Evolutionary ecology of host-parasite interactions in the Trinidadian guppy (Poecilia reticulata)

Felipe Pérez-Jvostov

Institute of Parasitology McGill University Montreal, Quebec

A thesis submitted to McGill University in partial fulfillment of the requirements of the Degree of Doctor of Philosophy!

!

© Felipe Pérez-Jvostov, 2015

All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author.!

! i

A mi familia, Felipe, Victoria, Karina y Sabrina.

! ii!

I shall always feel respect for every one who has written a [dissertation], let it be what it may, for I had no idea of the trouble which trying to write common English could cost one

–! CHARLES R. DARWIN, Letter to W. D. Fox, 1837

! iii! THESIS ABSTRACT

Some of the most extraordinary examples of adaptive evolution are the result of the interactions between hosts and their parasites. Although this so-called arms race between host defense and parasite virulence can have direct implications for host fecundity and survival, current research shows that parasitic infections can also change host behavioral and physiological traits. Hosts face multiple threats at any give time. Other natural enemies like predators and competitors can also reduce host survival. Thus, host-parasite adaptations may not only influence non-host species, but may, simultaneously, be influenced by their interaction with other natural enemies.

The role of parasites in modulating multi-species interactions has received increased attention recently, yet the majority of examples come from single case studies.

The aim of this dissertation is to fill this gap by using a model organism in evolutionary biology: the Trinidadian guppy. First, I start by broadly differentiating between generalist and specialist parasites, and review the mechanisms by which both can influence non- host species at ecological and evolutionary time scales. This distinction has been extensively studied in relation to the evolution of virulence, but its ecological relevance has been largely neglected. Second, I consider how the particular interactions between guppies and their specialist ectoparasite, Gyrodactylus, are influenced by guppy adaptation to predation (high vs. low) across multiple rivers in Trinidad. I found strong, and repeatable, differences between guppies from different rivers and their capacity to limit Gyrodactylus infections, but found no differences between predation environments.

iv ! Nonetheless, a consistent reduction in the growth of guppies exposed to Gyrodactylus was evident, potentially due to the energetic cost associated with the activation and maintenance of an immune response to fight Gyrodactylus infections. These results suggest a divergence in guppy-Gyrodactylus coevolutionary trajectories among rivers

(higher resistance and higher virulence vs. lower resistance and lower virulence), and highlight that population evolutionary history (i.e., genetic makeup, founder effects, genetic bottle necks, etc.), and not predation, strongly influence the intensity and direction of host-parasite coevolution in this system. Finally, I explore how Gyrodactylus can modify the interaction between guppies and their natural enemy, Rivulus hartii. I found that although Gyrodactylus can strongly reduce guppy growth, infection does not significantly modify guppy-Rivulus interactions, particularly because guppies can fine- tune their phenotype when simultaneously facing multiple enemies. These results demonstrate that hosts have the capacity to mediate, and even mitigate, any potential ecological effects that parasites could have on the broader community.

Understanding the mechanisms by which parasites can influence non-host species has become a major goal in evolutionary ecology. However, the complexity of natural communities, environmental heterogeneity, and host evolutionary history can have strong, and many times unexpected, effects on host-parasite interactions. The results I present here are a vivid example of such a natural community. Using a combination of experimental work and field manipulation, I demonstrate that Gyrodactylus can have strong effects on host life-history traits, but that these effects do not necessarily influence non-host species, largely because of host capacity to limit such effects. I believe my work

v ! has laid the foundations for future research on the ecological relevance of Gyrodactylus, as I was the first to investigate how guppy-Gyrodactylus coevolution could potentially influence the coexisting fauna.

vi ! RÉSUMÉ DE LA THÈSE

Les interactions entre les hôtes et leurs parasites représente un des exemples les plus extraordinaires de l’évolution adaptative. Bien que cette prétendue "course aux armements" entre la défense de l'hôte et la virulence parasitaire puisse avoir des répercussions directes sur la fécondité et la survie de l'hôte, la recherche actuelle démontre que les infections parasitaires peuvent également modifier les traits comportementaux et physiologiques des hôtes. À tout instant, les hôtes doivent affronter une abondance de menaces. Un ensemble varié d’ennemis naturels, tels que les prédateurs et les compétiteurs peuvent également menacer la survie de l'hôte. Ainsi, les adaptations hôte-parasite peuvent non seulement influencer les espèces non hôtes, mais peuvent à la fois être influencées par leur interaction avec d'autres ennemis naturels.

Au fil des dernières années, le rôle particulier des parasites qui modulent les interactions multiespèces a fait l'objet d'une attention accrue. Cependant, la majorité des exemples analysés proviennent d'études de cas simples. La présente thèse a pour objectif de combler cette lacune en employant un organisme modèle en biologie évolutive: le guppy trinidadien. Tout d'abord, je commence en caractérisant de manière générale les différences entre les parasites généralistes et spécialistes, pour ensuite envisager les mécanismes par lesquels chacun peut influencer des espèces non hôtes sur des échelles de temps écologiques et évolutives. Cette distinction a déjà été amplement étudiée par rapport à l'évolution de la virulence, mais sa pertinence écologique a été largement négligée. En second lieu, je considère la façon dont les interactions particulières entre les

vii ! guppys et leurs ectoparasites spécialistes - soit les Gyrodactylus - sont influencées par l'adaptation du guppy à la prédation (haute versus basse) au sein de plusieurs rivières du

Trinidad. Cette approche a dévoilée des différences importantes et répétées entre les populations guppy de différentes rivières et leur capacité de restreindre les infections

Gyrodactylus, sans toutefois permettre de constater de différences entre les environnements de prédation. Néanmoins, une diminution constante de la croissance des guppys exposés à Gyrodactylus était évidente, potentiellement en raison des coûts

énergétiques associés au déclenchement et à la maintenance d'une réponse immunitaire pour combattre les infections Gyrodactylus. Ces résultats suggèrent une divergence entre les trajectoires coévolutionnaires guppy-Gyrodactylus entre rivières (à savoir une plus grande résistance et une virulence plus élevée versus une faible résistance et un niveau de virulence inférieur), et mettent en évidence que l’histoire évolutive des populations

(c’est-à-dire la composition génétique, les effets fondateurs, les goulots d'étranglement génétiques, etc.), et non la prédation, influencent considérablement l'intensité et la direction de la coévolution hôte-parasite de ce système. Enfin, je considère comment le

Gyrodactylus peut modifier l'interaction entre le guppy et son ennemi naturel, Rivulus hartii. Bien que Gyrodactylus puisse fortement réduire la croissance guppy, j’ai trouvé que ce dernier ne modifie pas de manière significative les interactions guppy-Rivulus, notamment en raison du fait que les guppys peuvent affiner leur phénotype lorsqu’ils doivent simultanément affronter plusieurs ennemis. Ces résultats démontrent que les hôtes détiennent une capacité de médiation, et même d'atténuer les effets écologiques potentiels que les parasites pourraient avoir au sein de la communauté plus large.

viii ! La compréhension des mécanismes par lesquels les parasites peuvent influencer les espèces non hôtes est devenu un objectif capital dans le domaine de l’écologie évolutive.

Cependant, la complexité des communautés naturelles, l'hétérogénéité de l'environnement, et l’histoire évolutive de l’hôte peuvent chacun exercer des effets importants et parfois imprévus sur les interactions hôte-parasite. Les résultats de la présente thèse représentent une illustration frappante d'une telle communauté naturelle.

En déployant une combinaison de travail d’expérimentation et de manipulation sur le terrain je démontre que Gyrodactylus peut avoir des effets importants sur les traits d'histoire de vie des guppy, mais que ces effets n’influencent pas nécessairement les espèces non-hôtes, notamment en raison de la capacité de l’hôte de limiter ces effets.

Attendu qu’aucune étude actuelle ne propose une analyse de l’impact de la coévolution guppy-Gyrodactylus sur la faune coexistante, je crois que ce travail a posé les bases pour de futures recherches portant sur à la pertinence écologique de Gyrodactylus.

ix ! TABLE OF CONTENTS

THESIS ABSTRACT ...... iv

RÉSUMÉ DE LA THÈSE ...... vii

TABLE OF CONTENTS ...... x

LIST OF FIGURES ...... xiii

LIST OF TABLES ...... xv

ACKNOWLEDGEMENTS ...... xvii

PREFACE ...... xxi

THESIS FORMAT ...... xxii STATEMENT OF AUTHORSHIP ...... xxiii STATEMENT OF ORIGINALITY ...... xxiv

CHAPTER 1. GENERAL INTRODUCTION ...... 1

REFERENCES ...... 13

CHAPTER 2. ON THE ECOLOGICAL AND EVOLUTIONARY RELEVANCE OF SPECIALIST PARASITES:

LESSONS FROM GYRODACTYLUS ECTOPARASITES OF THE TRINIDADIAN GUPPY ...... 21

ABSTRACT ...... 22 INTRODUCTION ...... 23 Gyrodactylus ...... 25 Guppies ...... 26 Specialist parasites and mate choice ...... 27 Specialist parasites and predator-prey interactions ...... 29 Competition mediated by specialist parasites ...... 32 Gyrodactylus evolutionary implications ...... 34 General conclusions ...... 35 REFERENCES ...... 36

CONNECTING STATEMENT NO. 1 ...... 47

CHAPTER 3. ARE HOST–PARASITE INTERACTIONS INFLUENCED BY ADAPTATION TO PREDATORS?

A TEST WITH GUPPIES AND GYRODACTYLUS IN EXPERIMENTAL STREAM CHANNELS ...... 48

ABSTRACT ...... 49 INTRODUCTION ...... 50 Guppies, predators, and parasites ...... 51 Our experiment ...... 54 MATERIALS AND METHODS ...... 54

x ! Experimental design ...... 54 Guppy populations and collections ...... 56 Experimental infection ...... 58 Data evaluation and statistical analysis ...... 60 RESULTS ...... 63 Infection dynamics ...... 63 Guppy responses ...... 64 DISCUSSION ...... 65 Impact of predation regime on infection dynamics ...... 66 Guppy traits ...... 68 The way forward ...... 70 Conclusions ...... 71 ACKNOWLEDGMENTS ...... 72 REFERENCES ...... 73 TABLES ...... 81 FIGURES ...... 83

CONNECTING STATEMENT NO. 2 ...... 87

CHAPTER 4. TESTING FOR HOST-PARASITE LOCAL ADAPTATION: AN EXPERIMENT WITH

GYRODACTYLUS ECTOPARASITES AND GUPPY HOSTS ...... 88

ABSTRACT ...... 89 INTRODUCTION ...... 90 MATERIALS AND METHODS ...... 94 Fish collection and treatment ...... 94 Mesocosms ...... 95 Experimental design ...... 96 Experimental protocol ...... 96 Statistical analysis ...... 98 Models of Gyrodactylus performance ...... 98 Models of guppy performance ...... 100 RESULTS ...... 100 Gyrodactylus performance ...... 100 Guppy performance ...... 102 DISCUSSION ...... 102 ACKNOWLEDGMENTS ...... 109 REFERENCES ...... 110 FIGURES ...... 119 TABLES ...... 122

CONNECTING STATEMENT NO. 3 ...... 126

CHAPTER 5. AN EXPERIMENTAL TEST OF ANTAGONISTIC EFFECTS OF COMPETITION AND

PARASITISM ON HOST PERFORMANCE IN SEMI-NATURAL MESOCOSMS ...... 127

ABSTRACT ...... 128 INTRODUCTION ...... 129 Empirical system ...... 130

xi ! MATERIALS AND METHODS ...... 133 Fish collection and treatment ...... 133 The experiment ...... 134 Statistical analysis ...... 135 RESULTS ...... 136 DISCUSSION ...... 137 REFERENCES ...... 143 TABLES ...... 150 FIGURES ...... 153

CONNECTING STATEMENT NO. 4 ...... 154

CHAPTER 6. A DIRECT ASSESSMENT OF THE ECOLOGICAL IMPORTANCE OF GYRODACTYLUS

ECTOPARASITES IN TWO TRINIDADIAN STREAMS ...... 155

ABSTRACT ...... 156 INTRODUCTION ...... 158 Study system ...... 159 MATERIALS AND METHODS ...... 161 Experimental design ...... 161 Experimental protocol ...... 162 Statistical analyses ...... 164 RESULTS ...... 166 Population level effects ...... 167 Individual level effects ...... 168 Rivulus population and individual-level effects ...... 169 DISCUSSION ...... 169 Population-level effects ...... 170 Individual-level effects ...... 172 CONCLUSIONS ...... 174 ACKNOWLEDGMENTS ...... 175 REFERENCES ...... 176 FIGURES ...... 183 TABLES ...... 186

CHAPTER 7. GENERAL DISCUSSION ...... 189!

