of double-digest

RAD seq experiments. Molecular Ecology Resources 17: 230-246.

Moritz, C. 1994. Defining “evolutionarily significant units” for conservation. Trends in Ecology

and Evolution 9: 373-375.

Morris-Pocock, J.A., S.A. Taylor, T.P. Birt, M. Damus, J.F. Piatt, K.I. Warheit, 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.

Neubauer, G., M.M. Zagalska-Neubauer, J.-M. Pons, P.-A. Crochet, O. Chylarecki, A.

Przystalski, and L. Gay. 2009. Assortative mating without complete reproductive

isolation in a zone of recent secondary contact between Herring Gulls (Larus argentatus)

and Caspian Gulls (L. cachinnans). Auk 126: 409-419.

Neubauer, G., P. Nowicki, and M. Zagalska-Neubauer. 2014. Haldane’s rule revisited: Do hybrid

females have a shorter lifespan? Survival of hybrids in a recent contact zone between two

large gull species. Journal of Evolutionary Biology 27: 1248-1255.

Newman, S.H., A. Chmura, K. Converse, A.M. Kilpatrick, N. Patel, E. Lammers, and P. Daszak.

2007. Aquatic bird disease and mortality as an indicator of changing ecosystem health.

Marine Ecology Progress Series 352: 299-309.

O’Leary, S.J., J.B. Puritz, S.C. Willis, C.M. Hollenbeck, and D.S. Portnoy. 2018. These aren’t

the loci you’re looking for: Principles of effective SNP filtering for molecular ecologists.

Molecular Ecology 27: 3193-3206.

Pálsson, S., F. Vigfúsdóttir, and A. Ingólfsson. 2009. Morphological and genetic patterns of

hybridization of Herring Gulls (Larus argentatus) and Glaucous Gulls (L. hyperboreus)

in Iceland. Auk 126: 376-382.

Petersen, A. 1998. Íslenskir fuglar [Icelandic birds]. Vaka-Helgafell, Reykjavík. Pg. 312.

Petersen, A., D.B. Irons, H.G. Gilchrist, G.J. Robertson, D. Boertmann, H. Strøm, M. Gavrilo, Y.

Artukhin, D.S. Clausen, K.J. Kuletz, and M.L. Mallory. 2015. The status of Glaucous

Gulls Larus hyperboreus in the circumpolar Arctic. Arctic 68: 107-120.

Peterson, B.K., J.N. Weber, E.H. Kay, H.S. Fischer, 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: e37135.

Pierotti, R. 1987. Isolating mechanisms in seabirds. Evolution 41: 559-570.

Pongracz, J.D., D. Paetkau, M. Branigan, and E. Richardson. 2017. Recent hybridization

between a Polar Bear and Grizzly Bears in the Canadian Arctic. Arctic 70: 151-160.

Portenko, L. A. 1972. The Glaucous Gull, Larus hyperboreus. In: Birds of the Chukchi Peninsula

and Wrangel Island. Washington D.C.: Smithsonian Institute and the National Science

Fund.

Pritchard, J.K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using

multilocus genotype data. Genetics 155: 945-959.

Puechmaille, S.J. 2016. The program STRUCTURE does not reliably recover the correct

population structure when sampling is uneven: Subsampling and new estimators alleviate

the problem. Molecular Ecology Resources 16: 608-627.

R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for

Statistical Computing. Vienna, Austria. https://www.R-project.org/.

Rambaut, A., A.J. Drummond, D. Xie, G. Baele, and M.A. Suchard. 2018. Posterior

summarization in Bayesian phylogenetics using Tracer 1.7. Systematic Biology 67: 901-

904.

Randi, E. 2008. Detecting hybridization between wild species and their domesticated relatives.

Molecular Ecology 17: 285-293.

Raymond, M. and F. Rousset. 1995. GENEPOP (version 1.2): Population genetics software for

exact tests and ecumenicism. Journal of Heredity 86: 248-249.

Rhymer, J.M., and D. Simberloff. 1996. Extinction by hybridization and introgression. Annual

Review of Ecology, Evolution, and Systematics 27: 83-109.

Rohlf, F.J. 1972. An empirical comparison of three ordination techniques in numerical

taxonomy. Systematic Biology 21: 271-280.

Rousset, F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under

isolation by distance. Genetics 145: 1219-1228.

