Foraging Ecology, Sexual Selection, and Divergence in Sunfish (Lepomis Spp.)

Western University Scholarship@Western

Electronic Thesis and Dissertation Repository

12-13-2013 12:00 AM

Foraging ecology, sexual selection, and divergence in sunfish (Lepomis spp.)

Scott F. Colborne The University of Western Ontario

Supervisor Bryan Neff The University of Western Ontario

Graduate Program in Biology A thesis submitted in partial fulfillment of the equirr ements for the degree in Doctor of Philosophy © Scott F. Colborne 2013

Follow this and additional works at: https://ir.lib.uwo.ca/etd

Part of the Biology Commons

Recommended Citation Colborne, Scott F., "Foraging ecology, sexual selection, and divergence in sunfish (Lepomis spp.)" (2013). Electronic Thesis and Dissertation Repository. 1764. https://ir.lib.uwo.ca/etd/1764

This Dissertation/Thesis is brought to you for free and open access by Scholarship@Western. It has been accepted for inclusion in Electronic Thesis and Dissertation Repository by an authorized administrator of Scholarship@Western. For more information, please contact [email protected]. FORAGING ECOLOGY, SEXUAL SELECTION AND DIVERGENCE IN SUNFISH (LEPOMIS SPP.)

(Thesis format: Integrated Article)

Scott Fitzgerald Colborne

Graduate Program in Biology

A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

The School of Graduate and Postdoctoral Studies The University of Western Ontario London, Ontario, Canada

Abstract

The origins of novel traits and their contribution to biodiversity have long been of interest to biologists. My research focused on the links between foraging ecology and both natural and sexual selection, and how these mechanisms interact to shape the phenotypic diversification of natural populations. Using bluegill (Lepomis macrochirus) and pumpkinseed sunfish (L. gibbosus), I examined three major questions: 1) how are diet and morphological variation related to alternative reproductive tactics in bluegill; 2) are sexual selection and disruptive selection driving divergence between foraging ecomorphs in pumpkinseed; and 3) how are human-induced changes to prey communities affecting foraging ecology of pumpkinseed. By examining the alternative reproductive tactics of bluegill, I found that the deeper bodied parental males and females consumed more zooplankton as compared to the littoral diets of the more streamlined sneakers, satellites, and juveniles, the opposite relationship from what is typically observed in fish. However, the morphological and swim performance variation among reproductive groups may increase mating success, indicating that sexual selection may be an important factor in the phenotypic diversification of bluegill. I next examined the relationship between stable isotopes (δ13C and δ15N) in different female tissues (white muscle, liver, and eggs), and found that from pre-fertilization to the day of hatch eggs could be used as a proxy for female diet. By collecting nest guarding males and eggs from their nests it is possible to estimate assortative mating based on foraging tactic. Applying this technique to a wild population of pumpkinseed, I found evidence of phenotypic divergence and assortative mating between foraging tactics. However, there was no evidence of neutral genetic divergence, possibly related to strength of disruptive selection, gen flow by

ii cuckolders, or selection on functional loci. Finally, I found that native pumpkinseed sunfish in three populations have undergone substantial shifts in diet following the invasion of zebra mussels (Dreissena polymorpha), consuming considerably more zebra mussels and fewer benthic invertebrates than before the invasion. These results contribute to a growing body of evidence that foraging ecology and its relationships with both disruptive natural and sexual selection play a key role in the evolution and diversification of natural populations.

Keywords

Assortative mating, disruptive selection, ecomorphology, evolution, foraging ecology, invasive species, morphology, sexual selection, stable isotopes, sunfish, swim performance, sympatric speciation

iii

Co-Authorship Statement

A version of Chapter 2 was published in the Canadian Journal of Fisheries and Aquatic

Sciences (68: 1802-1810, 2011) with Marie-Christine Bellemare, Pedro Peres-Neto, and

Bryan Neff as co-authors (see Appendix E for permission to reproduce material). Funding for this project was received from both B.D. Neff and P.R. Peres-Neto. M.C. Bellemare contributed to fish collections and swim performance measurements. All co-authors contributed to experimental design, sampling, and contributed editorial comments to the manuscript. S.F. Colborne was responsible for sampling protocol, fish collection, data analysis and served as the primary author of this publication.

A version of Chapter 3 was published in the journal Behavioral Ecology (24:1339-1347,

2013) with Pedro Peres-Neto, Fred Longstaffe, and Bryan Neff as co-authors (see Appendix

E for permission to reproduce material). Funding for the project was received by grants to

B.D. Neff, P.R. Peres-Neto, F.J. Longstaffe, and S.F. Colborne (see chapter acknowledgements for details of funding sources to each individual). B.D. Neff and P.R.

Peres-Neto contributed to the experimental design and editorial comments. F.J. Longstaffe contributed to sample analysis as part of a collaboration with the Laboratory for Stable

Isotope Science (LSIS) at The University of Western Ontario, as well as data analysis, and provided editorial comments. S.F. Colborne was the individual primarily responsible for experimental design, sample collection, analysis, and preparation of written materials.

A version of Chapter 4 was submitted to the Canadian Journal of Fisheries and Aquatic

Sciences (under review) with T.J.A. Hain, F.J. Longstaffe, and B.D. Neff as co-authors.

Funding for this project was received through grants to B.D. Neff, P.R., Peres-Neto, and S.F.

Colborne (see chapter for funding source details). T.J.A. Hain contributed to the

iv experimental design, experimental crossing of fish, and editorial comments on the manuscript. F.J. Longstaffe contributed to data analysis and provided editorial comments as part of a collaborative project with the LSIS research group (see chapter 3 above). B.D. Neff contributed to the experimental design and editorial comments of the manuscript.

Chapter 5 was completed in collaboration with S.R. Garner, F.J. Longstaffe, and B.D. Neff as co-authors and is currently under review at the journal Evolution. Funding for this project was received through grants to B.D. Neff, F.J. Longstaffe, and S.F. Colborne (see chapter for details). S.R. Garner contributed to the laboratory work and genetic analysis components of this manuscript, as well as providing editorial comments to the chapter. F.J. Longstaffe contributed to the data analysis and provided editorial comments on the chapter. B.D. Neff contributed to the experimental design, statistical analysis, and provided editorial comments on the chapter.

A version of Chapter 6 has been submitted to the journal Freshwater Biology for publication and was co-authored with A.M. Clapp, F.J. Longstaffe, and B.D. Neff. Funding for this research was provided to B.D. Neff, F.J. Longstaffe, and S.F. Colborne (see chapter for funding source details). A.M. Clapp completed an undergraduate thesis research project (The

University of Western Ontario) which contributed to the data included in this chapter and initial analyses of stomach contents and fish isotopic composition. F.J. Longstaffe contributed to the data analysis and provided editorial comments on this chapter. B.D. Neff contributed to the study design and editorial comments on this chapter.

Acknowledgments

First, I would like to thank Bryan Neff, who welcomed me into his lab in May 2009 and has taught me about the scientific process and who I would like to become as a researcher. I have also received guidance from many other faculty members at Western. Fred Longstaffe started as a member of my advisory committee and then welcomed me as a member of the interdisciplinary Laboratory for Stable Isotope Science (LSIS) research group. Chris

Guglielmo has provided valuable feedback about my research since my first committee meeting and frequently since that time. I also thank Pedro Peres-Neto, who made my first inter-institutional collaboration a success and who helped me develop the morphological analysis skills used multiple times in my research.

I spent almost six months of my Ph.D. studies at the Queen’s University Biological Station

(QUBS), a world-class facility for ecology research. Many people contributed to my field work, including Chandra Rodgers, Marie-Christine Bellemare, Tim Hain, Michael Roubakha,

Ainsley Furlonger, Adrianna Clapp, Adrienne Berchtold, Malcolm Lau, Aimee Houde, and many students at QUBS who just wanted to come out and learn more about my research. In the lab, I am especially thankful for the help and guidance of Kim Law who taught me more about the technical aspects of isotope analysis then I ever expected. With Kim’s help I never could have run approximately 2000 isotope samples over three years.

The past and present members of the Neff lab – Ross Breckels, Melissa Evans, Shawn

Garner, Kayla Gradil, Tim Hain, Adam Henkel, Aimee Houde, Malcolm Lau, John Loggie,

Nico Muñoz, Charlyn Partridge, Chandra Rodgers, and Jessica Van Zwol – are a special group of people that have made me proud to be a “Neff-ster”. A special thanks to Shawn

Garner and Charlyn Partridge who provided thorough feedback on multiple sections of this thesis, especially when I was questioning my writing ability. Also, Chandra Rodgers who survived four field seasons with me and has become part of my extended family. When I first started at Western, the Collip lunch table consisted of the Neff lab, since that time the group has expanded to include grad students from many labs. These people are not only the Collip

Lunch Group, but amazing friends. To all these people who have provided counsel and much laughter – especially Ainsley Furlonger, Drew Moore, Laura Sanchez, Kim Schmidt, and

Chris Austin – I couldn’t imagine a better group of people.

Last, but certainly not least, Alex and Patsy Colborne have been better parents then I often deserved, especially during the last weeks of writing this thesis. Once everything is said and done, they have certainly earned the right to claim a large portion of my Ph.D. for themselves.

This research was supported by multiple sources to the various collaborators involved in each research project, detailed information is provided in the individual data chapters. I am grateful for the support provided to me by Western Graduate Thesis Research Fund awards, a

Department of Biology Travel Award, the Queen Elizabeth II Graduate Scholarship in

Science and Technology (QEIIGSST), and Ontario Graduate Scholarship (OGS) award programs for their support.

“Never doubt that a small group of thoughtful, committed citizens can change

the world. Indeed, it’s the only thing that ever has.”

-Margaret Mead

vii

Table of Contents

Abstract ...... ii

Co-Authorship Statement...... iv

Acknowledgments...... vi

Table of Contents ...... viii

List of Tables ...... xiii

List of Figures ...... xiv

List of Appendices ...... xvi

List of Abbreviations ...... xvii

Chapter 1 ...... 1

1 General Introduction ...... 1

1.1 Intrasexual selection and phenotypic diversification ...... 3

1.2 Phenotypic diversification and sympatric speciation ...... 4

1.3 Changing foraging dynamics within ecosystems ...... 7

1.4 Sunfish study system...... 9

1.5 Assessing foraging ecology of fish ...... 11

1.6 Thesis Structure ...... 13

1.6.1 Alternative reproductive tactics in bluegill ...... 13

1.6.2 Assortative mating and divergence of pumpkinseed sunfish ...... 14

1.6.3 Changing foraging dynamics in freshwater lakes ...... 15

1.6.4 Summary ...... 16

1.7 References ...... 16

Chapter 2 ...... 27

2 Morphological and swim performance variation among reproductive tactics of bluegill sunfish (Lepomis macrochirus) ...... 27

viii

2.1 Introduction ...... 28

2.2 Methods...... 31

2.2.1 Fish collection ...... 31

2.2.2 Experimental procedures ...... 31

2.2.3 Statistical analysis ...... 35

2.3 Results ...... 37

2.3.1 Morphological variation...... 37

2.3.2 Swim performance ...... 42

2.3.3 Morphology and swim performance ...... 43

2.4 Discussion ...... 43

2.5 Chapter acknowledgements ...... 48

2.6 References ...... 49

Chapter 3 ...... 55

3 Effects of foraging and sexual selection on ecomorphology of a fish with alternative reproductive tactics ...... 55

3.1 Introduction ...... 56

3.2 Methods...... 59

3.2.1 Fish collection ...... 59

3.2.2 Morphological analysis ...... 60

3.2.3 Stable isotope analysis ...... 60

3.2.4 Statistical analysis ...... 63

3.3 Results ...... 65

3.3.1 Bluegill morphological variation ...... 65

3.3.2 Diet variation among bluegill morphs ...... 65

3.3.3 Morphology and body length ...... 69

3.3.4 Bluegill Morphology and Stable Isotopes...... 70

3.4 Discussion ...... 70

3.5 Chapter acknowledgements ...... 80

3.6 References ...... 81

Chapter 4 ...... 86

4 Covariation in stable isotope composition between females and their eggs in a fish: the potential for less invasive inference of resource use ...... 86

4.1 Introduction ...... 87

4.2 Methods...... 89

4.3 Results ...... 93

4.4 Discussion ...... 99

4.5 Chapter acknowledgements ...... 104

4.6 References ...... 105

Chapter 5 ...... 110

5 Assortative mating but no evidence of genetic divergence in a species characterized by a trophic polymorphism ...... 110

5.1 Introduction ...... 111

5.2 Methods...... 115

5.2.1 Fish collection ...... 115

5.2.2 Morphological variation...... 117

5.2.3 Diet analysis ...... 118

5.2.4 Morphology, diet, and condition ...... 120

5.2.5 Assortative mating and genetic differentiation ...... 121

5.3 Results ...... 123

5.3.1 Morphological variation between habitats ...... 123

5.3.2 Diet Analysis ...... 123

5.3.3 Morphology, diet, and condition ...... 129

5.3.4 Assortative mating and genetic differentiation ...... 131 x

5.4 Discussion ...... 131

5.5 Chapter Acknowledgements ...... 138

5.6 References ...... 138

Chapter 6 ...... 147

6 Foraging ecology of native pumpkinseed sunfish (Lepomis gibbosus) following the invasion of zebra mussels (Dreissena polymorpha) ...... 147

6.1 Introduction ...... 148

6.2 Methods...... 152

6.2.1 Fish collection ...... 152

6.2.2 Stomach content analysis ...... 153

6.2.3 Stable isotope analysis ...... 154

6.2.4 Statistical analysis ...... 154

6.3 Results ...... 156

6.3.1 Prey resources ...... 156

6.3.2 Pumpkinseed diet analysis ...... 157

6.4 Discussion ...... 161

6.5 Chapter acknowledgements ...... 169

6.6 References ...... 170

Chapter 7 ...... 177

7 General discussion ...... 177

7.1 Intrasexual selection and alternative reproductive tactics ...... 177

7.2 Sympatric divergence and speciation ...... 179

7.3 Divergence and alternative reproductive tactics ...... 181

7.4 Impacts of changing ecosystems ...... 182

7.5 Advances in stable isotope analysis techniques ...... 184

7.6 Concluding remarks ...... 187

7.7 References ...... 188

Appendices ...... 193

Curriculum Vitae ...... 228

xii

List of Tables

Table 2.1 Summary of the mean (±S.E.) values of fish measurements taken on bluegill (Lepomis macrochirus) classified into one of five groups...... 38

Table 3.1 Summary of the mean (± standard error) values of size, shape, and diet measurements for juveniles and four reproductive morphs of bluegill (Lepomis macrochirus)...... 67

Table 3.2 Stepwise regression and AIC model selection for predictors of diet measured by δ13C values among bluegill morphs (Lepomis macrochirus)...... 76

Table 4.1 Summary of linear regression analyses of tissue carbon and nitrogen compositions in bluegill (Lepomis macrochirus)...... 96

Table 4.2 Summary of SIAR two-source mixing model estimates of diet based on δ13C and δ15N stable isotope values of bluegill (Lepomis macrochirus)...... 100

Table 5.1 Summary of the stomach contents of pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats...... 126

Table 6.2 Summary of the stable isotopic compositions of pumpkinseed sunfish (Lepomis gibbosus)...... 1612

Table 6.3 Summary of pumpkinseed sunfish (Lepomis gibbosus) diet estimates based on SIAR two-source mixing models ...... 163

xiii

List of Figures

Figure 2.1 Illustration of a bluegill (Lepomis macrochirus) with the 20 homologous landmarks used in the morphological analysis comparing body shape among five bluegill groups (juveniles, females, parentals, sneakers, and satellites)...... 33

Figure 2.2 Discriminant function analysis (DFA) of body shape among five groups of bluegill (Lepomis macrochirus)...... 40

Figure 2.3 Logistic regression of total body length and female likeness scores of sneaker (○) and satellite (●) bluegill (Lepomis macrochirus)...... 41

Figure 2.4 Relationship between morphology and (a,c) burst swim or (b,d) swimming endurance of five bluegill (Lepomis macrochirus) groups...... 44

Figure 3.1 Discriminant function analysis (DFA) of body shape among five morphs of bluegill (Lepomis macrochirus): juveniles, sneakers, satellites, females, and parental males...... 66

Figure 3.2 Boxplots of δ13C values (per mil deviation of measured 13C:12C relative to Vienna Pee Dee Belemnite) among five morphs of bluegill (Lepomis macrochirus): juveniles, sneakers, satellites, females, and parental males...... 72

Figure 3.3 Relationship between morphology (DFA 1) and body length (mm) of juvenile and satellite male bluegill (Lepomis macrochirus)...... 73

Figure 4.1 Stable isotopic composition of eggs during embryo development in bluegill (Lepomis macrochirus)...... 94

Figure 4.2 Isotopic compositions of eggs compared to two maternal tissues in bluegill (Lepomis macrochirus)...... 98 xiv

Figure 4.3 SIAR stable isotope mixing model estimates of the pelagic contribution to the diet of female bluegill (Lepomis macrochirus) based on different tissues sampled ...... 101

Figure 5.2 Boxplots of SIAR isotope-mixing model estimates of the littoral prey resource contribution to the diets of male and female pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats ...... 128

Figure 5.3 Relationship between body condition and ecomorph scores of (a) littoral- and (b) pelagic-caught pumpkinseed sunfish (Lepomis gibbosus) collected from Ashby Lake, Ontario ...... 130

Figure 5.4 Relationship between body condition and ecomorph scores of (a) littoral and (b) pelagic-caught pumpkinseed sunfish (Lepomis gibbosus) collected from Ashby Lake, Ontario ...... 132

Figure 5.5 Genetic clustering of individual pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats...... 133

Figure 6.1 Isotopic compositions of pumpkinseed sunfish (Lepomis gibbosus) and potential invertebrate prey types...... 160

Figure 6.2 SIAR mixing model estimates of the proportion of zebra mussels (Dreissena polymorpha) in the diet of pumpkinseed sunfish (Lepomis gibbosus)...... 166

List of Appendices

Appendix A: Chapter 3 supplemental material ...... 193

Appendix B: Chapter 4 supplemental material ...... 201

Appendix C: Chapter 5 supplemental material ...... 205

Appendix D: Chapter 6 supplemental material ...... 215

Appendix E: Permission to reproduce published material ...... 225

Appendix F: Animal care protocol approval documentation ...... 226

xvi

List of Abbreviations

AIC: Akaike information criteria ANCOVA: analysis of covariance ANOVA: analysis of variance BCI: Bayesian credibility interval BD: body depth BL: body length CP: caudal peduncle CPDF: caudal peduncle depth factor DFA: discriminant function axis or discriminant function analysis FL: female-likeness score GSI: gonadosomatic index K: Fulton’s condition factor LSIS: Laboratory for Stable Isotope Science MANOVA: multivariate analysis of variance PCA: principal component analysis

Pp: parametric probability

Pnp: non-parametric probability rsp: relative swim performance SD: standard deviation SE: standard error of the mean SIAR: stable isotope analysis in R TEF: trophic enrichment factor VPDB: Vienna Pee Dee Belemnite

xvii 1

Chapter 1 1 General Introduction

Selection acting on foraging tactics was a key contributor to the evolution of the tremendous biological diversity that exists today. Indeed, many of the most iconic examples of evolution have arisen as a result of resource competition acting on novel foraging phenotypes (Svanbäck and Eklöv 2004, Bolnick and Lau 2008). For example, the long necks of giraffes (Giraffa camelopardalis) evolved in response to competition among species for access to leaves on trees (Cameron and Toit 2007). Individuals with long necks had higher fitness because they were better able to consume leaves from the upper parts of the canopy, albeit this advantage came at the expense of reduced foraging efficiency for low lying grasses and shrubs (Cameron and Toit 2007). Alternatively, selection within species to exploit different food resources can lead to substantial phenotypic diversification, as was the case for Darwin’s finches (subfamily Geospizinae) in the Galapagos Islands (Grant and Grant 2006). These finch species were all descended from a single common ancestor that colonized these islands approximately 2 million years ago (Sato et al. 2001), but now show significant variation in beak morphology that has arisen from specialization to different resource niches (Kleindorfer et al. 2006).

Similarly, differences in foraging phenotypes have led to adaptive radiations of cichlid fish in the African rift lakes into species flocks that consist of hundreds of identifiable species within a single lake. For example, Lakes Malawi and Victoria in east Africa each have at least 700 identified cichlid species (reviewed by Turner et al. 2001). The cichlid species flocks are associated with specialist foraging phenotypes that differ in head shape, mouth orientation, jaw size and tooth characteristics (Albertson et al. 2003). A

2 particularly illustrative example of foraging related phenotypic divergence are scale- eating cichlids that have their mouth directed towards either the left or right side of their head based on the direction from which they approach their prey, a balance between which is maintained by frequency-dependent selection (Hori 1993). Overall, selection on foraging tactics is a significant driver of phenotypic diversification and can be linked to the biodiversity observed around the globe as populations adapt to their local conditions.

Variation in foraging tactics is not only found among species, but also individuals within species. These variants are known as foraging ecomorphs, and are common in fishes, amphibians, reptiles, birds, and mammals (reviewed by Skúlason and Smith 1995,

Svanbäck et al. 2008). For example, in many lakes Arctic charr (Salvelinus alpinus) have up to four discrete ecomorphs that differ in both overall body size and shape related to different foraging tactics (Snorrason et al. 1994). In cases such as this, disruptive selection for different foraging tactics has led to phenotypic diversification within populations (e.g. Rueffler et al. 2006, Kirkpatrick and Ravigné 2002). Given the right conditions, the initial phenotypic diversification associated with foraging ecomorphs can lead to speciation, making this process of considerable interest in evolutionary biology because it may explain the origins of new species without requiring geographic isolation

(e.g. Hatfield and Schluter 1999, Bolnick and Fitzpatrick 2007).

Natural selection may be the most commonly identified source of phenotypic diversification within populations, but it is not the only source of selection on morphological traits (Turner and Burrows 1995, Pfennig and Pfennig 2009). Sexual selection based on competition for mating opportunities can also result in phenotypic

3 diversification (e.g. Gross 1980, Garant et al. 2003, Sternalski et al. 2012). Indeed, disruptive natural selection based on ecological traits and selection for traits that increase mating success have been referred to as “two sides of the same coin” due to their roles in phenotypic diversification (Bolnick and Doebeli 2003), but research has tended to examine these effects in isolation (Pfennig and Pfennig 2009). However, many questions remain about the relationships between natural and sexual selection mechanisms and their role in phenotypic diversification (Matthews et al. 2010, Bonduriansky 2011).

My PhD research focused on three primary areas related to the foraging ecology, phenotypic diversification, and divergence of wild fish populations. More specifically, I examine foraging variation associated with multiple reproductive tactics within a species, the role of sexual selection and disruptive natural selection on divergence between ecomorphs and speciation, and human-induced changes to the foraging ecology of wild populations. In the remainder of this chapter, I provide background information about phenotypic diversification through natural and sexual selection, the methods of assessing foraging ecology that were used in my research, and a brief description of each data chapter.

1.1 Intrasexual selection and phenotypic diversification

Sexual selection can result in stable phenotypic differences between alternative reproductive tactics. In this case, individuals of one sex, typically males, adopt different reproductive strategies to optimize their mating success. These strategies are often associated with differences in the behavioural and morphological traits used to access mating opportunities (Moczek and Emlen 2000, Taborsky et al. 2008, Corlatti et al.

2013). A frequently observed pattern among alternative reproductive tactics involves large-sized males that attempt to guard access to females and other relatively small-sized males that attempt to “steal” fertilization opportunities by sneaking past guarding males

(Taborsky et al. 2008). Examples of these alternative mating tactics have been reported in invertebrates (Moczek and Emlen 2000), fish (Garant et al. 2003), and mammals

(Schradin et al. 2012). Similar to sneaking, males may also steal fertilization opportunities by “mimicking” female phenotypes and behaviours to deceive guarding males (e.g. Gross 1980, Corl et al. 2009, Sternalski et al. 2012). Both sneaking and female mimicry are associated with behavioural and morphological differences within populations (e.g. Ehlinger et al. 1997). However, relatively little is known regarding how these alternative male reproductive tactics relate to foraging tactics. Given the strong relationships between morphological traits and foraging ecology, it is possible that individuals adopting different reproductive tactics may also differ in resource use.

Indeed, the evolution of different reproductive phenotypes may result in trade-offs or synergistic interactions between intrasexual selection and natural selection. However, these possible interactions are not yet well understood.

1.2 Phenotypic diversification and sympatric speciation

Phenotypic diversification has frequently been examined in relation to disruptive natural selection rather than sexual selection because of the number of taxa that show phenotypic divergence within populations related to resource use (e.g. Schluter and McPhail 1992,

Gavilets and Losos 2009, Grant and Grant 2011). Foraging-related disruptive selection may result in relatively subtle changes in phenotype (e.g. Robinson and Wilson 1996, Sih et al. 2004) or be associated with population divergence to the point of speciation (e.g.

Seehausen 2006, Grant and Grant 2011). Speciation is commonly associated with allopatric conditions under which populations become isolated by geographic barriers, such as mountain ranges or oceanic islands, and subsequently evolve different phenotypes through adaptation to local selection pressures or genetic drift (Mayr 1963,

Dieckmann et al. 2004). However, speciation can also occur under sympatric conditions, where there are no identifiable geographic barriers (Skúlason and Smite 1995, Pfennig et al. 2007). Under sympatric conditions, disruptive selection can result in multiple phenotypes within a single population, primarily due to positive selection for extreme traits and negative selection for those with intermediate values. This process can eventually lead to sympatric speciation if, for example, individuals with these phenotypes become spatially isolated as a result of resource specialization. Sympatric speciation was widely contested during the early and mid-20th century due to a lack of theoretical and empirical evidence (e.g. Mayr 1963, Paterson 1978). Since that period, evidence for sympatric divergence and speciation has accumulated from both empirical data (e.g. Via

2011, Bolnick 2011) and population genetic models (e.g. Turner and Burrows 1995,

Thibert-Plante and Hendry 2011) that now include mechanisms of both natural and sexual selection, rather than just disruptive natural selection.

Current models of sympatric speciation have increased their focus on the development of reproductive isolation which stops or greatly reduces gene flow between individuals with different phenotypes (e.g. Johnson et al. 1996, Bolnick and Fitzpatrick 2007, Thibert-

Plante and Hendry 2011). One way that this isolation can occur is through strong assortative mating behaviours. For example, African rift lake cichlids evolved different foraging phenotypes related to mouth shape and size, but they also use their mouths to

6 generate sounds to attract mates (Salzburger 2009). In this example, disruptive natural selection on foraging tactics were correlated to changes in the sounds that could be produced by males that contributed to mate choice behaviours of females and reduced the gene flow between phenotypes through assortative mating. Thus, disruptive selection can increase the frequency of novel phenotypes within a population, but divergence to the point of speciation cannot occur unless there is an associated mechanism of reproductive isolation limiting gene flow between different phenotypes.

Disruptive natural selection based on foraging tactics has been identified as a key contributor to the early stages of divergence within populations (e.g. Robinson et al.

2000, Taylor and McPhail 2000). For example, threespine stickleback (Gasterosteus aculeatus) have repeatedly diverged between the shallow littoral and deeper pelagic habitats of freshwater lakes in British Columbia (Canada), which differ in the resources available (Lavin and McPhail 1985, Taylor and McPhail 2000). The divergence of stickleback into littoral and pelagic ecomorphs is the result of disruptive selection that decreases the fitness of intermediate foraging phenotypes (Bolnick 2004, Bolnick and

Lau 2008). However, it is recognized that for sympatric speciation to occur in these stickleback populations there must also be some form of assortative mating that reduces gene flow between them, which has now been tested in two populations of polymorphic stickleback (Snowberg and Bolnick 2008, 2012). As such, both natural and sexual selection are now both considered important components of sympatric speciation models

(e.g. Ritchie 2007, Schluter 2009, Thibert-Plante and Hendry 2011).

In summary, many studies of wild populations have established the presence of phenotypic diversification based on resource use (e.g. Skúlason and Smith 1995), but empirical evidence for assortative mating between foraging ecomorphs is still relatively rare in comparison because the potential role of sexual selection is a relatively recent addition to sympatric models and can be logistically challenging to examine in wild populations because of the temporal nature of reproductive periods (e.g. Snowberg and

Bolnick 2008, 2012, Martin 2013). In order to alleviate this issue, novel methods of testing for assortative mating between foraging ecomorphs must be developed and implemented in wild populations.

1.3 Changing foraging dynamics within ecosystems

Phenotypic diversification based on foraging tactic is likely to persist only if selection pressures remain relatively stable over long periods of time. If resource availability changes, the direction of selection is likely to be altered as well. For example, Darwin’s finches (Geospiza spp.) on Daphne Major (Galapagos Islands) experienced an El Niño event during 1982-1983 that dramatically decreased the availability of large seeds, one of the major food resources consumed by the finches (Grant and Grant 1993). As a result, individuals adapted to foraging on large seeds were removed from the population, and these foraging traits were not reestablished until large seeds became available again and the island was colonized by large-beaked individuals from a nearby island (Grant and

Grant 1993). The phenotypic diversity of this island population was greatly altered by environmental variation in resource availability and its effects on foraging related selection pressures. Similarly, freshwater ecosystems are experiencing a number of disturbances caused by humans, which have the potential to alter resource availability in

8 multiple ways. Climate change has the long-term potential to alter aquatic ecosystems through its effects on water temperature and chemistry, plus multiple secondary impacts on aquatic food webs (reviewed by Ficke et al. 2007). However, the movement and introduction of hundreds of species to non-native environments represents an immediate threat to freshwater food webs (Strayer 2009, 2010). Introduced species that establish themselves often lack natural predators or competitors that can naturally limit their abundances, and consequently invasive species can drastically alter ecosystems over relatively short periods of time (Griffiths 1991, Magoulick and Lewis 2002).

Of particular concern to the stability of ecosystems in North America are two invasive molluscs – zebra (Dreissena polymorpha) and quagga mussels (Dreissena rostriformis bugensis). These mussels were unintentionally introduced to North America through ballast water releases starting in the late 1980s (Brown and Stepien 2010). Unlike native mussels, invasive zebra mussels attach to hard substrates, including rocks, industrial pipes, docks and boats, causing extensive damage (Connelly et al. 2007, Brown and

Stepien 2010). The incalculable ecological impact of these mussels may represent the most severe consequences of these invasive species. Zebra mussels drastically alter the physical features of lake habitats, for example, by providing a predation refuge to many native benthic invertebrates (e.g. Zhu et al. 2006). Zebra mussels also feed on phytoplankton and small zooplankton suspended in the water column (Strayer 2010), consuming such large quantities that there are marked changes in the composition and overall abundance of pelagic phytoplankton following their establishment in lakes (Zhu et al. 2006, Strayer 2009, 2010). These shifts in the invertebrate community have subsequent effects on the resource base available for all species at higher trophic levels.

As a result of shifting resource bases, the foraging-related selection pressures within ecosystems may also change, having a variety of possible effects on the phenotypic variation observed currently and the evolutionary trajectory of phenotypic diversification within these ecosystems.

1.4 Sunfish study system

The Centrarchid family of fishes endemic to North America provides an excellent system to study natural and sexual selection in relation to phenotypic diversification and divergence. In particular, the sunfish genus (Lepomis spp.) consists of 15 currently recognized species that exhibit a wide range of foraging and reproductive tactics (Collar and Wainwright 2009, Neff and Knapp 2009).

Multiple sunfish exhibit alternative reproductive tactics among males, three of which exhibit all three tactics (parentals, sneakers, and satellite) (Neff and Knapp 2009).

“Parental” male sunfish build nests, court females, and provide sole parental care for the young until the fry leave the nest (Gross 1982). Alternatively, sunfish males can adopt a cuckolder reproductive life history that starts with the males maturing at a small size and stealing fertilization opportunities by “sneaking” into the nest of parental males (Gross and Charnov 1980, Gross 1982). As sneakers grow, they switch to a “satellite” tactic, which involves mimicry of female behaviour and appearance, presumable because larger cuckolders have decreased success sneaking into nests (Gross 1982).

The three sunfish species that exhibit these alternative reproductive strategies include bluegill (Lepomis macrochirus; Gross 1979, Gross 1982), longear (L. megalotis;

Keenleyside 1972), and pumpkinseed (L. gibbosus; Neff and Clare 2008). However, the

10 majority of the 30 years of research on sunfish mating systems has been conducted on bluegill and, in particular, the population in Lake Opinicon (Ontario, Canada) (e.g. Gross

1982, Knapp and Neff 2007, Rodgers et al. 2013). Previous studies of bluegill reproductive tactics have examined the behaviours, sperm characteristics, and hormones associated with each reproductive tactic (e.g. Dominey 1980, Neff et al. 2003, Rodgers et al. 2012). As a result of past research efforts, the role of intrasexual selection in the diversification of behaviours and phenotypes among reproductive tactics are well established, but the relationship between these reproductive phenotypes and their foraging tactics are not well understood. This is primarily the result of how studies of foraging ecology have traditionally viewed fish populations. In prior analyses of bluegill foraging ecology, “adult” fish were sampled, which consisted of only parental males and mature females with alternative reproductive phenotypes ignored (e.g. Keast 1978,

Deacon and Keast 1987). As a result, the foraging ecology of cuckolder males is not well understood. Thus, the first step to understanding possible interactions between sexual and natural selection in this system is to document the ecological and phenotypic variation between the reproductive tactics.

In addition to the alternative reproductive tactics found in sunfish, there are also three sunfish species that exhibit trophic polymorphisms: pumpkinseed (e.g. Robinson et al.

2000, Gillespie and Fox 2003, Jastrebski and Robinson 2004), bluegill (e.g. Ehlinger and

Wilson 1988, Ehlinger 1990, Ellerby and Gerry 2011), and orangespotted (Lepomis humilus; Hegrenes 2001). Most studies of trophic polymorphisms in sunfish have focused on pumpkinseed and bluegill, which show evidence of distinct foraging ecomorphs that specialize on littoral or pelagic resources. Fish feeding in the pelagic habitat have a more

11 streamlined body shape compared to littoral fish, which tend to have deeper body shapes

(Svanbäck and Eklöv 2002, 2003, Ojanguren and Braña 2003). These resource-related patterns of morphological variation have been linked to swim performance trade-offs that optimize foraging efficiency of each ecomorph. Streamlined bodies with deeper caudal- peduncle tail regions are associated with increased burst swim speeds compared to greater maneuverability and sustained swimming associated with a deeper body shape

(Webb 1984, Walker 1997). The phenotypic diversification between littoral and pelagic resources is typically attributed to disruptive selection related to foraging efficiency and resource competition.

Foraging ecomorphs within sunfish are generally consistent with the morphological patterns of variation reported across fish species with different foraging tactics (e.g.

Jastrebski and Robinson 2004, Weese et al. 2012). The differential phenotypes between sunfish foraging ecomorphs has been described as a component of the early stages of divergence and sympatric speciation (e.g. Jastrebski and Robinson 2004, McCairns and

Fox 2004, Ellerby and Gerry 2011). However, the potential for assortative mating within these ecomorphs has not been examined. Without examining whether reproductive traits limit gene flow between these morphs it is not possible to know if the second major component of current sympatric speciation models is being met.

1.5 Assessing foraging ecology of fish

There are several techniques that can be used to examine foraging ecology of wild populations. In my research I focused on a combination of stomach content and stable isotope analyses. Stomach contents, on one hand, provide accurate estimates of diet over

12 short periods of time, but cannot provide accurate long-term measurements or data during periods when fish alter their foraging patterns (Pinnegar and Polunin 1999; MacNeil et al.

2005). Stable isotope analysis, on the other hand, provide estimates of resource use over longer periods by measuring the ratios of different isotopes incorporated from diet into the tissues of organisms, most commonly carbon (12C, 13C) and nitrogen (14N, 15N)

(DeNiro and Epstein 1981, Pinnegar and Polunin 1999, Post 2002).

Stable isotopes are commonly used to study foraging ecology because they provide inferences about resource use over longer time scales than stomach contents alone

(reviewed by Boecklen et al. 2011). Carbon isotope ratios (δ13C) are frequently used to infer foraging patterns of freshwater fish because δ13C values vary predictably between the littoral and pelagic habitats due to differences in the primary producers (France 1995).

In addition, δ13C changes little between a consumer and its resources, typically < 1 ‰ with each trophic level (DeNiro and Epstein 1981, Inger and Bearhop 2008). In contrast, nitrogen isotopes (δ15N) increase predictably from one trophic level to the next, typically between 3 and 5 ‰, providing estimates of trophic position (Post 2002, Layman et al.

2012). As a result of the unique characteristics of carbon versus nitrogen isotopic ratios, we are able to simultaneously estimate direct links between consumers and primary producers and the trophic position of individuals.

Measuring the isotopic composition of different tissue types provides estimates of resource use over different time scales related to the metabolic activity of each specific tissue. Most tissues in living organisms remain metabolically active throughout the lifetime of an individual and will experience isotopic turnover at a rate dependent on the

13 tissue’s cell turnover rate. Tissues with high metabolic activity may turnover in days (e.g. blood plasma), while others take weeks (e.g. liver), or several months (e.g. white muscle)

(Perga and Gerdeaux 2005, Guelinckx et al. 2007). Thus, it is possible to estimate resource use over a variety of temporal periods using isotopic composition measurements of a variety of tissues from an individual.

1.6 Thesis Structure

The five data chapters in this thesis (Chapters 2 – 6) were prepared as independent units for publication. In addition to assessing the general patterns of foraging in sunfish populations, it was the goal of my doctoral research to expand our knowledge of natural and sexual selection in wild populations and develop techniques that can be used to incorporate foraging ecology into studies of selection mechanisms. Two of the data chapters have been published (Chapters 2 & 3) and the other three are currently under review (Chapters 4 – 6).

1.6.1 Alternative reproductive tactics in bluegill

The first two data chapters focus on morphological, swim performance, and diet variation among alternative male reproductive tactics in a population of bluegill sunfish. Chapter 2

(“Morphological and swim performance variation among reproductive tactics of bluegill sunfish (Lepomis macrochirus)”; Colborne et al. 2011) focuses on the relationship between morphology and swim performance among five bluegill groups (parental males, sneakers, satellites, females, and juveniles). This chapter focuses primarily on the functional relationships between body shape and swim performance under experimental conditions, providing the first comparisons of these traits in all five of these bluegill

14 groups. The morphological and swim performance comparisons presented here provide a base point from which ecological differences between reproductive tactics could also be examined.

Chapter 3 (“Effects of foraging and sexual selection on ecomorphology of a fish with alternative reproductive tactics”; Colborne et al. 2013) focuses on the possible trade-offs between natural and sexual selection by examining morphology and diet of the five bluegill groups described above. Using the general framework of ecomorphology which relates fish body shape to diet, I expected that if natural selection based on foraging tactic was the primary determinant of phenotype, then streamlined fish should consume pelagic resources and deeper-bodied fish should consume littoral benthic invertebrates. Deviation from these patterns would indicate a different form of selection, such as sexual selection, influencing phenotypic variation among reproductive tactics.

1.6.2 Assortative mating and divergence of pumpkinseed sunfish

For divergence and speciation to occur in sympatry there must be some form of reproductive barrier to reduce gene flow between sub-groups (e.g. foraging ecomorphs).

One potential mechanism is positive assortative mating. However, obtaining direct evidence for assortative mating among fish can be logistically challenging. Thus, the development of reliable methods to infer assortative mating will increase our ability to examine reproductive patterns in wild populations that may be under divergent selection.

In Chapter 4 (“Covariation in stable isotope composition between females and their eggs in a fish: the potential for less invasive inference of resource use”; Colborne et al. in review), I describe an experiment that tested the use of egg isotopic compositions to infer

15 the diet of females over the egg development period. Not only does the sampling of eggs have the potential to provide a technique of inferring female diet that has greatly reduced impact on the overall fitness of the female, but this technique may also be used to test for assortative mating based on foraging tactics. Given that male sunfish guard nests for extended periods of time, while females mate and then leave the area, it is relatively easy to collect males to assess diet from body tissues, but collecting females represents a challenge. However, if the eggs deposited in the nest by a female provide a reliable estimate of the female diet, then we have a method by which to infer female diet and to test for assortative mating.

Chapter 5 (“Assortative mating but no evidence of genetic divergence in a species characterized by a trophic polymorphism”; Colborne et al. in review), details an examination of foraging ecomorphs in a frequently studied population of pumpkinseed sunfish with littoral and pelagic foraging ecomorphs. I tested for assortative mating and genetic differentiation between foraging ecomorphs in a population of sunfish thought to be in the early stages of sympatric speciation.

1.6.3 Changing foraging dynamics in freshwater lakes

In Chapter 6 (“Foraging ecology of native pumpkinseed sunfish (Lepomis gibbosus) following the invasion of zebra mussels (Dreissena polymorpha)”; Colborne et al. in review), I detail a study to examine how zebra mussels may be altering the foraging habits of native pumpkinseed sunfish in three freshwater lakes near Kingston (Ontario,

Canada) and discuss the broader implications for freshwater ecosystems.

1.6.4 Summary

In Chapter 7, I discuss the implications of my thesis research in relation to our understanding of how natural and sexual selection interact with each other and how they may influence phenotypic diversification. Additionally, I discuss the potential impact of ecological changes caused by the introduction of non-native species on phenotypic diversification and divergence of native species. I also present directions for future research that can further explore the interactions between sexual and natural selection with respect to foraging ecology.

1.7 References

Albertson, R.C., J.T. Streelman, and T.D. Kocher. 2003. Directional selection has shaped the oral jaws of Lake Malawi cichlid fishes. Proceed. Natl. Acad. Sci. 100:5252- 5257.

Boecklen, W.J., C.T. Yarnes, B.A. Cook, and A.C. James. 2011. On the use of stable isotopes in trophic ecology. Ann. Rev. Ecol. Evol. Syst. 42:411-440.

Bolnick, D.I. 2004. Can intraspecific competition drive disruptive selection? An experimental test in natural populations of sticklebacks. Evolution 58:609-618.

Bolnick, D.I. 2011. Sympatric speciation in threespine stickleback: why not? Int. J. Ecol. 2011:Article ID 942847. doi: 10.1155/2011/942847.

Bolnick, D.I., and M. Doebeli. 2003. Sexual dimorphism and adaptive speciation: two side of the same ecological coin. Evolution 57:2433-2449.

Bolnick, D.I., and B.M. Fitzpatrick. 2007. Sympatric speciation: models and empirical evidence. Ann. Rev. Ecol. Evol. Syst. 38:459-487.

Bolnick, D.I., and O.L. Lau. 2008. Predictable patterns of disruptive selection in stickleback in postglacial lakes. Am. Nat. 172:1-11.

Bonduriansky, R. 2011. Sexual selection and conflict as engines of ecological diversification. Am. Nat. 178:729-745.

Brown, J.E., and C.A. Stepien. 2010. Population genetic history of the dreissenid mussel invasions: expansion patterns across North America. Biol. Invasions 12:3687-3710.

Cameron, E.Z., and J.T. du Toit. 2007. Winning by a neck: tall giraffes avoid competing with shorter browsers. Am. Nat. 169:130-135.

Collar, D.C., and P.C. Wainwright. 2009. Ecomorphology of centrarchid fishes. In Centrarchid fishes: diversity, biology and conservation. Edited by S.J. Cooke and D.P. Philipp. Wiley-Blackwell, West Sussex, U.K. pp. 70-89.

Connelly, N.A., C.R. O’Neil Jr., B.A. Knuth, and T.L. Brown. 2007. Economic impacts of zebra musssels on drinking water treatment and electric power generation facilities. Environ. Manage. 40:105-112.

Corl, A., A.R. Davis, S.R. Kuchta, T. Comendant, and B. Sinervo. 2009. Alternative mating strategies and the evolution of sexual size dimorphism in the side-blotched lizard, Uta stansburiana: a population-level comparative analysis. Evolution 64:79- 96.

Corlatti, L., B. Bassano, T.G. Valencak, and S. Lovari. 2013. Foraging strategies associated with alternative reproductive tactics in a large mammal. J. Zool. 291:111-118

Deacon, L.I., and J.A. Keast. 1987. Patterns of reproduction in two populations of pumpkinseed sunfish, Lepomis gibbosus, with differing food resources. Environ. Biol. Fish. 19:281–296.

DeNiro, M.J., and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta 45:341-351.

Dieckmann, U., M. Doebeli, J.A.J. Metz, and D. Tautz. 2004. Adaptive Speciation. Cambridge University Press, Cambridge USA.

Dominey, W.J. 1980. Female mimicry in male bluegill sunfish – a genetic polymorphism? Nature 284:546 – 548.

Ehlinger, T.J., and D.S. Wilson. 1988. Complex foraging polymorphism in bluegill sunfish. Proc. Natl. Acad. Sci. USA 85:1878-1882.

Ehlinger, T.J., M.R. Gross, and D.P. Philipp. 1997. Morphological and growth rate difference between bluegill males of alternative reproductive life histories. North Am. J. Fish. Man. 17:533-542.

Ellerby, D.J., and S.P. Gerry. 2011. Sympatric divergence and performance trade-offs of bluegill ecomorphs. Evol. Biol. 38:422-433.