MAJOR CONTRIBUTIONS ...... 195 IMPLICATIONS ...... 197 LIMITATIONS ...... 198 Phenotypic plasticity ...... 198 Gyrodactylus host range ...... 200 Cryptic species of Gyrodactylus ...... 202 REFERENCES ...... 203

xii ! LIST OF FIGURES

CHAPTER 1

Figure 1. Gyrodactylus life cycle. A) birth of first daughter (asexually developed), b) first-born daughter has a developing embryo in utero, c) maturation of the male reproductive system, d) second born daughter can develop either sexually or parthenogenetically. Adapted from Kearn (1994) with permission from publisher. © Elsevier...... 7!

CHAPTER 3

Figure 1. Geographical location of Arima tributary mesocosms used for the experiment and the eight sites from which guppies were collected: mesocosms (star), Marianne high-predation (A), El Cedro high-predation (B), Aripo high-predation (C), Quare high-predation (D), Marianne low-predation (E), El Cedro low-predation (F), Aripo low-predation (G) and Quare low-predation (H)...... 83

Figure 2. Epidemic dynamics in El Cedro and Marianne high-predation guppies in experiment 1 and experiment 2.a) Parasite prevalence in Marianne high-predation guppies, b) Gyrodactylus population dynamics in Marianne high-predation guppies, c) Parasite prevalence in El Cedro high-predation guppies, and d) Gyrodactylus population dynamics in El Cedro high-predation guppies...... 84

Figure 3. Gyrodactylus population dynamics in the eight experimental guppy populations from four rivers used in experiment 2: a) Aripo, b) El Cedro, c) Marianne and d) Quare. Parasite dynamics on high-predation guppies are represented by solid lines, dashed line for low-predation guppies...... 85

Figure 4. Differences in phenotypic traits between females from infected and control channels, and high-predation and low-predation experimental populations from experiment 2: a) growth. b) reproductive effort (proportion of female growth due to embryonic fresh mass ), c) embryo mass, and d) number of embryos. Only those populations in which the infection did establish and their respective controls were analyzed. Dashed line-open triangle, control treatment; solid line-closed square, infected treatment. Error bars represent ± 1 standard error...... 86

CHAPTER 4

Figure 1. A schematic of the experimental design and Gyrodactylus population dynamics when infectin each guppy population. Horizontal labels indicate the source of guppies, vertical labels indicate the source of the parasites used in the experiment.

xiii ! White squares represent allopatric combinations (guppies and Gyrodactylus from different locations) and gray squares represent sympatric combinations (guppies and Gyrodactylus from the same locations)...... 119

Figure 2. Gyrodactylus population dynamics on the four guppy host populations. Each figure represents the mean number of parasites on individual fish. Symbols represent guppy populations: Marianne HP (filled squares), Marianne LP (empty squares), Aripo HP (filled circles) and Aripo LP (empty circles). Error bars represent standard errors...... 120

Figure 3. LS Means for Gyrodactylus performance. a) mean number of parasites / fish / day when infecting the sympatric host vs. all allopatric hosts, b) mean duration of infection on sympatric vs all allopatric hosts c) mean number of parasites / fish / day according to the degree of similarity between the parasite strain and the guppy strain in the mesocosms. Marianne HP Gyro (filled squares), Marianne LP Gyro (empty squares), Aripo HP Gyro (filled circles) and Aripo LP Gyro (empty circles). Error bars represent standard errors...... 121

Figure 4. LS Means for guppy performance when infected with sympatric vs. allopatric hosts. A) female guppy growth during the length of the experiment, b) number of embryos per female. Symbols represent guppy populations: Marianne HP (filled squares), Marianne LP (empty squares), Aripo HP (filled circles) and Aripo LP (empty circles). Error bars represent standard errors...... 122

CHAPTER 5

Figure 1. Least Square means in growth for a) female and male guppies, and b) Rivulus pooled across genders. The five treatments as shown are: GO, guppy-only; GG, guppy-Gyrodactylus; GR, guppy-Rivulus; GGR, guppy-Gyrodactylus-Rivulus; RO, Rivulus-only. Error bars represent standard errors...... 153

CHAPTER 6

Figure 1. Normal approximations for 95% binomial confidence intervals on recapture probability using the Wilson Score Interval……………………………………...184

Figure 2. . LS Means for guppy density (a) and biomass (b). Error bars represent standard errors. Symbols represent Paria Control (), Paria Experimental (), Marianne Control (),Marianne experimental ()...... 184

Figure 3. LS Means for guppy weight and growth. Error bars represent standard errors. Symbols represent Paria Control (), Paria Experimental (), Marianne Control ( ), Marianne experimental ()...... 185

xiv ! CHAPTER 7

Figure 1. Schematic of this thesis. Chapters 3 and 4 explore how predation shape guppy- Gyrodactylus adaptation (dark blue), chapters 5 and 6 explore how guppy- Gyrodactylus interactions are influenced by Rivulus, and how Rivulus is influenced by guppy-Gyrodactylus interactions (light blue)...... 191! !

LIST OF TABLES

CHAPTER 3!

Table 1. Characterization of Gyrodactylus infection dynamics in the eight experimental stream channels from experiment 2...... 81

Table 2. Statistical analyses of female life-history traits in experiment 2. All traits were analyzed using general linear models with “infection” and “predation” as fixed effects. “River” was entered as a blocking factor. “Initial mass” and “Embryo development” were removed from female growth analysis due to non-significance...... 82

CHAPTER 4

Table 1. Statistical analysis for Gyrodactylus performance on sympatric vs. allopatric guppies populations. Analyses were performed using linear mixed effects model. P values and denominator degrees of freedom were obtained using Satterthwaite approximation for degrees of freedom...... 123

Table 2. Statistical analysis for Gyrodactylus performance on guppy populations, according to their degree of similarity (Same drainage and same predation environment, same drainage and different predation environment, different drainage and same predation environment, different drainage and different predation environment). Analysis were performed using general linear model...... 124! Table 3. Statistical analysis for guppy performance when infected with sympatric vs. allopatric Gyrodactylus. Analyses were performed using linear mixed effects models. P values and denominator degrees of freedom were obtained using Satterthwaite approximation for degrees of freedom...... 125

CHAPTER 5

Table 1. Descriptive statistics. Treatment abbreviations are as follows: GO, guppy-only; GG, guppy-Gyrodactylus; GR, guppy-Rivulus; GGR, guppy-Gyrodactylus-Rivulus; RO, Rivulus-only. Mixed species treatments consisted of 4 guppy females, 2 guppy

xv ! males and 5 Rivulus. The number of fish was doubled in the single species treatments. Initial and final parasite prevalence (percentage of infected individuals in the mesocosm) and abundance (total number of Gyrodactylus per mesocosms) were determined on days 1 and 20 of the experiment, respectively...... 150

Table 2. Statistical analysis for guppy growth across the different treatments (GO, guppy- only; GG, guppy-Gyrodactylus; GGR, guppy-Gyrodactylus-Rivulus; GR, guppy- Rivulus; RO, Rivulus-only). Analyses were performed using linear mixed effects models with replicate nested in treatment...... 151

Table 3. Specific Tukey’s HSD Contrasts for female and male guppies. GO, guppy-only; GG, guppy-Gyrodactylus; GGR, guppy-Gyrodactylus-Rivulus; GR, guppy-Rivulus; RO, Rivulus-only...... 152!

CHAPTER 6

Table 1. Descriptive statistics of collection parameters during the pre- and post- introduction stages in the Gyrodactylus and control treatments...... 186

Table 2. Analyses for density mediated effects of Gyrodactylus introduction. Density and biomass was calculated for each pool during the mark and the recapture collections. Significant p values highlighted in bold...... 187

Table 3. Analyses on changes in mean trait values of guppies using general linear mixed effects models...... 188!

xvi !

ACKNOWLEDGEMENTS

Marxists are more right than wrong when they argue that the problems scientists take up, the way they go about solving them, and even the solutions they are inclined to accept, are conditioned by the intellectual, social, and economic environments in which they live and work

– THEODOSIUS DOBZHANSKY, In Mankind Evolving, 1962

xvii ! I couldn’t be more thankful for all the support and friendship I have received throughout the amazing adventure that has been my Ph. D. I especially thank Marilyn Scott, Gregor

Fussmann and Andrew Hendry for giving me the opportunity to participate in such an exciting project as the NSERC-SRO “Ecology and evolution of parasite-host relationships in a real ecosystem”. I have been very fortunate to have them as mentors.

Each has a particular way of approaching science, and this has been reflected in my own intellectual development and ideas. I thank you for believing in me, and being so supportive and patient during all these years.

I thank David Marcogliese, who gave me countless suggestions as part of my supervisory committee. He always had great questions that challenged the boundaries of my understanding and his input greatly clarified my own ideas – although the debate between tolerance and resistance was always out of the question. I would also like to thank David Reznick, who kindly facilitated the use of experimental facilities in Trinidad, where I did a big portion of my experiments. His expertise and solidarity made my life easier during much of the experimental work.

During every single field season I had I encountered trials and tribulations.

Although many times they came in the shape of venomous snakes, in many others they did not. I thank Alfredo Marquez, Maryse Boisjolie, Christianne Aikins and Beth Turner, without who’s extraordinary help I would’ve never been able to successfully finish all the work. Alfredo joined me for my first season – which was the one where we had most

xviii ! drawbacks. I would like to especially thank him, for he dealt with my increasing frustration and my changing temper. Gracias Alfredo.

At McGill I met many people that enriched my doctoral experience in one way or another. I am happy that now I can call them my friends. Although this list is hopelessly incomplete, I would like to thank Kiyoko Gotanda, Ben Haller, Maryse Boisjolie, Gregor

Rolshausen, Victor Frankel, Beto Prado, Etienne Low-Decarie and Dieta Hansen, for they provided deep scientific discussion, as well as long and late conversations about the mysteries of life. I would also like to thank the members of the Hendry, Fussmann and

Scott labs, as they provided a friendly and intellectually stimulating environment. I feel very lucky for having you as colleagues.

During my Ph.D., I spent almost 16 months doing fieldwork on the beautiful island of Trinidad. I met many dedicated scientists – both young and old – whose passion for ecology and evolution motivated me to do my best. I would like to thank Brad

Lamphere, who took me under his wing and showed how to mark fish, as well as the great beauty of fishing at night in the tortuous streams of Trinidad. I also thank Doug

Fraser, Jim Gilliam, Andrés López-Sepulcre and Ronnie Hernandez, as they provided valuable support with logistics in the field. To Felipe Dargent, who has been a great lab mate, but also took care of me while I was agonizing with Dengue fever. And to Jack

Torresdal, Will Roberts and Jonathas Pereira with whom I explored some of the most remote parts of the island, and am now great friends.

xix ! I also thank Shirley Mongeau, Caroline LeBlond, Susan Bocti, and Susan Gabe for their tireless efforts on my behalf.

My parents, Victoria and Felipe Pérez, have always been there for me to give me advice, and have tirelessly cheered me to pursue my aspirations. I wouldn’t be here if it wasn’t for them. You are the best parents a son could ask for. I also thank my sister

Karina Pérez-Jvostova, who was my roommate for several years while she was studying in McGill. She made Montreal’s cold winters much warmer and pleasant. Although I dearly miss her, I’m very proud to see how much she has accomplished at her young age.

Finally, I would like to thank my wife Sabrina Vigneux. Sabrina has been an invaluable part of my life for the last eight years. She has always made me smile and be positive, even if she was thousands of miles away. I wouldn’t be here if it wasn’t for her. Thank you.

xx !

PREFACE

It may be conceit, but I believe the subject will interest the public, and I am sure that the views are original

–! CHARLES R. DARWIN, Letter to his publisher, John Murray, 1859

xxi ! THESIS FORMAT

The current thesis is conformed of a compilation of manuscripts that are currently published, under review or being prepared for submission to a peer-reviewed journal.