Rwahnih, M.A., A. Rowhani, N. Westrick, K. Stevens, A. Diaz-Lara, F.P. Trouillas, J. Preece, C.

Kallsen, K. Farrar, and D. Golino. 2018. Discovery of viruses and virus-like pathogens in

pistachio using high-throughput sequencing. Plant Disease 102: 1419-1425.

Sagerup, K., E.O. Henriksen, A. Skorping, J.U. Skaare, and G.W. Gabrielsen. 2000. Intensity of

parasitic nematodes increases with organochlorine levels in the Glaucous Gull. Journal of

Applied Ecology 37: 532-539.

Sambrook, J., E.F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A Laboratory Manual. 2nd

edition. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press.

Schuster, R., R.R. Germain, J.R. Bennett, N.J. Reno, and P. Arcese. 2019. Vertebrate

biodiversity on indigenous-managed lands in Australia, Brazil, and Canada equals that in

protected areas. Environmental Science and Policy 101: 1-6.

Schweyen, H., A. Rozenberg, and F. Leese. 2014. Detection and removal of PCR duplicates in

population genomic ddRAD studies by addition of a degenerate base region (DBR) in

sequencing adaptors. Biology Bulletin 227: 146-160.

Slatkin, M. 1991. Inbreeding coefficients and coalescence times. Genetics Research (Cambridge

Core) 58: 167-175.

Smith, N.G. 1966. Evolution of some Arctic gulls: An experimental study of isolating

mechanisms. Ornithological Monographs 4: 1-99.

Smith, W., and J.F. Grassle. 1977. Sampling properties of a family of diversity measures.

Biometrics 33: 283-292.

Smith, S. A., J. A. Whitney and G. I. Storm. 1994. Wildlife surveys in southwestern Foxe Basin,

1994. Ottawa, ON: Canada Parks Service.

Snell, R.R. 1991. Interspecific allozyme differentiation among North Atlantic white-headed

Larid gulls. Auk 108: 319-328.

Somanathan, S., C.P. Flynn and K.J. Szymanski. 2007. Feasibility of a sea route through the

Canadian Arctic. Maritime Economics and Logistics 9: 324-334.

Sonsthagen, S.A., R.T. Chesser, D.A. Bell, and C.J. Dove. 2012. Hybridization among Arctic

white-headed gulls (Larus spp.) obscures the genetic legacy of the Pleistocene. Ecology

and Evolution 2: 1278-1295.

Sonsthagen, S.A., R.E. Wilson, R.T. Chesser, J.-M. Pons, P.-A. Crochet, A. Driskell, and C.

Dove. 2016. Recurrent hybridization and recent origin obscure phylogenetic relationships

within the ‘white-headed’ gull (Larus sp.) complex. Molecular Phylogenetics and

Evolution 103: 41-54.

Spear, L.B. 1987. Hybridization of Glaucous and Herring Gulls at the Mackenzie Delta, Canada.

Auk 104: 123-125.

Stanley, R.R.E., N.W. Jeffery, B.F. Wringe, C. DiBacco, and I.R. Bradbury. 2016.

GENEPOPEDIT: A simple and flexible tool for manipulating multilocus molecular data

in R. Molecular Ecology Resources 17: 12-18.

Strang, C.A. 1977. Variation and distribution of Glaucous Gulls in Western Alaska. The Condor

79: 170-175.

Sternkopf, V., D. Liebers-Helbig, M.S. Ritz, J. Zhang, A.J. Helbig, and P. de Knijff. 2010.

Introgressive hybridization and the evolutionary history of the herring gull complex

revealed by mitochondrial and nuclear DNA. BMC Evolutionary Biology 10: 348-366.

Strang, C. A. 1976. Feeding behaviour and ecology of Glaucous Gulls in western Alaska. Ph.D.

thesis, Purdue University, Lafayette, Indiana.

Strøm, H. 2007. Bjørnøya. In: Anker-Nilssen, T., R.T. Barrett, J.O. Bustnes, K.E. Erikstad, P.

Fauchald, S.-H. Lorentsen, H. Steen, H. Strøm, G.H. Systad, and T. Tveraa. 2007.

SEAPOP studies in the Lofoten and Barents Sea area in 2006. NINA Report 249. Pg. 31-

32.

Swarth, H.S. 1934. Birds of Nunivak Island, Alaska. Pacific Coast Avifauna 22: 1-64.