Ficke, A.D, C.A. Myrick, and L.J. Hansen. 2007. Potential impacts of global climate change on freshwater fisheries. Rev. Fish. Biol. Fisheries 17:581-613.

France, R.L. 1995. Differentiation between littoral and pelagic food webs in lakes using stable carbon isotopes. Limnol. Oceanogr. 40:1310-1313.

Garant, D., J.J. Dodson, and L. Barnatchez. 2003. Differential reproductive success and heritability of alternative reproductive tactics in wild Atlantic salmon (Salmo salar L.). Evolution 57:1133-1141.

Gavrilets, S., and J.B. Losos. 2009. Adaptive radiation: contrasting theory with data. Science 323:732-737.

Gillespie, G.L., and M.G. Fox. 2003. Morphological and life-history differentiation between littoral and pelagic forms of pumpkinseed. J. Fish Biol. 62:1099-1115.

Grant, B.R., and P.R. Grant. 1993. Evolution of Darwin’s finches caused by a rare climatic event. Proc. R. Soc. Lond. B 251:111-117.

Grant, P.R., and B.R. Grant. 2006. Evolution of character displacement in Darwin’s finches. Science 313:224-226.

Grant, P.R., and B.R. Grant. 2011. How and why species multiply: the radiation of Darwin’s finches. Princeton University Press, Princeton USA.

Griffiths R.W., Schloesser D.W., Leach J.H. & Kovalak W.P. (1991) Distribution and dispersal of the zebra mussel (Dreissena polymorpha) in the Great Lakes region. Can. J. Fish. Aquat. Sci. 48:1381-1388.

Gross, M.R. Cuckoldry in sunfishes (Lepomis: Centrarchidae). Can. J. Zool. 57:1507- 1509.

Gross, M.R. 1980. Sexual selection and the evolution of reproductive strategies in sunfish (Lepomis: Centrarchidae). Ph.D. thesis, University of Utah.

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z. Tierpsychol. 60:1-26.

Gross, M.R., and E.L. Charnov. 1980. Alternative male life histories in bluegill sunfish. Proc. Natl. Acad. Sci. USA. 77:6937-6940.

Guelinckx, J., J. Maes, P. Van Den Driessche, B. Geysen, F. Dehairs, and F. Ollevier. 2007. Changes in δ13C and δ15N in different tissues of juvenile sand goby Pomatoschistus minutus: a laboratory diet-switch experiment. Mar. Ecol. Prog. Ser. 341:205-215.

Hatfield, T., and D. Schluter. 1999. Ecological speciation in sticklebacks: environment- dependent hybrid fitness. Evolution 53:866-873.

Hegrenes, S. 2001. Diet-induced phenotypic plasticity of feeding morphology in the orangespotted sunfish, Lepomis humilis. Ecol. Freshwat. Fish 10:35-42.

Hori, M. 1993. Frequency-dependent natural selection in the handedness of scale-eating cichlid fish. Science 260:216-219.

Inger, R. and S. Bearhop. 2008. Applications of stable isotope analyses to avian ecology. Ibis 150:447-461.

Jastrebski, C.J., and B.W. Robinson. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

Johnson, P.A., F.C. Hoppensteadt, J.J. Smith, and G.L. Bush. 1996. Conditions for sympatric speciation: a diploid model incorporating habitat fidelity and non-habitat assortative mating. Evol. Ecol. 10:187-205.

Keast, A. 1978. Trophic and spatial interrelationships in the fish species of an Ontario temperate lake. Environ. Biol. Fish. 3:7-31.

Keenleyside, M.H. 1972. Intraspecific intrusions into nests of spawning longear sunfish (Pisces: Centrarchidae). Copeia 1972:272-278.

Kirkpatrick, M., and V. Ravigné. 2002. Speciation by natural and sexual selection: models and experiments. Am. Nat. 159:S22-S35.

Kleindorfer, S., T.W. Chapman, H. Winkler, and F.J. Sulloway. 2006. Adaptive divergence in contiguous populations of Darwin’s small ground finch (Geospiza fulignosa). Evol. Evol. Res. 8:357-372.

Knapp, R., and B.D. Neff. 2007. Steroid hormones in bluegill, a species with male alternative reproductive tactics including female mimicry. Biol. Lett. 3:628-631.

Lavin, P.A., and J.D. McPhail. 1985. The evolution of freshwater diversity in the threespine stickleback (Gasterosteus aculeatus): site-specific differentiation of trophic morphology. Can. J. Zool. 63:2632-2638.

Layman, C.A.,M.S. Araujo, R. Boucek, C.M. Hammerschlag-Peyer, E. Harrison, Z.R. Jud, P. Matich, A.E. Rosenblatt, J.J. Vaudo, L.A. Yeager, D.M. Post, and S. Bearhop. 2012. Applying stable isotopes to examine food-web structure: an overview of analytical tools. Biol. Rev. 87:545-562.

MacNeil, M.A., G.B. Skomal, and A.T. Fisk. 2005. Stable isotopes from multiple tissues reveal diet switching in sharks. Mar. Ecol. Prog. Ser. 302:199-206.

Magoulick D.D. & Lewis L.C. (2002) Predation on exotic zebra mussels by native fishes: effects on predator and prey. Freshwater Biol. 47:1908-1918.

Martin, C.H. 2013. Strong assortative mating by diet, color, size, and morphology but limited progress toward sympatric speciation in a classic example: Cameroon crater lake cichlids. Evolution 67:2114-2123.

Matthews, B., K.B. Marchinko, D.I. Bolnick, and A. Mazumder. 2010. Specialization of trophic position and habitat use by sticklebacks in an adaptive radiation. Ecology 91:1025-1034.

Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge, U.S.A.

McCairns, R.J.S, and M.G. Fox. 2004. Habitat and home range fidelity in a trophically dimorphic pumpkinseed sunfish (Lepomis gibbosus) population. Oecologia 140:271-279.

Moczek, A.P., and D.J. Emlen. 2000. Male horn dimorphism in the scarab beetle, Onthophagus Taurus: do alternative reproductive tactics favour alternative phenotypes? Anim. Behav. 59:459-466.

Neff, B.D. 2003. Sperm investment and alternative mating tactics in bluegill sunfish (Lepomis macrochirus). Behav. Ecol. 14:634-641.

Neff, B.D., and E.L. Clare. 2008. Temporal variation in cuckoldry and paternity in two sunfish species (Lepomis spp.) with alternative reproductive tactics. Can. J. Zool. 86:92-98.

Neff, B.D., and R. Knapp. 2009. Alternative reproductive tactics in the Centrarchidae. In Centrarchid fishes: diversity, biology and conservation. Edited by S.J. Cooke and D.P. Philipp. Wiley-Blackwell, West Sussex, U.K. pp. 70-89.

Ojanguren, A.F., F. and Braña. 2003. Effects of size and morphology on swim performance in juvenile brown trout (Salmo trutta L.). Ecol. Fresh. Fish. 12:241- 246.

Paterson, H.E.H. 1978. More evidence against speciation by reinforcement. S. Afr. J. Sci. 74:369-371.

Perga, M.E., and D. Gerdeaux. 2005. ‘Are fish what they eat’ all year round? Oecologia 144:598-606.

Pfennig, K.S., and D.W. Pfennig. 2009. Character displacement: ecological and reproductive responses to a common evolutionary problem. Quart. Rev. Biol. 84:253-276.

Pfennig, D.W., A.M. Rice, and R.A. Marting. 2007. Field and experimental evidence for competition’s role in phenotypic divergence. Evolution 61:257 – 271.

Pinnegar, J.K., and N.V.C. Polunin. 1999. Differential fractionation of δ13C and δ15N among fish tissues: implications for the study of trophic interactions. Funct. Ecol. 13:225-231.

Post, D.M. 2002. Using stable isotopes to estimate trophic positions: models, methods, and assumptions. Ecology 83:703-718.

Ritchie, M.G. 2007. Sexual selection and speciation. Annu. Rev. Ecol. Evol. Syst. 38:79- 102.

Robinson, B.W., and D.S. Wilson. 1996. Genetic variation and phenotypic plasticity in a trophically polymorphic population of pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. 10:631-652.

Robinson, B.W., D.S. Wilson, and A.S. Margosian. 2000. A pluralistic analysis of character release in pumpkinseed sunfish (Lepomis gibbosus). Ecology 81:2799- 2812.

Rodgers, C.M.C., B.D. Neff, and R. Knapp. 2012. Effects of exogenous testosterone on parental care behaviours in male bluegill sunfish (Lepomis macrochirus). Ethology 118:636-643.

Rodgers, C.M.C., B.D. Neff, and R. Knapp. 2013. Androgen-mediated nurturing and aggressive behaviours during parental care in bluegill sunfish (Lepomis macrochirus). Horm. Behav. 63:454-461.

Rueffler, C., T.J.M. Van Dooren, and J.A.J. Metz. 2006. The evolution of resource specialization through frequency-dependent and frequency-independent mechanisms. Am. Nat. 167:81-93.

Salzburger, W. 2009. The interaction of sexually and naturally selected traits in the adaptive radiation of cichlid fishes. Mol. Ecol. 18:169-185.

Sato, A., H. Tichy, C. O’hUigin, P.R. Grant, B.R. Grant, and J. Klein. 2001. On the origin of Darwin’s finches. Mol. Biol. Evol. 18(3):299-311.

Schluter, D. 2009. Evidence for ecological speciation and its alternative. Science 323:737-741.

Schluter, D., and J.D. McPhail. 1992. Ecological character displacement and speciation in sticklebacks. Am. Nat. 140:85-108.

Schradin, C., S. Eder, and K. Müller. 2012. Differential investment into testes and sperm production in alternative male reproductive tactics of the African striped mouse (Rhabdomys pumilio). Hor. Behav. 61:686-695.

Seehausen, O. 2006. African cichlid fish: a model system in adaptive radiation research. Proc. R. Soc. B. 273:1987-1998.

Sih, A., A. Bell, and J.C. Johnson. 2004. Behavioral syndromes: an ecological and evolutionary overview. Trends Ecol. Evol. 19:372-378.

Skúlason, S., and T.B. Smith 1995. Resource polymorphisms in vertebrates. Trends Ecol. Evol. 10:366 – 370.

Snorrason, S.S., S. Skúlason, B. JOnsson, H.J. Malmquist, P.M. Jonasson, O.T. Sandlund, and T. Lindem. 1994. Trophic specialization in Arcitc charr Salvelinus alpinus (Pisces; Salmonidae): morphological divergence and ontogenetic niche shifts. Biol. J. Linn. Soc. 52:1-18.

Snowberg, L.K., and D.I. Bolnick. 2008. Assortative mating by diet in a phenotypically unimodal by ecologically variable population of stickleback. Am. Nat. 172:733- 739.

Snowberg, L.K., and D.I. Bolnick. 2012. Partitioning the effects of spatial isolation, nest habitat, and individual diet in causing assortative mating within a population of threespine stickleback. Evolution 66:3582-3594.

Sternalski, A., F. Mougeot, and V. Bretagnolle. 2012. Adaptive significant of permanent female mimicry in a bird of prey. Biol. Lett. 8:167-170.

Strayer, D.L. 2009. Twenty years of zebra mussels: lessons from the mollusk that made headlines. Front. Ecol. Environ. 7:135-141.

Strayer, D.L. 2010. Alien species in fresh waters: ecological effects, interaction with other stressors, and prospects for the future. Freshwater Biol. 55:152-174.

Svanbäck, R. and P. Eklöv. 2002. Effect of habitat and food resources on morphology and ontogenetic growth trajectories in perch. Oecologia. 131:61-70.

Svanbäck, R., and P. Eklöv. 2003. Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization? Oikos 102:273-284.

Svanbäck, R., and P. Eklöv. 2004. Morphology in perch affects habitat specific foraging efficiency. Func. Ecol. 18:503-510.

Svanbäck, R., P. Eklöv, R. Fransson, and K. Holmgren. 2008. Intraspecific competition drives multiple species resource polymorphism in fish communities. Oikos 117:114 – 124.

Taborsky, M., R.F. Oliveira, and H.J. Brockmann. 2008. The evolution of alternative reproductive tactics: concepts and questions. In: Oliveira, R.F., M. Taborsky, and H.J. Brockmann, eds. Alternative reproductive tactics. Cambridge University Press, Cambridge, M.A.

Taylor, E.B., and J.D. McPhail. Historical contingency and ecological determinism interact to prime speciation in sticklebacks, Gasterosteus. Proc. R. Soc. Lond. B. 267:2375-2384.

Thibert-Plante, X., and A.P. Hendry. 2011. Factors influencing progress toward sympatric speciation. J. Evol. Biol. 24:2186-2196.

Turner, G.F., and M.T. Burrows. 1995. A model of sympatric speciation by sexual selection. Proc. R. Soc. Lond. B. 260:287-292.

Turner, G.F., O. Seehausen, M.E. Knight, C.J. Allender, and R.L. Robinson. 2001. How many species of cichlid fishes are there in African lakes? Mol. Ecol. 10:793-806.

Via, S. 2001. Sympatric speciation in animals: the ugly duckling grows up. Trends Ecol. Evol. 16:381-390.

Walker, J.A. 1997. Ecological morphology of lacustrine threespine stickleback Gasterosteus aculeatus L. (Gasterosteidae) body shape. Biol. J. Linn. Soc. 61:3- 50.

Webb, P.W. 1984. Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24:107-120.

Weese, D.J., Ferguson, M.M., and B.W. Robinson. 2012. Contemporary and historical evolutionary processes interact to shape patterns of within-lake phenotypic

divergences in polyphonic pumpkinseed sunfish, Lepomis gibbosus. Ecol. Evol. 2:574-592.

Zhu, B., D.G. Fitzgerald, C.M. Mayer, L.G. Rudstam, and E.L. Mills. 2006. Alteration of ecosystem function by zebra mussels in Oneida Lake: impacts of submerged macrophytes. Ecosystems 9:1017-1028.

Chapter 2 2 Morphological and swim performance variation among reproductive tactics of bluegill sunfish (Lepomis macrochirus)1

Ecomorphology examines the relationship between morphology and ecological characteristics often in relation to foraging, predation, and habitat use. However, ecomorphology may also be linked to reproductive behaviour (“tactic”), but few studies have examined this relationship. I examined bluegill sunfish (Lepomis macrochirus), a species where some males become “parentals” while others adopt a parasitic “cuckolder” tactic. Parentals build nests, court females and care for the young. Cuckolders instead act as “sneakers”, darting into nests while females are releasing eggs, and then transition to

“satellites”, mimicking female appearance. I hypothesized that reproductive tactic would be associated with morphological variation and swimming performance. I collected bluegill parentals, sneakers, satellites, females and juveniles to compare morphology, burst swim, and swim endurance. I found significant morphological variation among the groups, with only satellites and females having similar body shapes. Interestingly, satellites did not overlap in shape with sneakers, despite representing a single ontogenetic life history, providing evidence for a relationship between reproductive tactic and morphology. I also found that swim performance varied among the groups, with sneakers

1 A version of this chapter has been published in the Canadian Journal of Aquatic Sciences and is presented here with permission.

Citation: Colborne, S.F., Bellemare, M.C., Peres-Neto, P.D., and Neff, B.D. 2011. Morphological and swim performance variation among reproductive tactics of bluegill sunfish (Lepomis macrochirus). Canadian Journal of Fisheries and Aquatic Sciences. 68: 1802 – 1810 doi: 10.1139/F2011-091

28 having the fastest burst swim and longest swim endurance. My results indicated that reproductive tactic is an important factor in the ecomorphology of fish.

2.1 Introduction

Ecomorphology is a field concerned with describing the covariation of morphological features and ecological factors within and among species (e.g., Winkler 1988, Motta et al.

1995, Aerts et al. 2000). Ecomorphology has been examined in a wide range of taxa, including birds, fish, reptiles, and mammals. These studies have been able to relate form and function while gathering insights into the ecological pressures that drive morphological variation within ecosystems. Most studies to date have focused on the relationship between morphological characteristics and foraging patterns (e.g.

Wainwright 1996, Bertrand et al. 2008, Figueirido et al. 2009), predator interactions (e.g.,

Hambright 1991), hormones (e.g., Moore et al. 1998, Knapp et al. 2003), or habitat use

(e.g., Schluter and Rambaut 1996, de Mederios and da Costa Ramos 2007, Velasco and

Herrel 2007).

In fish, one area of ecomorphology that is often examined is the covariation between body shape and swimming performance in relation to different ecological niches.

Swimming has been the focus of many studies of fishes because swimming is arguably the most important aspect of their fitness (Collar and Wainwright 2009), being essential for evading predators through to foraging efficiently. Additionally, the viscosity of their environment (i.e. water) results in unique demands for locomotion compared to terrestrial organisms. Swimming performance is tightly linked to morphology based on two general principles. First is the ability to generate thrust, which in many fishes is related to the

29 depth of the caudal peduncle (Webb 1984, Jastrebski and Robinson 2004). Deep caudal peduncles have been correlated with increased burst speed (initial swim speed; Webb

1984, Blake 2004, Fisher and Hogan 2007), whereas narrow caudal peduncles are associated with increased swimming endurance (sustained swimming period; Webb 1984,

Webb and Weihs 1986). The second principle is drag, with increased body depth associated with increased drag as a fish moves through water (Ojanguren and Brana

2003). Deeper bodied fish tend to have slower burst speeds and reduced swimming endurance. Previous studies have found that deeper body forms benefit fish that consume cryptic prey by increasing their manoeuvrability to capture prey quickly once it has been spotted (Svanbäck and Eklöv 2002, 2003). Conversely, streamlined fish with increased burst speed are more likely to escape a predator (Webb 1984).

Although most studies have focused on relating variation in body shape and swimming performance to foraging and predation, reproductive behaviour should also play a role in the relationship between body shape and swimming performance. Fish have some of the most diverse reproductive tactics of all organisms, with many species exhibiting two or more tactics within a sex. The so-called alternative reproductive tactics commonly involve one type of male that courts females and defends them from other males and a second type of male that steals mating opportunities by streaking between a spawning pair or by mimicking female appearance (Taborksy 1998). Streak-type males must stealthily gain close proximity to the female at the time of oviposition to fertilize eggs

(Blanchfield et al. 2003, Stoltz and Neff 2006). In comparison, female mimics usually gain access to spawning females by deceiving territorial males (Dominey 1980). Because of these striking differences in behaviour, fish species with alternative reproductive

30 tactics provide an exceptional opportunity to examine the influence of reproductive behaviour on morphology and swimming performance.

One common species that exhibits alternative reproductive tactics is the bluegill sunfish

(Lepomis macrochirus). Male bluegill exhibit alternative tactics termed “parental” and

“cuckolder” (e.g., Gross and Charnov 1980, Gross 1982). In Lake Opinicon (Ontario,

Canada, 44°34’N, 76°19’W), a well-studied population of bluegill, parentals sexually mature at approximately 7 years of age and then form breeding colonies, court females, and provide sole parental care for the young (Gross and Charnov 1980). The parental care period lasts up to 10 days, during which the parental does not leave the nest. Cuckolders mature at a much younger age and mate by stealing fertilizations from parentals (Gross

1982). Cuckolders first adopt a “sneaker” tactic (2-3 years old) and dart into nests while a female releases eggs. As the cuckolders grow, they switch to a “satellite” tactic (4+ years old) and mimic the appearance and behaviour of females (Gross and Charnov 1980,

Gross 1982). Although there have been some early studies examining body shape among the male tactics (e.g., Gross 1982, Ehlinger 1991, Ehlinger et al. 1997), no study in bluegill, or any other fish, has performed an analysis of covariation in body shape among all reproductive tactics within a species nor contrasted their swimming performances.

In this study, I set out to compare both the morphology and swimming performance

(burst swim and swimming endurance) of bluegill parentals, sneakers, satellites, females, and juveniles. Webb (1984) previously showed that a deeper body shape provides greater maneuverability than a streamlined form. Ehlinger (1991) subsequently predicted that parental bluegill would benefit from increased maneuverability when hovering over their

31 nest and protecting the young from brood predators; therefore I predicted that parentals would have the greatest body depth. I expected that sneakers would instead be streamlined in body shape but with a relatively large caudal peduncle to maximize their burst speed when darting into a nest to steal fertilizations. Additionally, satellite males were expected to be more similar to females in shape than compared to other tactics as part of their deception during spawning.

2.2 Methods

2.2.1 Fish collection

I collected 103 bluegill from Lake Opinicon during the period of June 8 - 26, 2009. Daily snorkel surveys of the littoral habitat were used to identify breeding colonies and fish were collected using dip-nets starting the first day that parentals formed nests until the day after spawning occurred. Initial classification of each fish was based on observation of their behaviour immediately prior to collection (see Knapp and Neff 2007).

Subsequently, juveniles and cuckolders (sneakers or satellites) were distinguished based on total body length (mm) and gonadosomatic index (GSI: ratio of gonad mass to body mass). Based on the findings of Gross (1982), I classified juvenile bluegill as those with minimal investment in their gonads, or a GSI of less than 1%, individuals with GSI of approximately 4% and total length less than 100 mm were classified as sneakers, and satellites were those with about 3% GSI and total length equal to or greater than 100 mm.

2.2.2 Experimental procedures

All collected fish were taken to the Queen’s University Biological Station and held in aquariums with flow-through water drawn directly from Lake Opinicon prior to

32 experimentation. Experiments were completed within 6 hours from the time of capture. I first assessed morphological variation in body shape among the five bluegill groups (n =

26 juveniles, 20 females, 19 parentals, 19 sneakers, 19 satellites). Photographs of the left side of each fish were taken with a Canon PowerShot A95 (5.0 mega pixels) digital camera and 20 homologous landmarks were placed on each image using tpsDig software

(Rohlf 2008, Fig. 2.1). In order to describe and compare body shape across bluegill groups I used partial warp analysis to compare variation across individuals and contrast differences among tactics (Sheets 2002, Zelditch et al. 2004). I also obtained measurements of total body length (nearest mm) from the mouth to end of the caudal fin, maximum body depth (mm), and caudal peduncle depth (mm) from the digital images of each fish using tpsDig. Measurements of body depth (BD) and caudal peduncle depth

(CP) were used to determine the caudal peduncle depth factor (CPDF) using the formula:

CPDF = CP / BD (Webb and Weihs 1986).

I assessed swim performance on a subset of the fish by obtaining a single burst speed and swimming endurance measurement for each fish (number of individuals were equal to 13 juveniles, 10 females, 10 parentals, 9 sneakers and 11 satellites). Burst speed was estimated because of its predicted importance in the reproductive success of sneakers. I examined swim endurance to provide an estimate of the total swimming capacity of each group. Swimming capacity is important for parental males because they spend several days providing nest care behaviours, including fanning and circling the nest, without rest,

34 and for cuckolders because they must search the littoral habitat for colonies to spawn in

(Gross 1991). Burst speed was tested first by placing the fish in a plexiglass aquarium

(100 cm × 30 cm × 30 cm) at one end and orienting them towards the far end. An adjustable wall was used to limit the ability of the fish from moving any direction other than forwards but with enough space to allow for their typical burst swimming behaviours (i.e. body coiled into an S-shape). A high-speed camera (InLine 500, Fastec

Imaging, California, USA) set to film at 250 frames•s-1 was used to film the first 3 s of the burst response following a mild electrical stimulus that was used to induce swimming.

The high speed video was used to estimate the burst speed (cm•s-1) of each individual during the initial 0.5 s after stimulation. After each trial I did a complete water change in the aquarium.

Immediately after the burst speed test, the fish was placed in a 32 L swim flume (Swim-

30, Loligo Systems, Tjele, Denmark) and left to acclimate for 5 min. The water flow was then set to 0.16 m•s-1 and increased by 0.04 m•s-1 every minute until the fish stopped swimming against the current. As in similar experiments (e.g., Boily and Magnan 2002) a mild electrical stimulus (5V) was applied when a fish stopped swimming and rested against a metallic grid to motivate the fish to swim. When fish became unresponsive to the stimulus for more than 5 s, the test was ended and the total time (nearest second) was recorded as a measure of swimming endurance (e.g., Winger et al. 1999, Hanna et al.

2008). The tank was equipped with a pump that circulated water and assured oxygen levels close to saturation. The temperature of the water was monitored to maintain a range of 21°C - 24°C and half of the water was changed every third trial to ensure a consistent environment across trials.

2.2.3 Statistical analysis

In order to compare morphological differences across tactics, I compared the partial warp scores of each group using discriminant function analysis (DFA). Multivariate analysis of variance (MANOVA) was used to test for multivariate differences across the five groups.

When significant differences were found among groups, post-hoc tests were completed using the Student’s t-test with Bonferroni correction for multiple comparisons to determine which groups differed (based on five groups being compared significance level became 0.01 for post-hoc comparisons based on a family-type I error of 0.05). The main patterns of shape differentiation across tactics were visualized using thin-plate splines produced by regressing each partial warp scores onto each of the first discriminant functions using the software tpsRegr (Rohlf 2009). Next, the morphological scores of

DFA 1 and DFA 2 for each bluegill were correlated to their total body length using

Pearson’s correlation coefficient. For each correlation I visually inspected the residuals of the relationship on a normal probability plot. If the residuals did not conform to a nearly linear pattern, I used a permutation model based on residuals as they are not prone to increased type I error and increase the power of the parametric t-test for Pearson’s correlation when residuals are not normal (Anderson and Legendre 1999). In order to differentiate results based on these two types of approaches, i.e., parametric versus non- parametric (permutation-based test), I reported their associated probabilities as Pp and

Pnp, respectively. Additionally, the sneakers and satellites were grouped together as cuckolders, because these individuals represent a single ontogenetic life history (Gross and Charnov 1980), and analyzed for a relationship between size and the morphological scores. Next, DFA 1 and DFA 2 scores of sneakers and satellites were compared to the

36 mean female morphological scores for both DFA axes to create a “female likeness” factor using the equation:

2 2 (1) FLi = DFA1i - DFA1f   DFA2i  DFA2f 

th where FLi is the female likeness of the i cuckolder, DFA1i and DFA2i are the scores for

th the i cuckolder on a given axis, and DFA1f and DFA2f are the mean female scores for the two DFA axes. The female likeness scores were used in a logistic regression to test if there was a shift in cuckolder body shape associated with the transition from the sneaker to satellite tactic.

In order to remove size effects on swim performance, I used the Aitchinson (1986) log- ratio transformation as follows (see also Peres-Neto and Magnan 2004):

(log spi + log sizei ) (2) rspi = log spi   K 2 where rspi is the relative swimming performance (endurance and burst were transformed

th separately) of the i individual, spi is the untransformed (original) swim performance measurement of the ith individual, size was determined based on the sum of all distances between each landmark and the body centroid (Zelditch et al. 2004) and K represents the smallest rsp across all individuals is included so that all rspi values are positive. The transformed values for burst speed and swimming endurance across the five tactics were compared using ANOVA. When groups were found to differ, pair-wise comparisons were made using Student’s t-test under the Bonferroni adjustment. The relationship between the two measures of swim performance was examined using Pearson’s correlation.

In order to examine the relationships between swimming and morphological differences across tactics I first examined for differences in scaling between length and maximum body depth among our groups using an ANCOVA analysis and examining the interaction term. After pooling the morphological scores I correlated the first discriminant axis with burst speed and swimming endurance. To determine if the caudal peduncle depth factor was related to burst speed and swimming endurance I used Pearson’s correlation; the probabilities were based on the same permutation procedure used earlier whenever residuals did not follow a normal distribution.

All analyses were performed in the statistical software package JMP v. 8.0 and the level of significance for all tests (α) was 0.05 for all tests without multiple comparisons. All means are reported plus or minus one standard error.

2.3 Results

2.3.1 Morphological variation

The mean length, mass, and gonadosomatic index (GSI) scores for the five bluegill groups are listed in Table 2.1. DFA of body shape revealed significant differences among the groups (MANOVA, Wilks’ λ = 0.004, P < 0.01) with 87% of the variation in body shape being explained by the first two discriminant axes: DFA 1 (63%) and DFA 2

(24%). Further examination of the DFA 1 scores found significant differences among the groups (ANOVA, F4,98 = 245.0, P < 0.01). Pair-wise comparisons indicated that juveniles, parentals and sneakers differed from each other and between all other groups

(Student’s t, P < 0.01 for each pair-wise comparison). Females were found to be similar in body shape to satellites (Student’s t, P = 0.91). Thin-plate spline analysis of DFA 1

Table 2.1 Summary of the mean (±S.E.) values of fish measurements taken on bluegill (Lepomis macrochirus) classified into one of five groups. Burst Speed Swimming Endurance Group n Length Mass GSI Observed Aitchinson Observed Aitchinson (mm) (g) (%) CPDF (cm•s-1) Transformed (s) Transformed Juveniles 26 123 ± 6 39 ± 6 0.3 ± 0.1 0.300 ± 0.002 43 ± 7 1.5 ± 0.1 240 ± 41 0.56 ± 0.09

Females 20 162 ± 7 85 ± 11 6.0 ± 2 0.280 ± 0.003 31 ± 5 1.3 ± 0.1 273 ± 78 0.55 ± 0.10

Parentals 19 196 ± 2 148 ± 8 1.3 ± 0.2 0.280 ± 0.003 40 ± 6 1.3 ± 0.1 445 ± 75 0.69 ± 0.09

Sneakers 19 78 ± 3 9.7 ± 1 4.0 ± 0.4 0.300 ± 0.003 43 ± 5 1.8 ± 0.1 266 ± 43 0.92 ± 0.09

Satellites 19 122 ± 6 38 ± 5 3.1 ± 0.3 0.300 ± 0.003 46 ± 6 1.7 ± 0.1 464 ± 113 0.89 ± 0.10

showed that parentals had a deeper body form, particularly in the cranial region and contrasted to the streamlined body form of sneakers (Fig. 2.2). We also found differences among the bluegill groups based on the DFA 2 scores (ANOVA, F4,98 = 93.0, P < 0.01).

Post-hoc comparisons found that the DFA 2 scores differed among all the bluegill groups

(Student’s t, all P < 0.001). Thin-plate splines of DFA 2 showed that most of the variation in body form was related to the ventral portion of the head, pectoral fin insertion and caudal peduncle shape (Fig. 2.2).

There was no relationship between body length and DFA 1 scores for females (Pearson’s r = -0.35, n = 20, Pnp = 0.15), parentals (Pearson’s r = -0.20, n = 19, Pnp = 0.39), or sneakers (Pearson’s r = -0.01, n = 19, Pnp = 0.94). However, there was a relationship between these variable for juveniles (Pearson’s r = -0.83, n = 26, Pnp < 0.01) and satellites (Pearson’s r = -0.61, n = 19, Pp < 0.01). Similarly, no relationship was found between body length and DFA 2 scores for females (Pearson’s r = 0.06, n = 20, Pnp =

0.79), parentals (Pearson’s r = -0.63, n = 19, Pp = 0.80), or sneakers (Pearson’s r = 0.13, n = 19, Pnp = 0.62). However, in this case there also were no relationships found for juveniles (Pearson’s r = -0.01, n = 26, Pp = 0.65) or satellites (Pearson’s r = 0.28, n = 19,

Pnp = 0.27). When sneakers and satellites were combined into a single cuckolder category

I found a significant relationship between body length and scores for DFA 1 (Pearson’s r

= -0.86, n = 38, Pp < 0.01) and DFA 2 (Pearson’s r = 0.75, n = 38, Pp < 0.01). Finally, I found a strong logistic relationship between the female likeness scores and cuckolder body length with the inflection point occurring at the transition between tactics (logistic

2 regression, R = 0.82, n = 38, F3,33 = 52.27, P < 0.01; Fig. 2.3).

females 3

juveniles 1 DFA 1 satellites

-7 -5 -3 -1 -1 1 3 5 7

-3 parentals sneakers -5

DFA 2

Figure 2.2 Discriminant function analysis (DFA) of body shape among five groups of bluegill (Lepomis macrochirus). The groups are juveniles (▼), females (●), parentals (■), sneakers (▲), and satellites (◊). The scatter plot depicts the first two discriminant function axes, (a) DFA 1 and (b) DFA 2. Thin-plate splines were used to visualize the body shape associated with the extreme values for both DFA 1 and DFA 2.

Female likeness scoreFemale 2

decreasing similarity

0 60 80 100 120 140 160 Total body length (mm)

Figure 2.3 Logistic regression of total body length and female likeness scores of sneaker (○) and satellite (●) bluegill (Lepomis macrochirus). Female likeness scores were determined based on discriminant function analysis axes 1 and 2, with higher numbers reflecting decreasing similarity in body shape between cuckolders and females.

The CPDF was significantly different among tactics (ANOVA, F4,96 = 16.9, P < 0.01;

Table 2.1). Post-hoc comparisons revealed that sneakers, juveniles and satellites had similar CPDFs (Student’s t, all P > 0.50). Also, satellite and sneaker CPDF was similar to females (Student’s t, both P > 0.02; corrected α = 0.01), but juveniles differed from females (Student’s t, P < 0.01). Parentals were similar to females and satellites (Student’s t, both P > 0.03), but differed from sneakers and juveniles (Student’s t, both P < 0.01).

2.3.2 Swim performance

The absolute and size-corrected (i.e., Aitchinson-transformed) values for swimming performance of each bluegill group showed limited differences in the rankings of burst speed, with only satellites and sneakers changing positions (Table 2.1). However, there were changes in the swimming endurance ranks, with sneakers and juveniles increasing in rank whereas parentals and females decreased when the transformation was applied

(Table 2.1). Comparison of Aitchinson burst speed showed significant differences among the five reproductive groups (ANOVA, F4,48=4.86, P < 0.01; Table 2.1), with sneakers having the fastest burst swim, satellites and juveniles intermediate values, whereas parentals and females were the slowest. Similarly, the Aitchinson transformed swimming endurance scores differed significantly among groups with satellites having the relatively longest swim duration, followed by sneakers and parentals, with juveniles and females the shortest swimming endurance (ANOVA, F4,48=3.34, P = 0.02; Table 2.1). Finally, across all groups combined there was no relationship between burst speed and swimming endurance of individuals (Pearson’s r = 0.09, n = 53, Pp = 0.51).

2.3.3 Morphology and swim performance

I pooled the morphological scores of all groups to test for a relationship between DFA scores and swim performance after an ANCOVA indicated that the scaling between body length and depth was similar in all groups (ANCOVA - interaction term; F3,95 = 0.68, P =

0.57). Burst speed was positively correlated with DFA 1 (Pearson’s r = 0.33, n = 53, Pnp

= 0.02; Fig. 2.4a), but no relationship was found between swimming endurance and DFA

1 (Pearson’s r = 0.05, n = 53, Pnp = 0.74; Fig. 2.4b). Burst speed was negatively

correlated with DFA 2 (Pearson’s r = 0.34, n = 53, Pnp = 0.02; Fig. 2.4c), but no relationship was found between DFA 2 and swimming endurance (Pearson’s r = -0.21, n

= 53, Pnp = 0.13; Fig. 2.4d). Examination of the caudal peduncle depth found no relationship with either burst speed (Pearson’s r = 0.16, n = 52, Pnp = 0.20) or swimming endurance (Pearson’s r = 0.02, n = 52, Pp = 0.89).

2.4 Discussion

I examined morphological and swim performance variation within bluegill by classifying individuals into five groups based on sex, life history stage, and reproductive tactic. Each of these groups are likely under different selection pressures to optimize both survival and reproductive success, which I predicted should result in morphological differences. I found that sneakers had the most streamlined body shape of all bluegill, likely contributing to them having the fastest burst speed and longest swimming endurance. The increased speed no doubt reflects their need to evade predators (Werner and Hall 1977) related to their smaller overall size, but likely also optimizes reproductive fitness as sneakers must quickly dart into the nest of parentals (Gross and Charnov 1980, Stoltz and

Neff 2006). Indeed, sneakers differed in shape as compared to juveniles of a similar size.

2.5 (a) (c)

2.0

1.5

1.0

Burst swim

0.5

0.0

1.5 (b) (d)

1.0

0.5

Swimming endurance

0.0

-8 -6 -4 -2 0 2 4 6 -4 -2 0 2 4 DFA 1 score DFA 2 score

Figure 2.4 Relationship between morphology and (a,c) burst swim or (b,d) swimming endurance of five bluegill (Lepomis macrochirus) groups. The groups are juveniles (▼), females (●), parentals (■), sneakers (▲), and satellites (◊). Swim performance of each fish was assessed by measuring the burst swim response during the first 0.5 s following a mild electrical stimulus and the swimming endurance time (s) of fish places against a current. All swim performance measurements were transformed using the Aitchinson (1986) log-ratio transformation to account for variability in the size of each fish and are therefore unitless. The DFA axis scores represent scores of morphological variation based on the discriminant function analysis of the five bluegill groups (see text for description).

Similarly, satellite body shape effectively mimicked female morphology, which presumably results in parentals misidentifying them as a second female and permitting them to enter their nest (Dominey 1980; Gross 1982; Neff and Gross 2001). Variation in body shape was not limited to cuckolders, as parentals had a deeper overall body shape than all other groups. The deeper bodies are predicted to facilitate parental care and defensive behaviours, such as hovering over and circling the nest, fanning eggs and rapid movements to chase away brood predators (Ehlinger 1991). Overall, the morphological variation among bluegill may be the result of a variety of selection pressures that vary with age, sex, environmental conditions and reproductive tactic.

The observed differences in body shape between sneakers and satellites are particularly interesting because the tactics represent a single ontogenetic life history. Sneaking has been described as the most common form of alternative reproductive tactic (Taborsky et al. 1987), but satellites, mimicking female appearance, have been well documented in several species (e.g., Thompson et al. 1993, Gonçalves et al. 1996, Hanlon et al. 2005,

Whiting et al. 2009). However, unlike bluegill, in these other species, sneaking and satellite tactics typically represent independent life histories. I found that bluegill satellite morphology was similar to that of females, the group that they would need to mimic in order to deceive parental males that defend nests, but sneakers were different in shape from all other groups, including satellites. Importantly, the shift to female-like morphology coincided with the transition in mating tactic, providing compelling support for the influence of reproductive behaviour on morphology. Indeed, my results indicate that there may be strong selection related to mating success in cuckolders, resulting in a change in morphology that leaves no detectable trace of sneaker morphology in satellite

46 males. These data provide some of the first evidence establishing a strong relationship between reproductive tactic and morphological variation.

Morphological adaptations related to reproductive tactic are not necessarily the result of a single selection pressure, but may be the product of multiple selection pressures.

Selection pressures influencing morphology may be antagonistic, resulting in trade-offs to maximize fitness, or reinforce each other and increase the fitness of certain morphological traits. Foraging ecology has been linked to morphology (e.g. Zweers et al.

1995, Figueirido et al. 2009) and therefore is a likely candidate for interacting with selection pressures related to reproductive tactic. In sunfish, morphological variation within some populations has been related to resource polymorphisms and population divergence between the littoral and pelagic habitats (e.g. Robinson et al. 1993, Jastrebski and Robinson 2004). In order for population divergence based on resource polymorphisms to be maintained, it is likely that reproductive behaviour is also involved

(see Chapter 5). Reproductive tactics may interact with resource polymorphisms to support the separation of populations through the evolution of reproductive isolation mechanisms, eventually leading to speciation, or may maintain gene flow among groups limiting the potential for speciation (Coyne and Orr 2004). The relationship between different selection pressures is an avenue of research that should be pursued to further our understanding of how various factors, such as reproductive tactic and foraging strategy, interact and contribute to morphological variation.

Morphology is also an important factor in the swimming dynamics of fish, influencing both the generation of thrust and the drag created when moving through a viscous

47 medium (Ojanguren and Braña 2003). A previous study of the closely related pumpkinseed sunfish associated a deep body form with a quick burst speed (Jastrebski and Robinson 2004), despite the increased drag that is associated with a deeper body form (e.g., Webb 1984). However, I found that burst speed was greatest in sneakers, the most streamlined bluegill group, and not the deeper bodied parentals. Jastrebski and

Robinson (2004) only looked at pumpkinseed that would have been classified as females and parentals in my study, limiting the range of body depths that were examined and may contribute to the apparent discrepancy across studies. My findings are consistent with studies of other fishes, including the Eurasian perch (Perca fluviatilis; Svanbäck and

Eklöv 2002, 2003) which found a strong relationship between streamlining and swim performance. Additionally, some studies have reported that swimming performance is less related to body shape, but instead is related to caudal peduncle depth, the region of the body responsible for the generation of thrust (e.g., Domenici and Blake 1997, Fisher and Hogan 2007). As such, it has been argued that the depth of the caudal peduncle relative to the total body depth, referred to as the caudal peduncle depth factor, may be a strong predictor of swim performance because it captures both thrust potential and drag.

However, I found no relationship between this factor and either burst swim or swimming endurance. Nevertheless, my results indicate that bluegill swimming performance is negatively related to body depth and thereby implicates drag as a key factor governing swim performance.

While these data support the prediction that reproductive tactic is related to morphology through swimming performance, it may not be the sole factor involved. I found no differences in the body shape of females and satellites on DFA 1, accounting for 63% of

48 the variation in body shape, but satellites were significantly better at both burst swim and swimming endurance. This may reflect the different internal conditions of the fish based on reproductive investment. The females collected had a mean GSI of 6% compared to

3% for satellite males. This increased investment in gamete production may influence female swim performance in multiple ways. First, a greater amount of space in the body cavity of females is being filled by the ovaries compared to the testes of the satellites.

The reduced space inside the body cavity may restrict the S-shaped movements required during burst swimming in bluegill (Jayne and Lauder 1993). Second, females may be in a lower energetic state after investing much of their energy into egg production. Koch and

Weiser (1983) found that in roach (Rutilus rutilus) the cost of producing gametes in females was recouped by decreasing swimming activity. If a similar relationship between reproductive investment and swimming exists in female bluegill it is possible that they had less energy available to allocate to swim performance.

In conclusion, I found variation among bluegill groups consistent with adaptations in morphology that may relate not just to survival and foraging, but also to reproductive tactic. My data suggest that considering the relationship between reproductive tactic and morphological variation can enhance our understanding of the various selection pressures involved in behaviour and morphology, and can contribute to the field of ecomorphology and more broadly functional ecology.

2.5 Chapter acknowledgements

I thank C. Rodgers for assistance in the field and helpful comments on this paper. I would also like to thank two anonymous reviewers and an associate editor for their informative

49 comments about the previous version of this manuscript. This work was supported by the

Natural Science and Engineering Research Council of Canada (grants to B.D. Neff and

P.R. Peres-Neto), a Canada Foundation for Innovation Leader's Opportunity Fund to P.R.

Peres-Neto, and by partial funding to S. Colborne from The University of Western

Ontario. M.C. Bellemare was funded by an NSERC-Undergraduate Summer Research

Award. This research was carried out with the approval of The University of Western

Ontario Council on Animal Care (animal use protocol #2006-062-05), the Université du

Québec à Montréal Council on Animal Care (0509-651-0510) and the Ontario Ministry of Natural Resources (license #1051389).

2.6 References

Aerts, P., R.V. Van Damme, B. Vanhooydonck, A. Zaaf, and A. Herrel. 2000. Lizard locomotion: how morphology meets ecology. Neth. J. Zool. 50:261-277.

Aitchinson, J. 1986. The statistical analysis of compositional data. Chapman and Hall, New York, N.Y.

Anderson, M.J., and P. Legendre. 1999. An empirical comparison of permutation methods for tests of partial regression coefficients in a linear model. J. Stat. Comp. Sim. 62:271-303.

Bertrand, M., D.J. Marcogliese, and P. Magnan. 2008. Trophic polymorphism in brook charr revealed by diet, parasites and morphometrics. J. Fish Biol. 72:555-572.

Blake, R.W. 2004. Fish functional design and swimming performance. J. Fish Biol. 65:1193-1222.

Blanchfield, P.J., M.S. Ridgway, and C.C. Wilson. 2003. Breeding success of male brook trout (Salvelinus fontinalis) in the wild. Mol. Ecol. 12:2417-2428.

Boily, P. and P. Magnan. 2002. Relationship between individual variation in morphological characters and swimming costs in brook charr (Salvelinus fontinalis) and yellow perch (Perca flavescens). J. Exp. Biol. 205:1031-1036.

Coyne, J.A., and H.A. Orr. 2004. Speciation. Sinauer, Sunderland, M.A. de Medeiros, P.R., and R.T. da Costa Ramos. 2007. Predicting ecomorphological patterns from morphology of a tropical estuarine fish assemblage. Biociências 15:40-46.

Domenici, P., and R.W. Blake. 1997. The kinematics and performance of fish fast-start swimming. J. Exp. Biol. 200:1165-1178.

Dominey, W.J. 1980. Female mimicry in male bluegill sunfish – a genetic polymorphism? Nature. 284:546-548.

Ehlinger, T.J. 1991. Allometry and anaylsis of morphometric variation in the bluegill, Lepomis macrochirus. Copeia. 1991:347-357.

Ehlinger, T.J., M.R. Gross, and D.P. Philipp. 1997. Morphological and growth rate difference between bluegill males of alternative reproductive life histories. North Am. J. Fish. Man. 17:533-542.

Figueirido, B., P. Plamqvist, and J.A. Pérez-Claros. 2009. Ecomorphological correlates of craniodental variation in bears and paleobiological implications for extinct taxa: an approach based on geometric morphometrics. J. Zool. 277:70-80.