Each of these manuscripts has been adapted to be included in this thesis as an independent chapter, according to the Thesis Specifications and the Thesis Formatting of

McGill University.

The corresponding bibliographic information for the manuscripts or the scientific journal they were prepared for is as follows:

Chapter 2. On the ecological and evolutionary relevance of specialist parasites: lessons from Gyrodactylus ectoparasites of the Trinidadian guppy. Pérez-Jvostov, F. and Scott,

M. E. TO BE SUBMITTED. Trends in Parasitology.

Chapter 3. Are host–parasite interactions influenced by adaptation to predators? A test with guppies and Gyrodactylus in experimental stream channels. Pérez-Jvostov, F.,

Hendry, A. P., Fussmann, G. F. and Scott, M. E. 2012. Oecologia, 170: 77-88. DOI:

10.1007/s00442-012-2289-9

Chapter 4. Testing for host-parasite local adaptation: an experiment with Gyrodactylus ectoparasites and guppy hosts. Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and

Scott, M. E. Int J Parasitol, 45: 409-417.

xxii ! Chapter 5. An experimental test of antagonistic effects of competition and parasitism on host performance in semi-natural mesocosms. Pérez-Jvostov, F., Hendry, A. P.,

Fussmann, G. F. and Scott, M. E. CONDITIONAL ACCEPTANCE, Oikos

Chapter 6. A direct assessment of the ecological importance of Gyrodactylus ectoparasites in Trinidadian streams. Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and Scott, M. E. TO BE SUBMITTED.

These chapters represent the core of my doctoral research. In addition to them, I have also conducted and participated in a number of other research projects that are not part of this thesis but are currently being prepared for publication.

STATEMENT OF AUTHORSHIP

With the guidance of Gregor Fussmann, Andrew Hendry and Marilyn Scott, I designed and conducted the experiments and fieldwork, analyzed the data and wrote the text contained herein. Although I am the principal author of all of the manuscripts included in this thesis, each of the co-authors assisted with the research, and provided extensive comments and edits to the various drafts of the manuscripts.

All co-authors have given permission to integrate the co-authored manuscripts into the current thesis.

xxiii ! STATEMENT OF ORIGINALITY

The manuscript chapters of this thesis represent distinct contributions to scientific knowledge. Chapter 2 is a review based on the work of others that serves as a literature review for the thesis. Chapters 3 to 6 are based on my own work in Trinidad. The four experimental chapters expand on previous work on guppy-Gyrodatylus interactions in multiple ways. First, all the experiments were performed in mesocosms that closely replicate natural streams, or directly in the streams themselves. This approach circumvents problems generally associated with laboratory experiments in which extrapolation to natural scenarios is troublesome. Second, the incorporation of guppy and

Gyrodactylus from multiple origins (e.g., rivers and predation environments) provides a broader understanding of the importance of evolutionary history, and further demonstrates that coevolutionary trajectories have greatly diverged across the geographical range. Finally, the experimental transplantation of Gyrodactylus into previously Gyrodactylus-free guppy populations is the first long-term study investigating the ecological implications associated with Gyrodactylus range expansion, as well as the resulting guppy adaptations.

The original scientific advancements of my work are listed below.

•! Quantify phenotypic changes (e.g., life-history traits) in guppies from high- and low-

predation environments when exposed to Gyrodactylus.

•! Perform the first fully reciprocal cross infection experiment using Gyrodactylus and

wild guppy populations from different rivers and predation environments.

xxiv ! •! Experimentally test how Gyrodactylus can affect competition between two fish

species.

•! Perform the first experimental transplantation of Gyrodactylus into previously

Gyrodactylus-free guppy populations.

•! Investigate the long-term effects of such introductions on guppies and Rivulus’

ecology.

The relevant findings that have stemmed from my work are the following:

•! Guppy phenotypic responses to the presence of Gyrodactylus were similar to those

previously reported for predators. Despite this observed parallelism in life-history

adjustment to both sources of mortality (e.g., higher reproductive allocation), which

would make adaptation to one of these sources of mortality advantageous for

adaptation to the other, my work suggests that guppy adaptation to predation does

not influence adaptation to parasitism by Gyrodactylus. !

•! I provide evidence that guppy-Gyrodactylus coevolution has undergone divergent

trajectories in different rivers, such as higher resistance and higher virulence in the

Aripo, and lower resistance and lower virulence in the Marianne. This highlights the

importance of incorporating evolutionary history of both host and parasite

populations into future studies on host-parasite coevolution.

•! Despite a strong reduction in growth observed in guppies exposed to Gyrodactylus,

guppy-Rivulus interactions were not significantly modified by the presence of the

xxv ! ectoparasite. While the reduction in growth in the presence of Gyrodactylus might free resources for the activation of the immune response, it could be disadvantageous in the presence of Rivulus, as large Rivulus can predate on small-size guppies. An intermediate growth rate would allow guppies to escape predation by the small- gaped Rivulus, while mounting an immune response. My results suggest that guppies are indeed capable of such fine-tuning in their phenotype in the presence of multiple enemies, and such response mitigates any potential effect that Gyrodactylus could have on guppy-Rivulus interactions.!

!

xxvi !

CHAPTER 1

General introduction

One never notices what has been done; one can only see what remains to be done

– MARIE CURIE Letter to her brother, 1894

1!! Common wisdom says that parasites have negative effects on their host fitness. These effects can be obvious – like pathology and mortality – but they can also be more cryptic

– like changes in host physiology and behaviour. This ecological notion of parasites, however, is relatively recent. While for a long time parasitologists mainly focused on describing and classifying the vast parasitic diversity, parasites were essentially neglected in ecological and evolutionary research, particularly because they are small, short-lived, and are commonly aggregated in, or on, only a few individuals (Poulin 2006).

Nonetheless, in the late 70’s and early 80’s a stream of research on the population biology of infectious diseases emerged as a new field within parasitology. Pioneered by

Andrew Dobson, Robert May, Robert Anderson and Peter Price, this field flourished, and the importance of parasites on the ecology and evolution of their host was slowly acknowledged outside parasitology. Empirical studies that demonstrated that parasites were capable of regulating host populations became the flagships of this transition

(Dobson et al., 1992; Hudson et al., 1998), but it wasn’t until late 1990’s, and early

2000’s, that the role of parasites in community and ecosystem ecology became a major research topic (Bush et al., 1997; Torchin et al., 2003; Hatcher et al., 2006; Hudson et al.,

2006; Lafferty et al., 2008; Dobson et al., 2008). Despite this increased attention to the ecological relevance of parasites, some of the most basic questions, such as how can parasites influence host evolution and, simultaneously, influence species coexistence, remain largely unanswered for all but a handful of systems. My goal in this thesis is to bring light to these questions by taking advantage of a model organism that has been extensively studied in evolutionary biology, and for which much is known about it’s ecology in its native geographical range.

2!! The guppy (Poecilia reticulata Peters, 1859) is a small, live-bearing fish that inhabits the majority of the streams of the Caribbean island of Trinidad (Guilliam et al.,

1993; Magurran and Phillips, 2001). Along these streams are several types of fish communities. In their most upstream range (i.e., low-predation environments or LP), guppies coexist with only one other fish species, the killifish (Rivulus hartii), which is predominantly a competitor, but has also been reported to consume smaller guppies

(Endler, 1980; Haskins et al., 1961). In their most downstream range (i.e., high-predation environments or HP), guppies coexist with numerous large piscivourous fish, including the pike cichlid (Crenichla alta), the blue acara (Aequidens pulcher), the two-spotted sardine (Astyanax bimaculatus), and the wolf fish (Hoplias malabaricus) (Haskins et al.

1961, Reznick and Endler, 1982, Gilliam et al., 1993). These differences in predatory fauna and their potential implications for guppy evolution were first documented by

Caryl Haskins (Haskins et al., 1961), and subsequently developed by John Endler and

David Reznick.

In the first comparative analysis of wild-caught guppies from several streams,

Endler and Reznick showed that guppy life-history traits greatly vary in dependence of the predatory environment (Reznick and Endler, 1982), and that these differences are largely genetically determined (Reznick, 1982). For example, HP guppies tend to be smaller (Haskins et al., 1961; Endler, 1983), grow and mature faster (Reznick and Endler,

1982), and have more and smaller embryos than LP fish(Reznick and Endler, 1983;

Reznick et al., 1996). In addition, the particular morphology of the streams in Trinidad, where large waterfalls that prevent upstream migration of fish separate HP and LP

3!! environments, has provided an ideal setup for experimental manipulations to study guppy adaptation to divergent predation environments. Introduction experiments in which guppies were released from predation by transplanting them from a HP locality to previously guppy and predator-free stream showed that not only does the LP phenotype can evolve as a consequence of the change in fish community (i.e., predation) (Bryga and

Reznick, 1987; Reznick and Bryga, 1996; Gordon et al., 2009), but that this change has happened independently in multiple watersheds (Alexander et al., 2006); and can do so in ecological timescales (Reznick et al., 1990, Bryga and Reznick, 1997, Ghalambor et al.,

2004).

The evolution of several other traits seems to also be influenced by predation. For example, bright LP males evolve lower coloration when the visual predator C. alta was introduced into mesocosms with guppies from the Aripo river (Endler, 1980). Similarly, guppy morphology (Liley and Seghers, 1975), courtship (Magurran and Seghers, 1990;

Houde, 1997) and schooling behaviours (Seghers, 1974; Magurran and Seghers, 1991), performance traits (Ghalambor et al., 2004), metabolic rates (Handelsman et al., 2013), and even dietary preferences (Zandonà et al., 2011) have been documented to co-vary with predation environment.

Much of the initial research interest on guppies gravitated around predation as the sole driver of guppy life-history evolution. However, recently, it has become evident that predation is not the only one. Both primary productivity (Reznick et al., 2001) and guppy density (Reznick et al., 2001; Reznick and Bryant, 2007; Bassar et al., 2012; Reznick et

4!! al., 2012; Bassar et al., 2013) have been shown to play an important role in guppy evolution through their effects on resource availability and intraspecific competition. This line of evidence suggests that the observed adaptive divergence between HP and LP guppies is the result of a complex interaction between predator abundance (Reznick and

Endler, 1982), size specific mortality rates (Reznick et al., 1996) and strong density- dependent selection in the LP environments (Reznick et al., 2012; Bassar et al., 2013).

Other factors, could also influence guppy adaptation to their local environment, but these have received comparatively little attention. One example is the specialist monogenean ectoparasites, Gyrodactylus: von Nordmann, 1832.

In Trinidad, two species of Gyrodactylus infect guppies: Gyrodactylus turnbulli and

G. bullatarudis (Harris and Lyles 1992). Laboratory and field studies have demonstrated that both species can greatly reduce guppy survival (Scott and Anderson, 1984; Cable and van Oosterhout, 2007a; Cable and van Oosterhout, 2007b; van Oosterhout et al., 2007), but can also influence guppy behaviours such as mating (Lopez, 1998; Lopez, 1999;

Houde and Torio, 1991) and feeding (van Oosterhout et al., 2003; Kolluru et al., 2006).

Gyrodactylus have grasped the interest of naturalists for almost two centuries (Kearn,

1994; Poulin, 2002), far before their importance in guppy ecology was known. The first studies dealt with the evolutionary origins of viviparity (von Siebold, 1849) and embryonic cell development (Wegner, 1861). Later work, however, had more direct applications. The devastating economic costs of G. salaris epidemics on wild and farmed

Atlantic salmon in Norway urged for extensive and thorough research on the evolution of host resistance, parasite range expansion and host switch (Johnsen, 1978; Johnsen and

5!! Jensen, 1986; Bakke et al., 1991; Bakke et al., 1993). As a result, we now understand in great detail the biology and epidemiology of this genus (for an in depth review of

Gyrodactylus biology see Kearn, 1994; Bakke et al., 2002; Poulin et al., 2002; Harris et al., 2004; Bakke et al., 2007).