Turner, R.R. 2019. Getting one’s ducks in a row: Conservation units of Common Eiders

(Somateria mollissima) throughout North America. Master’s thesis: Queen’s University.

Vendrami, D.L.J., L. Telesca, H. Weigand, M. Weiss, K. Fawcett, K. Lehman, M.S. Clark, F.

Leese, C. McMinn, H. Moore, and J.I. Hoffman. 2017. RAD sequencing resolves fine-

scale population structure in a benthic invertebrate: Implications for understanding

phenotypic plasticity. Royal Society Open Science 4: 160548.

Verhoeven, K.J., M. Macel, L.M. Wolfe, and A. Biere. 2011. Population admixture, biological

invasions and the balance between local adaptation and inbreeding depression.

Proceedings of the Royal Society B 278: 2-8.

Verreault, J., J.U. Skaare, B.M. Jenssen, and G.W. Gabrielsen. 2004. Effects of organochlorine

contaminants on thyroid hormone levels in Arctic breeding Glaucous Gulls, Larus

hyperboreus. Environmental Health Perspectives 112: 532-537.

Verreault, J., G.W. Gabrielsen, and J.O. Bustnes. 2010. The Svalbard Glaucous Gull as

bioindicator species in the European Arctic: Insight from 35 years of contaminants

research. Reviews of Environmental Contamination and Toxicology 205: 77-116.

Vigfúsdóttir, F., S. Pálsson, and A. Ingólfsson. 2008. Hybridization of glaucous gull (Larus

hyperboreus) and herring gull (Larus argentatus) in Iceland: Mitochondrial and

microsatellite data. Philosophical Transactions of the Royal Society B 363: 2851-2860.

Wang, J. 2019. A parsimonious estimator of the number of populations from STRUCTURE-like

analysis. Molecular Ecology Resources 19: 970-981.

Ware, C., J. Berge, A. Jelmert, S.M. Olsen, L. Pellissier, M. Wisz, D. Kriticos, G. Semenov, S.

Kwasniewski, and I.G. Alsos. 2016. Biological introduction risks from shipping in a

warming Arctic. Journal of Applied Ecology 53: 340-349.

Wegmann, D., C. Leuenberger, S. Neuenschwander, and L. Excoffier. 2010. ABCtoolbox: A

versatile toolkit for approximate Bayesian computation. BMC Bioinformatics 11: 116.

Weir, B.S. and C.C. Cockerham. 1984. Estimating F-statistics for the analysis of population

structure. Evolution 38: 1358-1370.

Weiser, E. and H. G. Gilchrist. 2020. Glaucous Gull (Larus hyperboreus), version 1.0. In: The

Birds of North America (S.M. Billerman, Editor). Cornell Lab of Ornithology, Ithaca,

NY, USA.

Wells, Z.R.R., T.A. Bernos, M.C. Yates, and D.J. Fraser. 2019. Genetic rescue insights from

population- and family-level hybridization effects in brook trout. Conservation Genetics

20: 851-863.

Wickham, H. 2016. ggplot2: Elegant graphics for data analysis. Springer-Verlag New York.

ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org.

Williamson, F.S.L., and L.J. Peyton. 1963. Interbreeding of Glaucous-winged and Herring Gulls

in the Cook Inlet Region, Alaska. Condor 65: 24-28.

Wilson, N.H. 1951. Hybrid Glaucous x Great Black-backed Gull at Limerick. British Birds 44:

286-287.

Wilson, G.A., and B. Rannala. 2003. Bayesian inference of recent migration rates using

multilocus genotypes. Genetics 163: 1177-1191.

Wringe, B.F., R.R.E. Stanley, N.W. Jeffery, E.C. Anderson, and I.R. Bradbury. 2017.

parallelnewhybrids: An R package for the parallelization of hybrid detection using

NEWHYBRIDS. Molecular Ecology Resources 17: 91-95.

Zarza, E., B.C. Faircloth, W.L.E. Tsai, R.W. Bryson Jr., J. Klicka, and J.E. McCormack. 2016.

Hidden histories of gene flow in highland birds revealed with genomic markers.

Molecular Ecology 25: 5144-5157.

Zhao, H., B. Beck, A. Fuller, and E. Peatman. 2020. EasyParallel: A GUI platform for

parallelization of STRUCTURE and NEWHYBRIDS analyses. PLOS ONE 15:

e0232110.