Fisher, R., and J.D. Hogan. 2007. Morphological predictors of swimming speed: a case study of pre-settlement juvenile coral reef fishes. J. Exp. Biol. 210:2436-2443.

Gonçalves, E.J., V.C. Almada, R.F. Oliveira, and A.J. Santos. 1996. Female mimicry as a mating tactic in males of the blennid fish Salaria pavo. J. Mar. Biol. Assoc. UK. 76:529-538.

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z. Tierpsychol. 60:1-26.

Gross, M.R. 1991. Evolution of alternative reproductive strategies: frequency-dependent sexual selection in male bluegill sunfish. Philos. Trans. R. Soc. Lond., Ser. B: Biol. Sci. 332:59-66.

Gross, M.R., and E.L. Charnov. 1980. Alternative male life histories in bluegill sunfish. Proc. Natl. Acad. Sci. USA. 77: 6937-6940.

Hambright, K.D. 1991. Experimental analysis of prey selection by largemouth bass: role of predator mouth width and prey body depth. Trans. Am. Fish. Soc. 120:500- 508.

Hanlon, R.T., M.-J. Naud, P.W. Shaw, and J.N. Havenhand. 2005. Transient sexual mimicry leads to fertilization. Nature 433:212.

Hanna, S.K., A.H. Haukenes, R.J. Foy, and C.L. Buck. 2008. Temperature effects on metabolic rate, swimming performance and condition of Pacific cod Gadus macrocephalus Tilesius. J. Fish. Biol. 72:1068-1078.

Jastrebski, C.J., and B.W. Robinson. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

Jayne, B.C., and G.V. Lauder. 1993. Red and white muscle activity and kinematics of the escape response of the bluegill sunfish during swimming. J. Comp. Physiol. A. 173:495-508.

Knapp, R., D.K. Hews, C.W. Thompson, L.E. Ray, and M.C. Moore. 2003. Environmental and endocrine correlates of tactic switching by nonterritorial male tree lizards (Urosaurus ornatus). Hor. Behav. 43:83-92.

Knapp, R., and B.D. Neff. 2007. Steroid hormones in bluegill, a species with male alternative reproductive tactics including female mimicry. Biol. Let. 3:628-632.

Koch, F., and W. Weiser. 1983. Partitioning of energy in fish: can reduction of swimming activity compensate for the cost of production? J. Exp. Biol. 107:141-146.

Moore, M.C., D.K. Hews, and R. Knapp, R. 1998. Hormonal control and evolution of alternative male phenotypes: generalizations of models for sexual differentiation. Amer. Zool. 38:133-151.

Motta, P.J., S.F. Norton, and J.J. Luczkovich. 1995. Perspective on the ecomorphology of bony fishes. Enviro. Biol. Fish. 44:11-20.

Neff, B.D., and M.R. Gross. 2001. Dynamic adjustment of parental care in response to perceived paternity. Pro. R. Soc. Lon. B. Biol. Sci. 268:1559-1565.

Ojanguren, A.F., and F. Braña. 2003. Effects of size and morphology on swim performance in juvenile brown trout (Salmo trutta L.), Ecol. Fresh. Fish. 12:241- 246.

Peres-Neto, P.R., and P. Magnan. 2004. The influence of swimming demand on phenotypic plasticity and morphological integration: a comparison of two polymorphic charr species. Oecologia 140:36-45.

Robinson, B.W., D.S. Wilson, A.S. Margosian, and P.T. Lotito. 1993. Ecological and morphological differentiation of pumpkinseed sunfish in lakes without bluegill sunfish. Evol. Ecol. 7:451-464.

Rohlf, F.J. 2008. TpsDig, Version 2.12. Stony Brook, New York, Department of Ecology and Evolution, State University of New York at Stony Brook. URL: http://life.bio.sunysb.edu/morph/

Rohlf, F.J. (2009) TpsRegr, Version 1.36. Stony Brook, New York, Department of Ecology and Evolution, State University of New York at Stony Brook. URL: http://life.bio.sunysb.edu/morph/

Schluter, D., and A. Rambaut. 1996. Ecological speciation in postglacial fishes. Phil. Trans. Biol. Sci. 351:807-814.

Sheets, H.D. 2002. IMP PCAGen6n. http://www2.canisius.edu/~sheet/morphsoft.html. Buffalo, New York, Canisius College.

Stoltz, J.A., and B.D. Neff. 2006. Male size and mating tactic influence proximity to females during sperm competition in bluegill sunfish. Behav. Ecol. Sociobiol. 59:811-818.

Svanbäck, R., and P. Eklöv. 2002. Effect of habitat and food resources on morphology and ontogenetic growth trajectories in perch. Oecologia 131:61-70.

Svanbäck, R., and P. Eklöv. 2003. Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization? Oikos 102:273-284.

Taborsky, M. 1998. Sperm competition in fish: bourgeois males and parasitic spawning. Trends Ecol. Evol. 13:222-227.

Taborsky, M., B. Hudde, and P. Wirtz. 1987. Reproductive behaviour and ecology of Symphodus (Crenilabrus) ocellatus, a European wrasse with four types of male behaviour. Behaviour 102:82-118.

Thompson, C.W., L.T. Moore, and M.C. Moore. 1993. Social, environmental and genetic factors in the ontogeny of phenotypic differentiation in a lizard with alternative male reproductive strategies. Behav. Ecol. Sociobiol. 33:137-146.

Velasco, J.A., and A. Herrel. 2007. Ecomorphology of Anolis lizards of the Choco’ region in Colombia and comparisons with Greater Antillean ecomorphs. Biol. J. Linn. Soc. 92:29-39.

Wainwright, P.C. 1996. Ecological explanation through functional morphology: the feeding biology of sunfishes. Ecology 77:1336-1343.

Webb, P.W. 1984. Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24:107-120.

Webb, P.W., and D. Weihs. 1986. Functional locomotor morphology of early life history stages of fishes. Tran. Am. Fish. Soc. 115:115-127.

Werner, E.E., and D.J. Hall. 1977. Competition and habitat shift in two sunfishes (Centrarchidae). Ecology 58:869-876.

Whiting, M.J., J.K. Webb, and J.S. Keogh. 2009. Flat lizard female mimics use sexual deception in visual but not chemical signals. Proc. R. Soc. Lon. B. Biol. Sci. 276:1585-1591.

Winger, P.D., P. He, and S.J. Walsh. 1999. Swimming endurance of American plaice (Hippoglossoides platessoides) and its role in fish capture. ICES J. Mar. Sci. 56:252-265.

Winkler, H. 1988. An examination of concepts and methods in ecomorphology. In Acta XIX International Ornithological Congress. Edited by H. Ouellet. National Museum of Natural Sciences, Ottawa, O.N. pp. 2246-2253.

Zelditch, M.L., D.L. Swiderski, H.D. Sheets, and W.L. Fink. 2004. Geometric morphometrics for biologists. Elsevier Academic Press, San Diego, CA

Zweers, G.A., F. de Jong, H. Berkhoult, and J.C. Cenden Berge. 1995. Filter-feeding in Flamingos (Phoenicopterus ruber). Condor 97:297-324.

Chapter 3 3 Effects of foraging and sexual selection on ecomorphology of a fish with alternative reproductive tactics2

The foraging ecology of fish is often considered to be the primary determinant of body shape due to tight links between morphology, swimming performance, and foraging efficiency. Fish foraging on littoral benthic macroinvertebrates typically have a deeper body shape than those foraging on pelagic zooplankton in the water column. However morphological traits often have multiple ecological functions, which could result in performance trade-offs between functions. Here, I provide the first examination of body shape and diet in a species with alternative reproductive tactics, in this case, bluegill sunfish (Lepomis macrochirus). Bluegill males mature into either “parental” or

“cuckolder” reproductive tactics. Parentals build nests and provide sole parental care and defense of the young. Cuckolders instead act as “sneakers” darting into the nests of parental males while mating is occurring and then later in life become “satellites”, mimicking female appearance and behavior. Using stable carbon and nitrogen isotopic analysis of diet I found that parentals and females consumed primarily pelagic zooplankton, yet were the deepest in body shape. Sneakers consumed more littoral resources, but were the most streamlined. Satellite males also consumed predominantly

2 A version of this chapter has been published in the journal Behavioral Ecology and is presented here with permission. Citation: Colborne, S.F., P.R. Peres-Neto, F.J. Longstaffe, and B.D. Neff. 2013. Effects of foraging and sexual selection on ecomorphology of a fish with alternative reproductive tactics. Behav. Ecol. 24:1339- 1347. doi: 10.1093/beheco/art072.

56 littoral resources, but had a deeper body form that was more similar to females than to size-matched juveniles. My results differ from past studies of foraging ecomorphology and suggest that other selection pressures, such as sexual selection in species with alternative reproductive tactics, may also be an important factor influencing shape.

3.1 Introduction

Understanding the diversity of phenotypes and their function has long interested biologists (reviewed by Bolnick et al. 2011). The field of ecomorphology examines the links between morphological characteristics and a variety of ecological factors including foraging, predation, reproduction, and locomotion (Collar and Wainwright 2009,

Cochran-Biederman and Winemiller 2010, Griffin and Mosblack 2011). It is arguably the relationship between morphology and foraging that has received the most attention, probably due to the strong selection pressures that are often associated with foraging efficiency. Studies of ecomorphology often examine variation in morphological features in relation to food acquisition tactics and overall resource use both within populations of a species and across multiple species within taxa. Several taxa, including mammals (e.g.

Saunders and Bailey 1992), invertebrates (e.g. Griffin and Mosblack 2011), reptiles (e.g.

Losos 2009), and fish (e.g. Cochran-Biederman and Winemiller 2010, Robinson et al.

2000), have shown strong, predictable relationships between specific morphological features and foraging. In some instances, the relationships between foraging and morphology have such consistent patterns that morphology has been accurately used as a predictor of resource use in previously unexamined populations and species (e.g.

Saunders and Bailey 1992, Jastrebski and Robinson 2004, Griffin and Mosblack 2011).

Studies of foraging ecomorphology in fish have shown that a streamlined body shape (i.e. decreased overall body depth) is generally associated with increased burst swimming speed due to reduced drag moving through water (Ojanguren and Braña 2003). By comparison, increased body depth is instead associated with increased maneuverability and is common among fish occupying a structurally complex environment (Svanbäck and

Eklöv 2002, 2003). Across species, streamlined fish are typically pelagic zooplanktivore feeders, whereas deep-bodied fish are found in littoral habitats feeding on a variety of benthic macroinvertebrates (Robinson et al. 2000, Collar and Wainwright 2009). Even within species that have specialized into littoral and pelagic foraging tactics (often referred to as “ecomorphs”), similar patterns of body shape variation exist between the tactics (Robinson et al. 2000, Robinson and Parsons 2002, Svanbäck et al. 2008).

Although many studies have focused on linking morphology to diet, a morphological trait may have multiple ecological functions. For example, cichlids use their fins not only for swimming and foraging, but also during parental care to fan developing eggs and larvae in nests (Cochran-Biederman and Winemiller 2010), likely resulting in selection pressures related to both functions and influencing the observed fin morphology. If the optimal morphology for a trait varies across different ecological functions then trade-offs between functions may exist. In mouth-brooding cichlids (Haplochromis spp.), for instance, there is variation in head morphology that has been linked to foraging tactic and mouth brooding function, resulting in a trade-off between biting performance and the volume of the buccal cavity to brood offspring (tkint et al. 2012). In situations when natural selection (e.g. foraging efficiency) and sexual selection (e.g. reproductive success) impose divergent selection pressures, successful individuals maximize fitness by

58 balancing the fitness costs and benefits of different morphological forms, resulting in an

“optimal” morphology (tkint et al. 2012). Thus, although morphology may often be related to foraging, it is possible that other ecological factors, such as sexual selection, play a role in the evolution of morphology and should be considered when studying the relationship between the features of organisms, their behaviours, and the environment.

In this chapter, I examined the interplay between different ecological functions and morphology by examining shape and diet variation among alternative morphs in bluegill

(Lepomis macrochirus). Bluegill alternative male reproductive tactics are referred to as

“parental” and “cuckolder” (Gross and Charnov 1980, Gross 1982). In the well-studied population in Lake Opinicon (Ontario, Canada, 44º34’N, 76º9’W), parental males mature at approximately 7 years of age, at which point they form breeding colonies and provide sole parental care for the offspring (Gross 1982). During the parental care period, the males exhibit a variety of behaviors, including fanning of the eggs and defensive opercular flares, lateral displays, and biting at a variety of brood predators (Gross and

MacMillan 1981, Côté and Gross 1993). Cuckolders, by comparison, mature at a younger age and “steal” fertilization opportunities from parentals (Gross 1982). Cuckolders first use a “sneaker” tactic (2 – 3 years old), which involves darting into nests while the female is releasing her eggs. As cuckolders grow, they switch to a “satellite” tactic (4+ years), and mimic female appearance and behavior (Dominey 1980, Gross 1982). In addition to the variation in behavior among these male tactics, there is documented variation in size and shape (Gross 1982, Chapter 2). Females follow a single life history, maturing at 4 years of age (Gross 1982).

The divergent reproductive tactics of bluegill provide an exceptional opportunity to examine the relative influence of reproduction versus foraging on shape. If shape is primarily related to foraging, then bluegill with a deeper body shape should feed predominantly on littoral benthic macroinvertebrates (e.g. snails), whereas streamlined fish should consume mostly pelagic zooplankton. I have previously shown that parental males and females have the deepest body shapes (Chapter 2) and therefore now predict that these morphs will consume mostly littoral resources. Conversely, sneakers are streamlined in shape and should consume pelagic zooplankton if general patterns of foraging ecomorphology hold. In order to directly test this set of predictions, I use stable isotope analysis to assess diet of the bluegill morphs and link variation in diet to the body shapes of each morph.

3.2 Methods

3.2.1 Fish collection

I collected a total of 142 bluegill from Lake Opinicon (Ontario, Canada, 44º34’N,

76º19’W). During the period of 8 – 26 June 2009, I used daily snorkel surveys of the littoral habitat to collect 102 bluegill using dip-nets; sampling began on the first day parental males began to form colonies and continued until the first day after spawning had occurred. These fish sampled in 2009 were previously used in an examination of bluegill morphology and swim performance (Chapter 2). An additional 40 fish (parental males and females only) were collected from multiple locations in Lake Opinicon by angling with a small piece (2-3 cm) of earthworm suspended from the side of the research boat during the period of 24 May – 30 June, 2010 and were added to the original

60 data set. Classification of fish into juveniles, females, parental males, satellites, or sneakers was initially based on observations of behavior in the field (i.e. immediately prior to collection; 2009 only) and the results of subsequent dissections. Juveniles have a gonadosomatic index (GSI; the ratio of gonad mass to body mass) of less than 1 %, sneakers have a total body length of less than 100 mm and GSI of about 4 %, and satellites have a total length of 100 mm or greater and 3 % GSI (Gross 1982; Chapter 2).

3.2.2 Morphological analysis

All collected fish were taken to the Queen’s University Biological Station and held in flow-through aquariums with water drawn directly from Lake Opinicon for no longer than 6 hours before being euthanized with clove oil. I first assessed morphological variation in body shape among the five bluegill groups (n = 25 juveniles, 43 females, 36 parentals, 18 satellites, and 20 sneakers) by taking photographs of the left side of each fish with either a Canon PowerShot A95 (5.0 megapixels) or Olympus Stylus Tough-

6000 (10 megapixels) digital camera. Using tpsDig software (Rohlf 2008) I placed 20 homologous landmarks on the image of each fish (see Fig 2.1 for landmark locations).

These homologous landmarks were then used to compare body shape, independent of size, across the five bluegill groups using partial warp analysis (e.g. Zelditch et al. 2004;

Chapter 2, Section 2.2.3).

3.2.3 Stable isotope analysis

Following photographing, I collected tissue samples for stable isotope analysis of diet. A sample of white muscle was removed from the right side of each fish immediately underneath the posterior portion of the dorsal fin. The liver was then removed and both

61 tissue samples were stored at -20ºC. White muscle and liver tissues were selected because they are commonly used in stable isotope studies of fish diet (reviewed in Boecklen et al.

2011) and provide the opportunity to infer diet over different time frames because of tissue specific differences in metabolic activity; white muscle reflects diet over a period of months, compared to several weeks for liver tissue (Hesslein et al. 1993, Perga and

Gerdeaux 2005, Guelinckx et al. 2007).

Reference prey samples of littoral benthic invertebrates and pelagic zooplankton were also collected at regular intervals from multiple sites around Lake Opinicon during both sampling years to provide a resource baseline from which to compare the bluegill tissues.

To verify that gastropods (snails) could be used as a general representative of the littoral habitat (Cabana and Rasmussen 1996, Post 2002), I collected the five most common littoral prey groups in Lake Opinicon (gastropods, amphipods, isopods, larval Odonata, and larval Ephemeroptera). The pelagic resource isotopic compositions reflect a pooled sample of shallow zooplankton collected using a 5 m vertical tow. The tow sample contained primarily copepods and cladocerans (e.g. Daphnia spp.), which were filtered to remove phytoplankton and algae.

The tissue and prey resource samples were prepared by freeze drying them at -50 ºC for

24 h and grinding them into a fine powder using a mortar and pestle. Stable isotope ratios of carbon (13C:12C) and nitrogen (15N:14N) were then determined using continuous-flow mass spectrometry (Costech elemental analyzer coupled to a Thermo Finnigan Deltaplus

XL mass spectrometer) in the Laboratory for Stable Isotope Science (LSIS) at The

University of Western Ontario (London, Ontario, Canada). Isotope ratios were expressed as the per mil (‰) difference from the standard reference material:

δX = (Rsample/Rstandard – 1) where X is 13C or 15N, R is the ratio of 13C:12C or 15N:14N and δ is the measure of heavy to light isotope in the sample. The isotopic compositions were calibrated to international standards for both δ13C and δ15N and assessed for sample precision and accuracy using internal laboratory standards and replicate samples (see Appendix A for details).

Tissues that are high in lipids (liver) may have lower δ13C values compared to the isotopic compositions of pure protein samples (Smyntek et al. 2007; Boecklen et al.

2011). To compensate for this lipid effect, I corrected the carbon isotopic composition of each liver tissue sample using the mass balance correction model developed by Fry et al.

(2003) and adapted by Smyntek et al. (2007) for freshwater organisms:

13 13 δ Cex = δ Cbulk + 6.3(C:Nbulk – 4.2/C:Nbulk)

13 13 13 where δ Cex is the predicted δ C value of the tissue sample without lipids, δ Cbulk is the

13 measured δ C value of each individual, 6.3 represents the mean ‰ discrimination factor between lipids and protein, C:Nbulk is the observed atomic ratio of C:N of each sample, and 4.2 represents the mean C:N ratio of lipid extracted tissues across a variety of taxa

(Smyntek et al. 2007). I did not mathematically correct the stable isotope composition of white muscle tissue because the lipid levels in white muscle are sufficiently low (C:N ratio < 4) that correction is considered not to be required (e.g. Pörtner 2002; Boecklen et al. 2011).

3.2.4 Statistical analysis

Partial warp scores for each bluegill were used in a discriminant function analysis (DFA) to compare shape variation among the morphs (juveniles, females, parentals, sneakers, and satellites). When significant differences among morphs were identified by the DFA, an analysis of variance (ANOVA) with Tukey’s post-hoc comparisons was used for each significant DFA axis to identify which morphs differed. In addition, patterns of body shape variation were visualized using thin-plate splines generated by tpsRegr software

(Rohlf 2009).

Stable isotope analysis of diet was completed by first comparing the prey resources to assess the suitability of δ13C and δ15N for differentiating between resources in the littoral and pelagic habitats. The isotopic compositions of five common littoral macroinvertebrate resource groups were compared using a one-factor ANOVA. I then compared the two sampling years for temporal variation using the two most abundant prey types (littoral snails and pelagic zooplankton) using two-tailed t-tests. Following this, the differences in stable isotope compositions of prey resources were compared between habitats using a two-tailed t-test. Due to the large degree of overlap in δ15N values between the littoral and pelagic resource groups (see results below) I did not consider δ15N further in my interpretation of the results for the bluegill tissues.

13 The differences in δ C values among bluegill morphs were then examined using a one- factor ANOVA with Tukey’s post-hoc comparisons, when appropriate, for each tissue type (white muscle and liver) separately. Additionally, estimates of the % littoral

64 contribution to the diet of each fish was made using a two-end-member-mixing model based on Post (2002):

13 13 13 13 % Littoral = (δ Cconsumer - δ Cbase 2) / (δ Cbase 1 - δ Cbase 2) × 100

13 13 13 where δ Cconsumer is the observed δ C value for each fish, δ Cbase 1 represents the mean

13 13 13 δ C value of littoral snails, and δ Cbase 2 is the mean δ C value of pelagic zooplankton.

To examine the possible relationship between diet (i.e. stable isotopic compositions) and morphology across the morphs (i.e. DFA axes) I used separate linear regression analyses for white muscle and liver tissues (dependent factor: δ13C value; independent factor: DFA axis score or body length). Only the first two DFA axes, which together represented 84% of the total variation in morphology (see Section 3.3.1 below), were considered in these analyses. Analysis of covariance (ANCOVA) was used to examine the relationship between body length (a surrogate of bluegill age; Neff et al. 2004) and body shape using bluegill morphs that overlapped in body length (dependent factor: DFA 1 or DFA 2; fixed factor: morph; covariate: body length).

To simultaneously compare the relationship among δ13C values (diet) and three possible explanatory variables (body length, DFA 1, DFA 2), I applied an all-possible-subset regression procedure and the best model was selected on the basis of the Akaike information criteria score recommended for small sample sizes (AICc; Burnham and

Anderson 1998). The model with the lowest score was considered to have the best fit unless the score was within 2 units of a simpler model (i.e. one with fewer explanatory variables), in which case the model with the fewest explanatory variables was considered the best fit (see Burnham and Anderson 1998, Berner et al. 2011).

All analyses were completed using JMP v. 10.0 (SAS Institute Inc., Cary, North

Carolina), at α-level = 0.05. All means are reported as plus or minus 1 standard error (± 1

SE), except for stable isotope compositions which are reported as means plus or minus 1 standard deviation (± 1 SD).

3.3 Results

3.3.1 Bluegill morphological variation

DFA revealed significant differences in body shape among the five morphs (MANOVA,

Wilks’ λ = 0.008, P < 0.001; Fig. 3.1), with 64% and 20% of the total variation being explained by the first two axes. The bluegill morphs were all significantly different in shape based on DFA 1 (ANOVA, F4,137 = 273.4, P < 0.001; Tukey’s, all P ≤ 0.01), except between females and satellite males, which were marginally non-significant (Tukey’s, P

= 0.06). Thin-plate spline visualizations of DFA 1 showed that parental males had the deepest body shape, followed by females, satellites, and juveniles, with sneakers being the most streamlined in shape (Fig. 1). The morphs also varied significantly on DFA 2

(ANOVA, F4, 138 = 87.7, P < 0.001), with the thin-plate splines showing that shape variation on this axis was localized to head and caudal peduncle areas (Fig. 3.1). Based on thin-plate splines of the DFA 2 scores, females had more streamlined heads and thicker caudal peduncles than all other morphs (Tukey’s: P < 0.05: Table 3.1).

3.3.2 Diet variation among bluegill morphs

Among the five primary littoral macroinvertebrate prey groups (Amphipoda, larval

Odonata, larval Ephemeroptera,I, and Gastropoda), there were similar δ13C values

Figure 3.1 Discriminant function analysis (DFA) of body shape among five morphs of bluegill (Lepomis macrochirus): juveniles, sneakers, satellites, females, and parental males. The first two discriminant function axes (DFA 1 and DFA 2) are depicted on the x and y-axes, respectively. Thin-plate splines on each axis display the overall body shape associated with the extreme values for DFA 1 and DFA 2.

Table 3.1 Summary of the mean (± standard error) values of size, shape, and diet measurements for juveniles and four reproductive groups of bluegill (Lepomis macrochirus). Different letters (A – D) correspond to significant differences between the reproductive groups (see methods to statistical analysis details).

Body Length DFA 1 DFA 2 White muscle Liver White muscle Liver Reproductive (mm) Score Score δ13C δ13C (% Littoral) (% Littoral) morph n Juvenile 25 117.4 ± 3.7 A 1.91 ± 0.19 A 0.08 ± 0.20 A –24.2 ± 0.3 A –24.6 ± 0.4 A 55 49 Sneaker 20 83.3 ± 4.2 B 4.43 ± 0.22 B -1.83 ± 0.22 B –23.8 ± 0.4 A –24.2 ± 0.4 A 61 55 Satellite 18 123.0 ± 4.4 A 0.51 ± 0.23 C -0.63 ± 0.24 AC –24.6 ± 0.4 A –25.1 ± 0.4 A 48 41 Female 40 164.4 ± 3.0 C -0.22 ± 0.15 C 2.21 ± 0.15 D –26.3 ± 0.3 B –27.2 ± 0.3 B 18 13 Parental 35 186.2 ± 3.1 D -3.87 ± 0.16 D -1.34 ± 0.17 BC –27.3 ± 0.3 B –27.2 ± 0.3 B 12 14

15 (ANOVA, F4, 32 = 2.04, P = 0.11). There was significant variation in δ N values among the common littoral prey groups (ANOVA, F4, 32 = 7.07, P < 0.001); however, post-hoc comparisons indicated that snails did not differ significantly from any of the other prey groups (Tukey’s, all P ≥ 0.10), except odonata (P = 0.04). Therefore, the gastropods

(snails) were representative of the littoral prey isotope values for both δ13C and δ15N (see

Appendix A, Table A.1). Across the two sampling years there was no significant

13 variation in δ C values for either littoral (t-test, t16 = 0.52, P = 0.61) or pelagic (t-test, t8

= -0.79, P = 0.45) prey groups, indicating that δ13C prey values were temporally consistent in my data set. In contrast, the δ15N prey values differed between years for both littoral (t16 = -3.19, P = 0.006) and pelagic (t8 = 6.84, P < 0.001) prey groups, indicating that δ15N values of prey resources varied between sampling years. Sampling years were subsequently pooled for further analysis of δ13C, but not δ15N. Across the habitat types, the mean δ13C value of littoral prey resources (represented by snails: –22.3

± 1.9 ‰, n = 18) was significantly higher compared to the value of pelagic zooplankton

13 prey (–29.2 ± 1.3 ‰, n = 10; t-test, t26 = -10.3, P < 0.001), indicating that δ C values could be used to compare littoral and pelagic resource use by bluegill. In contrast, the

15 δ N values of littoral prey resources were higher than pelagic prey in 2009 (t9 = -3.67, P

= 0.005), but lower than pelagic prey in 2010 (t15 = 6.35, P < 0.001). Thus, I excluded

δ15N from further interpretation, focusing instead on the δ13C values.

The five bluegill morphs showed significant differences in δ13C values among morphs in both white muscle (ANOVA, F4, 133 = 23.0, P < 0.001) and liver tissues (ANOVA, F4, 136

= 18.2, P < 0.001; Fig. 3.2). For both tissue types, parental males and females were

69 similar in δ13C (Tukey’s, all P ≥ 0.08), but had significantly lower δ13C values compared to the other three morphs (Tukey’s, all P ≤ 0.004; Table 3.1). The δ13C values of sneakers, satellites, and juveniles were not significantly different from each other

(Tukey’s, all P ≥ 0.52). Overall, parental males and females had mean diets composed of greater than 80 % pelagic resources whereas juveniles, sneakers, and satellites had mean diets composed of 40 – 60 % littoral resources (Table 3.1). See Appendix A (Table A.2) for a summary of the stable isotope compositions for each individual.

3.3.3 Morphology and body length

The morphs differed in body length (ANOVA, F4, 134 = 129.2, P < 0.001), with all morphs differing except satellite males and juveniles (Tukey’s, P = 0.87). Examining these latter two morphs, there was a significant relationship between DFA 1 values and body length

(ANCOVA, F1, 39 = -6.58, P < 0.001; Fig. 3.3). However, over the size range of these two groups, the DFA 1 scores were consistently higher for juveniles compared to satellite males (ANCOVA, F1, 39 = 5.14, P < 0.001). There was no interaction between morph and body length in this analysis (ANCOVA, F1, 39 = 0.01, P = 0.99). Similarly, for a given body length, juvenile DFA 2 scores were greater than satellites (ANCOVA, F1, 39 = 3.73,

P < 0.001), but there was no relationship between DFA 2 scores and body length

(ANCOVA, F1, 39 = 1.53, P = 0.13), and no interaction between morph and length

(ANCOVA, F1, 39 = -0.58, P = 0.57).

3.3.4 Bluegill Morphology and Stable Isotopes

Pooling data across all morphs, linear regression indicated that there was a positive

13 relationship between DFA 1 scores and δ C values in both white muscle (F1, 136 = 80.68,

2 2 R = 0.37, P < 0.001) and liver tissues (F1, 139 = 48.44, R = 0.26, P < 0.001; Fig. 3.4), with higher δ13C values (littoral diet) being associated with higher DFA 1 scores (i.e. streamlined body shape). In contrast, DFA 2 scores were not related to δ13C values in

2 white muscle (linear regression, F1, 136 = 2.66, R = 0.02, P = 0.11), but were related to

13 2 liver δ C (linear regression, F1, 139 = 8.01, R = 0.05, P = 0.005). Body length was

13 2 negatively related to δ C values of white muscle (linear regression, F1, 133 = 95.85, R =

2 0.42, P < 0.001) and liver (linear regression, F1, 136 = 77.54, R = 0.36, P = 0.005), with increased body length being associated with lower δ13C values (pelagic diet).

Model selection based on AICc scores indicated that, for white muscle, the model

13 predicting δ C values with the lowest AICc scores contained body length alone (Table

13 3.2). In comparison, the model for liver δ C values with the lowest AICc score included both body length and DFA 2; however, using the parsimony criterion, the best model included only body length for the liver δ13C values.

3.4 Discussion

Diets varied among the bluegill groups: juveniles, sneakers, and satellites consumed primarily littoral resources (benthic macroinvertebrates), whereas females and parentals consumed mostly pelagic resources (zooplankton). This pattern was stable over a period of weeks to months because the stable isotopic compositions were similar in the

Figure 3.2 Boxplots of δ13C values (per mil deviation of measured 13C:12C relative to

Vienna Pee Dee Belemnite) among five morphs of bluegill (Lepomis macrochirus): juveniles, sneakers, satellites, females, and parental males. The stable carbon isotope compositions of (a) white muscle and (b) liver tissues are presented separately. Boxes represent the interquartile range (inner 50% of observations) separated by the median, whiskers represent the 90th and 10th percentiles, and the dots correspond to the 95th and

5th percentiles. Capital letters above the boxplots denote homogeneous subsets as identified by ANOVA and Tukey’s post-hoc comparisons of morphs (α = 0.05).

Figure 3.3 Relationship between morphology (DFA 1) and body length (mm) of juvenile and satellite male bluegill (Lepomis macrochirus). The DFA 1 scores were generated using a discriminant function analysis of partial warp scores generated for each individual. The lines are based on regressions, but statistical analyses used ANCOVA (see text for details).

Figure 3.4 Relationship between δ13C values (per mil ratio of sample 13C:12C compared to Vienna Pee Dee Belemnite) and morphology (DFA 1) of (a) white muscle or (b) liver tissues of five morphs of bluegill (Lepomis macrochirus): juveniles, sneaker males, satellite males, females, and parental males. DFA 1 scores were generated using discriminant function analysis of partial warp scores generated for each individual (see text for description).

Table 3.2 Stepwise regression and AIC model selection for predictors of diet measured by δ13C values among bluegill morphs (Lepomis macrochirus). Predictors of δ13C values in either white muscle or liver comprised body length (BL), and two measures of shape from a discriminant function analysis (DFA 1, DFA 2). All possible model combinations were considered; only the three models with the greatest explanatory power based on 2 AICc scores for each tissue type are presented with their respective R values. The best model based on AICc and parsimony criteria are shown in bold (see text for details of selection criteria).

2 Model Variables AICc R δ13C White Muscle BL 514.20 0.42 BL + DFA 1 514.66 0.43 BL + DFA 2 516.54 0.42

δ13C Liver BL + DFA 2 551.79 0.37 BL 552.18 0.36 BL + DFA 1 + DFA 2 553.71 0.38

77 liver and white muscle tissues (Pinnegar and Polunin 1999, Dalerum and Angerbjorn

2005, Phillips 2012). Based on the general relationships between morphology and diet observed among a wide range of fish species (e.g. Robinson et al. 2000, Svanbäck et al.

2008, tkint et al. 2012), I expected a littoral diet to be associated with a deeper body shape. However, I found that bluegill parental males and females had the deepest body shapes, but had diets of primarily pelagic zooplankton. Conversely, the sneakers were the most streamlined of the morphs, but had some of the highest δ13C values (i.e. most littoral diet). Furthermore, the models comparing the relationship between diet (δ13C values) to length and shape (DFA 1 and DFA 2) did not identify the main axis of shape variation

(DFA 1) as a significant predictor of diet. Based on these results it appears that foraging tactic alone does not explain shape and resource use among bluegill morphs.

Many studies of fish ecomorphology have focused on individuals that have passed through the period of size-dependent predation (i.e. after the juvenile stage of life).

Predation may influence both habitat and resource use by restricting smaller fish to the littoral habitat until they reach a body size that exceeds the gape of most predators (e.g.

Chipps et al. 2004, Zandona et al. 2011). Indeed, in some species, distinct foraging ecomorphs develop only after the period of intense size-dependent predation has passed.

For example, arctic charr (Salvelinus alpinus) in Lake Thingvallavatn, Iceland, have four distinct ecomorphs that diverge in resource use starting at 70 mm body length, which is well above the 35 mm body length threshold for their most common predator (Malmquist et al. 1992). The extent to which predation delays the divergence into ecomorphs, however, is not yet well understood, and it is likely that other ecological factors, including early juvenile swim performance and foraging efficiency, also influence

78 morphology (Frischer-Rousseau et al. 2010). Nevertheless, previous studies of bluegill have found a predation threshold such that fish less than 100 mm in length are restricted to the littoral habitat (e.g. Mittelbach 1981). Unlike other studies of ecomorphology, the length variation among bluegill morphs includes fish that are well within the size- dependent predation risk threshold. For example, sneaker males in my samples had a mean body length of 83 mm. Sneakers also had a diet of approximately 60% littoral resources despite having a streamlined body shape. The resource use of sneakers is thus consistent with size-dependent predation driving habitat use.

Although body length may be an important predictor of diet for sneakers, it does not account for all of the variation that I observed among morphs. For example, satellite males and juveniles were similar in body length and showed increases in body depth with length, yet juveniles were consistently more streamlined than satellites. This difference in body shape indicates that, while allometric growth may increase body depth as overall size increases, it is not the only factor and at least some shape variation may instead be related to sexual selection. Satellite males gain access to mating opportunities by mimicking female appearance (Dominey 1980), and even though satellite males were on average smaller in length than the females in my sample, they were of a similar body shape. Thus, the shape variation between satellite males and juveniles may be the result of selection on satellite males to mimic the body shape of females (e.g. Dominey 1980).

Additionally, for sneaker males, burst swimming speed is critical to gain proximity to females during spawning and avoid detection by guarding parental males (Stoltz and Neff

2006). Of all the bluegill morphs, sneakers have the fastest burst swim speed, when adjusted for size, which is likely mediated by their streamlined body shape (Chapter 2).

Parental males had the deepest body form, a shape that increases maneuverability (Webb

1984, Collar and Wainwright 2009), which may increase their ability to thwart sneaker spawning attempts or brood predators during parental care (Gross 1982, Gross and

MacMillan 1981). The possibility that sexual selection plays an important role in body shape variation among bluegill morphs deserves further investigation.

My data also suggest that trade-offs exist between reproduction and foraging on female shape. I found that, even though females foraged primarily on pelagic zooplankton, their body shape was deeper than juveniles and sneakers, and consequently, females are probably less efficient at capturing plankton (e.g. Robinson et al. 2000). Females were larger than juveniles and sneakers in my sample, so the increased depth could be partly a result of allometric growth. However, the females had more streamlined head regions than any other group examined, suggesting that the body depth is not solely from allometry. Similar results have been found in perch (Perca fluviatilius; Svanbäck and

Eklov 2002). The increased overall body depth in females may instead be the result of selection on fecundity. Fish show non-linear increases between body size and the volume of eggs that can be carried (Bernardo 1996); therefore, increasing body depth can significantly increase fecundity. Indeed, bluegill females in Lake Opinicon produce an average of about 6000 eggs each breeding season (Gross and Charnov 1980) and larger body size is directly related to the number of eggs females produce (Gross 1980).

Selection for increased fecundity in bluegill might therefore lead to a slightly less optimal overall body shape when it comes to foraging.

In conclusion, the field of ecomorphology provides a solid foundation linking morphology and foraging ecology. Although I found morphological and resource use variation among the bluegill morphs, my results were inconsistent with previous studies of shape and foraging ecology. I instead argue that foraging ecomorphology interacts with other ecological selection pressures, including size-dependent predation and sexual selection. In order to understand the relationships between morphology, function, and adaptation, it is essential to consider the interplay between natural and sexual selection.

3.5 Chapter acknowledgements

I thank M.C. Bellemare, K. Law, C. Rodgers, M. Roubakha, A. Clapp, and T. Hain for assistance in the field and laboratory. I also thank three anonymous reviewers and B.

Wong for helpful comments on a previous version of this paper. This research was carried out with the approval of The University of Western Ontario Council on Animal

Care (animal use protocol No. 2006-062-05), the Université du Québec à Montréal

Council on Animal Care (0509-651-0510), and the Ontario Ministry of Natural Resources

(license numbers 1051389 and 1057026).

This work was supported by the Natural Sciences and Engineering Research Council of

Canada (NSERC) Discovery grants to B.D. Neff, P.R. Peres-Neto, and F.J. Longstaffe, a

Canada Foundation for Innovation (CFI) Leader’s Opportunity Fund award to P.R. Peres-

Neto, CFI and Ontario Research Fund (ORF) infrastructure awards to F.J. Longstaffe,

The University of Western Ontario Graduate Thesis Research Fund, Queen Elizabeth II

Graduate Scholarship in Science and Technology award (QEIIGSST), and Ontario

Graduate Scholarship (OGS) awards to S. Colborne. The work was also made possible, in

81 part, through release time provided to F.J. Longstaffe through the Canada Research

Chairs Program. This research represents Laboratory for Stable Isotope Science (LSIS)

Contribution #299.

3.6 References

Bernardo, J. 1996. The particular maternal effect of propagule size, especially egg size: patterns, models, quality of evidence and interpretations. Amer. Zool. 36:216-236.

Berner, D., R. Kaeuffer, A.C. Grandchamp, J.A.M. Raeymaekers, K. Räsänen, and A.P. Hendry AP. 2011. Quantitative genetic inheritance of morphological divergence in a lake-stream stickleback ecotype pair: implications for reproductive isolation. J. Evol. Biol. 24:1975-1983.

Boecklen, W.J., C.T. Yarnes, B.A. Cook, and A.C. James. 2011. On the use of stable isotopes in trophic ecology. Annu. Rev. Ecol. Evol. Syst. 42:411-440.

Bolnick, D.I., P. Amarasekare, M.S. Araújo, R. Bürger, J.M. Levine, M. Novak, V.H.W. Rudolf, S.J. Schreiber, M.C. Urban, and D.A. Vasseur. 2011. Why intraspecific trait variation matters in community ecology. Trends Ecol. Evol. 26:183-192.

Burnham, K.P., and D.R. Anderson. 1998. Model selection and inference: a practical information-theoretic approach. New York, NY: Springer.

Cabana, G., and J.B. Rasmussen. 1996. Comparison of aquatic food chains using nitrogen isotopes. Proc. Natl. Acad. Sci. USA 93:10844-10847.

Chipps, S.R., J.A. Dunbar, and D.H. Wahl. 2004. Phenotypic variation and vulnerability to predation in juvenile bluegill sunfish (Lepomis macrochirus). Oecologia 138:32-38.

Cochran-Biederman, J.L., and K.O. Winemiller. 2010. Relationships among habitat, ecomorphology and diets of cichlids in the Bladen River, Belize. Environ. Biol. Fish. 88:143-152.

Collar, D.C., and P.C. Wainwright. 2009. Ecomorphology of centrarchid fishes. In: Cooke, S.J., and D.P. Philipp, editors. Centrarchid fishes: diversity, biology, and conservation. Wiley-Blackwell, West Sussex, U.K.

Côté, I.M., and M.R. Gross. 1993. Reduced disease in offspring: a benefit of coloniality in sunfish. Behav. Ecol. Sociobiol. 33:269-274.

Dalerum, F., and A. Angerbjörn. 2005. Resolving temporal variation in vertebrate diets using naturally occurring stable isotopes. Oecologia 144:647-658.

Dominey, W.J. 1980. Female mimicry in male bluegill sunfish – a genetic polymorphism? Nature 284:546-548.

Fischer-Rousseau, L., K.P. Chu, and R. Cloutier. 2010. Developmental plasticity in fish exposed to a water velocity gradient: a complex response. J. Exp. Zool. (Mol. Dev. Evol.) 314B:67-85.

Fry, B.D., D.M. Baltz, M.C. Benfield, J.W. Fleeger, A. Gace, H.L. Hass, and Z.J. Quińones-Rivera. 2003. Stable isotope indicators of movement and residency for brown shrimp (Farfantepenaeus aztecus) in coastal Louisiana marshscapes. Estuaries 26:82-97.

Griffen, B.D., and H. Mosblack. 2011. Predicting diet and consumption rate differences between and within species using gut ecomorphology. J. Anim. Ecol. 80:854-863.

Gross, M.R. 1980. Sexual selection and the evolution of reproductive strategies in sunfish (Lepomis: Centrarchidae). Ph.D. thesis, University of Utah, U.S.A.

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z. Tierpsychol. 60:1-26.

Gross, M.R., and E.L. Charnov. 1980. Alternative male life histories in bluegill sunfish. Proc. Natl. Acad. Sci. USA. 77:6937-6940.

Gross, M.R., and A.M. MacMillan. 1981. Predation and the evolution of colonial nesting in bluegill sunfish (Lepomis macrochirus). Behav. Ecol. Sociobiol. 8:163-174.

Hesslein, R.H., K.A. Hallard, and P. Ramlal. 1993. Replacement of sulfur, carbon, and nitrogen in tissue of growing broad whitefish (Coregonus nasus) in response to a change in diet raced by δ34S, δ13C, and δ15N. Can. J. Fish. Aquat. Sci. 50:2071- 2076.

Jastrebski, C.J., and B.W. Robinson. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

Losos, J.B. 2009. Lizards in an evolutionary tree: ecology and adaptive radiation of Anoles. University of California Press, Los Angeles, C.A.

Malmquist, H.J., S.S. Snorrason, S. Skúlason, B. Jonsson, O.T. Sandlund, and P.M. Jonasson. 1992. Diet differentiation in polymorphic Arctic charr in Thingvallavatn, Iceland. J. Anim. Ecol. 61:21-35.

Mittelbach, G.G. 1981. Foraging efficiency and body size: a study of optimal diet and habitat use by bluegills. Ecology 62:1370-1386.

Neff, B.D., L.M. Cargnelli, and I.M. Côté. 2004. Solitary nesting as an alternative breeding tactic in colonial nesting bluegill sunfish (Lepomis macrochirus). Behav. Ecol. Sociobiol. 56:381-387.

Ojanguren, A.F., and F. Braña. 2003. Effects of size and morphology on swim performance in juvenile brown trout (Salmo trutta L.). Ecol. Freshwat. Fish. 12:241-246.

Perga, M.E., and D. Gerdeaux. 2005. ‘Are fish what they eat’ all year round? Oecologia 144:598-606.

Phillips, D.L. 2012. Converting isotope values to diet composition: the use of mixing models. J. Mammal. 93:342-352.

Pinnegar, J.K., and N.V.C. Polunin. 1999. Differential fractionation of δ13C and δ15N among fish tissues: implications for the study of trophic interactions. Funct. Ecol. 13:225-231.

Pörtner, H.O. 2002. Physiological basis of temperature-dependent biogeography: trade- offs in muscle design and performance in polar ectotherms. J. Exp. Biol. 205:2217-2230.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83:703-718.

Robinson, B.W., and K.J. Parsons. 2002. Changing times, spaces, and faces: tests and implications of adaptive morphological plasticity in the fishes of northern postglacial lakes. Can. J. Fish. Aquat. Sci. 59:1819-1833.

Robinson, B.W., D.S. Wilson, and A.S. Margosian. 2000. A pluralistic analysis of character release in pumpkinseed sunfish (Lepomis gibbosus). Ecology, 81:2799- 2812.

Rohlf, F.J. 2008. tpsDig, Version 2.12. Stony Brook, New York, Department of Ecology and Evolution, State University of New York at Stony Brook. URL: http://life.bio.sunysb.edu/morph/

Rohlf, F.J. 2009. tpsRegr, Version 1.36. Stony Brook, New York, Department of Ecology and Evolution, State University of New York at Stony Brook. URL: http://life.bio.sunysb.edu/morph/

Saunders, M.B., and R.M.R. Bailey. 1992. Ecomorphology of insectivorous bats: a test of predictions using two morphologically similar species. Ecology 73:1335-1345.