Most Gyrodactylus species are of similar morphology and all share a similar direct life cycle, feeding and reproducing directly on the integument of their fish host

(Kearn, 1979). Gyrodactylus are hyperviviparous; females give birth to a fully-grown daughter that in turn contains a developing embryo in utero (Cohen, 1977). They also show extreme progenesis, as they reproduce before either the female and male reproductive organs are fully developed. The female reproductive system becomes active before the male (Cable and Harris, 2002) (Fig. 1). This particular reproductive strategy results in the first-born daughter developing asexually, before the male reproductive system has matured, while subsequent daughters can develop either sexually or parthenogenetically (Kearn, 1994; Cable and Harris, 2002). Despite the presence of reproductive organs, the vast majority of Gyrodactylus reproduction is asexual (Harris et al., 1994; Kearn, 1994; Cable and Harris, 2002; (Schelckle et al., 2012).

6!! a

c

b

d

Figure 1. Gyrodactylus life cycle. A) birth of first daughter (asexually developed), b) first-born daughter has a developing embryo in utero, c) maturation of the male reproductive system, d) second born daughter can develop either sexually or parthenogenetically. Adapted from Kearn (1994) with permission from publisher. ©

Elsevier.

7!! In addition to their formidable life cycle, Gyrodactylus have been thoroughly studied due to their high host specificity. Although their morphological similarities have made their taxonomic classification quite troublesome, with over 400 described species from only ~200 hundred host taxa, they are inarguably some of the most specialized parasites (Poulin, 2002). Gyrodactylus thus provide an ideal opportunity to study how specialist parasites can mediate interspecific interactions. This is certainly important, as our knowledge on how specialist parasites can shape the ecology of biological communities has been overshadowed by research on generalist parasites and parasites with complex life cycles. This extensive gap in our knowledge is the motivation behind chapter 2, where I use the Gyrodactylus-guppy system to provide the empirical examples of how specialist parasites can influence non-host species at ecological and evolutionary time scales, and I differentiate them from what we currently know for generalist parasites.

If Gyrodactylus infections can have strong effects in guppy survival and behaviour, how do these effects interact with other sources of mortality such as predation? It is evident that the associated fitness costs of parasitic infections will greatly depend on whether or not infected individuals suffer higher mortality than healthy ones, and whether this mortality is the result of increased predation on infected individuals. Several examples of such parasite-induced predation exist in other systems. In their classic paper, Hudson et al. (1992) demonstrated that redgrouse (Lagopus lagopus scoticus) killed by predators had higher burdens of the nematode Trichostrongylus tenuis, and predator control lead to an increase in the numbers of heavily infected birds, indicating

8!! that predators selectively prey on infected birds (Hudson et al. 1992). Similarly, in a direct test for parasite-induced predation in field populations of voles (Microtus townsendii), voles treated with anti-helminthics suffered 17% less predation than infected individuals (Steen et al. 2002). Despite the vast empirical evidence for parasite-induced predation, to date there has been no direct evidence for increased mortality of

Gyrodactylus-infected guppies in high-predation environments. Nonetheless, a handful of studies have suggested that, indeed, HP guppies can limit the length and the intensity of

Gyrodactylus infections better than their LP counterparts (van Oosterhout et al., 2003;

Cable and van Oosterhout, 2007). This, however, is far from being a universal observation. The previous studies compared only one pair of HP and LP guppy populations from only one river, the Aripo – making generalizations difficult. For example, guppy populations from different rivers are genetically diverged, in part because they originated from independent colonization events (Alexander et al., 2006;

Crispo et al., 2006). In addition to these founder effects, different guppy evolutionary histories (i.e., selection, genetic bottlenecks and drift) could influence guppy-

Gyrodactylus coevolution in the various rivers in different ways. The predatory fauna in rivers from the northern and southern slopes is also drastically distinct: visual predators of the Cichlidae family are only present in the southern drainages, whereas the main predators in the northern drainages belong to the Eleotridae, a family of generalist ichtyophagous predators (Magurran and Phillips, 2010). Such differences in community composition could further influence the relative fitness costs associated with

Gyrodactylus infections in the different rivers and, hence, guppy-Gyrodactylus coevolution.

9!!

Several important questions are pertinent. For example, do guppy adaptations to predation influence adaptation to parasitism? And, is this adaptation consistent across different environments? These two questions represent the basis of chapter 3 and 4, where I experimentally test if HP guppies can cope with Gyrodactylus infections better than LP guppies. These chapters expand on previous studies by incorporating sympatric guppy-Gyrodactylus combinations from multiple rivers, and by using mesocosms - channels that closely replicate the natural conditions of Trinidadian streams, it demonstrates that adaptations to one source of mortality (i.e., predation) do not necessary favour adaptation to another (i.e., Gyrodactylus).

Guppies, however, are not the only species adapting to the predation environment.

Higher Gyrodactylus’ infectivity and transmission (i.e., virulence) are also expected to evolve under increased guppy mortality (Anderson and May, 1982; Gandon and

Michalakis, 2002), and this of course would in turn select for higher resistance in HP guppies. Given the particular experimental design in chapter 3, I was unable to disentangle the confounding effects between highly resistant hosts and highly virulent parasites, with those from low-resistant hosts and low-virulence parasites, which restricts potential inferences on host-parasite adaptation. Thus, chapter 4 builds on the previous experiment to assess local adaptation of guppies and Gyrodactylus, and to determine if such adaptation is influenced by the environment (i.e., predation) or evolutionary history

(i.e., river). Using fully reciprocal cross infections I demonstrate that, contrary to expectations, Gyrodactylus coevolution is not deterministic, and could be influenced by

10! ! historical demographic and evolutionary processes. This is an important aspect of host- parasite interactions that is commonly neglected in studies of local adaptation, and chapter 4 demonstrates the need to incorporate them into future research on host-parasite coevolution.

Host and parasites do not coevolve in isolation. As they adapt to each other, they inevitably influence the evolutionary ecology of coexisting non-host species. The reduced complexity of fish communities in the most upstream reaches of Trinidadian streams further provides the oportunity to better understand the mechanisms by which this can occur. In these reaches, guppies coexist with only one other fish species: its competitor,

Rivulus hartii (Haskins et al., 1961; Magurran and Phillip, 2001). Moreover, surveys on the geographical distribution of Gyrodactylus have demonstrated, that, while

Gyrodactylus are very common across the guppy range, some populations free of

Gyrodactylus also exist (Harris and Lyles, 1992; Martin and Johnsen, 2007; Gotanda et al., 2013; Stephenson et al., 2014). To better understand how Gyrodactylus can mediate guppy interspecific interactions, in chapter 5 and 6 I test guppy-Rivulus competition in the presence and absence of the coevolved Gyrodactylus. Chapter 5 is my first- experimental attempt to disentangle the mechanisms by which a specialist parasite can modify host’s interspecific interactions. The results obtained in this chapter suggest that, individually, Rivulus and Gyrodactylus elicit antagonistic guppy phenotypic responses, but when in combination, these effects are reduced. I was also unable to detect any effects of Gyrodactylus on guppy-Rivulus competition – something that surprised me and continues to intrigue me until today. Of course, this study is by no means free of

11! ! limitations. By using Gyrodactylus, guppies and Rivulus that have coevolved in the wild,

I was unable to determine if such results are an artefact of low parasite virulence, or whether Gyrodactylus effects cannot be measured in such short temporal scales (i.e., 20 days). Despite these limitations, I provide the first step in understanding the ecological importance of Gyrodactylus in Trinidadian streams.

The final chapter of this thesis is a test of how Gyrodactylus can influence community dynamics in the wild. Here, I directly compared changes in community dynamics before and after introducing Gyrodactylus into two previously Gyrodactylus- free wild guppy populations. This chapter is based on two years of work performed directly in the natural streams, compared to the experimental nature of the previous chapters. The results strongly reflect those of the previous experiment, and they further demonstrate the complexity of host-parasite interactions even in the context of a relatively simple community, as the one found in the upper reaches of Trinidadian streams. I demonstrate that Gyrodactylus can greatly reduce guppy survival and induce strong phenotypic changes, yet these effects do not necessarily influence Rivulus ecology.

I suspect that guppies can greatly reduce the ecological relevance of Gyrodactylus, because they are ahead of the coevolutionary arms race. Nonetheless, much of the work that I present here is just but a glimpse of a continuous battle between a parasite and its host, and what seems to be trivial today, could be a major driver of ecological dynamics tomorrow.

12! ! REFERENCES

1.! Alexander, H.J., Taylor, J.S., Wu, S.S.T. and Breden, F. (2006) Parallel evolution

and vicariance in the guppy (Poecilia reticulata) over multiple spatial and temporal

scales. Evolution, 60: 2352-2369.

2.! Anderson, R.M., and May, R.M. (1982) Coevolution of hosts and parasites.

Parasitology, 85: 411-426.

3.! Bakke, T.A. and MacKenzie, K. (1993) Comparative susceptibility of native Scottish

and Norwegian stocks of Atlantic salmon, Salmo salar L., to Gyrodactylus salaris

Malmberg: laboratory experiments. Fisheries Research, 17: 69-85.

4.! Bakke, T.A., Cable, J., and Harris, P.D. (2007) The biology of gyrodactylid

monogeneans: the “Russian-doll killers”. Advances in Parasitology, 64: 161-460.

5.! Bakke, T.A., Jansen, P.A. and Kennedy, C.R. (1991) The host specificity of

Gyrodactylus salaris Malmberg (Platyhelminthes, Monogenea): susceptibility of

Oncovhynchus mykiss (Walbaum) under experimental conditions. Journal of Fish

Biology, 39: 45-57.

6.! Bassar, R.D., Ferriere, R., Lopez-Sepulcre, A., Marshall, M.C., Travis, J., Pringle, C.

M. and Reznick, D.N. (2012) Direct and indirect ecosystem effects of evolutionary

adaptation in the Trinidadian guppy (Poecilia reticulata). American Naturalist, 180:

167-185.

7.! Bassar, R.D., Lopez-Sepulcre, A., Reznick, D.N. and Travis, J. (2013) Experimental

evidence for density-dependent regulation and selection on Trinidadian guppy life

histories. American Naturalist, 181: 25-38.

13! ! 8.! Bush, A.O., Lafferty, K.D., Lotz, J.M. and Shostak, A.W. (1997) Parasitology meets

ecology on its own terms: Margolis et al. revisited. Journal of Parasitology, 575-583.

9.! Haskins, C.P., Haskins, E.F., McLaughlin, J.J.A., Hewitt, R.E. (1961) Polymorphism

and population structure in Lebistes reticulatus, a population study. In: Blair, W. F.

(ed.) Vertebrate speciation, University of Texas Press, Austin, pp. 320–395.

10.!Hudson, P. J., Newborn, D., and Dobson, A. P. (1992). Regulation and stability of a

free-living host-parasite system: Trichostrongylus tenuis in red grouse. I. Monitoring

and parasite reduction experiments. Journal of Animal Ecology, 61: 477-486.

11.!Cable, J. and Van Oosterhout, C. (2007) The impact of parasites on the life history

evolution of guppies (Poecilia reticulata): the effects of host size on parasite

virulence, International Journal for Parasitology, 37: 1449-1458.

12.!Cable, J. and Van Oosterhout, C. (2007) The role of innate and acquired resistance in

two natural populations of guppies (Poecilia reticulata) infected with the

ectoparasite Gyrodactylus turnbulli. Biological Journal of the Linnean Society, 90:

647-655.

13.!Cable, J., and Harris, P.D. (2002) Gyrodactylid developmental biology: historical

review, current status and future trends. International Journal for Parasitology, 32:

255-280.

14.!Cohen, J. (1977) Reproduction. Butterworth, London.

15.!Crispo, E., Bentzen, P., Reznick, D. N., Kinnison, M. T., and Hendry, A. P. (2006)

The relative influence of natural selection and geography on gene flow in guppies.

Molecular Ecology, 15: 49-62.

14! ! 16.!Dobson, A.P. and Hudson, P.J. (1992) Regulation and stability of a free-living host-

parasite system: Trichostrongylus tenuis in red grouse. II. Population models.

Journal of Animal Ecology, 61: 487-498.

17.!Dobson, A., Lafferty, K.D., Kuris, A.M., Hechinger, R.F. and Jetz, W. (2008)

Homage to Linnaeus: How many parasites? How many hosts? Proceedings of the

National Academy of Sciences, 105: 11482-11489.

18.!Endler, J.A. (1980) Natural selection on color patterns in Poecilia reticulata.

Evolution, 34: 76-91.

19.!Gandon, S. and Michalakis, Y. (2002) Local adaptation, evolutionary potential and

host–parasite coevolution: interactions between migration, mutation, population size

and generation time. Journal of Evolutionary Biology, 15: 451-462.

20.!Ghalambor, C.K., Reznick, D.N. and Walker, J.A. (2004) Constraints on adaptive

evolution: the functional trade-off between reproduction and fast-start swimming

performance in the Trinidadian guppy (Poecilia reticulata). American Naturalist,

164: 38-50.

21.!Gilliam, J.F., Fraser, D.F. and Alkins-Koo, M. (1993) Structure of a tropical stream

fish community: a role for biotic interactions. Ecology, 74:1856-1870.

22.!Gordon, S.P., Reznick, D.N., Kinnison, M.T., Bryant, M.J., Weese, D.J., Räsänen, K.

and Hendry, A.P. (2009) Adaptive changes in life history and survival following a

new guppy introduction. American Naturalist, 174: 34-45.

23.!Gotanda, K.M., Delaire, L.C., Raeymaekers, J.A.M., Pérez-Jvostov, F., Dargent, F.,

Bentzen, P., Scott, M.E., Fussmann, F.G. and Hendry, A.P. (2013) Adding parasites

to the guppy-predation story: insights from field surveys. Oecologia 172: 155-166.

15! ! 24.!Handelsman, C.A., Broder, E.D., Dalton, C.M., Ruell, E.W., Myrick, C.A., Reznick,

D.N. and Ghalambor, C.K. (2013) Predator-induced phenotypic plasticity in

metabolism and rate of growth: rapid adaptation to a novel environment. Integrative

and Comparative Biology, 53: 975-988.

25.!Harris, P.D. and Lyles, A.M. (1992) Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. Journal of

Parasitology, 78: 912-914.

26.!Harris, P.D., Jansen, P.A. and Bakke, T.A. (1994) The population age structure and

reproductive biology of Gyrodactylus salaris Malmberg (Monogenea). Parasitology,

108: 167-173.

27.!Hatcher, M.J., Dick, J.T. and Dunn, A.M. (2006) How parasites affect interactions

between competitors and predators. Ecology Letters, 9: 1253-1271.

28.!Houde, A.E. and Torio, A.J. (1991) Effect of parasitic infection on male color pattern

and female choice in guppies. Behavioral Ecology, 3: 346-351.

29.!Hudson, P.J., Dobson, A.P. and Lafferty, K.D. (2006) Is a healthy ecosystem one

that is rich in parasites? Trends in Ecology and Evolution, 2: 381-385.

30.!Hudson, P.J., Dobson, A.P. and Newborn, D. (1998) Prevention of population cycles

by parasite removal. Science, 282: 2256-2258.

31.!Johnsen, B.O. (1978) The effect of an attack by the parasite Gyrodactylus salaris on

the population of Salmon parr in the river Lakselva, Misvaer in northern Norway.

Journal of Arctic Biology, 11: 7-9.

32.!Johnsen, B.O., and Jensen, A.J. (1986) Infestations of Atlantic salmon, Salmo salar,

by Gyrodactylus salaris in Norwegian rivers. Journal of Fish Biology, 29: 233-241.

16! ! 33.!Kearn, G.C. (1994) Evolutionary expansion of the Monogenea. International Journal

for Parasitology, 24: 1227-1271.

34.!Kolluru, G.R., Grether, G.F., South, S.H., Dunlop, E., Cardinali, A., Liu, L. and

Carapiet, A. (2006) The effects of carotenoid and food availability on resistance to a

naturally occurring parasite (Gyrodactylus turnbulli) in guppies (Poecilia reticulata).

Biological Journal of the Linnean Society, 89: 301-309.

35.!Lafferty, K.D., Allesina, S., Arim, M., Briggs, C.J., De Leo, G., Dobson, A.P. and

Thieltges, D.W. (2008) Parasites in food webs: the ultimate missing links. Ecology

Letters, 11: 533-546.

36.!Lopez, S. (1998) Acquired resistance affects male sexual display and female choice

in guppies. Proceedings of the Royal Society of London. Series B: Biological

Sciences, 265: 717-723.

37.!Lopez, S. (1999) Parasitized female guppies do not prefer showy males. Animal

Behaviour, 57: 1129-1134.

38.!Magurran, A.E. and Phillip, D.A. (2001) Implications of species loss in freshwater

fish assemblages. Ecography, 24: 645-650.

39.!Magurran, A.E. and Seghers, B.H. (1990) Risk sensitive courtship in the guppy

(Poecilia reticulata). Behaviour, 112: 194-201.

40.!Magurran, A.E., and Seghers, B.H. (1991) Variation in schooling and aggression

amongst guppy (Poecilia reticulata) populations in Trinidad. Behaviour, 118: 214-

234.

17! ! 41.!Martin, C.H. and Johnsen, S. (2007) A field test of the Hamilton–Zuk hypothesis in

the Trinidadian guppy (Poecilia reticulata). Behavioral Ecology and Sociobiology,

61: 1897-1909.

42.!Poulin, R. (2002) The evolution of monogenean diversity. International Journal for

Parasitology, 32: 245-254.

43.!Poulin, R. (2006) Evolutionary ecology of parasites. Princeton University Press.

44.!Reznick, D.N. (1983) The structure of guppy life histories: the trade-off between

growth and reproduction. Ecology, 64: 862-873.

45.!Reznick, D.N. (1982) The impact of predation on life history evolution in

Trinidadian guppies: genetic basis of observational life history patterns. Evolution,

36: 1236-1250.

46.!Reznick, D.N and Bryant, M. (2007) Comparative long-term mark-recapture studies

of guppies (Poecilia reticulata): differences among high and low predation localities

in growth and survival. Annales Zoologici Fennici, 44: 152-160.

47.!Reznick, D.N. and Endler, J.A. (1982) The impact of predation on life history

evolution in Trinidadian guppies (Poecilia reticulata). Evolution, 36: 160-177.

48.!Reznick, D.N. (1990) Plasticity in age and size at maturity in male guppies (Poecilia

reticulata): an experimental evaluation of alternative models of development.

Journal of Evolutionary Biology, 3: 185-203.

49.!Reznick, D.N. and Bryga, H. (1987) Life-history evolution in guppies (Poecilia

reticulata): 1. Phenotypic and genetic changes in an introduction experiment.

Evolution, 41: 1370-1385.

18! ! 50.!Reznick, D.N. and Bryga, H. A. (1996) Life-history evolution in guppies (Poecilia

reticulata: Poeciliidae). V. Genetic basis of parallelism in life histories. American

Naturalist, 147: 339-359.

51.!Reznick, D.N., Bassar, R.D., Travis, J., and Helen Rodd, F. (2012) Life-history

evolution in guppies VIII: the demographics of density regulation in guppies

(Poecilia reticulata). Evolution, 66: 2903-2915.

52.!Reznick, D.N., Butler, M.J.IV, and Rodd, H. (2001) Life-history evolution in

guppies. VII. The comparative ecology of high-and low-predation environments.

American Naturalist, 157: 126-140.

53.!Schelkle, B., Faria, P.J., Johnson, M.B., van Oosterhout, C., and Cable, J. (2012)

Mixed infections and hybridisation in monogenean parasites. PloS One, 7: e39506.

54.!Scott, M.E. and Anderson, R.M. (1984) The population dynamics of Gyrodactylus

bullatarudis (Monogenea) within laboratory populations of the fish host Poecilia

reticulata. Parasitology, 89: 159-194.!

55.!Steen, H., Taitt, M., and Krebs, C. J. (2002). Risk of parasite-induced predation: an

experimental field study on Townsend's voles (Microtus townsendii). Canadian

Journal of Zoology, 80: 1286-1292.!

56.!Stephenson, J.F., van Oosterhout, C., Mohammed, R.S., and Cable, J. (2015)

Parasites of Trinidadian guppies: evidence for sex-and age-specific trait-mediated

indirect effects of predators. Ecology. 96: 489–498.

57.!Torchin, M.E., Lafferty, K.D., Dobson, A.P., McKenzie, V.J. and Kuris, A.M.

(2003) Introduced species and their missing parasites. Nature, 421: 628-630.

19! ! 58.!Van Oosterhout, C., Harris, P.D. and Cable, J. (2003) Marked variation in parasite

resistance between two wild populations of the Trinidadian guppy, Poecilia

reticulata (Pisces: Poeciliidae). Biological Journal of the Linnean Society, 79: 645-

651.

59.!Van Oosterhout, C., Mohammed, R.S., Hansen, H., Archard, G.A., McMullan, M.,

Weese, D.J. and Cable, J. (2007) Selection by parasites in spate conditions in wild

Trinidadian guppies (Poecilia reticulata). International Journal for Parasitology, 37:

805-812.

60.! Von Siebold, C.T. (1849) Gyrodactylus, ein ammenartiges Wesen. Z. wiss. Zool.1:

347- 363.

61.!Wagner, G. (1861) Translations: On Gyrodactylus elegans, Nordmann. Quarterly

Journal of Microscopical Science, 2: 196-212.

62.!Zandona, E., Auer, S.K., Kilham, S.S., Howard, J.L., López-Sepulcre, A.,

O’Connor, M.P., Bassar, R.D., Osorio, A., Pringle, C. and Reznick, D.N. (2011) Diet

quality and prey selectivity correlate with life histories and predation regime in

Trinidadian guppies. Functional Ecology, 25: 964-973.

20! !

CHAPTER 2

On the ecological and evolutionary relevance of specialist

parasites: lessons from Gyrodactylus ectoparasites of the

Trinidadian guppy

Pérez-Jvostov, F. and Scott, M. E.

Specialists are always weird – the weirdness of the specialist are just the tools of its trade.

–! COLIN TUDGE! In The Bird, 2008!

21! ! ABSTRACT

The importance of parasites in biological communities has long been acknowledged in ecological and evolutionary research, and the mechanisms by which parasites can mediate ecological interactions have been particularly well studied. Yet, our understanding of the ecological importance of parasite host specialization (i.e. host specificity) is still lagging behind. In the ecological context, the main difference between generalist and specialist parasites is their capacity for host population regulation: generalists can reduce host density of many species, whereas specialists can do so, directly, through only one such host. Within an evolutionary context, however, specialist parasites are more likely to undergo strong co-evolutionary arms races with their host, and in this way affect the coexisting community in longer time scales. Gyrodactylus ectoparasites of fishes provide an ideal model system to understand how specialist parasites can modulate ecological interactions. Gyrodactylus-host interactions have been particularly well studied when infecting their Trinidadian guppy host, Poecilia reticulata.

Although most of the initial research on guppy-Gyrodactylus associations focused on infection dynamics and parasite-mediated sexual selection, our understanding of this system is now sufficiently mature to explore the role of this specialist parasite in the biological community. We thus synthesize current research on guppies, their environment and Gyrodactylus interactions, and provide comparative evidence on the differences between the ecological implications of specialist and generalist parasites.

22! ! Introduction

Ever since the publication of Roy Anderson and Robert May’s seminal papers on the population biology of infectious diseases (Anderson and May 1978; May and Anderson

1978), parasites have been recognized as important drivers of ecological and evolutionary processes in biological communities (Dobson and Hudson, 1987; Minchella and Scott,

1991; Marcogliese and Cone, 1997; Hatcher et al., 2006; Pedersen and Fenton, 2007;

Raffel et al., 2008; Johnson et al., 2010), ecosystem functioning (Lafferty and Morrison,

1996; Hudson et al., 2006; Lafferty et al., 2006; Lefèvre et al., 2008), biological invasions (Prenter et al., 2004; Dunn 2009; Dunn et al., 2012), and even animal personality (Moore, 2002; Barber and Dingemanse, 2010; Poulin, 2010). This may come as no surprise given that parasites can have profound effects on their host’s phenotype, physiology, and behaviour, and these effects can then easily spread through the community (Marcogliese and Cone, 1997; Hatcher et al., 2006). Theoretical and empirical evidence of these effects is abundant for many different taxa, but the majority of these examples involve parasites that can infect multiple host species (i.e., generalist parasites). However, parasites with narrow host ranges are likely to be just as common in natural communities (Agosta et al., 2010), and have even stronger potential for coevolution with their host (Lajeunesse and Forbes, 2003). Thus, they may also be as strong determinants of evolutionary and ecological processes as are generalist parasites.