Smyntek, P.M., M.A. Teece, K.L. Schula, and S.J. Thackeray. 2007. A standard protocol for stable isotope analysis of zooplankton in aquatic food web research using mass balance correction models. Limnol. Oceanogr. 52:2135-2146.

Stoltz, J.A.,and B.D. Neff. 2006. Male size and mating tactic influence proximity to females during sperm competition in bluegill sunfish. Behav. Ecol. Sociobiol. 59:811-818.

Svanbäck, R., and P. Eklöv. 2002. Effect of habitat and food resources on morphology and ontogenetic growth trajectories in perch. Oecologia 131:61-70.

Svanbäck, R., and P. Eklöv. 2003. Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization? Oikos 102:273-284.

Svanbäck, R., P. Eklöv, R. Fransson, and K. Holmgren. 2008. Intraspecific competition drives multiple species resource polymorphism in fish communities. Oikos 117:114-124.

Tkint, T., E. Verheyen, B. De Kegel, P. Helsen, and D. Adriaens. 2012. Dealing with food and eggs in mouthbrooding cichlids: structural and functional trade-offs in fitness related traits. PloS One 7:e31117.

Webb, P.W. 1984. Body form, locomotion, and foraging in aquatic vertebrates. Am. Zool. 24:107-120.

Zandonà, E., S.K. Auer, S.S. Kilham, J.L. Howard, A. Lopez-Sepulcre, M.P. O'Connor, R.D. Bassar, A. Osorio, C.M. Pringle, and D.N. Reznick. 2011. Diet quality and prey selectivity correlate with life histories and predation regime in Trinidadian guppies. Funct. Ecol. 25:964-973.

Zelditch, M.L., D.L. Swiderski, H.D. Sheets, and W.L. Fink. 2004. Geometric morphometrics for biologists. Elsevier Academic Press. San Diego, USA.

Chapter 4 4 Covariation in stable isotope composition between females and their eggs in a fish: the potential for less invasive inference of resource use3

Stable isotope analysis is frequently used to examine resource use in wild populations, but often involves invasive or lethal methods of collecting tissue samples. The development of less invasive or destructive sampling techniques will expand the possible uses of stable isotopes. Here, I examined if fish eggs meet two basic requirements to infer female resource use: (1) the isotope compositions of eggs must be similar to other maternal tissues known to be related to diet; and (2) isotope composition must remain constant over the egg development period. Using artificial crosses, I tested the relationship between eggs and two commonly sampled maternal tissues (white muscle and liver) in wild caught bluegill sunfish (Lepomis macrochirus). I found that egg isotope composition was strongly correlated to maternal tissues and remained constant from pre- fertilization to the day of hatch, with no change in δ13C and a biologically insignificant increase of 0.3 ‰ in δ15N values. Furthermore, despite lower overall δ13C values in eggs, the results of SIAR mixing models indicated a large degree of overlap in resource use estimates between eggs and the maternal tissues. Overall, eggs can be reliably used to infer pre-breeding foraging ecology of females throughout the egg development period.

3 A version of this chapter has been submitted to the journal Canadian Journal of Fisheries and Aquatic Sciences for publication. Citation: Colborne, S.F., T.J. Hain, F.J. Longstaffe, and B.D. Neff. In review. Covariation in stable isotope composition between females and theirs eggs in a fish: the potential for less invasive inference of resource use. Can. J. Fish. Aquat. Sci.

4.1 Introduction

Ecologists commonly use stable isotope analysis to study foraging patterns of wild populations. Isotopic compositions have been used to provide insight into resource use

(MacAvoy et al. 2008, Derbridge et al. 2012, Ravinet et al. 2013), migration patterns

(Hobson 1999, Rubenstein and Hobson 2004), and trophic relationships (Hobson et al.

1994, Vander Zanden and Rasmussen 1999). An important attribute of stable isotope analysis comes from its potential for inferences about resource use over long periods of time, from days to years depending on the tissues sampled (reviewed by Boecklen et al.

2011). When stable isotopes are applied to fish foraging ecology, samples of white muscle, liver, or blood are typically used, which reflect diet over time scales of several months (white muscle), weeks (liver and blood cellular components), or days (blood plasma; Perga and Gerdeaux 2005, Guelinckx et al. 2007, Boecklen et al. 2011).

However, collection of these tissue types is often through lethal methods or is associated with increased risk of mortality following sampling, which can limit the applicability of stable isotope analysis. There is, thus, interest in developing less invasive sampling techniques for stable isotope analysis.

Non-lethal sampling techniques are increasingly being used to investigate a variety of research questions. For example, samples of nails (Hobson et al. 1996, Cherel et al.

2007), scales (Wainright et al. 1993), feathers (Bearhop et al. 2002), and hair (Hobson et al. 1996, Derbridge et al. 2012) have all been used to examine diet during the growth period of those tissues. More recently, fish eggs have been collected to infer maternal resource use because they can be stripped from gravid females (e.g. Acolas et al. 2008) or retrieved shortly after spawning (e.g. Snowberg and Bolnick 2008, 2012). Because a

88 single female typically produces hundreds or thousands of eggs in most fishes, the impact on individual fitness of sampling a few eggs is much reduced. Certainly these different tissue types will expand the use of stable isotope analysis in foraging ecology.

Prior to using a tissue for stable isotope analysis the relationship between that tissue and diet must be understood. For example, it is well understood that between the two most commonly sampled fish tissues, white muscle has an isotopic composition that reflects diet over a longer time period than liver due to the greater metabolic activity of liver relative to muscle (Perga and Gerdeaux 2005, Buchheister and Latour 2010). Keratin- based tissues, such as hair and nails, can be used to infer past resource use because these tissues reflect diet only over the growth period (Hobson et al. 1999, Bearhop et al. 2002).

The relationship between egg isotopic composition and fish diet is still being developed.

Previous analyses comparing eggs to maternal tissues in two fish species have found that eggs are correlated to both white muscle (e.g. Grey 2001, Snowberg and Bolnick 2008) and liver (e.g. Snowberg and Bolnick 2012). In addition to examining this correlation in additional fish species, an outstanding question is whether development processes during embryogenesis affect egg isotopic composition. Eggs are closed to the entry of new nutrients during development (Kamler 2008); therefore, it is expected that the isotopic composition of the whole egg reflects female diet over the period leading up to breeding and will remain constant throughout the egg development period. To my knowledge, the isotopic composition of fish eggs has not been examined over their post-fertilization development period. For eggs to be reliably utilized to infer maternal diet post- fertilization, the composition must be relatively static over the development period.

Here I used artificial crosses of bluegill sunfish (Lepomis macrochirus) to determine the suitability of sampling eggs for foraging studies. Bluegill are in the Centrarchidae family, which include some of the most common freshwater species in North America (Scott and

Crossman 1998). The relationship between eggs and maternal tissues was examined over the entire egg development period, from pre-fertilization to the day of hatching, to quantify the effects developmental processes have on isotopic composition. I studied a northern population of bluegill. As with most temperate fishes, foraging is significantly reduced during the winter months, and spawning occurs within a few months after foraging resumes in the spring (e.g. Keast 1978). During this pre-breeding period, female bluegill synthesize a large amount of yolk that is deposited into the eggs. After fertilization, the yolk is used to sustain the developing embryo until exogenous feeding commences by larvae (Wiegand 1996, Kamler 2008). Over the winter period fish undergo extensive depletion of body reserves, thus it is assumed that egg yolk is produced directly from the resources consumed during the spring pre-breeding period

(Acolas et al. 2008); therefore, I expected the isotopic composition of eggs to reflect diet of females during the spring period leading up to reproduction.

4.2 Methods

Using daily snorkel surveys of the littoral habitat of Lake Opinicon (Ontario, Canada,

44º34’N, 76º19’W), 40 bluegill were collected (20 males and 20 females) using a dip net on the day of spawning. Collections were made from two separate colonies on June 7 and

8, 2011. Spawning pairs were identified when in a nest together and the female began

“dipping” behaviours, indicating spawning (Gross 1982). Within one hour of capture, the

90 males and females were transported to the Queen’s University Biological Station, located on the shore of Lake Opinicon, and housed in an on-site aquarium facility for no longer than an additional three hours.

Artificial crosses of males and females were completed using the in vitro fertilization technique described by Neff and Lister (2007). Each male and female collected were used once to make an artificial cross, for a total of 20 crosses. Briefly, the eggs were obtained from each female by applying gentle pressure on the abdomen and collecting the eggs in a 500 mL glass jar containing 50 mL of lake water drawn directly from Lake Opinicon. A subsample of eggs from each female was also placed in a 1.5 mL microcentrifuge tube and stored at -20 ºC for subsequent isotopic analysis. Sperm was then collected from each male by applying gentle pressure to the abdomen and gathering the released sperm in a 2 mL syringe. Sperm was then gently mixed with the eggs, following which each jar was filled with lake water. Air stones were attached to the top of each jar to maintain oxygen levels in the water. Each day, 50% of the water was replaced with freshly drawn lake water.

After the egg collection, females were euthanized with an overdose of clove oil and a sample of white muscle tissue was removed from below the posterior portion of the dorsal fin on the right side of the fish. The liver was also removed and both tissue samples were stored at -20 ºC. The males were then released at the site of their original collection. Subsamples of the eggs were taken at 24, 48, and 72 hours post-fertilization, representing the entire egg period of bluegill development (Gross 1982, Neff 2003).

These eggs were collected using a plastic pipette to gently loosen the eggs from the jar,

91 followed by retrieval of 50-75 eggs. The eggs were stored in 1.5 mL microcentrifuge tubes, drained of excess water, and frozen at -20 ºC. The overall health of the eggs in each cross was monitored; if the majority of the eggs in a jar changed from their usual grey colouration to white and lost their attachment to the jar surface they were considered unhealthy or dead, and collections were stopped for that cross.

The isotopic compositions of maternal tissues and egg samples were determined using the methods outlined in Chapter 3 (see Section 3.2.3). The isotopic measurements were also calibrated to international standards and assessed for accuracy and repeatability (see

Appendix B for details).

Atomic carbon to nitrogen ratios greater than 4 are considered to indicate high levels of lipids held within a given tissue, which may decrease δ13C values below those of pure tissue protein (Pörtner 2002, Fry et al. 2003, Boecklen et al. 2011). This ‘lipid effect’ can be compensated for by either chemically removing lipids (Bligh and Dyer 1959) or mathematically adjusting for lipid levels with a mass balance correction (e.g.

McConnaughey and McRoy 1979, Fry et al. 2003). Because chemical lipid removal can alter δ15N values (Smyntek et al. 2007, Boecklen et al. 2011), I chose to correct the δ13C values using the lipid normalization equation developed by McConnaughey and McRoy

(1979) and refined by Kiljunen et al. (2006) for aquatic organisms:

δ13C’ = δ13C + D ,

where

= ;

L is the estimated lipid content of the sample based on its measured atomic ratio of carbon to nitrogen (C:N), δ13C’ is the lipid-corrected value of a sample, δ13C is the measured value of a sample, D is the isotopic difference between lipid and pure protein

(7.02 ‰, Kiljunen et al. 2006), and I is a constant (0.05; Kiljunen et al. 2006).

To compare the δ13C or δ15N values of eggs over the four sampling periods (Day 0, Day

1, Day 2, Day 3), one-factor repeated measures analysis of variance (ANOVA) were used

(dependent factor: δ13C or δ15N value; within-subjects factor: collection day; random effect: female ID) with subsequent Tukey’s post-hoc comparisons, when appropriate. The repeated measure ANOVA included only crosses that had samples for all within-subject factors (i.e. viable eggs sampled from all four collection periods). This criterion removed three samples from the analysis due to egg mortality on Day 2 or Day 3. Repeated measures ANOVAs were also used to compare the three tissues (white muscle, liver, and eggs) collected from females on Day 0 (dependent factor: δ13C or δ15N value; within- subjects factor: tissue type; random effect: female ID). If the assumption of sphericity was violated for either of the repeated measures ANOVAs, the Greenhouse-Geisser (G-

G) correction was applied (Greenhouse and Geisser 1959). I then used analysis of covariance (ANCOVA) to test if the correlation between the isotopic compositions of the maternal tissues and the eggs differed across the sampling days (dependent factor: maternal tissue isotopic composition; co-variate: egg isotopic composition; independent factor: sampling point. If sampling day was determined to not have an effect, linear regression models were used to compare the compositions of the female tissues to the

Day 0 egg values only. Statistical analyses were completed using JMP version 10.0.0

(SAS Institute Inc., Cary, North Carolina, 2012) with α = 0.05 for all tests. Means are reported plus or minus one standard deviation (± 1 SD) unless otherwise stated.

A two-source mixing model in SIAR (Stable Isotope Analysis in R; R version 2.14.2; R

Development Core Team 2012) was also used to estimate the proportion of δ13C and δ15N from littoral and pelagic sources in the three tissue types sampled and all sampling periods for eggs (Parnell et al. 2010). “Consumer” values were represented in the model using the tissue δ13C and δ15N isotopic measurements of each female. The ‘source’ values for the mixing model were represented by the mean measured δ13C and δ15N-values (± 1

SD) of gastropods (snails) and zooplankton (cladocerans and copepods) to represent the littoral and pelagic resource pools, respectively (see Appendix B for details). Mean trophic enrichment factors (TEFs) of 0.47 ± 1.23 (δ13C) and 3.23 ± 0.41(δ15N) calculated across multiple freshwater fish species in northern temperate lakes (Vander Zanden and

Rasmussen 2001) were used as the “fractionation” values in the mixing model because specific TEFs for bluegill are unavailable. The results of the SIAR mixing model are reported using 95% Bayesian credibility intervals (95% BCI).

4.3 Results

Across the four egg collection times, there was no change in the δ13C values of the eggs

(repeated measures ANOVA, F1.31, 19.66 = 3.07, P = 0.09; Fig. 4.1a). In contrast, there was

15 a significant, albeit slight, change in δ N values (repeated measures ANOVA, F1.87, 28.00 =

8.43, P = 0.002; Fig. 4.1b); post-hoc comparisons indicated that Days 0, 1, and 2 were not different from each other (Tukey’s, all P > 0.05), but were all statistically different from

-20

-22

-24

-26

(‰ VPDB)

C -28



-30

Egg Egg

-32

-34

(‰ AIR)

N 9



Egg Egg

6 Day 0 Day 1 Day 2 Day 3 Sampling Period

Figure 4.1 Stable isotopic composition of eggs during embryo development in bluegill

(Lepomis macrochirus). Shown are measurements of δ13C and δ15N values for eggs sampled prior to fertilization (Day 0) and daily until hatching (Day 1, Day 2, Day 3).

Each line represents the stable isotopic compositions for eggs from a single female (n =

20).

Day 3 (all P < 0.05).

On Day 0 there were significant differences between the δ13C values of the tissues and the eggs within females (repeated measures ANOVA, F1.08, 19.44 = 16.5, P < 0.001). Based on post-hoc comparisons, δ13C of white muscle tissue was significantly higher than both liver (by 0.8 ± 1.3 ‰) and eggs (by 1.9 ± 2.0 ‰; Tukey’s, both P < 0.05) and liver δ13C was also significantly higher than eggs (by 1.1 ± 0.8 ‰; Tukey’s, P < 0.05). Similarly,

15 δ N values were significantly different among the tissues (ANOVA, F1.30, 23.36 = 66.1, P

< 0.001), with muscle being significantly higher in δ15N than both liver (by 1.9 ± 0.9 ‰) and eggs (by 0.68 ± 0.9 ‰), and eggs being significantly higher than liver (by 1.3 ± 0.4

‰; Tukey’s, all P < 0.05).

Analysis of covariance (ANCOVA) used to examine the association between the isotopic signature of eggs and the maternal tissues revealed no effect of sampling day (δ13C: both

15 F3,68 ≤ 0.10, P ≥ 0.95; δ N: both F3,68 ≤ 1.94, P ≥ 0.13) or the interaction between

13 15 sampling day and egg isotopic composition (δ C: both F3,68 ≤ 0.15, P ≥ 0.91; δ N: both

F3,68 ≤ 0.18, P ≥ 0.91; Table 4.1). Therefore only linear regression models were used on the Day 0 samples. These models indicated that δ13C values of maternal tissues

(dependent variable, y-axis) were related to those of the eggs (independent variable, x-

2 axis) (white muscle, F1, 18 = 13.6, R = 0.43, P = 0.002; y = 0.40x – 15.04; liver, F1, 18 =

280.3, R2 = 0.94, P < 0.001; y = 0.75x – 5.95; Figure 4.2a). Similarly, δ15N values of the maternal tissues were related to those of the eggs on Day 0 (white muscle, F1, 18 = 15.3,

2 2 R = 0.46, P = 0.001; y = 0.40x + 6.16; liver, F1, 18 = 179.9, R = 0.91, P < 0.001; y =

1.04x – 1.58; Figure 4.2b).

Table 4.1 Summary of linear regression analyses of tissue carbon and nitrogen compositions in bluegill (Lepomis macrochirus). Sample R2 F P Slope Intercept Mean R2 F P Slope Intercept Mean Day (± 95% C.I) (± 95% C.I) Difference (± 95% C.I) (± 95% C.I) Difference (ΔEgg - Tissue) (ΔEgg - Tissue) Carbon White muscle Liver Day 0 0.43 13.62 0.002 0.40 ± 0.2 -15.04 ± 6.1 –1.9 ± 2.0 ‰ 0.94 280.33 < 0.001 0.75 ± 0.1 -5.95 ± 2.5 –1.1 ± 0.8 ‰ Day 1 0.43 13.33 0.002 0.39 ± 0.2 -15.46 ± 6.0 –1.9 ± 2.0 ‰ 0.93 255.21 < 0.001 0.72 ± 0.1 -6.67 ± 2.6 –1.1 ± 0.9 ‰ Day 2 0.43 12.97 0.002 0.40 ± 0.2 -15.13 ± 6.2 –2.0 ± 2.0 ‰ 0.94 258.64 < 0.001 0.74 ± 0.1 -6.34 ± 2.6 –1.1 ± 0.9 ‰ Day 3 0.39 9.73 0.007 0.42 ± 0.3 -14.13 ± 7.6 –1.8 ± 1.9 ‰ 0.93 187.68 < 0.001 0.77 ± 0.1 -5.19 ± 3.2 –1.1 ± 0.8 ‰

Nitrogen White muscle Liver Day 0 0.46 15.28 0.001 0.40 ± 0.2 6.16± 1.2 –0.7 ± 0.9 ‰ 0.91 179.90 < 0.001 1.04 ± 0.1 -1.58 ± 1.4 1.3 ± 0.4 ‰ Day 1 0.51 18.77 < 0.001 0.42 ± 0.2 5.99 ± 1.8 –0.6 ± 0.9 ‰ 0.93 251.48 < 0.001 1.03 ± 0.1 -1.64 ± 1.2 1.3 ± 0.4 ‰ Day 2 0.53 18.96 < 0.001 0.43 ± 0.2 5.80 ± 1.9 –0.6 ± 0.9 ‰ 0.95 302.91 < 0.001 1.07 ± 0.1 -2.00 ± 1.1 1.4 ± 0.3 ‰ Day 3 0.47 13.43 0.002 0.40 ± 0.2 6.09 ± 2.1 –0.3 ± 1.0 ‰ 0.92 177.56 < 0.001 0.99 ± 0.1 -1.50 ± 1.4 1.6 ± 0.4 ‰

N.B. Regressions relate stable isotope values (δ13C or δ15N) of white muscle or liver (dependent variable) to eggs (independent variable). Degrees of freedom are 1, 18 for all regression analyses.

100

Figure 4.2 Isotopic compositions of eggs compared to two maternal tissues in bluegill (Lepomis macrochirus). The measurements of egg (a) δ13C and (b) δ15N are compared to the corresponding (●) white muscle and (○) liver maternal tissue samples from each female (n = 20). The dashed line indicates a 1:1 ratio between the isotopic compositions of eggs and females tissues.

(a) -22

-24

-26

-28

-30

C (‰ VPDB) Female Tissue Female VPDB) (‰ C

 -32

-34 -32 -30 -28 -26 -24 -22

13C (‰ VPDB) Eggs 12 (b)

N (‰ AIR) Female Tissue Female AIR) (‰ N



4 6 7 8 9 10 11

15  N (‰ AIR) Eggs

The two-source SIAR mixing model used to compare the proportion of littoral

(gastropods) and pelagic (zooplankton) contributions to diet indicated that, on average, pelagic resources comprised 86% (white muscle), 92% (liver), or 96% (eggs) of female diet. As expected, the 95% Bayesian credibility intervals revealed a stronger overlap in the estimated resource group proportions for liver and eggs as compared to muscle and eggs (Table 4.2, Fig. 4.3).

4.4 Discussion

In the field of fish ecology there has been interest in using stable isotope analysis of eggs to infer maternal diet. Such inference requires that the isotopic composition of eggs is relatively constant over the egg developmental period and correlates with the isotopic composition of other maternal tissues commonly used to infer diet (i.e. white muscle and liver). Here in bluegill I found that, over the three day egg development period, including from immediately prior to fertilization to the day of hatching, the stable isotopic composition of the eggs did not change in δ13C and increased only slightly in δ15N (by

0.3 ‰ between Day 2 and Day 3). This latter increase in δ15N is considered inconsequential compared to the differences found across trophic levels, which are typically more than 3 ‰ (Post 2002, Anderson and Cabana 2007). Indeed, the δ15N- increase in Day 3 eggs did not change the SIAR mixing model estimates of diet compared to the other egg sampling intervals. There were also strong correlations between egg isotopic composition and that of both the white muscle and liver tissues. The strength of these correlations were similar at each sampling point during egg

100

Table 4.2 Summary of SIAR two-source mixing model estimates of diet based on δ13C and δ15N stable isotope values of bluegill (Lepomis macrochirus). % Littoral % Pelagic White Muscle 18.8 (7.0 – 30.4) 81.2 (70.0 – 93.0) Liver 11.4 (0.5 – 22.0) 88.6 (78.0 – 99.5) Eggs 4.1 (0.0 – 11.3) 95.9 (88.7 – 100) Note: Data are presented as the mean estimate for each tissue type with 95% Bayesian credibility intervals in parentheses.

100

101

Eggs White muscle Liver Day 0 Day 1 Day 2 Day 3 1.00

0.95

0.90

0.85

0.80

Proportion pelagic diet pelagic Proportion 0.75

0.70 Fish tissue type Figure 4.3 SIAR stable isotope mixing model estimates of the pelagic contribution to the diet of female bluegill (Lepomis macrochirus) based on different tissues sampled. Estimates are presented for white muscle, liver, and eggs (Day 0) sampled from females immediately after collection. Also shown are the mixing model estimates for eggs over the three day period following fertilization until the day of hatching. The boxes represent the inner 50% of observations and the mean is represented by the line inside each box. The 90th and 10th percentiles are shown with the whiskers and the 95th and 5th percentiles are represented by the dots.

102

development. And, although there were absolute differences in egg stable isotope and maternal tissue compositions, these differences were relatively small and did not affect estimates of resource use; results of the SIAR mixing model produced similar estimates of diet across eggs and maternal tissues, particularly between liver and eggs. Thus, my data support the use of eggs to infer maternal diets during the breeding season.

Tissues with large amounts of lipid stored within them can bias stable isotope analysis of resource use. Lipids typically contain lower levels of 13C relative to other organic macromolecules (i.e. proteins and carbohydrates), in part because of isotopic fractionation during the initial steps of lipid synthesis (DeNiro and Epstein 1977,

Sweeting et al. 2006). This bias may lower the δ13C values of lipid-rich tissues leading to estimates of resource use that favour the pelagic food groups (e.g. zooplankton). Across fish species, the yolk not only constitutes the majority of egg mass, but also contains large quantities of lipid that is being stored for use during development (Hemming and

Buddington 1988, Kamler 2005). Consistent with past studies, I found that bluegill eggs had δ13C values that were 1 – 2 ‰ lower than the maternal white muscle and liver tissues

(see Grey 2001, Snowberg and Bolnick 2008). Because I mathematically corrected for lipid presence in the eggs, the offset in δ13C observed was likely the result of metabolic processes in the eggs that utilized δ13C-depleted lipids, which then carried the composition forward into the developing tissues. Regardless, the lower δ13C values observed in the eggs had little impact on the results of the SIAR mixing models used to estimate resource use. Thus, these data indicate that, although eggs have lower δ13C

103 values relative to maternal tissues, the effect on inferences of resource use is minimal, particularly when using a two-source mixing model that incorporates both δ13C and δ15N.

There was a notable difference in the strength of the correlations between the isotopic compositions of the eggs and the two maternal tissue types – the correlation with liver was much stronger than with white muscle. This difference may reflect the egg development process. Yolk production, referred to as vitellogenesis, involves the mobilization of lipids from the liver to the eggs which is then used in the production of egg tissues (Wiegand 1996, Dembski et al. 2006). In bluegill, vitellogenesis occurs in mid-May to June (Keast 1978), which coincides with when egg samples were collected.

Given the important relationship between liver and eggs, the strong isotopic correlations between these tissues should be expected. White muscle, in contrast, does not have an active role in providing energy to the eggs, and also reflects diet over a longer time period (Buchheister and Latour 2010). Both of these attributes could erode the correlation between white muscle and egg isotope composition, leading to the lower correlation coefficients.

Given that egg isotopic composition reflects female diet, these results also speak directly to the foraging ecology of bluegill during the few weeks prior to breeding over which eggs are developed and supplied with yolk (Keast 1978, Dempski et al. 2006). My analysis of maternal white muscle indicated greater consumption of littoral resources in the period prior to collection, i.e. during the previous fall. Indeed, stomach content analysis of the same population of bluegill has shown littoral invertebrate consumption to be greater during the fall months, probably related to an increase in littoral prey

104 abundance during that period (Keast 1978). In the spring, however, just before the onset of breeding, my analysis of eggs and liver indicated that bluegill females instead relied more on pelagic resources, and these resources provided the energy required for vitellogenesis. Given that pelagic plankton populations have been negatively impacted by invasive zebra mussels (Dreissena polymorpha) in a number of lakes in North America

(Strayer 2009), there could be severe impacts on reproduction in bluegill. I also found significant variation in the resources consumed by females, ranging from pelagic specialists to generalists consuming both littoral and pelagic resources. The SIAR mixing model indicated that bluegill females consumed primarily pelagic resources overall, but the relatively large 95% credibility intervals support the presence of more generalist females consuming a mixture of both littoral and pelagic resources. These results are consistent with those of other fishes, where individuals tend to specialize on a narrower range of the available resources (e.g. Bolnick et al. 2003, Araújo et al. 2007).

In conclusion, the results of this experiment indicate that the stable isotopic composition of fish eggs is a reliable method to infer female resource use. Additionally, I have inferred that female bluegill rely on pelagic food resource for yolk production during the pre-breeding period.

4.5 Chapter acknowledgements

I thank C. Rodgers, M. Lau, A. Berchtold, and K. Law for assistance in the field and laboratory. This research was carried out with the approval of the University of Western

Ontario Council on Animal Care (animal use protocol No. 2006-062-05) and the Ontario

Ministry of Natural Resources (license number 1062735). This work was supported by

105 the Natural Sciences and Engineering Research Council of Canada Discovery grants to

B.D. Neff and F.J. Longstaffe, Canada Foundation for Innovation and Ontario Research

Fund infrastructure awards to F.J. Longstaffe. Support for S.F. Colborne was provided by the Queen Elizabeth II Graduate Scholarship in Science and Technology and Ontario

Graduate Scholarship awards. The work was also made possible, in part, through release time provided to F.J. Longstaffe through the Canada Research Chairs Program. The data presented in this chapter is part of Laboratory for Stable Isotope Science (LSIS)

Contribution #302.

4.6 References

Acolas, M.L., J.M. Roussel, and J.-L. Baglinière. 2008. Linking migratory patterns and diet to reproductive traits in female brown trout (Salmo trutta L.) by means of stable isotope analysis of ova. Ecol. Freshwat. Fish 17:382-393.

Anderson, C., and G. Cabana. 2007. Estimating the trophic position of aquatic consumers in river food webs using stable nitrogen isotopes. J. North Amer. Benth. Soc. 26:273-285.

Araújo, M.S., D.I. Bolnick, G. Machado, A.A. Giaretta, and S.F. dos Reis. 2007. Using δ13C stable isotopes to quantify individual-level diet variation. Oecologia 152:643-654.

Bearhop, S., S. Waldron, S. Votier, and R.W. Furness. 2002. Factors that influence assimilation rates and fractionation of nitrogen and carbon stable isotopes in avian blood and feathers. Physiol. Biochem. Zool. 75:451-458.

Bligh, E.G., and W.J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37:911-917.

106

Boecklen, W.J., C.T. Yarnes, B.A. Cook, and A.C. James. 2011. On the use of stable isotopes in trophic ecology. Ann. Rev. Ecol. Evol. System. 42:411-440.

Bolnick, D.I., R. Svanbäck, J.A. Fordyce, L.H. Yang, J.M. Davis, D. Hulsey, and M.L. Forister. 2003. The ecology of individuals: incidence and implications of individual specialization. Am. Nat. 161:1-28.

Buchheister, A., and R.J. Latour. 2010. Turnover and fractionation of carbon and nitrogen stable isotopes in tissues of a migratory coastal predator, summer flounder (Paralichthys dentatus). Can. J. Fish. Aquat. Sci. 67:445-461.

Cherel, Y., K.A. Hobson, C. Guinet, and C. Vanpe. 2007. Stable isotopes document seasonal changes in trophic niches and winter foraging individual specialization in diving predators from the southern ocean. J. Anim. Ecol. 76:826-836.

Dembski, S., G. Masson, D. Monnier, P. Wagner, and J.C. Pihan. 2006. Consequences of elevated temperatures on life-history traits of an introduced fish, pumpkinseed Lepomis gibbosus. J. Fish Biol. 69:331-346.

DeNiro, M.J., and S. Epstein. 1977. Mechanism of carbon isotope fractionation associated with lipid synthesis. Science 197:261-263.

Derbridge, J., P.R. Krausman, and C.T. Dairmont. 2012. Using Bayesian stable isotope mixing models to estimate wolf diet in a multi-prey ecosystem. J. Wild. Manage. 76:1277-1289.

Greenhouse, S.W., and S. Geisser. 1959. On methods in the analysis of profile data. Psychometrika 24:95-112.

107

Grey, J. 2001. Ontogeny and dietary specialization in brown trout (Salmo trutta L.) from Lock Ness, Scotland, examined using stable isotopes of carbon and nitrogen. Ecol. Fresh. Fish 10:168-176.

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. J. Comp. Ethol. 60:1-26.

Hemming, T.A., and R.K. Buddington. 1988. Yolk absorption in embryonic and larval fishes. In: Hoar, W.S., and D.J. Randall., eds. Fish Physiology, Vol. 11A, pg. 407 – 446. Academic Press Inc., Boston, M.A.

Hobson, K.A. 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120:314-326.

Hobson, K.A., J.F. Piatt, and J. Pitocchelli. 1994. Using stable isotopes to determine seabird trophic relationships. J. Anim. Ecol. 63:786-798.

Hobson, K.A., D.M. Schell, D. Renouf, and E. Noseworthy. 1996. Stable carbon and nitrogen isotopic fractionation between diet and tissues of captive seals: implications for dietary reconstructions involving marine mammals. Can. J. Fish. Aquat. Sci. 53:528-533.

Kamler, E. 2005. Parent-egg-progeny relationships in teleost fishes: an energetics perspective. Rev. Fish Biol. Fisheries 15:399-421.

Kamler, E. 2008. Resource allocation in yolk-feeding fish. Rev. Fish Biol. Fisheries 18:143-200.

Keast, A. 1978. Trophic and spatial interrelationships in the fish species of an Ontario temperate lake. Environ. Biol. Fish. 3:7-31.

108

Kiljunen, M., J. Grey, T. Sinisalo, C. Harrod, H. Immonen, and R.I. Jones. 2006. A revised model for lipid-normalizing δ13C values from aquatic organisms, with implications for isotope mixing models. J. App. Ecol. 43:1213-1222.

MacAvoy, S.E., R.S. Carney, E. Morgan, and S.A. Macko. 2008. Stable isotope variation among the mussel Bathymodiolus childressi and associated heterotrophic fauna at four cold-seep communities in the Gulf of Mexico. J. Shellfish Res. 27:147-151.

McConnaughy, T., and C.P. McRoy. 1979. Food-web structure and the fractionation of carbon isotopes in the Bering Sea. Mar. Biol. 53:257-262.

Neff, B.D. 2003. Paternity and condition affect cannibalistic behavior in nest-tending bluegill sunfish. Behav. Ecol. Sociobiol. 54:377-384.

Neff, B.D., and J.S. Lister. 2007. Genetic life history effects on juvenile survival in bluegill. J. Evol. Biol. 20:517-525.

Parnell, A.C., R. Inger, S. Bearhop, and A.L. Jackson. 2010. Source partitioning using stable isotopes: coping with too much variation. PLoS One 5:e9672.

Pörtner, H.O. 2002. Physiological basis of temperature-dependent biogeography: trade- offs in muscle design and performance in polar ectotherms. J. Exper. Biol. 205:2217-2230.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83:703-718.

Ravinet, M., P.A. Prodöhl, and C. Harrod. 2013. Parallel and nonparallel ecological, morphological and genetic divergence in lake-stream stickleback from a single catchment. J. Evol. Biol. 26:186-204.

Rubenstein, D.R., and K.A. Hobson. 2004. From birds to butterflies: animal movement patterns and stable isotopes. Trends Ecol. Evol. 19:256-263.

109

Snowberg, L.K., and D.I. Bolnick. 2008. Assortative mating by diet in a phenotypically unimodal but ecologically variable population of stickleback. Am. Nat. 172:733- 739.

Snowberg, L.K., and D.I. Bolnick 2012. Partitioning the effects of spatial isolation, nest habitat, and individual diet in causing assortative mating within a population of threespine stickleback. Evolution 66:3582-3594.

Strayer, D.L. 2009. Twenty years of zebra mussels: lessons from the mollusk that made headlines. Fron. Evol. Environ. 7:135-141.

Sweeting, C.J., N.V.C. Polunin, and S. Jennings. 2006. Effects of chemical lipid extraction and arithmetic lipid correction on stable isotope ratios of fish tissues. Rapid Comm. Mass Spectrom. 20:595-601.

Vander Zanden, M.J., and J.B. Rasmussen. 1999. Primary consumer δ13C and δ15N and the trophic position of aquatic consumers. Ecology 80:1395-1404.

Vander Zanden, M.J., and J.B. Rasmussen. 2001. Variation in δ15N and δ13C trophic fractionation: implications for aquatic food web studies. Limnol. Oceanogr. 46:2061-2066.

Wainright, S.C., M.J. Fogarty, R.C. Greenfield, and B. Fry. 1993. Long-term changes in the Georges Bank food web: trends in stable isotopic compositions of fish scales. Mar. Biol. 115:481-493.

Wiegand, M.D. 1996. Composition, accumulation and utilization of yolk lipids in teleost fish. Rev. Fish Biol. Fisheries 6:259-286.

110

Chapter 5 5 Assortative mating but no evidence of genetic divergence in a species characterized by a trophic polymorphism4

Sympatric speciation requires divergent selection on phenotype and a mechanism of reproductive isolation reducing gene flow within a population. Species that exhibit foraging polymorphisms (“ecomorphs”) often show both behavioural and phenotypic variation mediated by disruptive selection. If the ecomorphs are also reproductively isolated, it is possible that speciation will occur. I examined a population of pumpkinseed sunfish (Lepomis gibbosus) with littoral and pelagic ecomorphs for disruptive selection based on diet and Fulton’s condition factor. Furthermore, I tested for assortative mating between ecomorphs using stable isotope analysis of nesting parental males and eggs from their nests (a proxy for the maternal diet) and genetic differentiation of ecomorphs using microsatellite loci. Although there was evidence of assortative mating based on diet, there was no evidence of genetic differentiation or reduced fitness of “intermediate” body shapes. My results support a growing body of evidence suggesting that genetic differentiation and sympatric speciation is dependent on not only the presence of phenotypic variation and reproductive isolation, but also on the strength of these selection mechanisms.

4 A version of this chapter has been submitted to the journal Evolution for publication. Citation: Colborne, S.F., S.R. Garner, F.J. Longstaffe, and B.D. Neff. In review. Assortative mating but no evidence of genetic differentiation in a species characterized by a trophic polymorphism. Evolution

111

5.1 Introduction

Speciation is the evolutionary process ultimately responsible for the tremendous biological diversity that exists today. Not surprisingly, for centuries biologists have placed considerable emphasis on understanding the mechanisms behind speciation.

Darwin not only titled his seminal work about speciation (‘On the Origin of Species’), but also devoted the sole figure to a schematic of the process (Darwin 1859, Archibald

2009). Traditionally, speciation has been thought to occur in allopatry, as geographic barriers (e.g. islands, mountain ranges) allowed genetic divergence to accumulate between populations through a combination of natural selection for different environmental conditions and passive genetic drift due to reproductive isolation (Mayr

1963, Thorpe et al. 2010, Blair et al. 2013). However, recent evidence indicates that speciation can also occur in sympatry, i.e. without a geographic barrier to gene flow (e.g.

Bolnick 2011, Thibert-Plante and Hendry 2011). Sympatric speciation is predicted to occur under a narrow range of conditions, requiring (1) disruptive natural selection resulting in multiple phenotypes within a population and (2) reproductive isolation that results in (3) genetic differentiation between the phenotypes (Dieckmann and Doebeli

1999, Bolnick and Fitzpatrick 2007). Consequently, understanding the ecological and behavioural mechanisms that contribute to phenotypic divergence within populations, and the conditions under which this divergence leads to sympatric speciation, is of considerable interest.

Foraging ecology is an important source of phenotypic divergence within populations

(e.g. Siwertsson et al. 2010, Schluter 1996). Often there are trade-offs such that generalist foragers, which consume a wide variety of prey items, are at a competitive disadvantage

112 as compared to specialists, which consume only a sub-set of the available prey items

(Schluter 1995, Rueffler et al. 2006). Because specialists have an advantage in acquiring or processing specific prey items, there is ultimately disruptive selection for foraging phenotypes that specialize on different prey resources (e.g. Ackermann and Doebeli

2004, Bernays et al. 2004, Svanbäck and Eklöv 2004). These foraging “ecomorphs” contribute to phenotypic divergence within populations, and, may thus be important precursors to sympatric speciation. However, for divergent selection between ecomorphs to result in sympatric speciation there must also be a mechanism of reproductive isolation that disrupts gene flow between ecomorphs, such as assortative mate choice. Assortative mating can be a passive process when, for example, disruptive selection changes habitat use or the timing of reproduction between ecomorphs; e.g. apple maggot fly (Rhagoletis pomonella; Feder et al. 1994, Filchak et al. 2000). Alternatively, assortative mating can be an active component of sexual selection when individuals differ in their mate preferences; e.g. colour-based mate choice in African Great Lake cichlids (Seehausen and van Alphen 1998, Gray and McKinnon 2007). Sympatric divergence and speciation requires both disruptive selection between phenotypes and related mechanisms that reduce gene flow between ecomorphs.

Freshwater lakes provide ideal environments to study divergence related to foraging ecology. In general, fish in the shallow littoral habitat of lakes consume a variety of benthic invertebrates and exhibit relatively deep body shapes, associated with increased maneuverability to capture cryptic prey in a structurally complex environment (e.g.

Robinson et al. 1996, Svanbäck and Eklöv 2003). In contrast, pelagic fish are relatively more streamlined in body shape, which increases their burst swim speed to catch prey

113 suspended in the water column (e.g. Schluter 1995, Collar and Wainwright 2009). Due to the strong functional relationships between morphology, swim performance, and foraging efficiency in fishes (e.g. Webb 1984, Fisher and Hogan 2007, Collar and Wainwright

2009), the development of trophic polymorphisms are likely related to divergent selection on phenotype. Indeed, trophic polymorphisms have been linked to phenotypic divergence within populations of threespine stickleback (Gasterosteus aculeatus; Schluter and

McPhail 1992, Svanbäck and Schluter 2012), Lake Malawi cichlids (Hulsey et al. 2013), and lake whitefish (Coregonus clupeaformis; Pigeon et al. 1997, Campbell and

Bernatchez 2004, Rogers and Bernatchez 2007).

Northern temperate lakes are of particular interest to studies of divergence and sympatric speciation because these lakes represent “young” environments, and have been colonized only within the last 12,000 years by fishes that were displaced into glacial refugia during the last ice age (Mandrak and Crossman 1992, Robinson et al. 2000). The process of colonization has resulted in resource specialization either by different species or, when heterospecific competitors are absent, specialization within a species (i.e., ecomorphs).

For example, in North America, bluegill (Lepomis macrochirus) and pumpkinseed sunfish (Lepomis gibbosus) co-exist in many lakes across their distribution, with pumpkinseed specializing on benthic invertebrates in the shallow littoral habitat and bluegill specializing on zooplankton in the deeper pelagic habitat (Keast 1978, Robinson et al. 1993). However, in lakes where only one of the sunfish species is present, littoral and pelagic ecomorphs develop. Presumably, when one of these competing species is absent from a lake, the other species can develop littoral and pelagic ecomorphs to occupy both resource niches. Indeed, trophic polymorphisms within populations of

114 bluegill have been identified in Michigan (USA) and introduced populations of bluegill in

Japan (Ehlinger and Wilson 1988, Ehlinger 1990, Yonekura et al. 2002) where pumpkinseed are absent. For pumpkinseed, almost 30 lakes with littoral and pelagic ecomorphs have been reported in the Canadian Shield (Ontario, Canada) and Adirondack

(New York, USA) regions alone (Robinson et al. 2000, Jastrebski and Robinson 2004,

Weese et al. 2012), all in the absence of bluegill.

Here, I sampled fish from a population of pumpkinseed in Ashby Lake (Ontario, Canada,

45º05’N, 77º21’W), a northern temperate lake located on the southern portion of the

Canadian Shield. Ashby Lake covers an area of approximately 260 ha and consists of a shallow littoral habitat, with a variety of benthic invertebrates, which quickly drops off into the deeper pelagic habitat, with abundant zooplankton surrounding islands and rock shoals in the central part of the lake (Jastrebski and Robinson 2004). Pumpkinseed colonized this lake after the glacial retreat some 9,000 to 12,000 years ago (Mandrak and

Crossman 1992). Previous studies have identified the presence of littoral and pelagic foraging ecomorphs in Ashby Lake (Jastrebski and Robinson 2004, Weese et al. 2012), morphology-dependent divergent selection based on growth rate (Jastrebski 2001), and some degree of body shape heritability using laboratory-based rearing experiments of juveniles from both the littoral and pelagic habitats (Parsons and Robinson 2007). A previous study of neutral genetic differentiation between pumpkinseed ecomorphs across

12 lakes, including Ashby Lake, found no evidence of genetic divergence between the ecomorphs, but could not utilize more powerful individual-based analyses using diet or morphology measurements because individual phenotypes were not measured (Weese et al. 2012). Based on these prior findings, the Ashby Lake pumpkinseed population was

115 considered to be a candidate system for sympatric divergence through divergent selection related to resource use.

To examine the potential mechanism for sympatric divergence and speciation I first tested for disruptive natural selection resulting in foraging ecomorphs that differed in morphology, diet, and overall body condition. Body condition is a measure of fitness that has been used in other sunfish populations to test for selection based on morphology within ecomorphs (e.g. Robinson et al. 1996; see also Neff and Cargnelli 2004). Next, I tested for reproductive isolation through assortative mating that could reduce gene flow between foraging ecomorphs by comparing the morphology and isotopic composition of nesting parental males to the isotopic composition of the eggs in their nests. The eggs were used as a proxy for the females’ diet, and presumably ecomorph, during the period leading up to mating (see Chapter 4; also Snowberg and Bolnick 2008, 2012). Finally, I tested for genetic differentiation between foraging ecomorphs to test for population structure based on neutral microsatellite loci. Overall, this approach allowed me to examine the three primary mechanisms predicted to be involved with sympatric speciation of fish populations.

5.2 Methods

5.2.1 Fish collection

In 2011, adult pumpkinseed were collected from Ashby Lake in the spring (May 26 –

June 15; n = 49) and summer (August 21 – 22; n = 37). Approximately equal numbers of fish from the littoral (n = 45) and pelagic habitats (n = 41) were collected either by angling with a piece of earthworm (2 – 3 cm) as bait or by dip-netting from the water

116 column. Immediately after collection, each fish was euthanized with clove oil and a picture of its left side was taken using an Olympus Stylus Tough-6000 (10 megapixel) digital camera. The wet mass (g) and total length (mm) of each fish was measured prior to removing the stomach contents and liver, which were stored at -20ºC for later analysis of diet. During the dissections, the sex and maturity of each individual was determined by examining the reproductive organs. Only reproductively mature fish were included in the analyses because niche shifts are known to occur between juvenile and adult life stages in pumpkinseed (e.g. Osenberg et al. 1988, Arendt and Wilson 1997).