The ability of parasites to have strong ecological effects on host populations depends on various factors, including how compatible the parasite is to the hosts and how often the parasite encounters the host species (Combes 1997). Specialist parasites have

23! ! evolved adaptations to the few host species that they exploit (Sasal et al., 1999), but specialization to one or a few host species may result in trade-offs: reduced ability to infect and reach transmission stages in other host species (Futuyma and Moreno 1988,

Pfennig, 2001). To appreciate how specialist parasites can influence community dynamics, we first need to understand the differences between the effects of generalist and specialist parasites. The general mechanisms by which parasites can affect non-host species have been classified in two. The first category includes density-mediated effects, which involve the direct effects of parasitic infection on host health, reproduction and survival (Price et al., 1988; Abrams 1992; Abrams 1996). The second category includes trait-mediated effects, which involve changes in host behaviour and life history traits to reduce the fitness costs associated with infection, but behavioural and phenotypic changes derived from the infection itself also fall in this category (Holt and Kotler, 1987;

Abrams, 1992; Abrams, 1996) Hatcher et al., 2006). In the case of competitive interactions, for example, under these mechanisms generalist parasites are predicted to enhance coexistence if superior competitors are more heavily impacted by the parasite, but the opposite effect is also possible if inferior competitors are more adversely affected.

Specialist parasites, on the other hand, are predicted to reduce coexistence as they can only infect one host (Fenton and Brockhurst, 2008). Infections with generalist parasites can further result in apparent competition if two species that do not otherwise interact host the same parasite species (Holt and Pickering, 1985). Specialist parasites’ restricted host range precludes them from such effects. Evidently generalist parasites can influence community dynamics through differential effects on multiple hosts – which potentially differ in their resistance to infection – while specialist parasites can only do so if

24! ! infection has strong implications for their host survival, or induces strong behavioural and phenotypic changes (Kawecki, 1998). These differences have received very little attention in the literature, and although they are subtle, they can have important implications for how and to what extent a parasite can influence community dynamics.

In this review we explore the ecological importance of the specialist ectoparasite,

Gyrodactylus, in the ecology of Trinidadian streams, and how they can influence community dynamics through density and trait mediated effects on their host, the

Trinidadian guppy, Poecilia reticulata. We start with a brief introduction to Gyrodactylus and guppy ecology, as well as their associations and implications for sexual and natural selection. Next, we explore the evidence for modified predator-prey and competitive interactions, as well as their implication for the stability and dynamics of the broader community. We conclude with the evolutionary implications of guppy-Gyrodactylus coevolution, and propose future directions for work on this and similar systems.

Gyrodactylus

Gyrodactylid monogeneans are a genus of direct-life-cycle ectoparasites with high host- specificity. The genus is a well-recognized pathogen of fish with 409 species infecting over

200 host species across 19 orders (Poulin, 1992; Kearn, 1994; Bakke et al., 2007). Some

Gyrodactylus species show a narrow host range, with an average of four host species, but at least 60% of Gyrodactylus are strict specialists (Poulin, 1992; Harris et al., 2004).

Gyrodactylus are viviparous and give birth to a fully-grown daughter that in turn contains a developing embryo in utero (Cohen, 1977). Newborn Gyrodactylus are fully-grown and

25! ! directly attach to the host alongside with their parents. Gyrodactylus feed on host mucus and epithelial cells in a manner similar to other surface-browsing monogeneans (Kearn, 1999).

They can move along the host body, and “jump” onto new hosts using their opisthaptor and adhesive glands (Cable and Harris, 2002). In this way, Gyrodactylus can rapidly spread through the host population, and have strong effects on host demographics through infection- associated mortality (Scott and Anderson 1984). Such epidemic dynamics can have strong implications for the demography of the host. This has been particularly well studied in association with the aquaculture industry (especially the generalist G. salaris infecting salmon and trout) and in the laboratory (especially on guppies). The two species known to infect guppies, G. turnbulli and G. bullatarudis, occur in natural guppy populations in their native range in Trinidad (Harris and Lyles, 1992).

Guppies

Guppies are a model organism in evolutionary ecology and have been extensively studied due to their short generation time and capacity for rapid adaptation to their local environment (reviews: Endler, 1995; Houde, 1997; Magurran, 2005). Across their native range, guppies show an amazing degree of phenotypic, life-history and behavioural variation.

This variation has been shown to be shaped by natural and sexual selection, but has been mainly studied with respect to adaptation to habitats with divergent degrees of predation by piscivorous fish, a pattern that is repeated across different rivers (Endler, 1995; Houde, 1997;

Magurran, 2005). Guppies that co-occur with predators (high-predation guppies) are younger and smaller at maturity (Reznick, 1982; Reznick et al., 2001), produce more offspring per reproductive event (Reznick and Endler, 1982), produce smaller offspring (Reznick, 1982), and allocate more resources to each reproductive event (Reznick, 1982; Reznick et al., 1996)

26! ! compared to guppies in communities that experience reduced predation pressure (low- predation guppies). Similarly, this divergence has also been reported for other traits including antipredator behaviour (Seghers, 1974; Liley and Seghers, 1975; Magurran et al., 1994;

Godin and Davis, 1995; Ghalambor et al., 2004; Magurran, 2005; Walker et al., 2005;

Botham et al., 2006), morphology (Langerhans and DeWitt, 2004; Alexander et al., 2006) and colouration (Endler, 1991; Magurran and Seghers, 1994; Endler and Houde, 1995;

Houde, 1997).

Specialist parasites and mate choice

Hamilton and Zuk (1982) proposed that males of many species exhibit ornamental traits such as brightly colouration or vigorous courtship displays. These ornamental traits are often suggested to be condition-dependent, meaning that their expression depends on environmental factors such as nutrition, social interactions, and especially disease or parasites. Thus, ornamentation should be expressed only by males who are resistant to parasites, and females should take decisions based on these traits in order to choose resistant males as mates (i.e., parasite-mediated mate choice). In their model, Hamilton and Zuk’s

(1982) proposed that parasites drive the evolution of ornamentation that signify resistance to parasites. Essentially, the greater the selective pressure by parasites, the higher the fitness benefit for females to base their choice of mate on traits that honestly indicate a male’s capacity to deal with local parasites. Additive genetic variance in fitness could be maintained when the traits under sexual selection are associated with genes for resistance to disease

(Borgia 1979). A host’s resistance genes will cycle with the parasite’s virulence genes, and thus the genetic variation associated with sexually selected traits would come from host– parasite co-evolution.

27! !

The original model of Hamilton and Zuk (1982) proposed that specialist parasites should be more likely to produce co-evolutionary cycles since the feedback from any one host species to a generalist parasite’s gene pool would be small. Not surprisingly, the majority of empirical examples of parasite-mediated mate choice come from specialist parasites (reviewed in Møller et al., 1999; and Ballenger and Zuk, 2014). Most recently, female colouration-based preference has also been reported for three-spine sticklebacks

(Gasterosteus aculeatus), where male colouration was associated with progeny resistance to the specialist cestode parasite Schistocephalus solidus (Barber et al., 2015). To our knowledge no such example exists for a generalist parasite – although female avoidance of males infected with generalist parasites has been extensively documented – and further empirical studies are needed to better understand whether generalist and specialist parasites indeed differentially affect host sexual selection.

In the case of guppies and Gyrodactylus, evidence exists that Gyrodactylus could have strong implications on sexual selection. Kennedy et al. (1987) found that female guppies preferred to mate with males with relatively few Gyrodactylus, which have a higher display rate during courtship behaviour than heavily parasitized males.

Interestingly, this mate preference is greatly reduced when females are heavily infected

(Lopez, 1998). It is possible that females use the negative correlation between parasite load and display rate as a cue to assess the quality of their mates (i.e., honest signal).

Indeed, Gyrodactylus infections result in significant increase of guppy foraging behavior

(van Oosterhout et al., 2003), and a reduction in male expression of carotenoid coloration

(Houde and Torio, 1991; but see Lopez, 1998). Carotenoids cannot be synthesized by

28! ! guppies, and must be obtained from their diet (Goodwin, 1984), particularly from unicellular algae, which are their main food source (Dussault and Kramer, 1981).

Therefore, infected males face a trade-off between allocating carotenoids to sexual displays versus combating parasite infection (Lozano, 1994), and must increase their food intake to compensate for the energetic costs of infections (Kolluru et al., 2006; 2008).

Specialist parasites and predator-prey interactions

Depending on whether they infect the predator or the prey, parasite effects on predator- prey interactions may be top-down or bottom-up. Parasites of predators may release predation pressure from the prey through parasite-induced mortality, and thus result in an increase in prey densities. However, parasitic infection can also change predator traits.

For example, infections by the generalist acanthocephalan Echinorhynchus truttae result in an increase of 30% in the predation rate of the freshwater amphipod Gammarus pulex, compared to uninfected individuals (Dick et al., 2010). Yet infected common periwinkle

(Littorina littorea) with the generalist digenean trematode Cryptocotyle lingua show a decrease of 40% in the consumption of ephemeral macroalgy, compared to healthy individuals (Wood et al., 2007). Although the previous case-studies involve generalist parasites, they exemplify how parasites can directly reduce predator densities and alter predator phenotype – which may increase or decrease the intensity of predation.

Specialist parasites are likely to also be able to induce top-down effects like those presented above, but to-date empirical evidence is lacking.

Parasites can also have strong bottom-up effects, particularly through parasite- induced predation. For example, redgrouse (Lagopus lagopus scoticus) heavily infected

29! ! with the generalist nematode Trichostrongylus tenuis are more commonly predated upon, and the control of the predator population leads to an increase in the numbers of heavily infected birds – indicating that predators selectively prey on infected birds (Hudson et al.,

1992). Similarly, in a direct test for parasite-induced predation in field populations of voles (Microtus townsendii), voles treated with anti-helminthics suffered 17% less predation than infected individuals (Steen et al., 2002). If parasitic infection impacts host health to the extent that it increases the chances of being predated, parasites should be able influence predator-prey dynamics, regardless of whether they are generalist and specialist. Nonetheless, specialist parasites might be more prone to extinction if prey density is greatly reduced, whereas generalist parasites might maintain population size by exploiting alternative hosts.

The extent to which Gyrodactylus plays a role in guppy predator-prey interaction is largely unknown. Predators may preferentially attack parasitized fish, which could present an easier prey, or they could prefer healthy individuals who present an obvious and attractive target. Nonetheless, to our knowledge there is currently no evidence exists for predatory discrimination between infected and uninfected guppies.

There are several potential mechanisms by which Gyrodactylus could influence guppy-predator interactions that are unrelated to sexual selection. First, Gyrodactylus could reduce guppy density through associated pathology, independently of predation, and thus reduce the net predator-prey encounter rates. Second, fin clamping and shimming are often observed in heavily parasitized fish, which may hinder their

30! ! swimming ability (van Oosterhout et al., 2003), and could reduce predator avoidance.

Third, healthy brightly coloured males undergo more sexual displays than infected ones

(Kolluru et al., 2006; 2008), which may further increase their conspicuousness to predators, and result in a disproportional predation on healthy males. Lastly, the increased feeding behavior of infected individuals (van Oosterhout et al., 2003; Kolluru et al., 2006) could expose them to higher risk of predation. Although these mechanisms have been shown under laboratory conditions, empirical evidence in the field is lacking.

Assuming that one or more of them also occur in nature, the effects of Gyrodactylus on guppy-predator interactions are likely to be context specific. Moreover, all of these mechanisms involve a change in the predator-prey interactions that could alter community dynamics. For example, if Gyrodactylus infection significantly reduces guppy density independently of predation, energy flow to predators could be reduced, yet decomposers or aquatic invertebrates may benefit from the increased decaying matter.

Alternatively, Gyrodactylus epidemic outbreaks could induce large energy flow from primary producers, being heavily grazed by infected guppies, to higher trophic levels where predators could find an increased abundance of easy accessible prey (Lefevre et al., 2008). This amplified guppy mortality could then increase predator density, and eventually primary producers through grazing release, until guppies eventually become abundant –which may in turn facilitate Gyrodactylus spread, and start the cycle again

(Lafferty et al., 2006; 2008). These processes could be an important aspect of community stability, but field studies are needed to better understand their relative importance in the local community.