In the spring of 2012 (June 11 – 22), nesting parental males that were actively guarding eggs were collected in the littoral (n=13) and pelagic (n=14) habitats. Prior to collection, each nest was visually monitored to confirm that the male was performing guarding and nest care behaviours. The male was then collected using a dip-net and approximately 100 eggs were sampled from the male’s nest and stored at -20ºC for stable isotope analysis.

As in 2011, the nesting males were euthanized immediately after collection, photographed, and the liver sampled as outlined above. Stomach contents were not collected in 2012 because nesting males do not actively forage and are therefore unlikely to have stomach contents that are representative of their diet (Gross and MacMillan

1981).

Samples of potential invertebrate prey were also collected during each of the three periods when pumpkinseed were sampled. Zooplankton samples were collected from pelagic rock shoals using a vertical tow net (mesh size 0.5 μm; depth of 3 – 4 m, repeated three times per site). Benthic invertebrates were collected from the littoral habitat using

117

D-net sweeps of the submerged macrophyte vegetation and the upper 1 – 2 cm of sediment. The D-net samples were then hand sorted through a series of nested sieves to collect littoral benthic invertebrates that were classified to the nearest order. For isotope analysis, gastropods (snails) were manually removed from their shells, but all other benthic invertebrate prey were analyzed intact.

5.2.2 Morphological variation

Using tpsDig software (Rohlf 2008), 15 homologous landmarks were placed on each of the pumpkinseed images (n = 113; see Jastrebski and Robinson 2004 for landmark locations). These landmarks were used to calculate partial warp coefficients for each individual, which allow body shape to be examined independently of body size (see

Chapter 2, Zelditch et al. 2004). Variation in warp coefficients was further partitioned into axes of major variation using a discriminant function analysis (DFA) comparing four groups based on collection habitat and sex: pelagic males (n = 35), pelagic females (n =

20), littoral males (n = 32), or littoral females (n = 26). Subsequent statistical analyses focused on only those DFA axes that explained at least 20% of the total variation in shape. For each significant DFA axis, two-factor ANOVA models were used to examine variation in DFA score (dependent variable) between sexes and collection habitats

(independent factors) and their interaction, with sampling period included as a random effect. Significant differences in body shape identified by these analyses were then visualized using thin-plate splines (Rohlf 2009).

118

5.2.3 Diet analysis

The preserved stomach content samples of each fish were thawed to room temperature and sorted using a dissection microscope into one of four prey groups: zooplankton

(copepods and cladocerans), molluscs (gastropods, bivalves), benthic prey

(ephemeroptera, trichoptera, odonates, and amphipods), and “other” (terrestrial insects, fish eggs, plant material, unidentifiable contents). Prey samples from each fish were then dried at 50ºC for 24 hours and the dry mass (mg) of each prey type was determined. The proportion of each prey type (dry mass of prey type/total dry mass of all stomach contents) from each fish was arcsine transformed and used in a two-factor MANOVA test of the four prey groups (dependent variables: proportion of each prey type in diet; independent variables: collection habitat and sex, and their interaction). If an independent factor (habitat or sex) was found to be significant in the MANOVA, separate t-tests were used for each prey group to compare that independent variable. Only fish with measurable stomach contents were included in the analyses.

Next, stable isotope analysis was conducted in the Laboratory for Stable Isotope Science

(LSIS) at The University of Western Ontario (London, Ontario Canada). The liver tissue samples of each fish and eggs (from the nests of parental males collected in Summer

2012), were prepared for stable isotope analysis following the procedure outlined in

Chapter 3 (see Section 3.2.3). The measured δ13C values were corrected for the presence of lipids using the mass balance correction for aquatic organisms outlined in Chapter 4

(see Section 4.2). Two-point curves were used to calibrate δ13C and δ15N values using international standards and precision and accuracy were monitored using internal laboratory standards and sample replicates (see Appendix C for details).

119

To create group estimates of the mean resource use between habitats (littoral and pelagic) and sexes in each sampling period I used a SIAR (Stable Isotope Analysis in R) two- member mixing model for estimating the contribution of gastropods (littoral prey) and zooplankton (pelagic prey) to diet (Parnell et al. 2010). The SIAR mixing model incorporates both δ13C and δ15N values of each fish and the variability both between and within the prey resources of each habitat (Parnell et al. 2010). The ‘source’ variables of the model were based on the mean (± 1 SD) isotopic composition of gastropods (littoral habitat) and zooplankton (pelagic habitat). Due to temporal variability in prey isotopic compositions, separate mixing models were used for each of the collection periods with unique ‘source’ values (see Appendix C for prey isotopic composition details, Table C.1).

Trophic fractionation was estimated using the mean trophic enrichment factors (TEFs) across multiple freshwater fish species in northern temperate lakes (δ13C: 0.47 ± 1.23 ‰;

δ15N: 3.23 ± 0.41 ‰; Vander Zanden and Rasmussen 2001) because species-specific

TEFs for pumpkinseed are unavailable.

To obtain individual estimates of diet, which are not available using the SIAR models, a two-end-member mixing model based on Post (2002) was used, where:

13 13 13 13 % Littoral = (δ Cconsumer - δ Cpelagic) / (δ Clittoral - δ Cpelagic) × 100

13 13 13 where δ Cconsumer is the measured δ C value of each fish, δ Cpelagic is the mean of

13 pelagic zooplankton, and δ Clittoral is the mean of littoral gastropods. The % Littoral proportion estimates were arcsine transformed and used in a two-factor ANOVA

120

(dependent variable: % Littoral; independent factors: collection habitat and sex, plus their interaction; random effect: sampling period).

5.2.4 Morphology, diet, and condition

The relationship between morphology and diet was first tested using a linear model that included the % Littoral estimates for each individual (dependent variable), total body length and DFA 1 scores of shape (independent factors), and sampling period (random effect). Next, I constructed an ecomorph “score” that combined the morphological (DFA

1) and diet (% Littoral) data using principal component analysis (PCA). Given that %

Littoral estimates ranged from low values for pelagic consumers to high values for littoral consumers, whereas DFA 1 scores ranged from high values for pelagic body shape to low values for littoral body shape (see results below), the DFA 1 scores for each individual were first multiplied by –1 before use in the PCA. This multiplication simply facilitated interpretation of the axis loadings. The first principal component (PCA 1) subsequently had positive loadings on both variables (see below) and was used as the ecomorph score.

Fulton’s condition factor was calculated for each fish using the wet mass (g) and total body length (mm) (K = mass/length3 × 105). The condition factor provides an estimate of overall energetic state for each fish (e.g. Neff and Cargnelli 2004, Magee et al. 2006).

Condition factor values (dependent variable) were then compared using ANCOVA models that included sex (independent factor) and ecomorph score (covariate) and sampling period (random effect). Separate models were run for fish collected from the littoral and pelagic habitats. Sex was determined to not be a significant variable in the

ANCOVA above (see below) and was removed from the analysis. Subsequently,

121 ecomorph score and Fulton’s condition factor (K) were correlated using separate

Pearson’s correlation coefficients for littoral and pelagic-caught fish.

5.2.5 Assortative mating and genetic differentiation

To test for assortative mating, the ecomorph scores for the parental males collected in

Spring 2012 were used in a Pearson’s correlation with the δ13C value of eggs collected from the male’s nest (a proxy for female diet and ecomorph). A positive correlation would indicate assortative mating within the littoral and pelagic foraging ecomorphs.

Next, to test for genetic differentiation, DNA from each of the 113 adult fish in this study was extracted using a proteinase K digestion (Neff et al. 2000). Each individual was then genotyped at nine previously described microsatellite loci (Colbourne et al. 1996: LMA

29, LMA 87; DeWoody et al. 1998: RB7, RB20; Neff et al. 1999: LMA 116, LMA 122,

LMA 124; Schable et al. 2002: LMAR 10, LMAR 14). The microsatellite products were visualized using a CEQ 8000 (Beckman Coulter) and manually scored in relation to a known size standard. Micro-checker was used to determine if microsatellite allele frequencies deviated significantly from the expectations of Hardy-Weinberg equilibrium

(Van Oosterhout et al. 2004). Only LMAR14 deviated significantly from Hardy-

Weinberg equilibrium, showing a homozygote excess consistent with the presence of a null allele. Consequently, this locus was included only in the genetic analyses that accommodate null alleles (Structure, Fst), but excluded from analyses that may be biased by null alleles (individual genetic distances).

The microsatellite dataset was first used to test for the presence of discrete genetic groups in Ashby Lake using the Bayesian clustering method implemented in the program

122

Structure v2.3.3 (Pritchard et al. 2000). Specifically, the presence of two genetic clusters was tested using a model that allowed for admixture and assumed independence among loci (uncorrelated allele frequencies). To ensure the results converged on a single solution, the model was run using five replicate simulations of 50,000 burn-in steps followed by 150,000 resampling steps. The results were then aggregated using Structure harvester and Clumpp (Earl and vonHoldt 2011, Jakobsson and Rosenberg 2007).

Next, the global Fst (Weir 1996) was calculated using the null allele correction implemented in FreeNA (Chapuis and Estoup 2007) to specifically examine genetic differentiation between fish collected from littoral and pelagic habitats. These comparisons were run both for all fish, and for just the subset of nesting males that were collected in 2012. For each test, significance was assessed by resampling over loci to generate 95% confidence intervals from 1000 bootstrap replicates.

Finally, a relationship between the pairwise genetic distance estimates between individuals and the difference in ecomorph score between those individuals was examined. Again, these comparisons were run both for all fish, and for just the nesting males that were collected in 2012. Genalex 6.4.1 was used to calculate the matrix of genetic distances between individuals, and to compare the genetic distance matrix to the ecomorph distance matrix using a Mantel test with 999 permutations to assess significance (Peakall and Smouse 2006).

123

5.3 Results

5.3.1 Morphological variation between habitats

Body shape differed significantly among the four pumpkinseed groups (pelagic males, pelagic females, littoral males, littoral females; DFA: Wilks’ λ = 0.17, P < 0.001; Fig.

5.1), with DFA 1 and DFA 2 accounting for 54% and 34%, respectively, of total variation in shape. Further examination of the DFA 1 and DFA 2 scores indicated males had higher values than females on both axes (DFA 1: ANOVA, F1, 41.4 = 22.11, P < 0.001; DFA 2:

F1, 48.05 = 48.05, P < 0.001). The DFA values also differed between collection habitats with pelagic caught fish having higher DFA 1 scores than littoral fish (ANOVA, F1, 108.7 =

108.19, P < 0.001), but littoral fish having higher DFA 2 values than pelagic fish

(ANOVA, F1, 107.1 = 25.98, P < 0.001). There were no interaction effects between sex and collection habitat for either DFA axis (DFA 1: ANOVA, F1, 105.8 = 3.07, P = 0.08; DFA 2:

F1, 108.1 = 0.0001, P = 0.99; Fig. 5.1). Visualization of DFA 1 using thin-plate splines showed that lower values (i.e. littoral females) were associated with decreased body depth in the mid-body and posterior region, whereas higher DFA 2 values (i.e. littoral males) were associated with a larger head region, reduced tail depth, and a more horizontal pectoral fin orientation (Fig. 5.1).

5.3.2 Diet Analysis

Of the fish collected for stomach content analysis, 74% (64 of 86 fish) had measureable contents. Comparisons across prey types indicated that overall there were significant differences in the stomach contents of pumpkinseed based on habitat or sex (MANOVA;

Wilks’ λ = 0.58, df = 12, 151.1, P = 0.001). Comparisons of the independent variables indicated that stomach contents of the prey groups differed between collection habitats

124

(F3, 58 = 0.40, P < 0.001; Table 5.1), but there were non-significant differences between the sexes (F3, 58 = 0.13, P = 0.07) and no significant interaction (F3, 58 = 0.05, P = 0.40).

Further comparisons of prey types between collection habitats indicated that pelagic caught fish consume more zooplankton and fewer benthic invertebrates than those caught in the littoral habitat (Table 5.1).

The isotopic compositions of potential prey types indicated that gastropods could be used as an accurate proxy for the littoral prey. Additionally, gastropods and zooplankton were consistently different in isotopic composition and therefore the resources from the littoral and pelagic habitat had the required differentiation between prey types for isotopic estimates of fish resource use (see Appendix C for details).

Isotopic compositions (δ13C and δ15N) of pelagic- and littoral-caught pumpkinseed varied among the sampling periods and collection habitats (Appendix C Fig. C.1). SIAR mixing model estimates of diet indicated that while there was variation between males and females across the sampling periods, littoral-caught fish consumed 33 - 69% littoral resources as compared to 3 – 54% littoral resources in the diets of pelagic-caught fish

(Table 5.2, Fig. 5.2). Individual-based estimates of % Littoral contribution to diet indicated that the fish caught in the littoral habitat consumed more littoral resources

(69%) than did pelagic fish (44%; ANOVA: F1, 107.5 = 16.85, P < 0.001). There were non- significant differences in % Littoral contribution between males (62%) and females

(51%; ANOVA: F1, 101 = 3.42, P = 0.07). There was, however, a significant interaction between sex and collection habitat, with males consuming more littoral resources than

125

Figure 5.1 Discriminant function analysis (DFA) of body shape variation among four groups of pumpkinseed sunfish (Lepomis gibbosus): littoral males, littoral females, pelagic males, pelagic females. The plot depicts the mean (± 1 SD) values of each pumpkinseed group for the first two discriminant function axes (DFA 1 and DFA 2). Thin-plate splines below the scatterplot depict the maximum and minimum observed values for each DFA axis at 3× magnification.

126

Table 5.1 Summary of the stomach contents of pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats. The mass and proportion of total stomach mass estimates for each other the four prey groups are presented as the mean ± 1 standard error. Test statistics (t-stat, df, and p-value) comparing the proportion of stomach content mass between collection habitats are also shown. Mean mass (mg) Mean proportion of diet Habitat comparison Littoral n Pelagic n Littoral Pelagic t-stat df P-value Zooplankton 0.65 ± 0.31 32 14.04 ± 5.91 32 0.09 ± 0.04 0.50 ± 0.08 4.44 62 < 0.001 Molluscs 3.23 ± 1.04 32 5.31 ± 4.54 32 0.20 ± 0.06 0.10 ± 0.30 -1.01 62 0.31 Benthic prey 21.08 ± 4.24 32 29.98 ± 25.70 32 0.39 ± 0.07 0.13 ± 0.06 -2.52 62 0.01 Other 7.24 ± 2.46 32 21.36 ± 13.15 32 0.32 ± 0.07 0.27 ± 0.06 -0.70 62 0.49

126

127

Table 5.2 Summary of the isotopic compositions and mixing model diet estimates of pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats over three sampling periods. Isotopic compositions of liver tissue were presented as the mean (± 1 SD). SIAR mixing model estimates of the proportion of diet from littoral resources are presented for each sex (and combined) for each sampling period, estimates are presented as the mean and 95% Bayesian credibility interval values. SIAR – Proportion Littoral Estimates Sampling Collection Period Habitat δ13C δ15N Males Females Sexes Combined Spring 2011 Littoral –25.5 ± 2.0 +6.4 ± 1.1 0.83 (0.69 – 0.99) 0.57 (0.42 – 0.72) 0.69 (0.56 – 0.81) Pelagic –26.6 ± 1.5 +7.5 ± 0.7 0.55 (0.43 – 0.68) 0.50 (0.31 – 0.69) 0.54 (0.44 – 0.64)

Summer 2011 Littoral –23.1 ± 1.9 +6.1 ± 0.7 0.39 (0.17 – 0.60) 0.28 (0.05 – 0.51) 0.33 (0.17 – 0.50) Pelagic –24.0± 1.7 +6.8 ± 0.5 0.10 (0 – 0.22) 0.04 (0 – 0.11) 0.03 (0 – 0.08)

Spring 2012 Littoral –23.0 ± 1.3 +6.7 ± 0.5 0.68 (0.48 – 0.88) -- -- Pelagic –25.7 ± 0.7 +7.5 ± 0.3 0.25 (0.12 – 0.39) -- -- Note: Only nesting parental males were collected in Spring 2012 (see methods for details).

127

128

Spring 2011 Summer 2011 Spring 2012 1.0

0.8

0.6

0.4

0.2 Proportion littoral diet littoral Proportion Males Females 0.0 Littoral Pelagic Littoral Pelagic Littoral Pelagic

Collection habitat Figure 5.2 Boxplots of SIAR isotope-mixing model estimates of the littoral prey resource contribution to the diets of male and female pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats. Stable isotopic compositions of liver tissues were used in independent mixing models for each sex and sampling period. The boxplots represent the inner 50% of observations, with the mean value indicated by the line within each box. The whiskers represent the 90th and 10th percentiles and dots are the 95th and 5th percentiles.

129 females in the littoral habitat, but more zooplankton than females in the pelagic habitat

(ANOVA: F1, 108.3 = 6.84, P = 0.01). Isotopic compositions of each fish can be found in

Appendix C (Tables C.2 and C.3).

5.3.3 Morphology, diet, and condition

Analysis of covariance found that the % Littoral diet estimates were not related to body length of the pumpkinseed (F1, 109.4 = 1.25, P = 0.27). However, the % Littoral contribution to diet was related to body shape, lower DFA 1 scores (i.e. increased body depth in the anterior region) were associated with a higher contribution of littoral resources to diet (F1, 109.3 = 7.22, P = 0.008).

The first axis (PCA 1) from the principal component analysis of DFA 1 and % Littoral values captured 60% of the variation among individuals and had positive loadings for both variables. This axis thus provided an ecomorph score, where higher PCA 1 values were associated with a littoral body shape and diet, whereas lower PCA 1 scores were associated with a pelagic body shape and diet. There were non-significant relationships between Fulton’s condition factor and ecomorph score in either littoral- (ANCOVA; F1,

53.14 = 0.38, P = 0.54) or pelagic-caught fish (ANCOVA; F1, 50.18 = 1.08, P = 0.30).

Additionally, condition factor did not differ between males and females collected in either the littoral (ANCOVA; F1, 54.36 = 0.21, P = 0.65) or pelagic habitat (ANCOVA; F1,

50.83 = 2.56, P = 0.11) and sex was removed from the analysis. There was a significant correlation between ecomorph score and condition factor in pelagic-caught fish

(Pearson’s r = -0.34, n = 54, P = 0.01), but not littoral-caught fish (Pearson’s r = -0.17, n

= 52, P = 0.44; Fig. 5.3).

130

(a) Littoral-caught (b) Pelagic-caught

2.8 2.8

2.6 2.6

2.4 2.4

2.2 2.2

2.0 2.0

1.8 1.8

Condition Factor Condition

Condition Factor Condition 1.6 1.6

1.4 1.4

1.2 1.2 -4 -2 0 2 4 -4 -2 0 2 4 Ecomorph score Ecomorph score

Figure 5.3 Relationship between body condition and ecomorph scores of (a) littoral and (b) pelagic-caught pumpkinseed sunfish (Lepomis gibbosus) collected from Ashby Lake, Ontario. Condition was estimated using Fulton’s condition factor. Ecomorph scores were generated for each individual based on DFA 1 scores and % Littoral diet values (see methods for details).

130

131

5.3.4 Assortative mating and genetic differentiation

There was a significant relationship between a nesting male’s ecomorph score and the

δ13C values of eggs from his nest; nesting males with a more littoral ecomorph score had eggs with higher (i.e. more littoral) δ13C values (Pearson’s r = 0.48, n = 27, P = 0.01; Fig.

5.4). The microsatellite data, however, did not provide any evidence of genetic differentiation between the ecomorphs. First, the Structure analysis did not identify discrete genetic clusters. When the data were fit to a model of two genetic clusters based on collection habitat, all individuals had intermediate membership in each cluster and the membership coefficients were not related to ecomorph scores (Fig. 5.5). The Fst values also did not indicate significant divergence between littoral and pelagic caught fish when comparing across fish from all sampling periods (Fst = 0.0004, n = 113, 95% CI: -0.0029

– 0.0038) or only the nesting males from Spring 2012 (Fst = 0.0002, n = 27, 95% CI: -

0.0095 – 0.0131). Finally, there was no relationship between genetic distance and the ecomorph score (all fish: r = -0.01, n = 113, P = 0.44; nesting males only: r = 0.04, n =

27, P = 0.26).

5.4 Discussion

In freshwater fishes, divergent selection in littoral versus pelagic habitats can result in foraging ecomorphs that have predictable differences in morphology and diet (e.g.

Skúlason and Smith 1995, Robinson et al. 2000). Here, I found that littoral-caught pumpkinseed had a deeper head region and were less streamlined overall than pelagic fish. Both stomach content and stable isotope analyses indicated that littoral-caught pumpkinseed consumed more gastropods and other benthic invertebrates, but fewer

132

-21

littoral -22

-23

-24

(‰ VPDB) -25

-26

C eggs C

13  -27

-28

pelagic -29 -3 -2 -1 0 1 2 pelagic littoral Male ecomorph score

Figure 5.4 Relationship between nesting male ecomorph scores and egg isotopic composition of pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats. Ecomorph scores for each nesting male were generated based on DFA 1 morphology scores and % Littoral resource use estimates (see methods for details). The δ13C values of eggs were used as a proxy for maternal diet and ecomorph.

133

1.0

0.8

0.6

0.4

0.2

0.0

Membershipcoefficient 20 40 60 80 100 pelagic littoral Ecomorph score (rank)

Figure 5.5 Genetic clustering of individual pumpkinseed sunfish (Lepomis gibbosus) collected from the littoral and pelagic habitats. Each vertical bar represents one individual, presented in rank order based on ecomorph scores, and indicating proportional membership coefficients in the two genetic clusters modelled by Structure (see methods for details).

134

zooplankton as compared to pelagic-caught fish. There was also a link between body shape and diet independent of habitat associations, indicating that pumpkinseed with deeper bodies consumed more littoral invertebrates and fewer zooplankton than pumpkinseed with shallower body shapes. These data thus support the presence of foraging ecomorphs in the Ashby Lake pumpkinseed population.

Foraging ecomorphs may exist as discrete phenotypes or as a phenotypic continuum in which individuals display varying degrees of specialization towards resource types

(Moles et al. 2010, Ellerby and Gerry 2011). Polymorphic pumpkinseed populations have been found to range from continuous variation in phenotypes between littoral and pelagic ecomorphs to nearly bimodal discrete foraging ecomorphs (e.g. Robinson et al. 1996).

Based on my data, littoral- and pelagic-caught pumpkinseed in Ashby Lake differed significantly in diet and body shape, but there was considerable variation and overlap between the ecomorphs. The high frequency of “intermediate” phenotypes in this population may result from weak selection for resource specialization. For example, I did not find evidence that Fulton’s condition factor, a correlate of energetic condition and fitness in sunfish (Neff and Cargnelli 2004, Magee et al. 2006), was related to ecomorph score in the littoral habitat, but these variable were related in pelagic-caught fish. Similar results were reported for the pumpkinseed sunfish of Paradox Lake (New York, USA), a population characterized by a roughly bimodal distribution of body shape variation between ecomorphs (Robinson et al. 1996). Taken together, these data suggest that selection for resource specialization may be similar across polymorphic pumpkinseed

135 populations, and, that selection pressures relating condition to ecomorph may be stronger in the pelagic habitat.

Whether divergent selection is strong or weak, for sympatric speciation to occur there must also be a mechanism of reproductive isolation, such as assortative mating to disrupt gene flow between ecomorphs. I found that there was a positive relationship between the ecomorph scores of nesting male pumpkinseed and the diet, and presumably ecomorph, of the females with whom he mated (inferred from the isotopic composition of the eggs in the nests). Such assortative mating may occur passively when ecomorphs forage and breed in different habitats (e.g. Feder et al. 1994, Snowberg and Bolnick 2008, 2012) as is likely in this study population – nesting males were generally separated by ecomorphs into the littoral and pelagic habitats during breeding. Assortative mating can also occur through active mate choice, as has been demonstrated in other sympatric populations of fishes (e.g. Seehausen and van Alphen 1998, Gray and McKinnion 2007). Although mate choice was not directly examined, I observed moderate variation in the isotopic composition of eggs collected from nests within each habitat, suggesting that females at least may not always breed in the same habitat in which they feed. These data thus indicate that there is assortative mating by ecomorph in Ashby Lake, probably as the result of spatial separation of the ecomorphs during breeding.

Regardless of the mechanism mediating assortative mating in the Ashby pumpkinseed, if gene flow was significantly reduced between littoral and pelagic ecomorphs, then I would expect to find genetic differentiation. However, using a variety of individual-based analyses of microsatellite loci, there was no evidence of neutral genetic differentiation

136 between littoral and pelagic ecomorphs. This lack of neutral genetic differentiation between ecomorphs could, in part, be a reflection of the relatively short amount of time that has passed since the lakes were re-colonized after the last ice age. Ashby Lake has been populated for at most 12,000 years (Mandrak and Crossman 1992). In comparison,

African rift lake cichlid species flocks have been diverging for between 2.5 to 4.5 million years in Lake Malawi and between 190,000 and 270,000 years in Lake Victoria (Genner et al. 2007). Genetic divergence in pumpkinseed may be found if we were to examine functional loci, such as though related to body shape, that are likely to be under selection during the evolution of foraging ecomorphs. Indeed, a crater lake population of Midas cichlids (Amphilophus spp.) showing similar littoral-pelagic resource divergence have been shown to differ at a few functional loci related to shape and fin placement rather than neutral loci after 22,000 years of divergence (Franchini et al. 2013). If disruptive natural selection and reproductive isolation is occurring in the Ashby Lake pumpkinseed,

I predict that functional loci related to morphological traits associated with foraging, e.g. body shape, will differ between littoral- and pelagic-caught fish before there is detectable variation in neutral loci.

The rate of gene flow between littoral and pelagic foraging ecomorphs is also a critical factor influencing the rate at which genetic differentiation could occur in sympatry. For example, northern temperate fishes have been shown to have considerable phenotypic plasticity related to foraging phenotypes. Foraging ecomorphs of both pumpkinseed and arctic charr (Salvelinus alpinus), for example, have been experimentally shown to arise largely related to phenotypic plasticity during development with only a low heritable component (Robinson and Wilson 1996, Adams and Huntingford 2004, Parsons and

137

Robinson 2006). Consequently, there may only be weak selection on “hybrids” (offspring with parents that differ in their ecomorphology) because offspring can basically develop into either ecomorph. Alternative reproductive tactics within sunfish may also maintain gene flow between foraging ecomorphs if cuckolders (i.e. sneaker and satellite males; see

Gross 1982) do not show the same patterns of assortative mating I reported in nesting parental males and females. Indeed, there is evidence that cuckolders have limited recognition of other species (Garner and Neff 2013), let alone foraging ecomorphs within their own species. Therefore, it is possible that cuckolders may be a major source of gene flow between foraging ecomorphs within lakes. Regardless of the mechanisms involved, without strong disruption to the gene flow between foraging ecomorphs, it is unlikely that genetic differentiation or sympatric speciation will occur (Thibert-Plante and Hendry

2011).

In conclusion, identifying cases of sympatric speciation, and more generally, the conditions under which the divergence begins, is a central goal of evolutionary biology.

Resource specialization is thus of particular interest because it provides a mechanism by which adaptive differentiation can occur even in the absence of geographic barriers to gene flow. Despite finding evidence of assortative mating by ecomorph in the Ashby pumpkinseed population, I found no evidence of genetic differentiation. Rather than diverging toward speciation, the Ashby Lake ecomorphs may be maintained within a single gene pool by weak selection against intermediate phenotypes or condition- dependent phenotypic plasticity. These data add to a growing number of examples in wild populations showing that sympatric speciation is only likely to occur under a limited range of conditions (e.g. Bolnick 2011, Martin 2013), and is rarely found without both

138 strong selection against intermediate phenotypes and reproductive isolation arising from assortative mating.

5.5 Chapter Acknowledgements

I thank C. Rodgers, M. Lau, A. Berchtold, A. Houde, and K. Law for assistance in both the field and laboratory. This research was carried out with the approval of The

University of Western Ontario Council on Animal Care (protocol number 2006-062-05) and the Ontario Ministry of Natural Resources (scientific collection permit numbers

1062675 and 1068338). Funding for this research was provided through the Natural

Sciences and Engineering Research Council of Canada Discovery Grants to B.D. Neff, and F.J. Longstaffe, the Canada Foundation for Innovation and Ontario Research Fund infrastructure awards to F.J. Longstaffe, Queen Elizabeth II Graduate Scholarship in

Science and Technology and Ontario Graduate Scholarship awards to S.F. Colborne. The data contained in this chapter represents Laboratory for Stable Isotope (LSIS) contribution #307.

5.6 References

Ackermann, M., and M. Doebeli. 2004. Evolution of niche width and adaptive diversification. Evolution 58:2599-2612.

Adams, C.E., and F.A. Huntingford. 2004. Incipient speciation driven by phenotypic plasticity? Evidence from sympatric populations of Arctic charr. Biol. J. Linn. Soc. 81:611-618.

Archibald, J.D. 2009. Edward Hitchcock’s pre-Darwinian (1840) “tree of life.” J. Hist. Biol. 42:561-592.

139

Arendt, J.D., and D.S. Wilson. 1997. Optimistic growth: competition and an ontogenetic niche-shift select for rapid growth in pumpkinseed sunfish (Lepomis gibbosus). Evolution 51:1946-1954.

Bernays, E.A., Singer, M.S., and D. Rodrigues. 2004. Foraging in nature: foraging efficiency and attentiveness in caterpillars with different diet breadths. Ecol. Entomol. 29:389-397.

Blair, M.E., Sterling, E.J., Dusch, M., Raxworthy, C.J., and R.G. Pearson. 2013. Ecological divergence and speciation between lemur (Eulemur) sister species in Madagascar. J. Evol. Biol. 26:1790-1801.

Bolnick, D.I. 2011. Sympatric speciation in threespine stickleback: why not? Int. J. Ecol. 2011:Article ID 942847. doi: 10.1155/2011/942847.

Bolnick, D.I., and B.M. Fitzpatrick. 2007. Sympatric speciation: models and empirical evidence. Ann. Rev. Ecol. Evol. Syst. 38:459-487.

Campbell, D., and L. Bernatchez. 2004. Generic scan using AFLP markers as a means to assess the role of directional selection in the divergence of sympatric whitefish ecotypes. Mol. Biol. Evol. 21:942-956.

Chapuis, M.-P., and A. Estoup. 2007. Microsatellite null alleles and estimation of population differentiation. Mol. Biol. Evol. 24:621-631.

Colbourne, J. K., B. D. Neff, J. M. Wright, and M. R. Gross. 1996. DNA fingerprinting

of bluegill sunfish (Lepomis macrochirus) using (GT)n microsatellites and its potential for assessment of mating success. Can. J. Fish. Aquat. Sci. 53:342-349.

Collar, D.C., and P.C. Wainwright. 2009. Ecomorphology of centrarchid fishes. In Centarchid fishes: diversity, biology and conservation. Edited by Cooke, S.J., and D.P. Philipp. Wiley-Blackwell, West Sussex, U.K.

140

Darwin, C.R. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. John Murray Publishing, London, U.K.

DeWoody, J. A., D. E. Fletcher, S. D. Wilkins, W. S. Nelson, and J. C. Avise. 1998. Molecular genetic dissection of spawning, parentage, and reproductive tactics in a population of redbreast sunfish, Lepomis auritus. Evolution 52:1802-1810.

Dieckmann, U., and M. Doebeli. 1999. On the origin of species by sympatric speciation. Nature 400:354-357.

Earl, D., and B. vonHoldt. 2011. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Res. 4:359-361.

Ehlinger, T.J. 1990. Habitat choice and phenotype-limited feeding efficiency in bluegill: individual differences and trophic polymorphism. Ecology 71:886-896.

Ehlinger, T.J., and D.S. Wilson. 1988. Complex foraging polymorphism in bluegill sunfish. Proc. Natl. Acad. Sci. USA 85:1878-1882.

Ellerby, D.J., and S.P. Gerry. 2011. Sympatric divergence and performance trade-offs of bluegill ecomorphs. Evol. Biol. 38:422-433.

Feder, J.L., Opp, S.B., Wlazlo, B., Reynolds, K., Go, W., and S. Spisak. 1994. Host fidelity is an effective premating barrier between sympatric races of the apple maggot fly. Proc. Natl. Acad. Sci. USA 91:7990-7994.

Filchak, K.E., Roethele, J.B., and J.L. Feder. 2000. Natural selection and sympatric divergence in the apple maggot Rhagoletis pomonella. Nature 407:739-742.

Fisher, R., and J.D. Hogan. 2007. Morphological predictors of swimming speed: a case study of pre-settlement juvenile coral reef fishes. J. Exp. Biol. 210:2436-2443.

141

Franchini, P., Fruciano, C., Spreitzer, M.L., Jones, J.C., Elmer, K.R., Henning, F., and A. Meyer. 2013. Genomic architecture of ecologically divergent body shape in a pair of sympatric crater lake cichlid fishes. Mol. Ecol. doi: 10.1111/mec.12590

Garner, S.R., and B.D. Neff. 2013. Alternative male reproductive tactics drive asymmetrical hybridization between sunfishes (Lepomis spp.). Biol. Lett.

Genner, M.J., Seehausen, O., Lunt, D.H., Joyce, D.A., Shaw, P.W., Carvalho, G.R., and G.F. Turner. 2007. Age of cichlids: new dates for ancient lake fish radiation. Mol. Biol. Evol. 24:1269-1282.

Gray, S.M., and J.S. McKinnon. 2007. Linking color polymorphism maintenance and speciation. Trends Ecol. Evol. 22:71-79.

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z. Tierpsychol. 60:1-26.

Gross, M.R., and A.M. MacMillan. 1981. Predation and the evolution of colonial nesting in bluegill sunfish (Lepomis macrochirus). Behav. Ecol. Sociobiol. 8:163-174.

Hulsey, C.D., Roberts, R.J., Loh, Y.-H.E., Rupp, M.F., and J.T. Streelman. 2013. Lake Malawi cichlid evolution along a benthic/limnetic axis. Ecol. Evol. 3:2262-2272.

Jakobsson, M., and N.A. Rosenberg. 2007. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23:1801-1806.

Jastrebski, C.J. 2001. Divergence and selection in trophically polymorphic pumpkinseed sunfish (Lepomis gibbosus). MSc Thesis, University of Guelph, Guelph, Canada.

Jastrebski, C.J., and Robinson, B.W. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

142

Keast, A. 1978. Trophic and spatial interrelationships in the fish species of an Ontario temperate lake. Env. Biol. Fish. 3:7-31.

Kiljunen, M., Grey, J., Sinisalo, T., Harrod, C., Immonen, H., and R.I. Jones. 2006. A revised model for lipid-normalizing δ13C values from aquatic organisms, with implications for isotope mixing models. J. Appl. Ecol. 43:1213-1222.

Magee, S.E., Neff, B.D., and R. Knapp. 2006. Plasma levels of androgens and cortisol in relation to breeding behavior in parental male bluegill sunfish, Lepomis macrochirus. Hor. Behav. 49:598-609.

Mandrak, N.E., and E.J. Crossman. 1992. Postglacial dispersal of freshwater fishes into Ontario. Can. J. Zool. 70:2247-2259.

Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge, U.S.A.

Moles, M.D., Robinson, B.W., Johnston, T.A., Cunjak, R.A., Jardine, T.D., Casselman, J.M., and W.C. Leggett. Morphological and trophic differentiation of growth morphotypes of walleye (Sander vitreus) from Lake Winnipeg, Canada. Can. J. Zool. 88:950-960.

Neff, B.D., and L.M. Cargnelli. 2004. Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Environ. Biol. Fish. 71:297-304.

Neff, B. D., P. Fu, and M. R. Gross. 1999. Microsatellite evolution in sunfish (Centrarchidae). Can. J. Fish. Aquat.Sci. 56:1198-1205.

Neff, B. D., P. Fu, and M. R. Gross. 2000. Microsatellite multiplexing in fish. Trans. Am. Fish. Soc. 129:584-593.

143

Osenberg, C.W., Werner, E.E., Mittelbach, G.G., and D.J. Hall. 1988. Growth patterns in bluegill (Lepomis macrochirus) and pumpkinseed (L. gibbosus) sunfish: environmental variation and the importance of ontogenetic niche shifts. Can. J. Fish. Aquat. Sci. 45:17-26.

Parnell, A.C., Inger, R., Bearhop, S., and A.L. Jackson. 2010. Source partitioning using stable isotopes: coping with too much variation. PloS One 5:e9672.

Parsons, K.J., and B.W. Robinson. 2006. Replicated evolution of integrated plastic responses during early adaptive divergence. Evolution 60:801-813.

Parsons, K.J., and B.W. Robinson. 2007. Foraging performance of diet-induced morphotypes in pumpkinseed sunfish (Lepomis gibbosus) favours resource polymorphism. Evol. Biol. 20:673-684.

Peakall, R. O. D., and P. E. Smouse. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6:288- 295.

Pigeon, D., Chouinard, A., and L. Bernatchez. 1997. Multiple modes of speciation involved in the parallel evolution of sympatric morphotypes of lake whitefish (Coregonus clupeaformis, Salmonidae). Evolution 51:196-205.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83:703-718.

Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945-959.

Robinson, B.W., and D.S. Wilson. 1996. Genetic variation and phenotypic plasticity in a trophically polymorphic population of pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. 10:631-652.

144

Robinson, B.W., Wilson, D.S., and A.S. Margosian. 2000. A pluralistic analysis of character release in pumpkinseed sunfish (Lepomis gibbosus). Ecology 81:2799- 2812.

Robinson, B.W., Wilson, D.S., and G.O. Shea. 1996. Trade-offs of ecological specialization: an intraspecific comparison of pumpkinseed sunfish phenotypes. Ecology 77:170-178.

Robinson, B.W., Wilson, D.S., Margosian, A.S., and P.T. Lotito. 1993. Ecological and morphological differentiation of pumpkinseed sunfish in lakes without bluegill sunfish. Evol. Ecol. 7:451-464.

Rogers, S.M., and L. Bernatchez. 2007. The genetic architecture of ecological speciation and the association with signatures of selection in natural lake whitefish (Coregonus sp. Salmonidae) species pairs. Mol. Biol. Evol. 24:1423-1438.

Rohlf, F.J. 2008. tpsDig, version 2.12. Department of Ecology and Evolution, State University of New York at Stony Brook. Stony Brook, U.S.A. Available at http://life.bio.sunysb.edu/morph/.

Rohlf, F.J. 2009. tpsRegr, version 1.36. Department of Ecology and Evolution, State University of New York at Stony Brook. Stony Brook, U.S.A. Available at http://life.bio.sunysb.edu/morph/.

Rueffler, C., Van Dooren, T.J.M., and J.A.J. Metz. 2006. The evolution of resource specialization through frequency-dependent and frequency-independent mechanisms. Am. Nat. 167:81-93.

Schable, N. A., R. U. Fischer, and T. C. Glenn. 2002. Tetranucleotide microsatellite DNA loci from the dollar sunfish (Lepomis marginatus). Mol. Ecol. Notes 2:509- 511.

Schluter, D. 1995. Adaptive radiation in sticklebacks: trade-offs in feeding performance and growth. Ecology 76:82-90.

145

Schluter, D. 1996. Ecological speciation in postglacial fishes. Phil. Trans. R. Soc. Lond. B 351:807-814.

Schluter, D., and J.D. McPhail. 1992. Ecological character displacement and speciation in sticklebacks. Am. Nat. 140:85-108.

Seehausen, O., and J.J.M. van Alphen. 1998. The effect of male coloration on female mate choice in closely related Lake Victoria cichlids (Haplochromis nyererei complex). Behav. Ecol. Sociobiol. 42:1-8.

Siwertsson, A., Knudsen, R., Kahilainen, K.K., Praebel, K., Primicerio, R., and P-A. Amundsen. 2010. Sympatric diversification as influenced by ecological opportunity and historical contingency in a young species lineage of whitefish. Evol. Ecol. Res. 12:929-947.

Skúlason, S., and T.B. Smith. 1995. Resource polymorphisms in vertebrates. Trends Ecol. Evol. 10:366-370.

Snowberg, L.K., and D.I. Bolnick. 2008. Assortative mating by diet in a phenotypically unimodal by ecologically variable population of stickleback. Am. Nat. 172:733- 739.

Svanbäck, R., and P. Eklöv. 2003. Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization? Oikos 102:273-284.

Svanbäck, R., and P. Eklöv. 2004. Morphology in perch affects habitat specific foraging efficiency. Func. Ecol. 18:503-510.

Svanbäck, R., and D. Schluter. 2012. Niche specialization influences adaptive phenotypic plasticity in the threespine stickleback. Am. Nat. 180:50-59.

146

Thibert-Plante, X., and A.P. Hendry. 2011. Factors influencing progress toward sympatric speciation. J. Evol. Biol. 24:2186-2196.

Thorpe, R.S., Surget-Groba, Y., and H. Johansson. 2010. Genetic tests for ecological and allopatric speciation in anoles on an island archipelago. Plos Genetics. 6:e1000929.

Van Oosterhout, C., W. F. Hutchinson, D. P. M. Wills, and P. Shipley. 2004. Micro- checker: software for identifying and correcting genotyping errors in microsatellite data. Mol. Ecol. Notes 4:535-538.

Vander Zanden, M.J., and J.B. Rasmussen. 2001. Variation in δ15N and δ13C trophic fractionation: implications for aquatic food web studies. Limnol. Oceanogr. 46:2061-2066.

Webb, P.W. 1984. Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24:107-120.

Weese, D.J., Ferguson, M.M., and B.W. Robinson. 2012. Contemporary and historical evolutionary processes interact to shape patterns of within-lake phenotypic divergences in polyphonic pumpkinseed sunfish, Lepomis gibbosus. Ecol. Evol. 2:574-592.

Weir, B.S. 1996. Genetic Data Analysis II. Sinauer Associates, Sunderland

Yonekura, R., Nakai, K., and M. Yuma. 2002. Trophic polymorphism in introduced bluegill in Japan. Ecol. Res. 17:49-57.

Zelditch, M.L., Swiderski, D.L., Sheets, H.D., and W.L. Fink. 2004. Geometric morphometrics for biologists. Elsevier Academic Press, San Diego, U.S.A.

147

Chapter 6

6 Foraging ecology of native pumpkinseed sunfish (Lepomis gibbosus) following the invasion of zebra mussels (Dreissena polymorpha)5

Invasive species are a major concern both biologically and economically due to the rates at which they spread and the alterations they make to ecosystems. Zebra mussels

(Dreissena polymorpha) are of particular concern in North America. Following their introduction in 1988, this invasive species has cost several billion dollars in damage to industrial equipment and control efforts. However, their greatest cost may be the ecological changes they are causing in lakes. I tested if the native pumpkinseed sunfish

(Lepomis gibbosus) from three study populations in southern Ontario, Canada, were adapting to consume zebra mussels. I used a combination of stomach content and stable isotope analyses of diet. I found that pumpkinseed diet was composed of primarily zebra mussels in all three populations, a major shift from their native diet of benthic invertebrates, prior to the invasion of zebra mussels. It is possible that pumpkinseed may begin to act as a natural biological control, decreasing the overall zebra mussel abundance and rate of spread. However, the shifting diet of pumpkinseed may also have other important ecological consequences for lakes as many native invertebrates now appear to be released from pumpkinseed predation pressure. Zebra mussels are having a

5 A version of this chapter has been submitted to the journal Freshwater Biology for publication and is currently under review.

Citation: Colborne, S.F., A.D.M. Clapp, F.J. Longstaffe, and B.D. Neff. In review. Foraging ecology of native pumpkinseed sunfish (Lepomis gibbosus) following the invasion of zebra mussels (Driessena polymorpha). Freshwater Biol.

148 number of direct and indirect effects on native species, many of which we are just beginning to understand and which may have significant implications for aquatic ecosystems.

6.1 Introduction

Invasive species are among the top threats to global biodiversity and ecosystem stability

(Ricciardi et al. 1998; Sousa et al. 2013), particularly in aquatic environments that have been subject to hundreds of human-introduced species (Strayer 2010). Bivalves are some of the most frequently studied invasive species because of their wide range of impacts following introduction, and because of the number of bivalve species that have become invasive around the world (reviewed by Sousa et al. 2013). In North America, much attention has focused on zebra mussels (Dreissena polymorpha), which were introduced to the Great Lakes region in 1988 and have since spread south through the Mississippi watershed and as far west as California (Strayer 2009). Zebra mussels are attributed with causing $4 billion (US) damage to the Great Lakes region alone during the first decade following their introduction due to their effects on a variety of industrial systems, including piping for drinking water and power generation plants (Connelly et al. 2007,

Strayer 2009). Today, they are estimated to still cause $1 billion (US) in economic losses each year in the United States alone (Pimentel et al. 2005). Zebra mussels also have a significant ecological impact. Once they have colonized a lake, zebra mussels can reach densities between 10,000 to 700,000 individuals per m2, up to 100-fold higher than in their native habitats, likely due to the lack of predators and native competing species

(Griffiths et al. 1991; Ricciardi et al. 1997; Magoulick and Lewis 2002). At these high densities, zebra mussels alter the physical characteristics of littoral habitats simply by

149 their presence (Ricciardi et al. 1997) and reduce phytoplankton density and alter species composition (e.g. Jaeger Miehls et al. 2009; Fishman et al. 2010; Gergs et al. 2011). The reduction in phytoplankton suspended in the water column affects the densities of zooplankton (Pace et al. 2008, Higgins and Vander Zanden 2010), which in turn impacts the food availability for zooplanktivores, as well as increasing overall water clarity (Zhu et al. 2006).

It is thought that the primary biological mechanism contributing to the spread and high densities of zebra mussels in North America is the release from natural predation that exists in their native Ponto-Caspian range (Molloy et al. 1997, Watzin et al. 2008). Native predators of zebra mussels include roach (Rutilus rutilus) and round goby (Neogobius melanostomus) (Molloy et al. 1997). In North America, however, there is now evidence that some native species of birds (e.g. Petrie and Knapton 1999), turtles (e.g. Bulte and

Blouin-Demers 2008), and fish (e.g. Magoulick and Lewis 2002, Watzin et al. 2008) consume zebra mussels as part of their diet. If native species alter their diet and consume zebra mussels, this may represent a natural form of biological control. Such biological control can be highly effective at reducing the rate that zebra mussels spread and the overall impact they have on ecosystems (Carlsson et al. 2009, 2011). More research is therefore needed to better understand how native species are adjusting their diet to the presence of zebra mussels.

Here I focus on pumpkinseed sunfish (Lepomis gibbosus) as a potential predator of zebra mussels in North America. Pumpkinseed have the most northerly distribution of all sunfishes (Lepomis spp.) (Laughlin and Werner 1980) and are native to the Great Lakes

150 region, where zebra mussels were first introduced (Hebert et al. 1989). Pumpkinseed are one of the few sunfishes capable of crushing hard-shelled prey items, including molluscs, particularly gastropods (snails) and juvenile native bivalves (mussels) (Mittelbach 1984;

Wainwright 1996; Mittelbach et al. 1999).

Pumpkinseed diet was assessed using two frequently used methods of assessing fish foraging ecology: stomach content and stable isotope analyses. Stomach contents provide quantitative measurements of what has been consumed within hours to a day or two, but do not provide reliable estimates of diet over longer periods of time (Pinnegar and

Polunin 1999; MacNeil et al. 2005). To infer diet over longer time periods, I used stable isotope analysis of carbon (δ13C) and nitrogen (δ15N). In brief, such estimates of foraging habits are possible because the δ13C and δ15N values of the consumers tissues (i.e. fish) are reliably related to those of the resources they consume (e.g. benthic invertebrates).

Stable carbon isotope ratios change by only relatively small amounts (< 1‰) between a consumer and its resources (DeNiro and Epstein 1981, Peterson and Fry 1987), but consistently differ between the littoral and pelagic primary producers (usually by greater than 5 ‰), with those differences then persisting through the trophic levels (see France

1995). In contrast to the small changes in δ13C values between trophic levels, much larger increases in δ15N (3 – 5‰) predictably occur from a resource to its consumer (Peterson and Fry 1987, Post 2002), thereby also providing insight into the trophic position of consumers. Furthermore, stable isotopic compositions can be used to infer resource use over different time periods based on the tissue type that is sampled. Tissue-specific turnover rates are related to the metabolic activity of the tissue (Perga and Gerdeaux

2005, Guelinckx et al. 2007, Boecklen et al. 2011). Tissues with lower metabolic activity

151

(e.g. white muscle) take longer to change in isotopic composition in response to isotopic changes in diet than those with higher metabolic rates (e.g. liver; reviewed by Boecklen et al. 2011). The combination of stomach content and stable isotope analyses therefore provided us the opportunity to examine pumpkinseed diet over multiple temporal scales.

Pumpkinseed have been found to consume zebra mussels in Lake Champlain (44º 32’N,

73º 20’W; Watzin et al. 2008), but these estimates were based on fish stomach/gut content analyses from a single bay within the lake. Additionally, stable isotope analysis of pumpkinseed diet in Opinicon Lake (Ontario, Canada, 44º34’N, 76º19’W) indicated that zebra mussels were a major contributor to fish diet (Locke et al. 2013). However,

Locke et al. (2013) did not consider the possible contribution of zooplankton to diet, the major pelagic prey type found in sunfish diets (e.g. Keast 1978, Jastrebski and Robinson

2004), and also the prey type likely to have similar isotopic compositions to zebra mussels because of their similar diets (Bridgeman et al. 1995). Although these studies provide promising evidence that pumpkinseed consume zebra mussels, the temporal and prey base information necessary to estimate the degree to which pumpkinseed are consuming zebra mussels has not been fully addressed. Additionally, comparisons of pre- invasion pumpkinseed diet to post-invasion diet will provide an indication of how trophic relationships may have changed due to the presence of an invasive species, a factor often unexamined in invasive species research (Maguire and Grey 2006). To address these issues, I tested the foraging habits of pumpkinseed from three lakes in southern Ontario,

Canada, including two lakes for which pre-invasion diet has been examined (Keast 1978,

Deacon and Keast 1987). The isotopic composition of zebra mussels were also compared to other prey types found in both the littoral and pelagic habitats to assess the utility of

152 stable isotope analysis in differentiating zebra mussels from other prey types. I also measured the current diet of pumpkinseed in each lake using both stomach content and stable isotope analyses to provide estimates of post-invasion resource use of these fish.

6.2 Methods

6.2.1 Fish collection

Pumpkinseed were collected from three lakes over multiple sampling periods during

2011 and 2012 (n = 203). A total of 78 pumpkinseed were collected from Opinicon Lake

(Ontario, Canada, 44º34’N, 76º19’W) over two sampling periods: Spring 2011 (25 – 30

May 2011, n = 50) and Summer 2011 (17 – 21 August 2011, n = 28). Seventy-five fish were sampled from Lower Beverley Lake (Ontario, Canada, 44º36’N, 76º08’W) in

Spring 2011 (24 – 28 June 2011, n = 50) and Summer 2011 (18 – 24 August 2011, n =

25). The following year, 50 pumpkinseed were collected from Indian Lake (Ontario,

Canada, 44º36’N, 76º20’W) during a single Spring 2012 (19 May – 3 June 2012) sampling period. Fish were collected from multiple sites distributed around each lake using a combination of dip-netting by swimmers and angling using a small (2–3 cm) piece of earthworm suspended directly from the side of a research boat. Only fish greater than 100 mm in body length were retained for analysis to reduce the likelihood of collecting individuals that were undergoing an ontogenetic niche shift from their juvenile to adult diet (see Mittelbach 1981; Chipps et al. 2004). All fish were immediately euthanized with clove oil and stored on ice for transport to the Queen’s University

Biological Station for dissection.

153

Dissections were used to collect the stomach contents of each fish, as well as tissue samples of white muscle and liver for diet analysis. The stomach was removed from the body cavity, the contents were removed and placed in a 1.5 mL microcentrifuge tube

(multiple tubes were used as required) and stored at –20ºC for later sorting. A sample of white muscle was collected from the right side of each fish underneath the posterior portion of the dorsal fin, as well as the entire liver. Muscle and liver samples were stored in separate 1.5 mL microcentrifuge tubes at –20ºC for stable isotope analysis. I also determined the sex of each individual through visual examination of the gonads; males were identified by having cream coloured gonads which were bifurcated into left and right sections, females were identified by ovaries that did not show distinct division into sides and were filled with numerous eggs visible through the ovary wall, and juveniles were identified by having small, translucent gonads (Deacon and Keast 1987).

6.2.2 Stomach content analysis

Of the 203 fish collected, 187 had contents in their stomachs. The contents were thawed and sorted under a dissecting microscope and assigned to one of five prey types: (1) snails (i.e. gastropods); (2) zebra mussels; (3) non-mollusc littoral benthic invertebrates

(e.g. native bivalves, isopoda, amphipoda, larval emphemeroptera); (4) zooplankton

(cladocera and copepoda; e.g. Daphnia spp.); and (5) ‘other’ (terrestrial insects, fish eggs, plant matter, and unidentifiable prey items). Once sorted, the stomach contents were dried for 24 hours at 50ºC after which the dry mass of each prey type was measured to determine the proportion that it contributed to the total dry mass.

154

6.2.3 Stable isotope analysis

Reference prey samples of littoral benthic invertebrates and pelagic zooplankton were collected during each sampling period. These prey samples were used to develop a resource baseline. Common littoral benthic invertebrates (snails, isopoda, amphipoda, and larval odonata) were collected using D-net sweeps of the macrophytes and upper 1–2 cm of sediment, sorted to taxonomic order, and then frozen at –20 ºC. Pelagic zooplankton (primarily copepods and cladocerans) were also collected using 5 m vertical tows, which were then filtered to remove phytoplankton and algae.

Fish tissue and reference prey samples were prepared for stable isotope analysis following the procedure outlined in Chapter 3 (see Section 3.2.3). The presence of lipids in tissues may result in lower δ13C values compared to the isotopic composition of pure protein samples and were corrected using the lipid correction equation from Chapter 4

(see Section 4.2). Isotopic measurements were calibrated to international standards using two-point curves and monitored for precision and accuracy using internal lab standards and replicate samples (see Appendix D for details).

6.2.4 Statistical analysis

Stomach contents were analyzed by transforming the proportion prey type mass data using an arcsine transformation (e.g. Jastrebski and Robinson 2004) and comparing within each lake and sampling period using one-factor ANOVAs with Tukey’s post-hoc comparisons, when appropriate.

To determine if stable isotopes could be reliably used to identify zebra mussels in the diets of the fish, the isotopic compositions (δ13C and δ15N) of the five prey types

155

(molluscs, represented by snails; zebra mussels; other benthic littoral invertebrates, amphipods and isopods separately; and zooplankton) were compared for each lake and sampling period using two-factor analysis of variance models (dependent variable: δ13C or δ15N values of prey; independent variables: lake/sample period (five levels), invertebrate type (five levels), plus their interaction) with Tukey’s post-hoc comparisons, when appropriate.

The contribution of littoral prey and zebra mussels to the diets of pumpkinseed were then estimated for each lake and sampling period individually (n = 6 models) using two-source

SIAR mixing models (Stable Isotope Analysis in R; R version 2.14.2; R Development

Core Team 2012) were used because isotopic composition of prey can vary temporally and spatially (e.g. Vizzini and Mazzola 2003; Syväranta, Hämäläinen & Jones 2006; also see results). Separate models for each lake and sampling period were calculated using the mean measured δ13C and δ15N (± 1 SD) values to estimate the contribution to diet for two prey categories: (1) littoral benthic invertebrates, consisting of the combined isotope values for gastropods (snails), amphipods, and isopods, and (2) zebra mussels.

Zooplankton were not incorporated into the SIAR models due to the consistent absence of zooplankton from the stomach samples (see below). The two prey group values were then used as the “source” estimates for each lake and sampling period within a lake.

Mean trophic enrichment factors (TEFs) were based on multiple freshwater fish species from northern temperate lakes (δ13C: 0.47 ± 1.23 ‰, δ15N: 3.23 ± 0.41 ‰; Vander

Zanden and Rasmussen 2001) because species-specific TEFs were unavailable for pumpkinseed.

156

The results of the SIAR models were also used to estimate the probability that pumpkinseed in a particular sample (lake/sample period) consumed a greater proportion of zebra mussels in their diet as compared to the other groups. To do this, I took the

10,000 simulations generated by the SIAR models and developed a resampling routine that randomly selected a mean value from the SIAR model for each of the five lake/sampling periods and repeated this procedure over one million simulations. All possible pairwise comparisons of the five groups of fish were made (n = 10 comparisons), so I included a Bonferroni correction for multiple comparisons (α = 0.005) for this analysis.

Statistical analyses were completed using JMP version 10.0.0 (SAS Institute Inc., Cary,

North Carolina, 2012) and the Stable Isotope Analysis in R (SIAR) software package (R

Development Core Team 2012, Parnell et al. 2010). Unless otherwise stated, means are presented ± 1 SD.

6.3 Results

6.3.1 Prey resources

The isotopic composition of the five most common prey types (molluscs, amphipods, isopods, zooplankton, and zebra mussels) differed significantly among lakes and

13 15 sampling periods in δ C (ANOVA; F4, 67 = 8.63, P < 0.001) and δ N values (ANOVA;

13 F4, 67 = 82.11, P < 0.001). There were also significant differences in the δ C values of the

13 different prey types (ANOVA; F4, 67 = 51.46, P < 0.001). Amphipods had higher δ C values than any other prey type (Tukey’s, all P ≤ 0.03), whereas snails and isopods had similar values (P = 0.75). Zooplankton and zebra mussel δ13C values were similar

157

(Tukey’s, P = 0.53), but lower than all other groups (all P < 0.001). The prey types also

15 differed in δ N values (ANOVA; F4, 67 = 24.07, P < 0.001). Zooplankton and zebra mussels had similar δ15N values (Tukey’s, P > 0.99), but were significantly higher in value than all other prey types (P ≤ 0.01). Snails had lower δ15N values than zooplankton and zebra mussels but higher values than amphipods (Tukey’s, P = 0.02) and isopods (P

< 0.001). Amphipods and isopods had similar δ15N values (P = 0.32). There was no

13 evidence of an interaction effect for either δ C (ANOVA; F16, 67 = 1.34, P = 0.20) or

15 δ N values (F16, 67 = 1.45, P = 0.15). The isotopic composition values of all the prey types for each lake and sampling period are presented in Supplementary Table D.1 (also see Fig. 6.1).

6.3.2 Pumpkinseed diet analysis

On average, across the three lakes and all sampling periods, zebra mussels accounted for

68% of stomach content dry mass compared to 19% for molluscs, 3% for benthic invertebrates (excluding molluscs), and 9% for the ‘other’ prey type group. Zooplankton were not identified in any of the stomach contents and were removed as a possible prey type from subsequent analyses. Within each lake and sampling period, there was significant variation in the proportion of stomach content mass falling into the four prey types (all P < 0.001, see Table 6.1). Post-hoc comparisons indicated that zebra mussels represented the largest proportion of stomach content mass in each lake and sampling period (Tukey’s, all P < 0.001). The proportion of snails was similar to that of the non- mollusc benthic invertebrates and ‘other’ prey types in both Opinicon Lake and Lower

Beverley Lake were not identified in any of the stomach contents and were removed as a possible prey type from subsequent analyses. Within each lake and sampling period, there

158

160

Figure 6.1 Isotopic compositions of pumpkinseed sunfish (Lepomis gibbosus) and potential invertebrate prey types. Fish and invertebrates were sampled from three lakes: Opinicon (a,b), Lower Beverley (c,d), or Indian (e), over multiple sample periods. Samples of (●) white muscle and (○) liver were collected from each fish for stable isotope analysis of carbon (δ13C) and nitrogen (δ15N). The isotopic compositions of prey baseline samples are also presented for molluscs (snails), zebra mussels, and other benthic invertebrates (combined samples of amphipods and isopods).

159

160

Table 6.1 Summary of stomach content analysis of pumpkinseed sunfish (Lepomis gibbosus) from three lakes. Stomach contents were collected from fish sampled from three lakes (Opinicon, Lower Beverley, and Indian) over sampling periods during 2011 and 2012. Stomach contents were sorted into five prey types (molluscs, zebra mussels, non-mollusc littoral benthic invertebrates, zooplankton, and ‘other’) and are shown as the proportion of the total dry mass of the contents.

Proportion of total stomach content mass Sampling Total dry Zebra Benthic Lake Period n mass (mg) Snails mussels invertebrates ‘Other’ F df P Opinicon Spring 2011 44 (of 50) 462.67 ± 80.04 0.20 ± 0.04A 0.69 ± 0.06B 0.06 ± 0.03A 0.05 ± 0.03A 53.82 3, 172 < 0.001 Summer 2011 27 (of 28) 123.39 ± 22.46 0.18 ± 0.07A 0.59 ± 0.09B 0.01 ± 0.01A 0.22 ± 0.07A 11.67 3, 104 < 0.001

A B A A Beverley Spring 2011 46 (of 50) 348.47 ± 50.85 0.14 ± 0.04 0.76 ± 0.05 0.01 ± 0.004 0.09 ± 0.04 70.61 3, 180 < 0.001 Summer 2011 25 (of 25) 283.81 ± 63.25 0.07 ± 0.05A 0.82 ± 0.07B 0.04 ± 0.04A 0.07 ± 0.04A 53.00 3, 96 < 0.001

A B C C Indian Spring 2012 45 (of 50) 243.44 ± 27.22 0.32 ± 0.06 0.57 ± 0.07 0.04 ± 0.03 0.06 ± 0.03 22.19 3, 176 < 0.001 Note: Fish with empty stomach contents were not included in statistical analyses and are indicated in the sample size (n) column. Mean are expressed as plus or minus one standard error.

160

161

was significant variation in the proportion of stomach content mass falling into the four prey types (all P < 0.001, see Table 6.1). Post-hoc comparisons indicated that zebra mussels represented the largest proportion of stomach content mass in each lake and sampling period (Tukey’s, all P < 0.001). The proportion of snails was similar to that of the non-mollusc benthic invertebrates and ‘other’ prey types in both Opinicon and Lower

Beverley during all sampling periods (Tukeys, all P ≥ 0.45). Indian lake pumpkinseed, in comparison, consumed significantly more snails than either non-mollusc benthic invertebrates or the ‘other’ prey type (Tukey’s, both P ≤ 0.002).

Isotopic compositions of the pumpkinseed tissue samples varied across the lakes and sampling periods for both δ13C and δ15N (Fig. 6.1; Table 6.2; Appendix Table D.2).

Based on the SIAR models, mean estimates of fish resource use indicated that white muscle and liver tissues differed by less than 6%, with the zebra mussel contribution to diet estimated at 72 – 96% of the pumpkinseed diet (Table 6.3, Fig. 6.2). Using my resampling routine of the SIAR model estimates, I found that pumpkinseed collected from the different lakes and sampling periods consumed similar levels of zebra mussels

(Fig. 6.2). The only exception was the Summer 2011 fish from Lower Beverley, which consumed more zebra mussels than Spring 2011 Opinicon fish (based on white muscle and liver tissues) and Summer 2011 Opinicon fish (based on white muscle tissue only).

6.4 Discussion

Assessing the foraging ecology of native fishes can help one to evaluate the impact of zebra mussels on freshwater communities throughout North America. Here, I examined the diet of pumpkinseed sunfish, which are native to the Great Lakes – the region of

162

Table 6.2 Summary of the stable isotopic compositions of pumpkinseed sunfish (Lepomis gibbosus). The isotopic compositions are based on white muscle and liver samples collected from fish sampled from three lakes (Opinicon, Lower Beverley, and Indian) over multiple samplings periods from 2011 to 2012. White muscle Liver Sampling Lake Period n δ13C δ15N n δ13C δ15N Opinicon Spring 2011 50 –25.2 ± 1.7 +9.6 ± 0.5 50 –24.9 ± 2.0 +8.3 ± 0.9 Summer 2011 28 –24.3 ± 1.4 +9.6 ± 0.3 27 –23.6 ± 1.4 +8.2 ± 0.5

Lower Beverley Spring 2011 50 –26.9 ± 1.2 +11.7 ± 1.0 50 –26.7 ± 1.4 +10.2 ± 0.9 Summer 2011 25 –26.6 ± 1.1 +11.7 ± 0.6 25 –25.9 ± 0.9 +10.0 ± 0.5

Indian Spring 2012 50 –25.6 ± 1.4 +10.1 ± 0.7 50 –25.3 ± 1.6 +8.7 ± 0.9 Note: Mean values are presented (±1 SD). The liver isotopic composition for one Summer 2011 fish was unavailable due to a technical error during freeze drying of the sample.

162

163

Table 6.3 Summary of pumpkinseed sunfish (Lepomis gibbosus) diet estimates based on SIAR two-source mixing models. The isotopic composition values (δ13C and δ15N) of fish white muscle and liver samples were used to estimate the proportion of littoral benthic invertebrates and zebra mussels to the diets of fish sampled from three lakes (Opinicon, Lower Beverley, and Indian) over multiple sampling periods during 2011 and 2012. White muscle Liver Sampling Proportion Proportion Proportion Proportion Lake Period Littoral Zebra Mussels Littoral Zebra Mussels Opinicon Spring 2011 0.24 (0.16 – 0.32) 0.76 (0.68 – 0.84) 0.23 (0.19 – 0.37) 0.72 (0.63 – 0.81) Summer 2011 0.21 (0.11 – 0.31) 0.79 (0.69 – 0.89) 0.27 (0.17 – 0.37) 0.73 (0.63 – 0.83)

Lower Beverley Spring 2011 0.19 (0.05 – 0.32) 0.81 (0.68 – 0.95) 0.24 (0.12 – 0.36) 0.76 (0.64 – 0.88) Summer 2011 0.04 (0 – 0.11) 0.96 (0.87 – 1.0) 0.06 (0 – 0.13) 0.94 (0.87 – 1.0)

Indian Spring 2012 0.17 (0.09 – 0.24) 0.83 (0.76 – 0.91) 0.19 (0.11 – 0.27) 0.81 (0.73 – 0.89) Note: Means are presented with the 95% Bayesian credibility intervals (BCI) in parentheses.

163

164

166

Figure 6.2 SIAR mixing model estimates of the proportion of zebra mussels (Dreissena polymorpha) in the diet of pumpkinseed sunfish (Lepomis gibbosus). Estimates are presented for (a) white muscle and (b) liver fish tissue samples collected for fish from three lakes (Opinicon, Lower Beverley, Indian). Boxes represent the inner 50% of observations, the line inside each box represents the overall mean value, the whiskers are the 90th and 10th percentiles, and the dots represent the 95th and 5th percentiles. For lakes with multiple sample periods (Opinicon and Lower Beverley) the line across the SIAR estimates represents the mean zebra mussel consumption across the sample periods. The * symbol indicates sample periods that had significantly different contributions of zebra mussels to diet based on a resampling simulation of the SIAR results (see methods for details).

104

165

(a) White muscle * 1.0

0.8

0.6

0.4

0.2

Proportion of zebra mussels in diet musselszebra in of Proportion

0.0 (b) Liver * * 1.0

0.8

0.6

0.4

0.2

Proportion of zebra mussels in diet musselszebra in of Proportion

0.0

Spring 2011 Spring 2011 Spring 2012 Summer 2011 Summer 2011

Opinicon Lower Beverley Indian

166 initial zebra mussel introduction. Using stomach content and stable isotope analyses, which allow for inferences of diet over a period of hours to months, I tested for the incorporation of zebra mussels into pumpkinseed diet. I found that pumpkinseed from all my study lakes and at all sampling periods were consuming large quantities of zebra mussels as compared to other prey types. Zebra mussels were found in the stomachs of

81% of the pumpkinseed we collected and represented more than 50% of diet by mass in all three populations. The stable isotope estimates indicated that zebra mussels represented 70% or more of the diet across the lakes. A previous study of Opinicon pumpkinseed similarly estimated that zebra mussels composed 61% of their diet (Locke et al. 2013). Studies of several other native fishes in North America have found the presence of zebra mussels in stomach contents, e.g. freshwater drum (Aplodinotus grunniens), yellow perch (Perca flavescens), and redear sunfish (Lepomis microlophus)

(Morrison, Lynch & Dabrowski 1997; Magoulick and Lewis 2002). These studies, however, did not estimate the total proportion of diet composed of zebra mussels. The proportion of zebra mussels in the diet of round gobies (Neogobius melanostomus), another invasive species in the Great Lakes, was found to be 40% (Brush et al. 2012). My analysis of diet not only indicates that zebra mussels have become the primary resource group for native pumpkinseed, but that the degree to which pumpkinseed are consuming zebra mussels is, to my knowledge, greater than any other fish, native or exotic, in the

Great Lakes region.

I also assessed the usefulness of isotopic compositions for differentiating zebra mussels from other common invertebrate prey types. I found that the benthic invertebrates sampled from the littoral habitat of our lakes separated into two groups based on δ13C

167 isotopic composition; snails, isopods and amphipods grouped together with higher δ13C values compared to the grouping of zebra mussels and zooplankton. The similar isotopic composition of zooplankton and zebra mussels is consistent with their similar diets of phytoplankton that are suspended in the water column (Strayer 2009). Indeed, other native filter-feeding bivalves that consume phytoplankton have been found to have similar isotopic compositions to zooplankton (e.g. Vander Zanden and Rasmussen 1999).

This significant isotopic overlap between some prey types (e.g. zooplankton and zebra mussels), however, could limit the ability to distinguish them based on isotopic composition alone. In this study, the isotopic analysis was refined by including dietary information from stomach content analysis. Specifically, zooplankton were not present in any of the pumpkinseed stomachs collected. Technical issues related to detecting zooplankton in stomach contents were unlikely because I successfully identified zooplankton in the stomach of pumpkinseed in other lakes using the same methods

(Chapter 5; also see Keast 1978, Robinson et al. 1993, Jastrebski and Robinson 2004).

The absence of zooplankton from the pumpkinseed stomachs analyzed in this study is also consistent with resource partitioning in sympatric populations of bluegill (Lepomis macrochirus) and pumpkinseed (e.g. Keast 1978, Mittelbach 1984) – all three lakes that were studied also contained bluegill. Thus, based on stomach contents I was able to eliminate zooplankton as a pelagic resource and use isotopic composition to directly assess zebra mussel consumption by pumpkinseed compared to other littoral benthic invertebrates.

In addition to demonstrating that pumpkinseed are consuming zebra mussels, this study also afforded the opportunity to compare their current diet to that prior to the invasion of

168 zebra mussels. For two of the study lakes, Opinicon and Lower Beverley, pumpkinseed diet has been studied within the twenty years prior to the introduction of zebra mussels.

Stomach content estimates of Opinicon pumpkinseed, sampled between 1969 and 1971, estimated 33% snails, 65% other benthic invertebrates, and 2% zooplankton (Keast

1978). Similar estimates of Lower Beverley pumpkinseed, sampled in 1981, estimated

11% snails and 89% other littoral benthic invertebrates (Deacon and Keast 1987).

Compared to these historical diet analyses, my analyses show that pumpkinseed have decreased benthic invertebrate consumption by more than 60% and snail consumption by

14%. The pumpkinseed thus have shifted from a diet of primarily snails and soft-bodied invertebrates to a diet composed primarily of zebra mussels.

It is conceivable that pumpkinseed, along with several other native fishes in North

America that have been found to consume zebra mussels (e.g. Magoulick and Lewis

2002, Carlsson et al. 2011), may begin to act as natural biological controls decreasing both the overall densities and rates at which zebra mussels spread. However, these diet shifts in native species may further change lake food webs, adding to the indirect effects of zebra mussels. For example, the drop in consumption of other littoral benthic invertebrates by the pumpkinseed studied here is a form of predator release, which can cause changes in foraging patterns of species (Fraser et al. 2004, Burkpile and Hay 2007), overall species composition (Duffy 2002; Sieben et al. 2011b), and trophic cascades throughout food webs (Eriksson et al. 2009, Sieben et al. 2011a). Reduced predation pressure on benthic invertebrates specifically has been documented to result in an increase in algae consumption by those invertebrates (e.g. Råberg and Kautsky 2007).

Despite the promising biological control implications that come with the shift of native

169 species such as pumpkinseed to consumption of zebra mussels, there are other ecological shifts in trophic relationships that may negatively impact ecosystem and warrant further study.

In summary, I have found that pumpkinseed in three lakes have changed their foraging patterns from consuming a mixture of snails and other soft-bodied benthic invertebrates to consuming primarily zebra mussels. This shift in diet has important ecological consequences for not only the pumpkinseed and zebra mussels, but also other species within the lakes. As native species respond to the presence of invasive species, assessing the consequences of changing foraging patterns for food web dynamics will provide insight into both the direct and indirect effects of these introduced species.

6.5 Chapter acknowledgements

I thank C. Rodgers, M. Roubakha, T. Hain, M. Lau, and A. Berchtold for assistance in the field. I also thank K. Law for assistance with the stable isotope analysis in the laboratory. This research was carried out with approval of The University of Western

Ontario Council on Animal Care (animal use protocol No. 2006-062-05) and the Ontario

Ministry of Natural Resources (license numbers 1057026 and 1062735). This work was supported by the Natural Sciences and Engineering Research Council of Canada

(NSERC) Discovery grants to B.D. Neff and F.J. Longstaffe; Canada Foundation for

Innovation (CFI) and Ontario Research Fund (ORF) infrastructure awards to F.J.

Longstaffe; The University of Western Ontario Graduate Thesis Research Fund, Queen

Elizabeth II Graduate Scholarship in Science and Technology (QEIIGSST), and Ontario

Graduate Scholarship (OGS) awards to S.F. Colborne. This work was also partially

170 supported through release time provided to F.J. Longstaffe through the Canada Research

Chairs program. The data presented in this chapter is Laboratory for Stable Isotope

Science (LSIS) contribution #306.

6.6 References

Boecklen, W.J., C.T. Yarnes, B.A. Cook, and A.C. James. 2011. On the use of stable isotopes in trophic ecology. Ann. Rev. Ecol. Evol. System. 42:411-440.

Bridgeman, T.B., G.L. Fahnenstiel, G.A. Lang, and T.F. Nalepa. 1995. Zooplankton grazing during zebra mussel (Dreissena polymorpha) colonization of Saginaw Bay, Lake Huron. J. Gt. Lakes Res. 21:567-573.

Brush, J.M., A.T. Fisk, N.E. Hussey, and T.B. Johnson. 2012. Spatial and seasonal variability in the diet of round goby (Neogobius melanostomus): stable isotopes indicate that stomach contents overestimate the importance of dreissenids. Can. J. Fish. Aquat. Sci. 69:573-586.

Bulté, G., and G. Blouin-Demers. 2008. Northern map turtles (Graptemys geographica) derive energy from the pelagic pathway through predation on zebra mussels (Dreissena polymorpha). Freshwater Biol. 53:497-508.

Burkepile, D.E., and M.E. Hay. 2007. Predator release of the gastropod Cyphoma gibbosum increases predation on gorgonian corals. Oecologia 154:167-173.

Caraco, N.F., J.J. Cole, and D.L. Strayer. 2006. Top-down control from the bottom: regulation of eutrophication in a large river by benthic grazing. Limnol. Oceanogr. 51:664-670.

Carlsson, N.O.L., and H. Bustamante, D.L. Strayer, and M.L. Pace. 2011. Biotic resistance on the increase: native predators structure invasive zebra mussel populations. Freshwater Biol. 56:1630-1637.

171

Carlsson, N.O.L., O. Sarnelle, and D.L. Strayer 2009. Native predators and exotic prey – an acquired taste? Front. Ecol. Environ. 7:525 -532.

Chipps, S.R., J.A. Dunbar, and D.H. Wahl. 2004. Phenotypic variation and vulnerability to predation in juvenile bluegill sunfish (Lepomis macrochirus). Oecologia 138:32- 38.

Connelly, N.A., C.R. O’Neil Jr., B.A. Knuth, and T.L. Brown. 2007. Economic impacts of zebra mussels on drinking water treatment and electric power generation facilities. Environ. Manage. 40:105-112.

Deacon, L.I., and J.A. Keast. 1987. Patterns of reproduction in two populations of pumpkinseed sunfish, Lepomis gibbosus, with differing food resources. Environ. Biol. Fishes 19:281-296.

DeNiro, M.J., and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta 45:341-351.

Duffy, J.E. 2002. Biodiversity and ecosystem function: the consumer connection. Oikos 99:201-219.

Eriksson, B.E., L. Ljunggren, A. Sandstrom, G. Johansson, J. Mattila, A. Rubach, S. Raberg, and M. Snickars. 2009. Decline in predatory fish promote bloom-forming macroalgae. Ecol. App. 19:1975-1988.

Fishman, D.B., S.A. Adlerstein, H.A. Vanderploeg, G.L. Fahnenstiel, and D. Scavia. 2010. Phytoplankton community composition of Sagniaw Bay, Lake Huron, during the zebra mussel (Dreissena polymorpha) invasion: a multivariate analysis. J. Gt. Lakes Res. 36:9-19.

France, R.L. 1995. Differentiation between littoral and pelagic food webs in lakes using stable carbon isotopes. Limnol. Oceanogr. 40:1310-1313.

172

Fraser, D.F., J.F. Gilliam, J.T. Akkara, B.W. Albanese, and S.B. Snider. 2004. Night feeding by guppies under predator release: effects on growth and daytime courtship. Ecology 85:312-319.

Gergs, R., J. Grey J., and K.-O. Rothhaupt. 2011. Temporal variation in zebra mussel (Dreissena polymorpha) density structure the benthic food web and community composition on hard substrates in Lake Constance, Germany. Biol. Invasions 13:2727-2738.

Griffiths, R.W., D.W. Schloesser, J.H. Leach, and W.P. Kovalak. 1991. Distribution and dispersal of the zebra mussel (Dreissena polymorpha) in the Great Lakes region. Can. J. Fish. Aquat. Sci. 48:1381-1388.

Guelinckx, J., J. Maes, P. Van Den Driessche, B. Geyson, F. Dehairs, and F. Ollevier. 2007. Chanes in δ13C and δ15N in different tissues of juvenile sand goby Pomatoschistus minutus: a laboratory diet-switch experiment. Mar. Ecol. Prog. Ser. 341:205-215.

Hebert, P.D.N., B.W. Muncaster, and G.L. Mackie. 1989. Ecological and genetic studies on Dreissena polymorpha (Pallas): a new mollusc in the Great Lakes. Can. J. Fish. Aquat. Sci. 46:1587-1591.

Higgins, S.N., M.J. Vander Zanden. 2010. What a difference a species makes: a meta- analysis of dreissenid mussel impacts on freshwater ecosystems. Ecol. Monogr. 80:179-196.

Jaeger Miehls, A.L., D.M. Mason, K.A. Frank, A.E. Krause, S.D. Peacor, and W.W. Taylor. 2009. Invasive species impacts on ecosystem structure and function: a comparison of Oneida Lake, New York, USA, before and after zebra mussel invasion. Ecol. Model. 220:3194-3209.

Jastrebski, C.J., and B.W. Robinson. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

173

Keast, A. 1978. Trophic and spatial interrelationships in the fish species of an Ontario temperate lake. Environ. Biol. Fish. 3:7-31.

Kiljunen, M., J. Grey, T. Sinisalo, C. Harrod, H. Immonen & R.I. Jones. 2006. A revised model for lipid-normalizing δ13C values from aquatic organisms, with implications for isotope mixing models. J. Appl. Ecol. 43:1213-1222.

Laughlin, D.R., and E.E. Werner. 1980. Resource partitioning in two coexisting sunfish: pumpkinseed (Lepomis gibbosus) and northern longear sunfish (Lepomis megalotis peltastes). Can. J. Fish. Aquat. Sci. 37:1411-1420.

Locke, S.A., G. Bulté, M.R. Forbes, and D.J. Marcogliese. 2013. Estimating diet in individual pumpkinseed sunfish Lepomis gibbosus using stomach contents, stable isotopes and parasites. J. Fish Biol. 82:522-537.

Lodge, D.M. 1993. Biological invasions: lessons for ecology. Trends Ecol. Evol. 8:133- 137.

MacNeil, M.A., G.B. Skomal, and A.T. Fisk. 2005. Stable isotopes from multiple tissues reveal diet switching in sharks. Mar. Ecol. Prog. Ser. 302:199-206.

Magoulick, D.D., and L.C. Lewis. 2002. Predation on exotic zebra mussels by native fishes: effects on predator and prey. Freshwater Biol. 47:1908-1918.

Maguire, C.M., and J. Grey. 2006. Determination of zooplankton dietary shift following a zebra mussel invasion, as indicated by stable isotope analysis. Freshwater Biol. 51:1310-1319.

Mittelbach, G.G. 1981. Foraging efficiency and body size: a study of optimal diet and habitat use by bluegills. Ecology 62:1370-1386.

Mittelbach, G.G. 1984. Predation and resource partitioning in two sunfishes (Centrarchidae). Ecology 65:499-513.

174

Mittelbach, G.G., C.W. Osenberg, and P.C. Wainwright. 1999. Variation in feeding morphology between pumpkinseed populations: phenotypic plasticity or evolution? Evol. Ecol. Res. 1:111-128.

Molloy, D.P, A.Y. Karatayev, L.E. Burlakova, D.P. Kurandina, and F. Laruelle. 1997. Natural enemies of zebra mussels: predators, parasites and ecological competitors. Rev. Fish. Sci. 5:27-97.

Morrison, T.W., W.E. Lynch Jr., and K. Dabrowski. 1997. Predation on zebra mussels by freshwater drum and yellow perch in western Lake Erie. J. Gt. Lakes Res. 23:177- 189.

Naddafi, R., and L.G. Rudstam. 2013. Predation on invasive zebra mussel, Dreissena polymorpha, by pumpkinseed sunfish, rusty crayfish, and round goby. Hydrobiologia doi: 10.1007/s10750-013-1653-z.

Pace, M.L., D.L. Strayer, D. Fischer, and H.M. Malcom. 2010.Recovery of native zooplankton associated with increased mortality of an invasive mussel. Ecosphere 1:doi: 10.1890/ES10-00002.1.

Perga, M.E., and D. Gerdeaux. 2005. ‘Are fish what they eat’ all year round? Oecologia 144:598-606.

Peterson, B.J., and B. Fry. 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol. Syst. 18:293-320.

Petrie, S.A, and R.W. Knapton. 1999. Rapid increase and subsequent decline of zebra and quagga mussels in Long Point Bay, Lake Erie: possible influence of waterfowl predation. J. Grt. Lakes Res. 25:772-782.

Pimentel, D., R. Zuniga, and D. Morrison. 2005. Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecol. Econom. 52:273-288.

175

Pinnegar, J.K., and N.V.C. Polunin. 1999. Differential fractionation of δ13C and δ15N among fish tissues: implications for the study of trophic interactions. Funct. Ecol. 13:225-231.

Post, D.M. 2002. Using stable isotopes to estimate trophic positions: models, methods, and assumptions. Ecology 83:703-718.

Råberg, S., and L. Kautsky. 2007. Consumers affect prey biomass and diversity through resource partitioning. Ecology 88:2468-2473.

Ricciardi, A., R.J. Neves, and J.B. Rasmussen. 1998. Impending extinctions of North American freshwater mussels (Unionida) following the zebra mussel (Driessena polymorpha) invasion. J. Anim. Ecol. 67:613-619.

Ricciardi, A., F.G. Whoriskey, and J.B. Rasmussen. 1997. The role of the zebra mussel (Dreissena polymorpha) in structuring macroinvertebrate communities on hard substrata. Can. J. Fish. Aquat. Sci. 54:2596-2608.

Robinson, B.W., D.S. Wilson D.S., A.S. Margosian, and P.T. Lotito. 1993. Ecological and morphological differentiation of pumpkinseed sunfish in lakes without bluegill sunfish. Evol. Ecol. 7:451-464.

Sieben, K., L. Ljunggren, U. Bergström, and B.K. Eriksson. 2011a. A meso-predator release of stickleback promotes recruitment of macroalgae in the Baltic Sea. J. Exp. Mar. Biol. Ecol. 397:79-84.

Sieben, K., A.D. Rippen, and B.K. Eriksson. 2011b. Cascading effects from predator removal depend on resource availability in a benthic food web. Mar. Biol. 158:391- 400.

Sousa, R., A. Novais, R. Costa, and D.L. Strayer. 2013. Invasive bivalves in fresh waters: impacts from individuals to ecosystems and possible control strategies. Hydrobiologia doi:10.1007/s10750-012-1409-1

176

Strayer, D.L. 2009. Twenty years of zebra mussels: lessons from the mollusk that made headlines. Fron. Ecol. Environ. 7:135-141.

Strayer, D.L. 2010. Alien species in fresh waters: ecological effects, interaction with other stressors, and prospects for the future. Freshwater Biol. 55:152-174.

Syväranta, J., H. Hämäläinen, and R.I. Jones. 2006. Within-lake variability in carbon and nitrogen stable isotope signatures. Freshwater Biol. 51:1090-1102.

Vander Zanden, M.J., and J.B. Rasmussen. 1999. Primary consumer δ13C and δ15N and the trophic position of aquatic consumers. Ecology 80:1395-1404.

Vander Zanden, M.J., and J.B. Rasmussen. 2001. Variation in δ15N and δ13C trophic fractionation: implication for aquatic food web studies. Limnol. Oceanogr. 46:2061-2066.

Vizzini, S., and A. Mazzola. 2003. Seasonal variations in the stable carbon and nitrogen isotope ratios (13C/12C and 15N/14N) of primary producers and consumers in a western Mediterranean coastal lagoon. Mar. Biol. 142:1009-1018.

Wainwright, P.C. 1996. Ecological explanation through functional morphology: the feeding biology of sunfishes. Ecology 77:1336-1343.

Watzin, M.C., K. Joppe-Mercure, J. Rowder, B. Lancaster, and L. Bronson. 2008. Significant fish predation on zebra mussels Dreissena polymorpha in Lake Champlain, U.S.A. J. Fish Biol. 73:1585-1599.

Zhu, B., D.G. Fitzgerald, C.M. Mayer, L.G. Rudstam, and E.L. Mills. 2006. Alteration of ecosystem function by zebra mussesls in Oneida Lake: impacts on submerged macrophytes. Ecosystems 9:1017-1028.

177

Chapter 7 7 General discussion

Foraging ecology has frequently been associated with phenotypic diversification because morphological features (e.g. beak shape, fin orientation, neck length, mouth shape) are often tightly linked to foraging efficiency and competitive ability to access resources

(Mittelbach 1981, Svanbäck and Eklöv 2003). However, sexual selection can also drive patterns of phenotypic diversification (e.g. Seehausen and van Alphen 1999, Ritchie

2007), particularly when there is strong assortative mating related to different phenotypes. The interactions between natural and sexual selection may either promote or limit phenotypic diversification, but these interactions are not yet well understood

(Ritchie 2007, Pfennig and Pfennig 2009). Using foraging ecology as the central focus of my doctoral research, I identified three primary research questions: 1) how are diet and morphological variation related to the alternative reproductive tactics of bluegill; 2) what are the roles of disruptive and sexual selection in the divergence of pumpkinseed foraging ecomorphs; and 3) how are human-induced changed to ecosystems affecting the foraging ecology of fish.

7.1 Intrasexual selection and alternative reproductive tactics

The first two data chapters in my thesis used bluegill to examine the relationship between different reproductive tactics and variation in overall body shape and diet. Morphological variation in body shape and differences in body size among male reproductive morphs

(parental males and sneakers) were first suggested by Ehlinger et al. (1997). However,

178 they did not examine the association between phenotypic differences and diet among all the reproductive morphs. My research examined the morphological variation among five bluegill groups (parental males, sneakers, satellites, females, and juveniles) in relation to measures of both burst and endurance swimming (Chapter 2), establishing that there is a functional relationship between body shape and swim performance measures. I also found trade-offs between morphology and diet that were not consistent with the expected relationships between phenotype and resource use in fish (Chapter 3). It is thus likely that factors other than foraging ecology serve a primary role in the evolution of phenotypic diversification among bluegill reproductive tactics (Chapter 3). One such process that may be involved in the divergence of these morphs is selection on swim performance related to the different reproductive tactics found in bluegill. I found that sneaker males, the most streamlined fish, had the fastest burst swim speeds of all the bluegill groups, which is consistent with the expected relationship between shape and swim performance

(Webb 1984). However, sneakers primarily consumed littoral benthic invertebrates, the opposite diet typically associated with a streamlined body shape (Svanbäck et al. 2008).

Part of their diet may be explained by selection related to predation that restricts small- sized fish to the littoral habitat (Mittelbach 1981), but the swim performance measures may also support the idea that sexual selection has led to increased speed for darting into nests quickly (Gross 1982). The body shape of sneakers is likely the result of both natural selection (predation) and sexual selection (mating success). Additionally, I found that satellite males effectively mimicked female body shape (Chapters 2 & 3), which should increase their mating success, but consumed significantly more littoral resources than females. If selection based on foraging efficiency was the only mechanism influencing

179 bluegill phenotypes, then satellites and females should have been consuming similar resources. It is possible that sexual selection for a deeper female body shape, related to the volume of eggs (Bernardo 1996), may have resulted in a sub-optimal foraging morphology for female bluegill by reducing the streamlining that increases prey capture success in pelagic-foraging fish.

The data from Chapters 2 & 3 provide evidence that sexual and natural selection result in phenotypic trade-offs within the bluegill reproductive groups. At present, it would appear that phenotypic differences among groups are the product of sexual selection, as opposed to disruptive natural selection related to foraging tactic. However, it is possible that some degree of morphological compensation occurs, such that trade-offs between morphology and foraging efficiency are completely compensated or reduced in intensity, an area of study that needs to be explored further. Overall, my research has provided some of the first evidence of phenotypic trade-offs between foraging-related natural selection and sexual selection among the bluegill alternative reproductive tactics. Alternative reproductive tactics represent an interesting source of phenotypic diversification typically associated with sexual selection (e.g. Taborsky et al. 2008), but disruptive selection based on foraging has been largely ignored in these populations. Further examination of these selection mechanisms and their relationships will provide insights into the evolution and maintenance of alternative reproductive tactics.