31! ! Competition mediated by specialist parasites

A question that inevitably comes to mind is how important specialist parasites are in host competitive interactions and community stability? Specialist parasites have been previously suggested to promote species coexistence and increase biodiversity, but with the exception of a few examples (Hudson et al., 1992), we still lack enough evidence for this. Based on Anderson and May’s (1978) original model, Fenton and Brockhurst (2008) demonstrated that the direct effect of specialist parasites on competition, through increased host mortality, greatly depends on the nature of this interaction in the absence of the parasite. For example, if competitive exclusion does not occur, the addition of the specialist parasite has no effect on the stability of the system. However, if competitive exclusion is present, specialist parasites can increase species coexistence or even change the outcome of competition if they infect the better competitor (Fenton and Brockhurst,

2008). Although this model does not account for trait-mediated effects on the host, it does provide evidence for the potential importance of specialist parasites in mediating interspecific competition and influencing community structure and stability. Studies comparing competition in the presence and absence of a specialist parasite are needed to understand this in the wild, and the guppy-Gyrodactylus system can be helpful to study such questions.

Across their native range in Trinidad, guppies are commonly found coexisting with only one other fish species: Rivulus hartii. Adult Rivulus are much larger than guppies (> 40 mm and >14 mm, respectively), and are strict predators, foraging mainly on aquatic and terrestrial invertebrates and small fish, including juvenile guppies (Gilliam

32! ! et al., 1993; Mattingly and Butler, 1994; Fraser et al., 1999). Juvenile Rivulus, on the other hand, are of similar size and ecology as guppies, and compete for resources like shelter and food (i.e., aquatic invertebrates) (Dussault and Kramer, 1981; Grether et al.,

2001). Previous studies have shown that guppies have the capacity to alter the ecology of

Rivulus in a variety of ways. First, direct competition between sized-matched Rivulus and guppies results in a reduced growth rate for Rivulus, but not for guppies (Gilliam et al.,

1993). Second, adult guppies can prey on Rivulus fry, and in this way reduce intraspecific competition of adults (Walsh and Reznick, 2009).

If Gyrodactylus infections are indeed energetically costly, it is, thus not a stretch to suggest that guppy-Rivulus competitive interactions could be affected through behavioural and phenotypic changes (i.e. trait mediated effects), or even associated mortality (i.e. density mediated effects). For instance, Gyrodactylus-induced changes in guppy health, behaviour and/or life history might indirectly benefit Rivulus by reducing the competitive ability of guppies or by reducing guppy population density, and thus releasing Rivulus from competitive pressure. However, Gyrodactylus could also increase the competitive pressure of guppies on Rivulus if infected guppies increase their foraging to compensate for the energetic costs of fighting infection, as has been previously documented in laboratory studies (van Oosterhout et al., 2003). These trait-mediated effects of Gyrodactylus infections on community stability and dynamics could be much higher than accounted for simply through infection-induced guppy mortality, but future work is needed if we are to fully understand this.

33! ! Gyrodactylus evolutionary implications

An aspect of host-parasite interactions that has received particular interest is the co- evolutionary dynamic between parasite virulence and host defenses (Bremermann and

Pickering, 1983; May and Anderson, 1983; Antonovics and Thrall, 1994; Bowers et al.,

1994; Boots and Haraguchi, 1999). The selective pressures exerted by the parasite can influence the genetic makeup of the host, and in turn change host-parasite dynamics, which can then easily spill over to the coexisting community (Hatcher et al., 2006).

Extensive evidence of this exists for Daphnia water fleas (Miner et al., 2012), but is probably common across other systems. In the case of Gyrodactylus, recent work has focused in understanding how different environments – specifically those that differ in predation pressure – have shaped the co-evolutionary history between guppies and

Gyrodactylus (van Oosterhout et al., 2003; Cable and van Oosterhout, 2007a; Cable and van Oosterhout, 2007b; Gotanda et al., 2013). Two field surveys (Martin and Johnsen,

2007; Gotanda et al., 2013), however, failed to find an association between parasite levels

(e.g. intensity of infection and prevalence), predation environment and male colouration in guppies – although there were strong differences between rivers. One possible explanation is that in high-predation environments guppy-Gyrodactylus interactions are embedded in a much larger and more complicated network, and it is thus difficult to detect signals of parasite-mediated selection. Another explanation is that guppies are well adapted to their Gyrodactylus parasites, and thus the fitness cost of infection is reduced and has no consequences for predation. Both mechanisms have direct ecological implications. For example, if alternate prey species are present in the community, parasite induced mortality may result in reduced host density without translating in changes in the

34! ! predator population (Holt, 1977). In the latter case, host-local adaptation can prevent parasites from having an ecological impact on the broader community by maintaining low infection levels. The importance of Gyrodactylus infections on guppies, thus, will greatly depend on the ecological context where guppies and Gyrodactylus coevolve.

General conclusions

Understanding how parasites can influence the ecology and evolution of their host is crucial for successfully predicting the impact of parasites in community processes.

Equally important, however, is to build a comprehensive framework that incorporates parasite host range. Although much evidence is still lacking, evidence exists that generalist and specialist parasites do not equally mediate host’s interspecific interactions, and thus differentiating between them will increase our predictive power and decision- making capacity. Here, we have explored the potential importance of the specialist monogenean ectoparasites Gyrodactylus in the Trinidadian guppy system. The extensive history of research on guppies, in the context of predation, and Gyrodactylus, in the context of epidemiology, facilitates to study host-parasite interactions inclusively with the broader selective pressures. Although, to-day, the evidence for Gyrodactylus mediated selection on guppies is scarce, we propose that Gyrodactylus have the capacity to strongly modify guppy’s competitive and predator-prey interactions through their direct effects on guppy survival and behaviour. Given the importance of gyrodactylids as pathogens in the aquaculture industry and freshwater rivers, research on this system could have direct implications for conservation and the management of invasive species.

35! ! REFERENCES

1.! Abrams, P.A. (1992) Predators that benefit prey and prey that harm predators:

unusual effects of interacting foraging adaptations. American Naturalist, 140: 573–

600.

2.! Abrams, P.A. and Matsuda, H. (1996) Positive indirect effects between prey species

that share predators. Ecology, 77: 610–616.

3.! Agosta, S.J., Janz, N., and Brooks, D.R. (2010) How specialists can be generalists:

resolving the "parasite paradox" and implications for emerging infectious disease.

Zoologia (Curitiba), 27: 151-162.

4.! Alexander, H.J., Taylor, J.S., Wu, S.S.T., and Breden, F. (2006) Parallel evolution

and vicariance in the guppy (Poecilia reticulata) over multiple spatial and temporal

scales. Evolution, 60: 2352-2369.

5.! Anderson, R.M. and May, R.M. (1978) Regulation and Stability of Host-Parasite

Population Interactions: I. Regulatory Processes. Journal of Animal Ecology, 47:

219-247.

6.! Andersson, M.B. (1994) Sexual selection. Princeton University Press.

7.! Antonovics, J., and Thrall, P.H. (1994) The cost of resistance and the maintenance of

genetic polymorphism in host-pathogen systems. Proceedings of the Royal Society of

London. Series B: Biological Sciences, 257: 105-110.

8.! Bakke, T.A., Cable, J., and Harris, P.D. (2007) The biology of gyrodactylid

monogeneans: the “Russian-doll killers”. Advances in Parasitology, 64: 161-460.

36! ! 9.! Barber, I., and Dingemanse, N.J. (2010) Parasitism and the evolutionary ecology of

animal personality. Philosophical Transactions of the Royal Society B: Biological

Sciences, 365: 4077-4088.!

10.! Balenger, S.L., and Zuk, M. (2014) Testing the Hamilton–Zuk hypothesis: past,

present, and future. Integrative and Comparative Biology, icu059.

11.! Boots, M., and Sasaki, A. (1999) ‘Small worlds’ and the evolution of virulence:

infection occurs locally and at a distance. Proceedings of the Royal Society of

London. Series B: Biological Sciences, 266: 1933-1938.

12.! Borgia, G. (1979) Sexual selection and the evolution of mating systems. In. M.S.

Blum and N.A. Blum, eds. Sexual selection and reproductive competition in insects.

Academic Press, New York.

13.! Botham, M.S., Kerfoot, C.J., Louca, V., and Krause, J. (2006) The effects of

different predator species on antipredator behavior in the Trinidadian guppy,

Poecilia reticulata. Naturwissenschaften, 93: 431-439.!

14.! Bowers, R.G., Boots, M. and Begon, M. (1994) Life-history trade- o¡s and the

evolution of pathogen resistance: competition between host strains. Proceedings of

the Royal Society of London: Biological Sciences, 257: 247-253.

15.! Bremermann, H.J., and Pickering, J. (1983) A game-theoretical model of parasite

virulence. Journal of Theoretical Biology, 100: 411-426.

16.! Cable, J., and Harris, P.D. (2002) Gyrodactylid developmental biology: historical

review, current status and future trends. International Journal for Parasitology, 32:

255-280.

37! ! 17.! Combes, C. (1997) Fitness of parasites: Pathology and selection. International

Journal for Parasitology, 27: 1-10. !

18.! Dick, J.T., Armstrong, M., Clarke, H.C., Farnsworth, K.D., Hatcher, M.J., Ennis, M.,

Kelly, A. and Dunn, A.M. (2010) Parasitism may enhance rather than reduce the

predatory impact of an invader. Biology Letters, rsbl20100171.

19.! Dunn, A.M., Torchin, M.E., Hatcher, M.J., Kotanen, P.M., Blumenthal, D.M., Byers,

J. E., and Perkins, S.E. (2012) Indirect effects of parasites in invasions. Functional

Ecology, 26: 1262-1274.

20.! Dunn, A.M. (2009) Parasites and biological invasions. Advances in Parasitology, 68:

161-184.

21.! Dussault, G.V., and Kramer, D.L. (1981) Food and feeding behavior of the guppy,

Poecilia reticulata (Pisces: Poeciliidae). Canadian Journal of Zoology, 59: 684-701.

22.! Endler, J.A. (1980) Natural selection on color patterns in Poecilia reticulata.

Evolution, 34: 76-91.

23.! Endler, J.A. (1982) Convergent and divergent effects of natural selection on color

patterns in two fish faunas. Evolution, 36: 178-188.

24.! Endler, J.A. (1991) Variation in the appearance of guppy color patterns to guppies

and their predators under different visual conditions. Vision Research, 31: 587-608.

25.! Endler, J.A. (1992) Signals, signal conditions, and the direction of evolution.

American Naturalist, 139: S125-S153.

26.! Endler, J.A. (1995) Multiple-trait coevolution and environmental gradients in

guppies. Trends in Ecology and Evolution, 10: 22-29.

38! ! 27.! Endler, J.A., and Houde, A.E. (1995) Geographic variation in female preferences for

male traits in Poecilia reticulata. Evolution, 49: 456-468.

28.! Fenton, A., and Brockhurst, M.A. (2008) The role of specialist parasites in

structuring host communities. Ecological research, 23:795-804.

29.! Fraser, D.F, Gilliam, J. F., MacGowan, M.P., Arcaro, C.M. and Guillozet, P.H.

(1999) Habitat quality in a hostile river corridor. Ecology, 80: 597-607.!

30.! Futuyma, D.J., and Moreno, G. (1988) The evolution of ecological specialization.

Annual Review of Ecology and Systematics, 19: 207-233.!

31.! Ghalambor, C.K., Reznick, D.N., and Walker, J.A. (2004) Constraints on adaptive

evolution: the functional trade-off between reproduction and fast-start swimming

performance in the Trinidadian guppy (Poecilia reticulata). American Naturalist,

164: 38-50.

32.! Gilliam, J.F., Fraser, D.F. and Alkins-Koo, M. (1993) Structure of a tropical stream

fish community: a role for biotic interactions. Ecology, 74: 1856-1870.

33.! Goodwin, T.W. (1986) Metabolism, nutrition, and function of carotenoids. Annual

Review of Nutrition, 6: 273-297.

34.! Gotanda, K.M., Delaire, L.C., Raeymaekers, J.A.M., Pérez-Jvostov, F., Dargent, F.

Bentzen, P., Scott, M.E., Fussmann, G.F. and Hendry, A.P. (2013) Adding parasites

to the guppy-predation story: insights from field surveys. Oecologia, 17: 155-166.