7.2 Sympatric divergence and speciation

Phenotypic diversification related to differing foraging tactics within populations has been of interest to biologists since Darwin (1859). This is primarily due to the potential

180 for disruptive selection based on foraging phenotypes to lead to speciation (McKinnon and Rundle 2002, Kocher 2004). My research focused on a polymorphic population of pumpkinseed sunfish that has developed littoral and pelagic foraging ecomorphs over the past 12,000 years (Jastrebski and Robinson 2004, Parsons and Robinson 2006). Despite finding evidence of phenotypic divergence related to foraging tactic and condition, particularly in the pelagic habitat, and positive assortative mating based on foraging tactic, I found no evidence of genetic differentiation between the littoral and pelagic- caught pumpkinseed of Ashby Lake. Similarly, Weese et al. (2012) found that at least 11 other populations of polymorphic pumpkinseed show no evidence of neutral genetic divergence between habitats, suggesting that divergence is still early in the process.

Indeed, it may be productive to use these populations to examine the factors that influence the rate at which divergence may be occurring within these populations. Some effort has already been put into examining the role of phenotypic plasticity in response to foraging conditions during juvenile development and the strength of disruptive natural selection based on body shape (e.g. Robinson et al. 1996, Parsons and Robinson 2006).

The patterns of assortative mating observed between ecomorphs also warrant further examination to determine if they are due to passive isolation based on differences in habitat use or to active behavioural mate choice mechanisms between these groups.

Understanding the conditions that affect the rates of divergence and speciation under sympatric conditions may provide as much insight into these processes as studies of populations that already show clear divergence into separate species.

181

7.3 Divergence and alternative reproductive tactics

In my thesis, I examined phenotypic diversification among reproductive tactics and divergence between foraging ecomorphs as independent subjects. However, alternative reproductive tactics could have important implications for divergence between foraging phenotypes. Indeed, the role of cuckolders could be especially important in the establishment of assortative mating that reduces gene flow between ecomorphs.

Cuckolders may facilitate gene flow between foraging ecomorphs if they do not show similar levels of mate discrimination as parental males and females. Neff (2004) found that parental male and female bluegill in Lake Opinicon (Ontario) showed evidence of non-random mating (i.e. assortative mating) based on an optimal level of genomic divergence between the pair, but this was not found in the offspring of cuckolder males.

Presumably, these patterns of genetic divergence are the result of cuckolder males being opportunistic for any reproductive opportunity rather than discriminating among the nests available for cuckoldry. Indeed, there is evidence that cuckolders often fail to spawn even in the nests of the appropriate sunfish species (Garner and Neff 2013). If polymorphic populations also have cuckolders that do not distinguish mating opportunities based on foraging tactic, then it is possible for gene flow to be maintained between foraging ecomorphs even if some members of the population show strong positive assortative mating.

Additionally, gene flow resulting from cuckolders may not be equal between foraging ecomorphs due to other natural selection pressures that affect habitat use of fishes. I found that the small bluegill (sneakers) consume more littoral resources than large bluegill (parental males and females) (Chapter 3), likely due to size-based predation risk

182

(Mittelbach 1981). Both bluegill and pumpkinseed sunfish go through similar niche shifts over the course of early development related to changing predation risk at different sizes.

When sunfish leave the nest, they become pelagic zooplanktivores until they are approximately 35 mm in length (Lemly and Dimmick 1982, Keast 1988), they then move to the littoral habitat and consume benthic invertebrates (Werner et al. 1983) until a final ontogenetic niche shift to their adult diet at approximately 100 mm in length (Mittelbach

1981). It is possible that due to predation risk at smaller body sizes the offspring of both littoral and pelagic parents will reside in the littoral habitat if they adopt a cuckolder life history tactic. If cuckolders of parents from both foraging ecomorphs live and breed in the same habitat, then directional gene flow from the pelagic to littoral habitats may occur, preventing full reproductive isolation. The possible interactions between reproductive tactics and population divergence present an interesting avenue to explore because of the increasing number of fish populations that seem to meet both the expectations of disruptive selection and positive assortative mating based on foraging tactic, but which are not showing evidence of speciation (e.g. Bolnick 2011, Martin

2013).

7.4 Impacts of changing ecosystems

In my research, I focused on shifts in pumpkinseed foraging following the invasion of zebra mussels (Chapter 6). However, this is just one possible example of how native species are responding to invasive species and their associated changes to freshwater communities. In addition to the shifts from native prey species to invasive zebra mussels, the observed diet shifts of pumpkinseed may impact resource-related phenotypic variation in a number of ways. In freshwater lakes, the most common resource divisions

183 are between littoral benthic invertebrates and pelagic zooplankton (e.g. Schluter 1993,

Svanbäck et al. 2008, Willacker et al. 2010). However, the abundance of zooplankton resources for pelagic fish has decreased significantly in the presence of zebra mussels, while at the same time the littoral invertebrates are becoming more abundant overall

(Strayer 2009, 2010). These changes in resource availability could have major effects on the foraging ecomorphs of fish. For example, the development of the pelagic ecomorph of pumpkinseed has been linked to the availability of the pelagic resource niche in the absence of bluegill and high levels of resource competition in the littoral habitat

(Robinson and Wilson 1996, Robinson et al. 2000). As zebra mussels decrease the abundance of pelagic zooplankton and increase the abundance of littoral benthic invertebrates, disruptive selection pressures related to foraging tactic are likely to be altered. If these changes in selection pressures are extreme, there may no longer be disruptive selection between littoral and pelagic foraging tactics. Indeed, disruptive selection may occur within the littoral habitat instead. For example, foraging ecomorphs may develop between individuals that specialize on crushing hard-shelled gastropods and zebra mussels while others specialize on soft-bodied benthic invertebrates. Invasive species, and other human-induced changes, may thus alter the current patterns of disruptive selection in freshwater ecosystems, having important implications for populations that may be in the process of diverging in resource use.

Even though invasive species may alter current patterns of selection, they could also result in new foraging opportunities for native species. As I observed in pumpkinseed

(Chapter 6), some native species may adapt to directly consume invasive species that have become established. There may also be indirect changes to foraging patterns of

184 native species as a result of invasions. For example, if pumpkinseed within a lake are able to shift their resource use as a group to consume zebra mussels, as I have observed in three lakes, then the other littoral benthic invertebrates could in effect represent an abundant niche available for other fish to occupy because they have been released from their typical source of predation. If this occurs, we could see other native species change their resource use patterns to utilize the abundant native littoral benthic invertebrates.

Bluegill are a prime candidate fish species to shift towards a littoral foraging tactic because not only do they have the ability to consume littoral benthic invertebrates

(Ellerby and Gerry 2011), but the pelagic zooplankton they typically consume are decreasing in abundance. The selection pressures in bluegill related to foraging in a pelagic habitat with fewer zooplankton prey available may result in either directional selection, bringing this formerly pelagic feeding species into the littoral habitat as a whole group, or disruptive selection resulting in pelagic and littoral ecomorphs even in the presence of pumpkinseed. Regardless of how ecosystems may change in response to invasive species, it is important that biologists capitalize on this opportunity to study how rapidly changing resource availabilities alter selection regimes and the subsequent effects on phenotypic diversity within populations.

7.5 Advances in stable isotope analysis techniques

Throughout the research presented in my thesis I used carbon and nitrogen stable isotope analysis of “bulk” tissue samples to test my hypotheses regarding foraging ecology of sunfish. Bulk stable isotope analysis refers to measuring the mean isotopic composition of all the components within a given tissue at the time of its collection. The analysis of bulk tissue samples for stable isotope ratios is frequently used to assess the diet of wild

185 organisms (e.g. Peterson and Fry 1987, Hobson 1999, Anderson and Cabana 2007), based

13 on the assumptions that δ Cbulk values differ minimally between consumers and their

15 resources (DeNiro and Epstein 1981, Fry et al. 1978), while δ Nbulk increases with trophic level (3 to 5 ‰ per trophic level; DeNiro and Epstein 1981, Post 2002). However, multiple assumptions related to the use of bulk isotopic composition have been identified as potentially confounding variables in isotopic studies. For example, temporal and spatial variability in the isotopic composition of the primary producers at the base of food webs has been identified as a source of variation in isotopic measurements of consumers

(Post 2002). To address this source of variation, samples of potential resource groups are routinely collected at the same time as the consumers of interest to provide “baseline” measurements (Vander Zanden and Rasmussen 1999), a protocol that was followed in my research. However, even this sampling protocol is subject to temporal prey variability.

Temporal variability in isotopes is most likely to affect the isotopic composition of tissues that reflect diet over long periods of time, e.g. white muscle. By sampling both consumers and resources simultaneously we are able to address many assumptions of bulk tissue analysis and achieve accurate estimates of resource use. However, there have been efforts to continue developing techniques of isotopic analysis that would reduce the number of prey samples that have to be collected without reducing the accuracy of the inferences being made.

Recent advances in stable isotope analysis of individual amino acids have the potential to allow inference of diet using fewer assumptions than bulk isotopic compositions

(Boecklen et al. 2011). Both δ13C and δ15N values of essential amino acids (e.g. glycine, phenylalanine) have been experimentally shown to not fractionate, i.e. have unchanged

186 isotopic compositions, between primary producers and consumers across multiple trophic levels, even more so than bulk δ13C values. Essential amino acids remain consistent across trophic levels because they are generally not synthesized or altered in metabolic processes occurring within consumers (e.g. McMahon et al. 2010, Olson et al. 2010).

However, non-essential amino acids (e.g. glutamic acid) are often altered by processes at each trophic level, resulting in predictable increases in isotopic measurements with trophic level. For example, δ15N has been shown to increase by 7 ‰ with each trophic level for multiple non-essential amino acids, providing an accurate estimate of trophic level in comparison to the baseline values provided by essential amino acids (Popp et al.

2007, Olson et al. 2010). Using amino acids, it is possible to estimate resource base lines values without directly sampling the prey items and trophic position using the same isotope, i.e. δ15N, for all measurements (Olson et al. 2010, Boecklen et al. 2011).

While compound specific isotope analysis is an interesting field of future development in studies of foraging ecology, it does not completely replace bulk analysis of tissues.

Currently, one of the main limitations to compound specific isotopic analyses is the reproducibility of measurements. There is sample variability ranging between 0.1 to 4.4

‰ for compound specific analyses compared to ≤ 0.2 ‰ variation for traditional bulk tissue analyses (Popp et al. 2007). It is likely that in the near future, as compound specific isotope analysis is refined and further developed, there will be increased emphasis on combining bulk and compound specific isotopic measures of diet, much like stomach contents and bulk stable isotope analyses are currently used in combination for many studies (Chapters 5 and 6). Compound specific isotopic analysis has the potential to increase the accuracy of inferences about foraging ecology over multiple temporal

187 periods. Additionally, this technique may be used to study the role of foraging in other processes, such as assortative mating and phenotypic divergence, in addition to analyses of bulk isotopic composition.

7.6 Concluding remarks

The field of foraging ecology is not simply about describing what organisms eat and how natural selection in the past resulted in the diversity of foraging tactics and phenotypes observed today. Rather, foraging ecology is a dynamic field of research that can be used to examine the mechanisms that create, maintain, and even reduce phenotypic diversity within populations. Phenotypic diversification is most frequently examined in the context of foraging-based disruptive selection related to competition for resources. However, adaptations related to mating success may also result in phenotypic diversification through sexual selection. To date, much research has focused on these factors separately even though phenotypes may be a product of strong interactions between these two mechanisms. I have added to the growing evidence for interactions between natural and sexual selection by examining phenotypic diversification within populations associated with alternative reproductive tactics and divergence based on foraging ecomorphs.

Many important questions remain to be asked about the interactions between natural and sexual selection in relation to foraging ecology, especially in the context of changing resource conditions resulting from human-induced alterations to the environment. The research presented in this thesis represents the beginning stages of a growing field of work that is bridging the divide that has long existed between selection mechanisms and diversification within populations. By understanding phenotypic diversification within

188 populations and the interactions of the selection mechanisms involved, we can gain insight into how these processes have resulted in the world’s vast biological diversity.

7.7 References

Anderson, C., and G. Cabana. 2007. Estimating the trophic position of aquatic consumers in river food webs using stable nitrogen isotopes. J. North Amer. Benth. Soc. 26:273-285.

Bernardo, J. 1996. The particular maternal effect of propagule size, especially egg size: patterns, models, quality of evidence and interpretations. Amer. Zool. 36:216-236.

Boecklen, W.J., C.T. Yarnes, B.A. Cook, and A.C. James. 2011. On the use of stable isotopes in trophic ecology. Ann. Rev. Ecol. Evol. Syst. 42:411-440.

Bolnick, D.I. 2011. Sympatric speciation in threespine stickleback: why not? Int. J. Ecol. 2011:Article ID 942847. doi: 10.1155/2011/942847.

Darwin, C.R. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. John Murray Publishing, London, U.K.

DeNiro, M.J., and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Ac. 45:341-351.

Ehlinger, T.J., M.R. Gross, and D.P. Philipp. 1997. Morphological and growth rate difference between bluegill males of alternative reproductive life histories. North Am. J. Fish. Man. 17:533-542.

Ellerby, D.J., and S.P. Gerry. 2011. Sympatric divergence and performance trade-offs of bluegill ecomorphs. Evol. Biol. 38:422-433.

Garner, S.R., and B.D. Neff. 2013. Alternative male reproductive tactics drive asymmetrical hybridization between sunfishes (Lepomis spp.). Biol. Lett. doi: 10.1098/rsbl.2013.0658

189

Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z. Tierpsychol. 60:1-26.

Hobson, K.A. 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120:314-326.

Jastrebski, C.J., and B.W. Robinson. 2004. Natural selection and the evolution of replicated trophic polymorphisms in pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. Res. 6:285-305.

Keast, A. 1988. Planktivory in larval, juvenile, and adult planktivores: resource division in a small lake. Verh. Int. Verein. Limnol. 23:1692-1697.

Kocher, T.D. 2004. Adaptive evolution and explosive speciation: the cichlid fish model. Nat. Rev. Genet. 5:288-298.

Lemly, A.D., and J.F. Dimmick. 1982. Growth of young-of-the-year and yearling Centrarchids in relation to zooplankton in the littoral zone of lakes. Copeia 1982:305-321.

McKinnon, J.S., and H.D. Rundle. 2002. Speciation in nature: the threespine stickleback model systems. Trends Ecol. Evol. 17:480-488.

McMahon, K.W., M.L. Fogel, B.J. Johnson, L.A. Houghton, and S.R. Thorrold. 2010. A new method to reconstruct fish diet and movement patterns from δ13C values in otolith amino acids. Can. J. Fish. Aquat. Sci. 68:1330-1340.

Mittelbach, G.G. 1981. Foraging efficiency and body size: a study of optimal diet and habitat use by bluegills. Ecology 62:1370-1386.

190

Neff, B.D. 2004. Stabilizing selection on genomic divergence in a wild fish population. Proc. Natl. Acad. Sci. U.S.A. 101:2381-2385.

Olson, R.J., B.N. Popp, B.S. Graham, G.A. Lopez-Ibarra, F. Galvan-Magaña, C.E. Lennert-Cody, N. Bocanegra-Castillo, N.J. Wallsgrove, E. Gier, V. Alatorre- Ramirez, L.T. Balance, and B. Fry. 2010. Food-web inferences of stable isotope spatial patterns in copepods and yellowfin tuna in the pelagic eastern Pacific Ocean. Prog. Oceanogr. 86:124-138.

Parsons, K.J., and B.W. Robinson. 2006. Replicated evolution of integrated plastic responses during early adaptive divergence. Evolution 60:801-813.

Peterson, B.J., and B. Fry. 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol. Syst. 18:293-320.

Pfennig, K.S., and D.W. Pfennig. 2009. Character displacement: ecological and reproductive responses to a common evolutionary problem. Quart. Rev. Biol. 84:253-276.

Popp, B.N., B.S. Graham, R.J. Olson, C.C.S. Hannides, M.J. Lott, G.A. Lopez-Ibarra, F. Galván-Magaña, and B. Fry. 2007. Insight into the trophic ecology of yellowfin tuna, Thunnus albacares, from compound-specific nitrogen isotope analysis of proteinaceous amino acids. In: Dawson, T.D., and R.T.W. Siegwolf, eds. Stable isotopes as indicators of ecological change. Elsevier Academic Press, San Diego, C.A.

Post, D.M. 2002. Using stable isotopes to estimate trophic positions: models, methods, and assumptions. Ecology 83:703-718.

Ritchie, M.G. 2007. Sexual selection and speciation. Annu. Rev. Ecol. Evol. Syst. 38:79- 102.

191

Robinson, B.W., and D.S. Wilson. 1996. Genetic variation and phenotypic plasticity in a trophically polymorphic population of pumpkinseed sunfish (Lepomis gibbosus). Evol. Ecol. 10:631-652.

Robinson, B.W., D.S. Wilson, and A.S. Margosian. 2000. A pluralistic analysis of character release in pumpkinseed sunfish (Lepomis gibbosus). Ecology 81:2799- 2812.

Robinson, B.W., Wilson, D.S., and G.O. Shea. 1996. Trade-offs of ecological specialization: an intraspecific comparison of pumpkinseed sunfish phenotypes. Ecology 77:170-178.

Schluter, D. 1993. Adaptive radiation in sticklebacks: size, shape, and habitat use efficiency. Ecology 74:699-709.

Seehausen, O., and J.J.M. van Alphen. 1999. Can sympatric speciation by disruptive sexual selection explain rapid evolution of cichlid diversity in Lake Victoria? Ecol. Lett. 2:262-271.

Strayer, D.L. 2009. Twenty years of zebra mussels: lessons from the mollusk that made headlines. Front. Ecol. Environ. 7:135-141.

Strayer, D.L. 2010. Alien species in fresh waters: ecological effects, interaction with other stressors, and prospects for the future. Freshwater Biol. 55:152-174.

Svanbäck, R., and P. Eklöv. 2003. Morphology dependent foraging efficiency in perch: a trade-off for ecological specialization? Oikos 102:273-284.

Svanbäck, R., P. Eklöv, R. Fransson, and K. Holmgren. 2008. Intraspecific competition drives multiple species resource polymorphism in fish communities. Oikos 117:114 – 124.

Taborsky, M., R.F. Oliveira, and H.J. Brockmann. 2008. The evolution of alternative reproductive tactics: concepts and questions. In: Oliveira, R.F., M. Taborsky, and

192

H.J. Brockmann, eds. Alternative reproductive tactics. Cambridge University Press, Cambridge, M.A.

Vander Zanden, M.J., and J.B. Rasmussen. 1999. Primary consumer δ13C and δ15N and the trophic position of aquatic consumers. Ecology 80:1395-1404.

Webb, P.W. 1984. Body form, locomotion and foraging in aquatic vertebrates. Am. Zool. 24:107-120.

Werner, E.E., J.F. Gilliam, D.J. Hall, and G.G. Mittelbach. 1983. An experimental test of the effects of predation risk on habitat use in fish. Ecology 64:1540-1548.

Willacker, J.J., F.A. von Hippel, P.R. Wilton, and K.M. Walton. 2010. Classification of threespine stickleback along the benthic-limnetic axis. Biol. J. Linn. Soc. 101:595- 608.

193

Appendices Appendix A: Chapter 3 supplemental material Stable isotope calibration curves and standards

The δ13C values were calibrated to VPDB (Vienna Pee Dee Belemnite) using a two-point curve anchored by international standards, either the combinations of NBS-22 (accepted value –30.03 ‰) and ANU-sucrose (IAEA-CH-6; accepted value –10.45 ‰) or USGS-40

(accepted value –26.39 ‰) and USGS-41 (accepted value +37.63 ‰). An internal keratin standard (accepted value –24.04 ‰) and USGS-40, USGS-41, and IAEA-CH-6 (when not part of the calibration curve) were used to monitor precision and accuracy. The δ13C values obtained for standards were: keratin, –24.1 ± 0.1 ‰ (n = 77); USGS-40, –26.4 ±

0.04 ‰ (n = 21); USGS-41, +37.6 ± 0.2 ‰ (n = 22); and IAEA-CH-6, –10.4 ± 0.1 ‰.

Additionally, the mean δ13C sample reproducibility for all tissue and resource samples was ± 0.1 ‰ across 31 replicated samples. The δ15N values were calibrated to atmospheric nitrogen using a two-point curve calculated using either USGS-40 (accepted value –4.52 ‰) and IAEA-N2 (accepted value +20.30 ‰) or USGS-40 and USGS-41

(accepted value +47.57 ‰). To monitor the precision and accuracy, internal keratin

(accepted value +6.36 ‰) and either USGS-41 or IAEA-N2 (when not part of the calibration curve) were used. The mean δ15N values of the standards were: keratin, +6.4 ±

0.2 ‰ (n = 76); USGS-41, +47.2 ± 0.4 (n = 22); and IAEA-N2, +20.5 ± 0.2 ‰ (n = 12).

The δ15N reproducibility for fish tissues and resource samples combined was ± 0.2 ‰ (n

= 31), within the ± 0.2 ‰ range expected for sample reproducibility.

194

Table A. 1 Summary of common invertebrate prey resource isotopic compositions for both the littoral and pelagic habitats of Lake Opinicon, Ontario. Inverterbate samples were collected in May – June 2009 and 2010. (See Chapter 3). Isotopic compositions are presented as the mean (± 1 SD) for each of the invertebrate groups.

Invertebrate Group n Collection Habitat δ13C δ15N Amphipoda 8 Littoral –21.5 ± 1.9 +3.6 ± 0.7 Ephemeroptera 3 Littoral –23.5 ± 1.4 +5.7 ± 1.8 Isopoda 5 Littoral –22.8 ± 3.1 +3.7 ± 0.3 Odonata 3 Littoral –24.3 ± 1.7 +5.8 ± 0.4 Gastropoda (snails) 18 Littoral –22.3 ± 1.9 +4.4 ± 0.8 Zooplankton 10 Pelagic –29.2 ± 1.3 +4.9 ± 1.3 (cladocera and copepoda)

195

Table A.2 Summary of the mass, length, DFA scores, and stable isotope compositions for each bluegill (Lepomis macrochirus) collected from Lake Opinicon, Ontario, during May – June 2009 and 2010. (See Chapter 3). White Muscle Liver Body Fish Collection Length Mass Number Morph Year (mm) (g) DFA 1 DFA 2 δ13C δ15N δ13C δ15N 1 Female 2009 192.2 113.5 0.16 0.32 –26.4 +8.9 –26.8 +7.6 2 Female 2009 186.9 145.0 -0.43 2.69 –26.1 +9.9 –27.3 +7.8 3 Female 2009 183.6 117.2 0.34 2.41 –27.2 +10.2 –28.5 +8.5 4 Female 2009 152.5 70.6 0.66 1.81 –27.6 +8.4 –29.9 +6.8 5 Female 2009 195.9 168.8 -2.28 2.24 -- -- –27.5 +7.3 6 Female 2009 151.7 61.3 0.02 1.19 –26.5 +8.8 –28.2 +7.5 7 Female 2009 128.9 41.5 1.04 3.12 –27.0 +7.9 –28.5 +6.2 8 Female 2009 135.6 49.6 0.23 2.19 –23.6 +8.8 –24.0 +7.5 9 Female 2009 177.0 105.4 0.15 1.68 –25.2 +9.7 –25.5 +8.6 10 Female 2009 153.7 63.3 0.48 1.46 –25.6 +8.7 –27.3 +7.0 11 Female 2009 149.0 63.6 0.75 2.59 –23.4 +9.0 –25.1 +7.4 12 Female 2009 133.2 44.0 0.86 1.51 –23.2 +7.4 –23.8 +5.4 13 Female 2009 141.7 48.2 0.86 1.14 –28.7 +7.0 –28.0 +5.4 14 Female 2009 151.7 66.4 1.24 3.07 –24.3 +9.1 –25.1 +7.0 15 Female 2009 148.4 62.4 0.64 1.22 –22.3 +8.2 –22.6 +6.0 16 Female 2009 152.6 82.5 0.56 2.91 –24.5 +9.4 –24.2 +7.1 17 Female 2009 117.8 28.6 1.28 2.53 –26.5 +9.3 –26.4 +7.4 18 Female 2009 168.2 104.5 -0.23 2.62 –24.2 +9.5 –24.9 +8.2 19 Female 2009 142.4 80.8 -0.46 0.81 –26.3 +9.1 –27.6 +7.3 20 Female 2009 155.1 110.4 0.15 1.35 –27.3 +9.3 –28.9 +7.7 21 Female 2010 185.0 90.5 0.11 2.57 -- -- –26.9 +7.7

195

196

22 Female 2010 178.0 77.9 -0.64 0.31 -- -- –29.0 +7.7 23 Female 2010 165.0 80.4 -0.40 2.33 –26.5 +9.1 –26.2 +7.3 24 Female 2010 170.0 107.6 -0.96 4.47 –26.8 +10.0 –27.6 +8.2 25 Female 2010 166.0 90.5 -0.85 2.80 –27.8 +10.2 –29.1 +8.9 26 Female 2010 153.0 73.2 0.41 3.06 –28.4 +9.9 –29.2 +8.4 27 Female 2010 184.0 82.2 0.25 1.12 –24.0 +8.7 –24.9 +9.2 28 Female 2010 176.0 83.0 -1.02 2.06 –25.1 +8.7 –25.8 +7.2 29 Female 2010 175.0 64.5 -0.40 1.87 –25.8 +10.3 –27.0 +8.4 30 Female 2010 156.0 103.2 -1.39 2.72 –25.7 +10.0 –25.7 +8.5 31 Female 2010 184.0 93.3 -0.47 1.93 –27.3 +9.2 –29.9 +6.8 32 Female 2010 165.0 89.0 -0.64 3.56 –27.7 +9.8 –28.5 +7.9 33 Female 2010 190.0 67.6 0.36 4.32 –27.2 +9.5 –27.2 +7.1 34 Female 2010 150.0 106.7 -2.04 1.17 –25.0 +9.8 –26.4 +7.9 35 Female 2010 165.0 83.2 -1.04 2.90 –26.2 +8.9 –27.5 +7.9 36 Female 2010 175.0 114.7 -0.07 2.23 –27.6 +10.4 –28.3 +8.7 37 Female 2010 187.0 61.1 0.20 3.54 –27.7 +10.0 –28.7 +8.3 38 Female 2010 185.0 77.8 -0.94 3.02 –27.9 +9.7 –30.5 +6.9 39 Female 2010 158.0 93.8 -1.01 2.65 –28.6 +9.5 –29.2 +7.9 40 Female 2010 191.0 118.6 -1.71 0.80 –26.7 +9.6 –28.2 +7.8 41 Female 2010 -- 118.0 -1.48 2.74 –27.6 +9.6 –29.0 +7.8 42 Female 2010 -- 69.3 -0.59 0.93 –28.2 +9.7 –29.1 +8.0 43 Female 2010 -- 139.9 -1.23 3.43 –28.1 +9.8 –23.7 +8.0 44 Juvenile 2009 101.1 19.5 2.97 0.13 –23.4 +9.4 –24.5 +8.0 45 Juvenile 2009 98.2 15.5 3.15 -0.22 –23.7 +9.7 –23.7 +8.3 46 Juvenile 2009 101.7 16.4 2.63 0.36 –22.6 +9.2 –23.1 +8.1 47 Juvenile 2009 99.6 15.2 3.20 -2.03 –22.3 +9.8 –22.9 +7.8 48 Juvenile 2009 88.3 11.5 2.36 -0.17 –22.4 +9.2 –23.1 +7.1

196

197

49 Juvenile 2009 105.4 22.7 1.84 -0.22 –23.5 +9.1 –23.2 +7.8 50 Juvenile 2009 107.2 19.8 2.84 -1.36 –23.6 +8.9 –23.9 +7.9 51 Juvenile 2009 115.4 22.6 0.97 -0.24 –22.2 +9.1 –21.6 +7.6 52 Juvenile 2009 97.4 15.9 2.26 0.26 –20.6 +9.0 –21.8 +7.5 53 Juvenile 2009 94.8 13.9 2.54 -1.40 –21.9 +8.2 –22.0 +7.9 54 Juvenile 2009 97.3 14.2 3.88 -0.11 –23.7 +8.4 –23.7 +7.7 55 Juvenile 2009 116.8 27.4 2.70 0.56 –23.8 +8.7 –23.3 +7.6 56 Juvenile 2009 147.8 69.2 1.11 -0.70 –23.9 +8.8 –24.3 +7.7 57 Juvenile 2009 171.2 92.0 -0.15 1.07 –25.6 +9.0 –26.7 +8.0 58 Juvenile 2009 146.8 52.9 1.72 0.58 –25.9 +9.2 –25.5 +7.7 59 Juvenile 2009 127.6 39.3 0.53 0.32 –26.0 +8.9 –27.0 +8.1 60 Juvenile 2009 121.9 39.0 1.88 0.97 –25.7 +9.9 –26.9 +8.4 61 Juvenile 2009 100.2 17.3 2.83 1.64 –25.4 +9.9 –26.3 +8.3 62 Juvenile 2009 91.4 14.3 2.62 0.40 –25.2 +9.8 –26.8 +8.4 63 Juvenile 2009 129.2 41.9 1.13 1.11 –23.0 +9.4 –23.5 +7.0 64 Juvenile 2009 92.0 14.7 3.02 0.72 –26.6 +9.7 –25.9 +8.0 65 Juvenile 2009 122.2 33.7 1.72 -1.03 –26.6 +9.9 –26.5 +9.3 66 Juvenile 2009 135.5 39.6 1.42 1.00 –25.4 +8.5 –25.1 +6.8 67 Juvenile 2009 142.4 56.6 0.15 0.44 –23.9 +8.8 –24.8 +7.9 68 Juvenile 2009 184.1 124.2 -1.58 0.01 –27.5 +9.1 –27.8 +8.7 69 Parental 2009 206.3 203.3 -3.46 -1.54 –28.2 +10.6 –28.4 +9.7 70 Parental 2009 180.7 135.4 -3.97 0.71 –24.4 -- –23.5 +9.3 71 Parental 2009 194.4 144.0 -4.40 -2.84 –25.7 +9.7 –25.5 +9.9 72 Parental 2009 191.0 129.5 -5.58 -2.64 –25.4 +10.4 –25.2 +9.9 73 Parental 2009 191.5 117.6 -4.10 -3.13 –25.7 +9.1 –25.1 +9.1 74 Parental 2009 180.9 111.6 -3.15 -1.74 –28.9 +8.0 –28.8 +7.5 75 Parental 2009 195.0 132.5 -3.97 -1.57 –26.6 +9.8 –25.4 +9.7

197

198

76 Parental 2009 209.4 153.9 -3.00 -1.02 –27.0 +9.2 –26.4 +9.4 77 Parental 2009 200.8 143.9 -3.83 -1.46 –22.9 +10.3 –22.6 +10.1 78 Parental 2009 198.3 152.2 -4.22 -2.45 –26.6 +10.0 –27.0 +10.4 79 Parental 2009 201.5 169.4 -3.35 -2.02 –27.3 +9.6 –26.2 +10.3 80 Parental 2009 198.5 148.0 -4.65 -1.07 –25.6 +9.9 –24.4 +9.4 81 Parental 2009 220.2 154.4 -3.53 -0.65 –28.0 +10.1 –28.0 +9.7 82 Parental 2009 184.6 134.5 -3.70 -1.73 –27.3 +8.5 –26.7 +7.8 83 Parental 2009 202.7 151.9 -3.06 -0.87 –28.1 +10.0 –28.4 +9.8 84 Parental 2009 201.3 143.2 -4.83 -2.82 –27.8 +9.9 –27.9 +9.8 85 Parental 2009 198.0 132.9 -3.47 -1.98 –27.2 +9.9 –27.8 +9.7 86 Parental 2009 194.8 128.2 -4.66 -4.26 –26.4 +10.7 –27.1 +10.7 87 Parental 2009 184.6 124.4 -3.21 -0.33 –30.2 +8.9 –30.4 +8.5 88 Parental 2010 175.0 91.6 -3.15 -0.62 –30.3 +8.5 –29.1 +7.6 89 Parental 2010 187.0 94.4 -3.68 0.34 –25.1 +9.3 –25.6 +8.7 90 Parental 2010 184.0 150.6 -4.51 -0.13 –28.7 +10.3 –29.6 +9.6 91 Parental 2010 190.0 113.5 -1.95 0.02 -- -- –29.4 +8.9 92 Parental 2010 203.0 119.3 -4.17 -1.80 –28.7 +9.3 –27.9 +9.0 93 Parental 2010 197.0 144.4 -3.67 -0.43 –28.6 +8.7 –29.5 +7.7 94 Parental 2010 158.0 128.7 -4.72 -1.51 –28.1 +9.5 –27.7 +9.3 95 Parental 2010 160.0 150.8 -4.18 -0.90 –28.1 +10.4 –28.6 +10.0 96 Parental 2010 180.0 131.9 -3.58 -1.54 –26.9 +9.7 –26.2 +9.5 97 Parental 2010 144.0 111.5 -2.01 -0.35 –25.3 +7.9 –25.5 +7.1 98 Parental 2010 155.0 137.9 -3.89 -1.20 –27.3 +10.0 –26.5 +9.6 99 Parental 2010 156.0 138.9 -4.59 -1.27 –27.9 +8.4 –27.5 +8.0 100 Parental 2010 171.0 156.3 -3.49 -2.13 –29.5 +9.3 –29.7 +9.0 101 Parental 2010 184.0 140.1 -3.77 -0.86 –27.9 +10.3 –28.4 +10.0 102 Parental 2010 185.0 138.5 -3.62 0.60 –26.7 +9.8 –27.4 +8.7

198

199

103 Parental 2010 178.0 165.1 -4.05 -1.88 –27.8 +9.4 –28.0 +10.1 104 Parental 2010 165.0 126.2 -6.00 -1.29 –28.2 +9.0 –27.3 +9.1 105 Satellite 2009 115.3 31.8 0.19 -0.70 –24.3 +8.1 –24.3 +6.9 106 Satellite 2009 118.8 32.3 0.19 0.87 –22.6 +8.8 –23.5 +8.3 107 Satellite 2009 141.2 53.4 -0.14 -1.20 –27.6 +9.1 –27.5 +8.9 108 Satellite 2009 120.5 33.8 1.78 0.41 –24.3 +8.4 –25.0 +7.9 109 Satellite 2009 125.7 37.7 -0.72 1.28 –24.2 +8.8 –23.3 +7.9 110 Satellite 2009 115.8 28.4 0.26 -0.21 –23.8 +9.4 –25.3 +8.4 111 Satellite 2009 104.1 22.0 0.21 -1.61 –23.5 +9.8 –23.9 +9.1 112 Satellite 2009 130.0 43.4 1.60 0.18 –25.2 +9.2 –25.9 +8.6 113 Satellite 2009 157.2 76.0 -0.65 -0.51 –26.7 +8.9 –26.4 +8.5 114 Satellite 2009 147.4 65.0 -1.40 0.03 –28.0 +9.7 –28.5 +9.5 115 Satellite 2009 121.5 31.0 0.49 -0.13 –24.1 +7.7 –24.6 +7.8 116 Satellite 2009 114.9 24.8 0.74 -0.50 –22.1 +8.4 –21.5 +7.2 117 Satellite 2009 129.9 44.7 -0.61 -1.90 –24.7 +10.4 –24.4 +8.6 118 Satellite 2009 121.4 32.5 1.98 -1.50 –26.7 +10.0 –27.5 +7.7 119 Satellite 2009 100.9 27.4 1.51 -2.20 –25.9 +10.2 –26.8 +8.2 120 Satellite 2009 111.3 23.7 1.41 -1.89 –20.9 +9.5 –23.6 +8.4 121 Satellite 2009 106.8 25.0 2.21 -0.64 –22.0 +9.1 –22.2 +8.3 122 Satellite 2009 130.4 42.3 0.22 -1.04 –26.4 +9.6 –27.8 +9.0 123 Sneaker 2009 85.5 9.9 3.84 -3.23 –24.5 +10.3 –25.1 +8.5 124 Sneaker 2009 104.5 21.7 2.89 -0.81 –21.8 +9.7 –22.9 +8.4 125 Sneaker 2009 82.9 9.5 4.33 -1.59 –23.8 +8.1 –24.6 +7.8 126 Sneaker 2009 78.6 10.2 4.58 -1.63 –23.5 +9.3 –24.3 +8.5 127 Sneaker 2009 94.0 12.1 4.55 -2.90 –26.6 +9.5 –26.6 +8.7 128 Sneaker 2009 85.3 9.9 4.96 -2.88 –24.1 +9.0 –24.4 +7.6 129 Sneaker 2009 64.6 4.8 4.68 -2.10 –23.9 +8.9 –25.4 +8.3

199

200

130 Sneaker 2009 80.8 8.9 3.70 -1.67 –23.2 +9.6 –24.8 +8.3 131 Sneaker 2009 100.3 17.9 3.06 -2.23 –24.0 +10.1 –24.8 +8.7 132 Sneaker 2009 104.4 21.5 2.43 -0.62 –25.2 +10.1 –25.5 +8.7 133 Sneaker 2009 83.1 9.9 5.31 -2.04 –23.0 +9.8 –21.2 +8.0 134 Sneaker 2009 76.1 7.0 5.43 -0.61 –22.2 +9.7 -- -- 135 Sneaker 2009 70.5 5.2 5.43 -2.05 –26.2 +9.3 –25.1 +7.8 136 Sneaker 2009 83.7 9.8 5.51 -1.13 –23.2 +9.8 –23.2 +8.2 137 Sneaker 2009 70.9 4.9 5.61 -1.88 –25.0 +9.7 –24.7 +7.7 138 Sneaker 2009 83.8 8.4 5.16 -3.94 –23.1 +9.9 –23.5 +8.7 139 Sneaker 2009 67.3 4.6 3.53 -1.23 –22.7 +9.5 –23.2 +8.4 140 Sneaker 2009 82.8 10.4 5.38 -2.42 –23.9 +9.3 –23.8 +7.8 141 Sneaker 2009 81.2 8.5 3.55 -2.45 –23.8 +8.6 –24.5 +8.8 142 Sneaker 2009 84.8 10.6 4.62 0.82 –23.1 +9.3 –21.6 +7.5 Note: -- denotes an entry for which the data is unavailable and was excluded from all statistical analyses

200

201

Appendix B: Chapter 4 supplemental material

Stable isotope calibration curves and standards

The δ13C values were calibrated to VPDB (Vienna Pee Dee Belemnite) using a two-point curve calibrated using the international standards USGS-40 (accepted value –26.39 ‰) and USGS-41 (accepted value +37.63 ‰). Internal standards of keratin (accepted value –

24.04 ‰) and ANU-sucrose (IAEA-CH-6; accepted value –10.45 ‰) were used to monitor precision and accuracy of the isotope measurements. The δ13C values of the keratin standard were –24.1 ± 0.05 ‰ (n = 34) and IAEA-CH-6 has a mean measured value of –10.4 ± 0.04 ‰ (n = 15). The mean δ13C sample reproducibility of the fish and resource samples was ± 0.11 ‰ (n = 12). These measurements for δ13C values of the standards and sample reproducibility were within the acceptable ± 0.20 ‰ range. The international standards USGS-40 (accepted value –4.52 ‰) and USGS-41 (accepted value +47.57 ‰) were also used to calibrate δ15N values to atmospheric nitrogen. Internal keratin (accepted value +6.36 ‰) and international IAEA-N2 (accepted value +20.30 ‰) reference standards were used to verify measurement precision and accuracy. The keratin standards had a mean δ15N of +6.4 ± 0.20 ‰ (n = 34) and IAEA-N2 had a mean δ15N of

+20.5 ± 0.4 ‰. Additionally, the reproducibility of fish and resource samples was ± 0.16

‰ across 12 replicate samples. Therefore, as with δ13C values, the δ15N standards and sample replicates were generally within the acceptable ± 0.20 ‰ range for accuracy and reproducibility.

202

Isotopic composition of littoral prey groups

The littoral prey ‘source’ values for the SIAR mixing model were represented by the mean measured δ13C and δ15N-values (± 1 SD) of Gastropods (snails), which have been used previously to represent that overage littoral prey isotopic composition in northern temperate lakes (see Vander Zanden and Rasmussen 1999, Colborne et al. 2013). To verify the use of Gastropods as a representative sample of the littoral resource groups during our sampling period, a one-factor analysis of variance (ANOVA) was used to compare the four primary resource groups sampled from the littoral habitat. There were non-significant differences between the four common littoral prey types collected

13 (Amphipoda, Odonata, Isopoda, and Gastropoda) in both δ C (ANOVA; F3,10 = 1.20, P =

15 0.36) and δ N (ANOVA; F3,10 = 1.27, P = 0.34). Therefore, Gastropods were used as a representative of the littoral prey resource groups in the SIAR mixing model estimates of diet.

203

Table B.1 Summary of mass, length, and stable isotopic compositions of female bluegill (Lepomis macrochirus) and eggs collected over multiple sampling periods. The isotopic compositions of eggs collected directly from the females (Day 0), as well as over three days following fertilization were determined.

δ13C δ15N

Female Mass Length Eggs Eggs Eggs Eggs Eggs Eggs Eggs Eggs Muscle Liver Muscle Liver ID (g) (mm) (Day 0) (Day 1) (Day 2) (Day 3) (Day 0) (Day 1) (Day 2) (Day 3) 1 50.6 145 –21.1 –21.2 –22.1 –21.9 –22.0 –21.9 +8.6 +7.5 +9.5 +9.1 +9.2 +9.7 2 56.8 143 –28.1 –28.8 –29.6 –29.6 –29.5 –29.8 +7.7 +4.8 +6.6 +6.8 +6.9 +6.7 3 62.5 157 –25.9 –26.1 –26.2 –26.3 –26.3 –26.3 +10.4 +9.6 +10.6 +10.6 +10.7 +11.0 4 64.4 157 –25.8 –28.9 –31.1 –31.1 –31.2 –31.5 +9.9 +6.2 +7.6 +7.3 +7.5 +7.5 5 68.0 160 –26.7 –26.6 –26.9 –26.9 –26.8 –26.9 +10.4 +9.5 +10.5 +11.0 +11.0 +11.1 6 69.0 161 –27.5 –30.0 –32.9 –33.2 –33.1 -- +9.4 +6.5 +7.7 +7.7 +7.7 -- 7 69.2 161 –27.0 –26.2 –26.8 –26.8 –26.8 –26.8 +10.4 +8.7 +10.2 +10.2 +10.3 +10.7 8 73.5 170 –25.5 –26.1 –27.5 –27.7 –27.5 –27.7 +10.4 +8.9 +10.1 +10.0 +10.2 +10.1 9 74.4 160 –24.6 –27.4 –28.8 –28.8 –28.8 –28.9 +10.1 +7.4 +9.0 +9.0 +9.1 +9.2 10 83.0 173 –26.4 –27.0 –27.9 –28.0 –27.9 –28.0 +10.1 +8.8 +9.8 +10.2 +10.2 +10.3 11 84.6 167 –28.2 –30.5 –32.5 –32.9 –32.7 –32.9 +8.8 +5.8 +7.7 +7.7 +7.6 +7.6 12 87.1 162 –27.7 –29.0 –31.0 –31.2 –31.3 -- +9.1 +6.6 +7.8 +7.8 +8.0 -- 13 94.9 174 –26.1 –26.5 –27.3 –27.3 –27.4 –27.3 +10.4 +9.2 +10.3 +10.4 +10.3 +10.8 14 95.8 183 –27.7 –27.8 –28.4 –28.3 –28.4 –28.6 +10.1 +9.3 +10.9 +10.7 +10.5 +10.5 15 100.6 175 –25.9 –27.6 –28.5 –28.7 –28.7 –28.6 +10.2 +7.2 +7.8 +8.5 +8.5 +8.8 16 108.3 181 –25.3 –25.6 –25.6 –25.7 –25.6 –27.4 +9.9 +8.4 +9.6 +9.7 +9.8 +10.4 17 115.0 180 –26.2 –26.2 –26.9 –26.6 -- -- +10.2 +8.7 +10.1 +10.3 -- -- 18 118.1 185 –26.8 –28.5 –30.8 –30.9 –30.3 –30.7 +9.9 7.3 +7.8 +7.9 +8.0 +8.0 19 126.0 191 –26.6 –26.6 –26.6 –26.6 –26.7 –26.6 +10.4 +8.8 +10.0 +10.0 +10.1 +10.1

203

204

20 130.4 181 –27.7 –25.8 –26.5 –26.5 –26.5 –26.5 +10.4 +9.0 +9.5 +9.6 +9.7 +10.1 Note: Eggs that were no longer viable on a given day of collection were not used for stable isotope analysis and are denoted by the dashes.

204

205

Appendix C: Chapter 5 supplemental material

Stable isotope calibration curves and standards

Isotope measurements of δ13C and δ15N were calibrated to the international standards USGS-40 (δ13C accepted value –26.39 ‰; δ15N accepted value –4.52 ‰) and

USGS-41 (δ13C accepted value +37.63 ‰; δ15N accepted value +47.57 ‰) using two- point curves. To monitor the precision and accuracy of the δ13C isotope measurements internal laboratory standards of keratin (accepted value –24.04 ‰) and ANU-sucrose

(IAEA-CH-6; accepted value –10.45 ‰) were used. The mean (± 1 standard deviation) measured δ13C values of keratin standards was –24.0 ± 0.1 ‰ (n = 52) and –10.5 ± 0.07

‰ (n = 22) for IAEA-CH-6. Additionally, the δ13C sample reproducibility among all tissue types sampled was ± 0.08 ‰ (n = 9). The δ15N measurements were monitored for precision and accuracy using internal standards of keratin (accepted value +6.36 ‰) and

IAEA-N2 (accepted value +20.30 ‰). The keratin standards had a mean δ15N of +6.36 ±

0.13 ‰ (n = 52) and IAEA-N2 standards were +20.46 ± 0.19 ‰ (n = 22). Sample reproducibility of δ15N across the tissue types sampled was ± 0.15 ‰. Therefore, all standards and duplicate samples were within the acceptable ± 0.20 ‰ range for δ13C and

δ15N measurements.

Stable isotopes of potential prey resources

Isotopic compositions of the three most common littoral invertebrate prey types

(gastropods, amphipods, and odonates) were compared using one-factor analyses of variance (ANOVA) with Tukey’s post-hoc comparisons, when appropriate. Insufficient

206 quantities of odonates and amphipods were collected during Spring 2011 for isotopic measurements and statistical analysis; therefore the littoral prey types were compared only for Summer 2011 and Summer 2012 sampling periods. The comparison of littoral prey groups allowed us to evaluate the use of gastropods (snails) as representative samples of the overall littoral prey resources (e.g. Post 2002). The isotopic compositions of gastropods and zooplankton were also compared between the three sampling intervals using ANOVA with Tukey’s post-hoc comparisons.

Littoral prey types (gastropods, amphipods, and odonates) had similar δ13C values in both Summer 2011 (ANOVA; F2,5 = 0.06, P = 0.95) and Spring 2012 (F2,7 = 1.16, P =

0.37). The δ15N values of the littoral prey types were similar in Spring 2012 (ANOVA;

F2,7 = 2.57, P = 0.15), but differed in Summer 2011 (F2,5 = 15.25, P = 0.01) due to higher

δ15N values in odonates compared to amphipods (Tukey’s, P = 0.01) and gastropods (P =

0.04). However, given that δ13C values were similar across all littoral prey types and gastropods were the only prey type collected during all sampling intervals, we used gastropods as a proxy for all littoral invertebrates in subsequent analyses. See Appendix

C for prey isotopic composition values (Table C.1).

Isotopic composition of gastropods and zooplankton, the respective representatives of littoral and pelagic resources, indicated that gastropods did not differ in

13 15 either δ C (ANOVA; F2,7 = 1.32, P = 0.33) or δ N (ANOVA; F2,7 = 3.41, P = 0.09)

13 across the three collection periods. However, zooplankton differed in δ C (ANOVA; F2,8

= 33.49, P < 0.001) across all three collection periods, with the highest values being found during Summer 2011 (Tukey’s, all P ≤ 0.04). Similarly, δ15N values of zooplankton differed across the sampling periods (ANOVA; F2,8 = 17.32, P = 0.001),

207 with higher δ15N in Summer 2011 compared to the two other collection periods (Tukey’s, both P ≤ 0.005). Due to this variation in the prey samples, the isotopic compositions of the fish were analyzed separately for each collection period.

208

Table C.1 Summary of the isotopic compositions of invertebrate prey groups collected from the littoral and pelagic habitats. The mean ± 1 standard deviation isotopic compositions of carbon (δ13C) and nitrogen (δ15N) are presented for common littoral (gastropods and other benthic invertebrates) and pelagic (zooplankton) prey groups. Spring 2011 Summer 2011 Spring 2012 Prey type δ13C δ15N δ13C δ15N δ13C δ15N Gastropods –24.2 ± 1.7 +1.6 ± 0.1 –22.0 ± 0.3 +1.7 ± 0.5 –22.9 ± 2.2 +1.0 ± 0.4 Benthic invertebrates -- -- –21.8 ± 0.8 +2.0 ± 1.0 –24.7 ± 1.4 +1.7 ± 0.9 Zooplankton –30.4 ± 0.3 +2.0 ± 0.3 –24.3 ± 0.8 +3.4 ± 0.3 –27.6 ± 0.4 +2.5 ± 0.3 Note: Benthic invertebrates are not included for the Spring 2011 sampling period because insufficient quantities were available for stable isotope and statistical analyses.

208

209

Table C.2 Stable isotope measurements of male and female pumpkinseed sunfish (Lepomis gibbosus) collected from littoral and pelagic habitats during 2011. Fish were collected over two time periods: May – June 2011 (Spring 2011) or August 2011 (Summer 2011). Isotopic compositions (δ13C and δ15N) of liver tissues are shown, as well as % Littoral resource estimates (see methods for details). Liver Fish Collection Mass Length % ID Period Habitat Sex (g) (mm) δ13C δ15N Littoral 1 Spring Littoral Female 11.3 86 –25.9 +6.1 47 2 Spring Littoral Female 12.8 94 –26.0 +5.3 59 3 Spring Littoral Female 16.1 97 –23.9 +6.1 68 4 Spring Littoral Female 18.7 100 –30.4 +7.1 35 5 Spring Littoral Female 19.4 102 –26.2 +7.2 78 6 Spring Littoral Female 20.9 100 –26.1 +4.7 70 7 Spring Littoral Female 21.1 100 –28.6 +4.8 60 8 Spring Littoral Female 22.2 107 –22.6 +4.9 62 9 Spring Littoral Female 27.3 114 –26.5 +6.0 67 10 Spring Littoral Female 29.5 118 –22.7 +5.2 0 11 Spring Littoral Female 29.6 123 –26.8 +7.6 100 12 Spring Littoral Female 31.9 119 –26.2 +5.5 71 13 Spring Littoral Female 34.1 127 –26.7 +6.5 72 14 Spring Littoral Female 34.2 127 –25.6 +8.2 100 15 Spring Littoral Female 35.9 126 –28.2 +7.8 100 16 Spring Littoral Female 44.2 139 –26.1 +7.4 70 17 Spring Littoral Female 48.5 141 –27.5 +7.3 29 18 Spring Littoral Male 9.9 81 –24.4 +6.1 100 19 Spring Littoral Male 16.5 100 –22.5 +6.2 81 20 Spring Littoral Male 22.1 109 –25.8 +7.1 78 21 Spring Littoral Male 23.8 111 –21.3 +6.1 67 22 Spring Littoral Male 30.2 122 –26.3 +7.9 100 23 Spring Littoral Male 31.5 122 –23.2 +6.2 91 24 Spring Littoral Male 31.6 122 –25.4 +6.8 100 25 Spring Littoral Male 32.5 122 –24.2 +5.7 75 26 Spring Littoral Male 45 134 –24.8 +5.7 100 27 Spring Littoral Male 47.9 141 –25.6 +8.4 97 28 Spring Pelagic Female 35 122 –27.2 +7.1 36 29 Spring Pelagic Female 35.1 139 –23.1 +6.2 41 30 Spring Pelagic Female 38.2 133 –27.6 +7.2 45 31 Spring Pelagic Female 48 143 –26.5 +6.4 49 32 Spring Pelagic Female 48.4 153 –27.4 +6.9 52

210

33 Spring Pelagic Female 56.5 146 –27.8 +6.3 100 34 Spring Pelagic Female 61.3 156 –26.5 +7.0 63 35 Spring Pelagic Female 65.8 150 –28.2 +8.5 63 36 Spring Pelagic Male 48.4 137 –27.5 +8.5 59 37 Spring Pelagic Male 48.6 139 –27.4 +8.5 46 38 Spring Pelagic Male 49.9 146 –26.9 +7.4 56 39 Spring Pelagic Male 52.1 144 –27.1 +7.0 90 40 Spring Pelagic Male 56.4 146 –27.3 +8.2 50 41 Spring Pelagic Male 60.6 153 –24.4 +7.0 100 42 Spring Pelagic Male 60.8 149 –27.0 +7.0 48 43 Spring Pelagic Male 63.9 154 –22.7 +8.0 77 44 Spring Pelagic Male 67 156 –26.9 +7.9 37 45 Spring Pelagic Male 67.4 143 –28.1 +7.5 50 46 Spring Pelagic Male 69 160 –24.8 +8.2 56 47 Early 2011 Pelagic Male 71.5 163 –26.8 +8.0 96 48 Early 2011 Pelagic Male 80.6 162 –27.3 +7.0 55 49 Early 2011 Pelagic Male 83.1 164 –25.6 +8.4 53 50 Late 2011 Littoral Female 14.5 94 –24.7 +6.3 34 51 Late 2011 Littoral Female 18 104 –26.7 +7.2 85 52 Late 2011 Littoral Female 18.2 97 –23.3 +5.5 42 53 Late 2011 Littoral Female 19.5 106 –23.5 +6.1 0 54 Late 2011 Littoral Female 23.8 114 –23.7 +6.7 45 55 Late 2011 Littoral Female 31 120 –22.4 +5.5 80 56 Late 2011 Littoral Female 43.8 134 –21.5 +6.2 100 57 Late 2011 Littoral Female 46.5 140 –23.3 +4.9 0 58 Late 2011 Littoral Female 67.8 159 –22.5 +6.9 26 59 Late 2011 Littoral Male 19.2 103 –24.7 +5.2 0 60 Late 2011 Littoral Male 21.8 105 –25.8 +6.2 0 61 Late 2011 Littoral Male 22.4 108 –20.4 +6.1 43 62 Late 2011 Littoral Male 27.9 120 –25.5 +6.9 0 63 Late 2011 Littoral Male 29.7 112 –23.3 +5.5 95 64 Late 2011 Littoral Male 42.4 132 –21.3 +6.4 100 65 Late 2011 Littoral Male 60 150 –22.1 +5.5 68 66 Late 2011 Littoral Male 74.8 163 –22.7 +6.8 100 67 Late 2011 Littoral Male 99.1 174 –19.3 +5.9 100 68 Late 2011 Pelagic Female 13 94 –23.7 +6.6 0 69 Late 2011 Pelagic Female 14.7 96 –22.2 +6.4 0 70 Late 2011 Pelagic Female 14.8 101 –22.8 +7.0 0 71 Late 2011 Pelagic Female 19 103 –25.1 +6.8 0 72 Late 2011 Pelagic Female 19.2 105 –26.0 +6.8 0 73 Late 2011 Pelagic Female 19.5 106 –21.7 +6.7 94 74 Late 2011 Pelagic Female 21.2 109 –24.7 +6.9 0

211

75 Late 2011 Pelagic Female 27.1 116 –24.5 +7.2 92 76 Late 2011 Pelagic Female 36.5 132 –25.3 +6.5 0 77 Late 2011 Pelagic Female 37.2 130 –26.1 +8.4 28 78 Late 2011 Pelagic Female 50.7 141 –22.1 +6.6 100 79 Late 2011 Pelagic Female 62.5 161 –25.6 +7.2 67 80 Late 2011 Pelagic Male 18.8 105 –25.0 +6.7 0 81 Late 2011 Pelagic Male 18.8 105 –24.3 +6.8 100 82 Late 2011 Pelagic Male 23.9 112 –24.4 +6.5 31 83 Late 2011 Pelagic Male 27.6 124 –23.5 +6.5 0 84 Late 2011 Pelagic Male 33.4 125 –19.7 +6.1 0 85 Late 2011 Pelagic Male 34.7 130 –25.4 +7.3 0 86 Late 2011 Pelagic Male 35.6 127 –23.6 +6.8 35

212

Table C.3 Stable isotope measurements of nesting male pumpkinseed sunfish (Lepomis gibbosus) and eggs collected from the males’ nests in the littoral and pelagic habitats. Nesting males were collected during Spring 2012 (11 – 22 June). Isotopic compositions (δ13C and δ15N) of nesting male liver tissue and egg samples from the male’s nests are shown. Diet estimates of nesting males are shown based on % Littoral resource contributions to diet for each individual (see methods for details).

Liver Eggs Male Collection Mass Length %

ID Habitat (g) (mm) δ13C δ15N δ13C δ15N Littoral 1 Littoral 73.3 142 –23.9 +6.9 –24.7 +7.4 78 2 Littoral 36.6 115 –23.9 +6.2 –23.8 +7.4 79 3 Littoral 47.3 122 –22.3 +6.8 –23.9 +6.8 100 4 Littoral 62.2 141 –25.2 +7.5 –27.4 +7.8 51 5 Littoral 28.8 115 –24.5 +6.7 –26.9 +7.4 67 6 Littoral 21.7 102 –22.9 +6.6 –22.8 +7.6 100 7 Littoral 20.8 102 –23.2 +6.5 –24.4 +7.0 94 8 Littoral 21.4 106 –23.5 +6.9 –26.4 +7.5 88 9 Littoral 41.6 125 –21.1 +6.0 –27.6 +8.1 100 10 Littoral 31.4 115 –22.2 +6.8 –24.7 +7.8 100 11 Littoral 28.8 110 –21.3 +6.1 –24.0 +7.0 100 12 Littoral 29.5 114 –24.0 +6.7 –24.1 +6.8 77 13 Littoral 81.2 165 –21.4 +7.7 –21.9 +6.2 100 14 Pelagic 61.7 140 –26.3 +7.9 –26.8 +7.0 29 15 Pelagic 66.3 150 –26.9 +7.1 –28.5 +6.7 16 16 Pelagic 88.7 164 –25.0 +8.0 –27.5 +6.8 56 17 Pelagic 63.5 140 –26.0 +7.1 –24.2 +7.2 35 18 Pelagic 81.3 157 –25.5 +7.3 –26.8 +7.0 46 19 Pelagic 55.1 138 –25.4 +7.5 –27.7 +7.1 46 20 Pelagic 64.1 141 –26.5 +7.6 –24.7 +7.0 24 21 Pelagic 52.7 135 –25.6 +7.5 –26.8 +7.8 43 22 Pelagic 56.0 140 –26.6 +7.2 –26.8 +7.9 22 23 Pelagic 96.2 163 –24.3 +7.5 –25.9 +7.0 70 24 Pelagic 47.1 130 –24.8 +8.0 –27.3 +7.6 60 25 Pelagic 58.2 145 –25.7 +7.2 –26.2 +8.1 40 26 Pelagic 92.3 164 –25.1 +7.7 –25.6 +7.0 53 27 Pelagic 50.1 134 –26.3 +7.3 –27.2 +8.2 27

213

215

Figure C.1 Isotopic compositions of individual pumpkinseed sunfish (Lepomis gibbosus) liver tissue and eggs collected from the littoral and pelagic habitats over three sampling periods (Spring 2011, Summer 2011, Spring 2012). Isotopic compositions (δ13C and δ15N) of (●) littoral- and (○) pelagic-caught fish are shown for each individual. Also shown, are the mean (± 1 SD) values of littoral gastropods and pelagic zooplankton for each sampling period.

214

(a) Spring 2011 - Liver (b) Summer 2011 - Liver

10 10

8 8

6 6

(‰ AIR) (‰ AIR)

N N

15 4 15 4  

2 2 Zooplankton Zooplankton Snails Snails

0 0 -32 -30 -28 -26 -24 -22 -20 -18 -32 -30 -28 -26 -24 -22 -20 -18

13C (‰ VPDB) 13C (‰ VPDB) (c) Spring 2012 - Liver (d) Spring 2012 - Eggs 10 10

8 8

6 6

(‰ AIR)

15 15 4 4  

2 2 Zooplankton Zooplankton

Snails Snails 0 0 -32 -30 -28 -26 -24 -22 -20 -18 -32 -30 -28 -26 -24 -22 -20 -18 13 13C (‰ VPDB)  C (‰ VPDB)

214

215

Appendix D: Chapter 6 supplemental material

Stable isotope calibration curve and standards

The measured δ13C values of each sample were calibrated to Vienna Pee Dee Belemnite

(VPDB) using two-point curves with international standards of either NBS-22 (accepted value –30.03 ‰) and ANU-sucrose (IAEA-CH-6; accepted value –10.45 ‰) or USGS-40

(accepted value –26.39 ‰) and USGS-41 (accepted value +37.63 ‰). Precision and accuracy was monitored with additional internal standards of keratin (accepted value –

24.04 ‰) and USGS-40, USGS-41, and IAEA-CH-6 (when these standards were not included in the calibration curve). The δ13C values of keratin (–24.1 ± 0.1 ‰; n = 118),

USGS-40 (–26.3 ± 0.4 ‰; n = 79), USGS-41 (37.7 ± 0.5 ‰; n = 76), and IAEA-CH-6 (–

10.5 ± 0.1 ‰; n = 54) are within the expected ranges for these standards. The reproducibility of fish tissue and resource sample δ13C values were ± 0.1 ‰ across 50 replicate samples. The measured δ15N values were calibrated to atmospheric nitrogen using a two-point curve calibrated with the combination of either USGS-40 (accepted value –4.52 ‰) and IAEA-N2 (accepted value +20.30 ‰) or USGS-40 and USGS-41

(accepted value +47.57 ‰). Precision and accuracy of measurements were also monitored using internal standards of keratin (accepted value +6.36 ‰) and either USGS-

41 or IAEA-N2 (when not used as part of the calibration curve). The mean δ15N values of the keratin (+6.4 ± 0.2 ‰; n = 118), USGS-41 (+47.4 ± 0.6 ‰), and IAEA-N2 (+20.4 ±

0.3 ‰) were within the acceptable ranges for these standards. The reproducibility of δ15N measurements for fish tissues and reference prey samples was ± 0.2 ‰, meeting the expected range for sample reproducibility.

216

Table D.1 Stable isotopic composition measurements of potential invertebrate prey resource groups. Benthic Snails invertebrates Zebra mussels Zooplankton Lake Sampling period δ13C δ15N n δ13C δ15N n δ13C δ15N n δ13C δ15N n –20.8 ± +3.9 –21.5 ± +3.4 –27.2 ± +4.1 –27.0 ± +5.0 ± Opinicon Spring 2011 5 6 4 3 2.9 ± 0.4 2.8 ± 0.5 0.8 ± 0.5 1.9 0.4 –20.3 ± +4.1 –18.0 ± +2.7 –26.6 ± +4.9 –28.5 ± +4.3 ± Summer 2011 6 5 3 4 2.1 ± 0.4 3.5 ± 0.8 2.0 ± 0.3 2.0 0.6

–24.7 ± +6.0 –24.7 ± +5.9 –27.9 ± +7.0 –29.0 ± +7.4 ± Lower Beverley Spring 2011 4 6 4 4 2.5 ± 0.4 0.9 ± 0.4 1.9 ± 0.9 0.6 0.6 –22.1 ± +5.7 –21.9 ± +4.7 –25.9 ± +6.3 –27.7 ± +6.0 ± Summer 2011 6 7 4 3 2.1 ± 0.6 2.1 ± 0.5 2.2 ± 0.7 0.6 0.9

–22.0 ± +4.3 –20.6 ± +3.6 –26.9 ± +4.4 –27.4 ± +4.2 ± Indian Summer 2012 5 6 4 3 1.5 ± 0.3 2.3 ± 0.3 1.6 ± 0.4 0.3 0.7

216

217

Table D.2 Summary of the mass, length, and stable isotope compositions for each pumpkinseed sunfish (Lepomis gibbosus). White muscle Liver Fish Sampling Mass Length ID Lake Period (g) (mm) Sex δ13C δ15N δ13C δ15N 1 Opinicon Spring 2011 69 149 Female –24.9 +8.7 –24.8 +8.4 2 Opinicon Spring 2011 90.9 172 Female –26.3 +9.7 –26.3 +8.7 3 Opinicon Spring 2011 99.3 185 Female –25.3 +8.5 –25.5 +7.9 4 Opinicon Spring 2011 109.3 169 Female –24.4 +9.6 –25.0 +7.2 5 Opinicon Spring 2011 122 175 Female –28.9 +9.2 –30.6 +6.2 6 Opinicon Spring 2011 126.9 183 Female –27.4 +9.9 –28.6 +8.8 7 Opinicon Spring 2011 127 180 Female –25.7 +9.9 –25.9 +7.4 8 Opinicon Spring 2011 140.2 184 Female –21.8 +9.2 –23.7 +7.6 9 Opinicon Spring 2011 199.8 202 Female –25.6 +9.0 –26.9 +6.4 10 Opinicon Spring 2011 209.6 214 Female –23.9 +9.8 –23.1 +7.8 11 Opinicon Spring 2011 210 211 Female –23.7 +9.6 –24.8 +6.6 12 Opinicon Spring 2011 42.9 127 Juvenile –25.6 +9.2 –25.8 +7.7 13 Opinicon Spring 2011 85 162 Juvenile –22.0 +9.4 –21.7 +8.3 14 Opinicon Spring 2011 65.4 150 Male –24.1 +9.7 –23.4 +8.8 15 Opinicon Spring 2011 73.9 155 Male –25.0 +10.0 –24.3 +8.7 16 Opinicon Spring 2011 87.8 163 Male –23.8 +8.6 –23.5 +7.5 17 Opinicon Spring 2011 100.6 167 Male –23.1 +9.1 –22.6 +8.0 18 Opinicon Spring 2011 112.3 170 Male –24.7 +9.4 –24.5 +7.6 19 Opinicon Spring 2011 113.1 169 Male –25.0 +9.8 –24.6 +6.9 20 Opinicon Spring 2011 113.7 169 Male –25.4 +9.2 –24.4 +7.9 21 Opinicon Spring 2011 113.7 171 Male –24.6 +9.3 –24.9 +8.6 22 Opinicon Spring 2011 116.6 168 Male –25.0 +9.4 –24.8 +7.1 23 Opinicon Spring 2011 123.2 185 Male –26.5 +9.3 –25.9 +9.5

217

218

24 Opinicon Spring 2011 124.1 179 Male –22.2 +9.6 –22.3 +8.5 25 Opinicon Spring 2011 126.8 180 Male –25.5 +9.8 –25.4 +8.4 26 Opinicon Spring 2011 128.2 179 Male –24.1 +8.4 –23.7 +7.6 27 Opinicon Spring 2011 135.1 203 Male –23.9 +10.5 –23.0 +9.7 28 Opinicon Spring 2011 140.1 182 Male –26.6 +10.1 –27.0 +8.2 29 Opinicon Spring 2011 140.9 183 Male –24.9 +10.0 –23.0 +8.0 30 Opinicon Spring 2011 142.1 190 Male –26.4 +9.9 –25.4 +8.9 31 Opinicon Spring 2011 143.9 197 Male –25.7 +10.1 –24.1 +9.0 32 Opinicon Spring 2011 144.9 189 Male –24.9 +8.5 –23.4 +7.8 33 Opinicon Spring 2011 150.7 186 Male –26.8 +10.3 –27.0 +10.3 34 Opinicon Spring 2011 153.6 190 Male –28.3 +9.5 –28.2 +7.6 35 Opinicon Spring 2011 157.8 196 Male –26.6 +10.3 –26.3 +9.0 36 Opinicon Spring 2011 165 192 Male –26.2 +9.2 –25.4 +8.3 37 Opinicon Spring 2011 165.8 195 Male –28.4 +9.4 –28.8 +7.4 38 Opinicon Spring 2011 168.1 201 Male –25.1 +9.6 –25.3 +8.7 39 Opinicon Spring 2011 169.5 190 Male –27.0 +8.9 –25.6 +7.9 40 Opinicon Spring 2011 171.4 187 Male –23.5 +9.3 –22.4 +9.0 41 Opinicon Spring 2011 173.6 195 Male –27.7 +9.9 –26.9 +8.4 42 Opinicon Spring 2011 178.1 208 Male –22.2 +10.3 –20.3 +9.0 43 Opinicon Spring 2011 178.7 205 Male –24.7 +9.8 –23.7 +9.3 44 Opinicon Spring 2011 178.8 206 Male –26.9 +10.2 –25.8 +9.4 45 Opinicon Spring 2011 186.2 203 Male –24.7 +10.2 –23.0 +9.1 46 Opinicon Spring 2011 188.3 207 Male –27.5 +10.2 –27.3 +9.4 47 Opinicon Spring 2011 189.4 205 Male –23.0 +9.2 –21.6 +8.3 48 Opinicon Spring 2011 189.9 200 Male –25.2 +10.0 –25.0 +9.0 49 Opinicon Spring 2011 191.5 198 Male –26.0 +10.3 –25.9 +10.1

218

219

50 Opinicon Spring 2011 195 204 Male –23.8 +9.6 –23.8 +8.1 51 Opinicon Summer 2011 58.5 145 Female –23.4 +9.9 –23.6 +8.1 52 Opinicon Summer 2011 59.3 146 Female –26.1 +9.7 –25.5 +8.0 53 Opinicon Summer 2011 64.7 149 Female –25.0 +9.2 –24.3 +7.6 54 Opinicon Summer 2011 77.4 152 Female –23.7 +9.7 –23.8 +7.9 55 Opinicon Summer 2011 85.9 167 Female –22.0 +9.6 –20.2 +7.8 56 Opinicon Summer 2011 94.8 186 Female –24.6 +10.0 –24.0 +9.1 57 Opinicon Summer 2011 114.6 200 Female –25.5 +9.3 –24.3 +9.0 58 Opinicon Summer 2011 118.7 191 Female –25.4 +9.6 –24.8 +9.1 59 Opinicon Summer 2011 127.3 184 Female –26.2 +9.9 -- -- 60 Opinicon Summer 2011 145.9 201 Female –24.0 +8.7 –22.7 +7.7 61 Opinicon Summer 2011 63.2 146 Juvenile –23.4 +9.4 –23.6 +8.1 62 Opinicon Summer 2011 36.1 122 Male –24.9 +9.9 –24.1 +7.9 63 Opinicon Summer 2011 54 139 Male –23.6 +9.3 –23.7 +8.0 64 Opinicon Summer 2011 54.5 141 Male –24.9 +9.0 –24.0 +7.9 65 Opinicon Summer 2011 67.6 149 Male –23.8 +9.5 –23.6 +7.9 66 Opinicon Summer 2011 69 147 Male –25.5 +9.6 –24.1 +7.9 67 Opinicon Summer 2011 71.5 151 Male –24.4 +9.6 –24.2 +8.4 68 Opinicon Summer 2011 74.3 156 Male –25.2 +9.2 –23.3 +7.7 69 Opinicon Summer 2011 74.6 154 Male –22.6 +9.5 –23.3 +8.0 70 Opinicon Summer 2011 76.7 160 Male –22.1 +9.2 –21.8 +8.4 71 Opinicon Summer 2011 77.8 159 Male –23.2 +10.0 –23.3 +8.3 72 Opinicon Summer 2011 79.9 153 Male –23.4 +9.6 –23.7 +7.9 73 Opinicon Summer 2011 96.4 175 Male –24.3 +9.5 –24.0 +7.8 74 Opinicon Summer 2011 118 182 Male –28.7 +9.6 –27.5 +9.2 75 Opinicon Summer 2011 156.7 201 Male –24.8 +10.1 –23.5 +9.4 76 Opinicon Summer 2011 165.4 209 Male –24.3 +9.9 –22.5 +9.1

219

220

77 Opinicon Summer 2011 197.3 216 Male –24.3 +9.7 –22.5 +8.0 78 Opinicon Summer 2011 204.8 215 Male –22.5 +9.9 –20.9 +7.6 79 L. Beverley Spring 2011 36.4 129 Female –27.7 +10.9 –28.3 +9.3 80 L. Beverley Spring 2011 62.9 141 Female –27.8 +11.6 –28.5 +10.0 81 L. Beverley Spring 2011 84.1 152 Female –26.5 +11.8 –27.0 +10.2 82 L. Beverley Spring 2011 94.2 157 Female –25.4 +11.8 –25.2 +10.0 83 L. Beverley Spring 2011 98.3 162 Female –26.9 +11.3 –26.4 +9.3 84 L. Beverley Spring 2011 113.9 174 Female –27.2 +12.1 –27.1 +10.3 85 L. Beverley Spring 2011 123.4 174 Female –27.4 +12.0 –26.4 +9.4 86 L. Beverley Spring 2011 130.2 172 Female –25.9 +12.1 –25.9 +9.8 87 L. Beverley Spring 2011 137.9 201 Female –27.0 +12.5 –26.7 +11.2 88 L. Beverley Spring 2011 142.1 190 Female –28.8 +12.8 –27.3 +10.5 89 L. Beverley Spring 2011 148.7 182 Female –27.3 +12.6 –28.3 +10.4 90 L. Beverley Spring 2011 166.4 195 Female –26.7 +12.1 –27.1 +9.2 91 L. Beverley Spring 2011 37.6 116 Juvenile –26.8 +10.7 –27.0 +9.5 92 L. Beverley Spring 2011 40.5 125 Juvenile –27.3 +10.8 –27.5 +9.7 93 L. Beverley Spring 2011 44.2 130 Juvenile –23.8 +11.4 –25.2 +10.0 94 L. Beverley Spring 2011 49.2 132 Juvenile –27.4 +12.1 –27.3 +10.7 95 L. Beverley Spring 2011 50.2 135 Juvenile –24.0 +10.9 –24.4 +9.1 96 L. Beverley Spring 2011 50.8 134 Juvenile –23.7 +11.8 –25.8 +10.7 97 L. Beverley Spring 2011 56.1 131 Juvenile –28.1 +11.5 –29.2 +10.2 98 L. Beverley Spring 2011 56.4 140 Juvenile –27.4 +9.5 –25.9 +8.5 99 L. Beverley Spring 2011 57.5 139 Juvenile –27.6 +9.2 –26.6 +8.3 100 L. Beverley Spring 2011 61 137 Juvenile –28.6 +9.8 –28.0 +8.6 101 L. Beverley Spring 2011 63.5 142 Juvenile –26.0 +11.7 –25.9 +10.1 102 L. Beverley Spring 2011 66.6 143 Juvenile –26.9 +11.0 –25.9 +9.8 103 L. Beverley Spring 2011 71.8 151 Juvenile –28.1 +11.6 –29.2 +10.6

220

221

104 L. Beverley Spring 2011 76 145 Juvenile –26.8 +11.4 –27.7 +10.3 105 L. Beverley Spring 2011 79.5 150 Juvenile –28.2 +11.6 –28.6 +10.4 106 L. Beverley Spring 2011 81.8 157 Juvenile –24.4 +12.0 –24.4 +10.6 107 L. Beverley Spring 2011 88.1 160 Juvenile –27.8 +11.5 –27.6 +9.8 108 L. Beverley Spring 2011 90.9 158 Juvenile –26.7 +10.9 –26.5 +10.4 109 L. Beverley Spring 2011 92.1 158 Juvenile –27.7 +10.8 –27.8 +9.9 110 L. Beverley Spring 2011 158.9 192 Juvenile –25.0 +11.1 –24.5 +9.1 111 L. Beverley Spring 2011 75.8 151 Male –26.9 +13.0 –26.1 +12.0 112 L. Beverley Spring 2011 80.2 149 Male –27.1 +10.6 –27.0 +10.8 113 L. Beverley Spring 2011 83.6 153 Male –28.0 +12.9 –27.6 +11.3 114 L. Beverley Spring 2011 98.5 160 Male –24.6 +12.0 –22.8 +9.9 115 L. Beverley Spring 2011 105.2 172 Male –27.1 +10.1 –26.3 +9.3 116 L. Beverley Spring 2011 110.2 171 Male –28.1 +12.9 –27.1 +11.5 117 L. Beverley Spring 2011 113.9 164 Male –28.0 +13.2 –27.6 +12.3 118 L. Beverley Spring 2011 117.9 175 Male –26.5 +11.6 –24.5 +10.2 119 L. Beverley Spring 2011 124 172 Male –26.2 +13.3 –25.0 +11.2 120 L. Beverley Spring 2011 124.5 175 Male –27.1 +11.6 –26.1 +10.4 121 L. Beverley Spring 2011 125.7 179 Male –26.9 +12.3 –26.2 +11.6 122 L. Beverley Spring 2011 130.7 175 Male –27.2 +11.9 –26.3 +10.3 123 L. Beverley Spring 2011 132.4 169 Male –26.9 +11.4 –26.0 +10.4 124 L. Beverley Spring 2011 137.7 180 Male –27.2 +12.4 –26.8 +10.7 125 L. Beverley Spring 2011 140 179 Male –28.7 +13.1 –27.7 +11.2 126 L. Beverley Spring 2011 142.2 171 Male –28.9 +13.7 –29.3 +12.4 127 L. Beverley Spring 2011 155.2 186 Male –27.1 +13.1 –25.5 +11.5 128 L. Beverley Spring 2011 228.2 174 Male –28.0 +12.7 –26.8 +11.4 129 L. Beverley Summer 2011 37.5 120 Female –26.6 +11.2 –25.9 +9.4 130 L. Beverley Summer 2011 41.1 122 Female –29.2 +9.7 –27.4 +8.9

221

222

131 L. Beverley Summer 2011 84.8 160 Female –26.3 +12.2 –25.1 +10.4 132 L. Beverley Summer 2011 93.2 164 Female –27.1 +11.9 –26.4 +10.1 133 L. Beverley Summer 2011 94.2 165 Female –27.1 +11.9 –25.7 +9.7 134 L. Beverley Summer 2011 104.3 161 Female –26.6 +11.8 –26.1 +9.9 135 L. Beverley Summer 2011 107.9 175 Female –23.4 +12.1 –23.2 +10.2 136 L. Beverley Summer 2011 111.3 176 Female –25.9 +12.2 –25.8 +9.8 137 L. Beverley Summer 2011 111.5 167 Female –27.0 +12.1 –26.8 +10.2 138 L. Beverley Summer 2011 112.7 170 Female –27.0 +11.8 –26.6 +10.3 139 L. Beverley Summer 2011 135.5 187 Female –26.9 +12.4 –26.5 +10.5 140 L. Beverley Summer 2011 44.3 129 Male –28.4 +11.7 –28.3 +9.8 141 L. Beverley Summer 2011 48.8 131 Male –25.8 +11.1 –25.4 +9.5 142 L. Beverley Summer 2011 50.2 135 Male –27.3 +11.0 –26.0 +9.4 143 L. Beverley Summer 2011 72.6 151 Male –26.3 +11.8 –25.7 +10.1 144 L. Beverley Summer 2011 74.4 149 Male –27.7 +11.0 –26.7 +9.6 145 L. Beverley Summer 2011 93.2 160 Male –24.6 +11.9 –25.2 +10.4 146 L. Beverley Summer 2011 103.9 170 Male –26.3 +12.2 –25.1 +10.6 147 L. Beverley Summer 2011 105 169 Male –25.6 +12.7 –25.2 +11.0 148 L. Beverley Summer 2011 105.7 170 Male –26.3 +11.3 –25.9 +9.7 149 L. Beverley Summer 2011 108.2 165 Male –27.0 +12.1 –25.5 +10.2 150 L. Beverley Summer 2011 115.4 170 Male –26.6 +11.8 –26.2 +10.1 151 L. Beverley Summer 2011 118.2 178 Male –26.7 +12.0 –26.1 +10.8 152 L. Beverley Summer 2011 121.6 178 Male –26.5 +11.6 –25.0 +9.8 153 L. Beverley Summer 2011 136 151 Male –25.8 +12.2 –25.3 +10.7 154 Indian Spring 2012 47.7 135 Female –23.8 +9.8 –24.8 +9.6 155 Indian Spring 2012 69.6 150 Female –26.0 +10.1 –26.6 +8.6 156 Indian Spring 2012 99 176 Female –25.6 +9.5 –26.5 +7.1 157 Indian Spring 2012 108.4 173 Female –25.8 +9.6 –26.4 +7.9

222

223

158 Indian Spring 2012 108.9 177 Female –26.5 +11.0 –27.1 +7.9 159 Indian Spring 2012 137.4 191 Female –26.6 +11.4 –26.9 +9.1 160 Indian Spring 2012 139 190 Female –27.8 +10.0 –28.4 +8.7 161 Indian Spring 2012 152.5 191 Female –25.7 +10.2 –24.2 +7.4 162 Indian Spring 2012 157.1 200 Female –24.5 +10.0 –22.5 +7.4 163 Indian Spring 2012 232.8 212 Female –24.9 +10.6 –22.7 +8.1 164 Indian Spring 2012 46.3 123 Juvenile –23.7 +10.1 –23.8 +9.2 165 Indian Spring 2012 47.6 133 Juvenile –24.4 +10.7 –23.5 +8.0 166 Indian Spring 2012 102.1 176 Juvenile –23.7 +9.6 –24.7 +8.5 167 Indian Spring 2012 104.5 173 Juvenile –26.1 +10.6 –26.3 +9.7 168 Indian Spring 2012 31 122 Male –24.7 +10.4 –23.1 +9.3 169 Indian Spring 2012 41.4 129 Male –24.7 +9.6 –26.8 +8.7 170 Indian Spring 2012 50.8 139 Male –25.5 +10.1 –26.1 +9.3 171 Indian Spring 2012 52.5 140 Male –25.0 +9.5 –26.9 +8.2 172 Indian Spring 2012 70.1 148 Male –24.9 +9.4 –24.2 +8.0 173 Indian Spring 2012 74.3 150 Male –25.7 +9.5 –27.1 +7.7 174 Indian Spring 2012 74.8 160 Male –21.9 +10.0 –21.0 +9.6 175 Indian Spring 2012 81.3 161 Male –23.3 +9.8 –23.4 +8.8 176 Indian Spring 2012 90.7 166 Male –26.1 +10.6 –26.1 +9.3 177 Indian Spring 2012 92.3 168 Male –26.9 +9.6 –27.7 +8.0 178 Indian Spring 2012 96.2 168 Male –26.9 +10.3 –26.5 +8.9 179 Indian Spring 2012 99.3 170 Male –26.2 +10.0 –26.4 +8.7 180 Indian Spring 2012 102.5 164 Male –26.8 +9.4 –25.8 +8.2 181 Indian Spring 2012 106 166 Male –25.9 +9.7 –24.8 +8.3 182 Indian Spring 2012 108.3 175 Male –24.8 +11.1 –24.7 +9.9 183 Indian Spring 2012 112.4 175 Male –24.5 +9.5 –23.6 +8.0 184 Indian Spring 2012 112.5 174 Male –26.1 +10.6 –25.0 +9.9

223

224

185 Indian Spring 2012 113.7 175 Male –24.0 +9.6 –23.8 +7.8 186 Indian Spring 2012 115.5 169 Male –27.5 +8.4 –26.6 +8.0 187 Indian Spring 2012 116.4 178 Male –23.8 +10.4 –23.9 +9.7 188 Indian Spring 2012 116.7 167 Male –27.3 +8.7 –28.0 +6.8 189 Indian Spring 2012 127.6 180 Male –24.9 +9.7 –24.6 +8.5 190 Indian Spring 2012 133.2 184 Male –27.6 +8.6 –26.7 +8.2 191 Indian Spring 2012 133.3 188 Male –25.3 +10.9 –24.5 +10.0 192 Indian Spring 2012 134.6 184 Male –27.3 +10.2 –25.2 +8.6 193 Indian Spring 2012 135.9 188 Male –26.6 +11.1 –24.8 +10.1 194 Indian Spring 2012 137.5 192 Male –26.7 +9.7 –26.1 +8.6 195 Indian Spring 2012 141.7 191 Male –24.4 +11.0 –22.9 +8.3 196 Indian Spring 2012 143.7 184 Male –25.6 +9.8 –25.3 +9.2 197 Indian Spring 2012 158.2 198 Male –26.0 +9.8 –25.8 +7.9 198 Indian Spring 2012 173.8 202 Male –25.5 +11.1 –25.0 +10.6 199 Indian Spring 2012 200.2 220 Male –27.4 +10.2 –26.2 +9.7 200 Indian Spring 2012 211.8 218 Male –28.0 +11.3 –26.5 +10.2 201 Indian Spring 2012 227.3 219 Male –25.9 +10.6 –25.3 +9.5 202 Indian Spring 2012 239.1 221 Male –27.2 +11.9 –26.8 +11.0 203 Indian Spring 2012 241.9 218 Male –23.3 +10.0 –22.7 +9.1

224

225

Appendix E: Permission to reproduce published material

The content contained within Chapter 2 has been published in the Canadian Journal of Fisheries and Aquatic Sciences (2011). As of 2009, the authors of all publications in this journal retain the copyright permission for the published content. As such, authors:

“may reuse all or part of their manuscript in other works created by them for noncommercial purposes, provided the original publication in an NRC Research Press journal is acknowledged through a note or citation.”

http://www.nrcresearchpress.com/page/authors/information/rights

The contents of Chapter 3 have been published in the journal Behavioral Ecology which retains copyright only for publication rights of the submitted works. Authors have the retain the copyright and permission to “include the article in full or in part in a thesis or dissertation, provided that this is not published commercially.”

http://www.oxfordjournals.org/access_purchase/publication_rights.html

226

Appendix F: Animal care protocol approval documentation

227

228

Curriculum Vitae Post-secondary University of Guelph Education and Guelph, Ontario, Canada Degrees: 2002-2006 B.Sc. Zoology

University of Guelph Guelph, Ontario, Canada 2006-2008 M.Sc. Zoology Supervisor: Prof. Beren Robinson

The University of Western Ontario London, Ontario, Canada 2009-Present Ph.D. Biology Supervisor: Prof. Bryan Neff

Honours and Province of Ontario Graduate Scholarship Awards: 2012-2013

Western University Biology Graduate Student Teaching Award 2013

Dr. Irene Uchida Fellowship in Life Sciences 2012

Robert and Ruth Lumsden Graduate Award in Science 2012

Queen Elizabeth II Graduate Scholarship in Science and Technology 2011-2012

Related Work Teaching Assistant (2006 – 2008) Experience University of Guelph Courses: Biology of Polluted Waters, Environmental Biology of Fishes, Humans in the Natural World, Ecological Methods

Teaching Assistant (2009 – 2013) The University of Western Ontario Courses: Animal Behaviour (2009-2012), Behavioural Ecology (2011-2013), Environmental Biology (2010)

Volunteer Co-chair Ontario Ecology, Ethology and Evolution Colloquium Experience (The University of Western Ontario) 2012 – 2013

229

Graduate Representative: Faculty of Science Decanal Review and Selection Committee (The University of Western Ontario) 2010 – 2011

Outreach Representative: Society of Biology Graduate Students (The University of Western Ontario) 2009 – 2010

Ontario Ecology and Ethology Colloquium (University of Guelph) 2007 – 2008

Graduate Student Representative: Faculty Hiring Committee (University of Guelph) 2007

Publications

Peer-reviewed publications

(1 ) Colborne, S.F., Peres-Neto, P.R., Longstaffe, F.J., and Neff, B.D. 2013. Effects of foraging and sexual selection on ecomorphology of a fish with alternative mating tactics. Behav. Ecol. 24(6):1339-1347.

(2) Colborne, S.F., and Robinson, B.W. 2013. Effect of nutritional condition on variation in δ13C and δ15N stable isotope values in pumpkinseed sunfish (Lepomis gibbosus) fed different diets. Enviro. Biol. Fishes 96(4):543-554.

(3) Colborne, S.F., Bellemare, M.C., Peres-Neto, P.R., and Neff, B.D. 2011. Morphological and swim performance variation among reproductive tactics of bluegill sunfish (Lepomis macrochirus). Can. J.Fish. Aquat. Sci. 68(10):1802-1810.

In preparation

(4) Colborne, S.F., Hain, T.J.A, Longstaffe, F.J., and Neff, B.D. In review. Covariation in stable isotope composition between females and their eggs in a fish: potential for less invasive inference of resource use. Can. J. Fish. Aquat. Sci.

(5) Colborne, S.F., Garner, S.R., Longstaffe, F.J., and Neff, B.D. In review. Assortative mating but no evidence of genetic divergence in a species characterized by a trophic polymorphism. Evolution

230

(6) Colborne, S.F., Clapp, A.D.M., Longstaffe, F.J., and Neff, B.D. In review. Foraging ecology of native pumpkinseed sunfish (Lepomis gibbosus) following the invasion of zebra mussels (Driessena polymorpha). Freshwater Biol.

Conferences Ontario Ecology, Ethology and Evolution Colloquium (Oe3C and Meetings 2013) - May 2013 (University of Western Ontario, London, Ontario) [Oral presentation]

Biology Graduate Research Forum – October 2012 (University of Western Ontario, London, Ontario) [Poster presentation]

8th International Conference on Applications of Stable Isotope Techniques to Ecological Studies (IsoEcol 2012) – August 2012 (Brest, France) [Poster presentation]

First Joint Congress in Evolution – July 2012 (Ottawa, Ontario) [Oral presentation]

Canadian Society of Zoologists Annual Meeting – May 2009 (University of Toronto – Scarborough, Ontario) [Oral presentation]

Canadian Society for Ecology and Evolution Annual Meeting – May 2008 (University of British Columbia; Vancouver, British Columbia) [Oral presentation]

Ontario Ethology and Ecology Colloquium – April 2008 (University of Guelph; Guelph, Ontario) [Oral presentation]

Graduate Speciation Concepts (The University of Western Ontario) Courses Famous Biologists (The University of Western Ontario)

Teach the Controversy (The University of Western Ontario)

Behavioural Ecology (The University of Western Ontario)

Advances in Evolutionary Biology (University of Guelph)

Statistical Methods for the Life Sciences (University of Guelph)

Scientific Communication I & II (University of Guelph)