35.! Grether, G.F., Hudon, J., and Endler, J.A. (2001) Carotenoid scarcity, synthetic

pteridine pigments and the evolution of sexual coloration in guppies (Poecilia

reticulata). Proceedings of the Royal Society of London B: Biological Sciences, 268:

1245-1253.

39! ! 36.! Hamilton, W.D. and M. Zuk. (1982) Heritable true fitness and bright birds: a role for

parasites? Science, 218: 384-387.

37.! Harris, P.D., and Lyles, A.M. (1992) Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. Journal of

Parasitology, 78: 912-914.

38.! Harris, P.D., Shinn, A.P., Cable, J. and Bakke, T. (2004) Nominal species of the

genus Gyrodactylus von Nordmann 1832 (Monogenea: Gyrodactylidae), with a list

of principal host species. Systematic Parasitology, 59: 1-27.

39.! Hatcher, M.J., Dick, J.T., and Dunn, A.M. (2006) How parasites affect interactions

between competitors and predators. Ecology Letters, 9: 1253-1271.

40.! Holt, R.D. (1977) Predation, apparent competition, and the structure of prey

communities. Theoretical Population Biology, 12: 197-229.

41.! Holt, R.D., and Pickering, J. (1985) Infectious disease and species coexistence: a

model of Lotka-Volterra form. American Naturalist, 126: 196-211.

42.! Houde, A.E. (1997) Sex, color, and mate choice in guppies. Princeton University

Press.

43.! Houde, A.E., and Torio, A.J. (1991) Effect of parasitic infection on male color

pattern and female choice in guppies. Behavioral Ecology, 3: 346-351.

44.! Hudson, P.J., Dobson, A.P., and Newborn, D. (1992) Do parasites make prey

vulnerable to predation? Red grouse and parasites. Journal of Animal Ecology, 61:

681-692.

45.! Johnson, P.T.J., Dobson, A., Lafferty, K.D., Marcogliese, D.J., Memmott, J.,

Orlofske, S.A., Poulin, R. and Thieltges, D.W. (2010) When parasites become prey:

40! ! ecological and epidemiological significance of eating parasites. Trends in Ecology

and Evolution, 25: 362-371.

46.! Kawecki, T.J. (1998) Red queen meets Santa Rosalia: arms races and the evolution

of host specialization in organisms with parasitic lifestyles. American Naturalist,

152: 635-651.

47.! Kearn, G.C. (1994) Evolutionary expansion of the Monogenea. International Journal

for Parasitology, 24:1227-1271.

48.! Kearn, G.C. (1999) The survival of monogenean (Platyhelminth) parasites on fish

skin. Parasitology, 119: S57-S88.

49.! Kennedy, C.E.J., Endler, J.A., Poynton, S.L., and McMinn, H. (1987) Parasite load

predicts mate choice in guppies. Behavioral Ecology and Sociobiology, 21: 291-295.

50.! Kolluru, G.R., Grether, G.F., South, S.H., Dunlop, E., Cardinali, A., Liu, L., and

Carapiet, A. (2006) The effects of carotenoid and food availability on resistance to a

naturally occurring parasite (Gyrodactylus turnbulli) in guppies (Poecilia reticulata).

Biological Journal of the Linnean Society, 89: 301-309.

51.! Lafferty, K.D., Allesina, S., Arim, M., Briggs, C.J., De Leo, G., Dobson, A.P., and

Thieltges, D.W. (2008) Parasites in food webs: the ultimate missing links. Ecology

Letters, 11: 533-546.

52.! Lafferty, K.D., Dobson, A.P., and Kuris, A.M. (2006) Parasites dominate food web

links. Proceedings of the National Academy of Sciences, 103: 11211-11216.

53.! Lajeunesse, M.J., and Forbes, M.R. (2003). Variable reporting and quantitative

reviews: a comparison of three meta-analytical techniques. Ecology Letters, 6: 448-

454.

41! ! 54.! Langerhans, R.B., and DeWitt, T.J. (2004). Shared and unique features of

evolutionary diversification. American Naturalist, 164: 335-349.

55.! Lefevre, T., Lebarbenchon, C., Gauthier-Clerc, M., Misse, D., Poulin, R., and

Thomas, F. (2009) The ecological significance of manipulative parasites. Trends in

ecology and evolution, 24: 41-48.

56.! Lefevre, T., Roche, B., Poulin, R., Hurd, H., Renaud, F., and Thomas, F. (2008)

Exploiting host compensatory responses: the ‘must’ of manipulation? Trends in

Parasitology, 24: 435-439.

57.! López, S. (1998) Acquired resistance affects male sexual display and female choice

in guppies. Proceedings of the Royal Society of London. Series B: Biological

Sciences, 265: 717-723.

58.! Lozano, G.A. (1994) Carotenoids, parasites, and sexual selection. Oikos, 70: 309-

311.

59.! Magurran, A.E. (2005) Evolutionary ecology: the Trinidadian guppy. Oxford

University Press.

60.! Magurran, A.E., and Seghers, B.H. (1994) A cost of sexual harassment in the guppy,

Poecilia reticulata. Proceedings of the Royal Society of London. Series B: Biological

Sciences, 258: 89-92.

61.! Magurran, A.E., Seghers, B.H., Shaw, P.W., and Carvalho, G.R. (1994) Schooling

preferences for familiar fish in the guppy, Poecilia reticulata. Journal of Fish

Biology, 45: 401-406.

62.! Marcogliese, D.J., and Cone, D.K. (1997) Food webs: a plea for parasites. Trends in

Ecology and Evolution, 12: 320-325.

42! ! 63.! Martin, C.H., and Johnsen, S. (2007) A field test of the Hamilton–Zuk hypothesis in

the Trinidadian guppy (Poecilia reticulata). Behavioral Ecology and Sociobiology,

61: 1897-1909.

64.! Mattingly, H.T., and Butler IV, M.J. (1994) Laboratory predation on the Trinidadian

guppy: implications for the size-selective predation hypothesis and guppy life history

evolution. Oikos, 69: 54-64.

65.! May, R.M., and Anderson, R.M. (1978) Regulation and stability of host-parasite

population interactions: II. Destabilizing processes. Journal of Animal Ecology, 47:

249-267.

66.! May, R.M., and Anderson, R.T. (1983) Epidemiology and genetics in the

coevolution of parasites and hosts. Proceedings of the Royal society of London.

Series B. Biological sciences, 219: 281-313.

67.! Miner, B.E., De Meester, L., Pfrender, M.E., Lampert, W., and Hairston, N.G.

(2012) Linking genes to communities and ecosystems: Daphnia as an ecogenomic

model. Proceedings of the Royal Society B: Biological Sciences, 279: 1873-1882.

68.! Moller, A. P., Christe, P., and Lux, E. (1999) Parasitism, host immune function, and

sexual selection. Quarterly Review of Biology, 74: 3-20.

69.! Moore, J. (2002) Parasites and the behavior of animals. Oxford University Press.

70.! Pedersen, A.B., and Fenton, A. (2007) Emphasizing the ecology in parasite

community ecology. Trends in ecology and evolution, 22(3), 133-139.!

71.! Pfennig, K. S. (2001) Evolution of pathogen virulence: the role of variation in host

phenotype. Proceedings of the Royal Society of London. Series B: Biological

Sciences, 268: 755–760.

43! ! 72.! Poulin, R. (1992) Determinants of host-specificity in parasites of freshwater fishes.

International Journal for Parasitology, 22: 753-758.

73.! Poulin, R. (2010) Parasite manipulation of host behavior: an update and frequently

asked questions. Advances in the Study of Behavior, 41: 151-186.

74.! Prenter, J., MacNeil, C., Dick, J.T., and Dunn, A.M. (2004) Roles of parasites in

animal invasions. Trends in Ecology and Evolution, 19: 385-390.!

75.! Price, P. W., Westoby, M., and Rice, B. (1988). Parasite-mediated competition: some

predictions and tests. American Naturalist, 131: 544-555.

76.! Raffel, T.R., Martin, L.B., and Rohr, J.R. (2008) Parasites as predators: unifying

natural enemy ecology. Trends in Ecology and Evolution, 23: 610-618.

77.! Reznick, D.N. (1982) The impact of predation on life history evolution in

Trinidadian guppies: genetic basis of observed life history patterns. Evolution, 36:

1236-1250.

78.! Reznick, D.N., and Bryga, H.A. (1996) Life-history evolution in guppies (Poecilia

reticulata: Poeciliidae). V. Genetic basis of parallelism in life histories. American

Naturalist, 147: 339-359.

79.! Reznick, D.N., Butler IV, M.J., and Rodd, H. (2001) Life-history evolution in

guppies. VII. The comparative ecology of high-and low-predation environments.

American Naturalist, 157: 126-140.!

80.! Sasal, P., Trouvé, S., Müller-Graf, C., and Morand, S. (1999) Specificity and host

predictability: a comparative analysis among monogenean parasites of fish. Journal

of Animal Ecology, 68: 437-444.

44! ! 81.! Scott, M.E., and Anderson, R.M. (1984) The population dynamics of Gyrodactylus

bullatarudis (Monogenea) within laboratory populations of the fish host Poecilia

reticulata. Parasitology, 89: 159-194.!

82.! Steen, H., Taitt, M., and Krebs, C. J. (2002) Risk of parasite-induced predation: an

experimental field study on Townsend's voles (Microtus townsendii). Canadian

Journal of Zoology, 80: 1286-1292.

83.! Van Oosterhout, C., Harris, P.D., and Cable, J. (2003) Marked variation in parasite

resistance between two wild populations of the Trinidadian guppy, Poecilia

reticulata (Pisces: Poeciliidae). Biological Journal of the Linnean Society, 79: 645-

651.

84.! Van Oosterhout, C., Mohammed, R.S., Hansen, H., Archard, G.A., McMullan, M.,

Weese, D.J., and Cable, J. (2007) Selection by parasites in spate conditions in wild

Trinidadian guppies (Poecilia reticulata). International Journal for Parasitology, 37:

805-812.

85.! Walker, J.A., Ghalambor, C.K., Griset, O.L., McKenney, D., and Reznick, D.N.

(2005) Do faster starts increase the probability of evading predators? Functional

Ecology, 19: 808-815.

86.! Walsh, M.R., and Reznick, D.N. (2009) Phenotypic diversification across an

environmental gradient: a role for predators and resource availability on the

evolution of life histories. Evolution, 63: 3201-3213.!

87.! Wood, C. L., Byers, J. E., Cottingham, K. L., Altman, I., Donahue, M. J., and

Blakeslee, A. M. (2007) Parasites alter community structure. Proceedings of the

National Academy of Sciences, 104: 9335-9339.

45! ! 88.! Zuk, M., and Kolluru, G.R. (1998) Exploitation of sexual signals by predators and

parasitoids. Quarterly Review of Biology, 73: 415-438.

46! ! CONNECTING STATEMENT NO. 1

The importance of parasites as drivers of ecological and evolutionary processes has had extensive theoretical and empirical evidence. Nonetheless, not all parasites are the same.

The previous chapter is a review of the differences between specialist and generalist parasites, but it is also a plea to acknowledge parasite host range as an important ecological and evolutionary filter. Parasites with multiple hosts can more easily affect community dynamics through changes in population densities of different host species.

On the other hand, parasites with one host species can only do so through one such host, but are more likely to engage in strong coevolutionary arms races. The trajectory and intensity of this coevolutionary arms race between specialist parasites and their host can be strongly dependent on other ecological factors such as predation and competition.

The effects of predation on host evolution have been particularly well documented on guppies, where extensive evidence of adaptive divergence between high and low predation environments also exists. But how would these adaptations alter host- parasites coevolution? In the following chapter I answer this question by comparing the interactions of guppies from different predation environments and their sympatric

Gyrodactylus parasite.

47! !

CHAPTER 3

Are host-parasite interactions influenced by adaptation to predators? A test

with guppies and Gyrodactylus in experimental stream channels.

Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and Scott, M. E. 2012. Oecologia,

170: 77-88. DOI: 10.1007/s00442-012-2289-9

I want to suggest that the struggle against disease, and particularly infectious disease, has been a very important evolutionary agent, and that some of its results have been rather unlike those of the struggle against […] predators.

– JOHN B. S. HALDANE Disease and Evolution. Ric. Sci. Suppl. 1949

48! ! ABSTRACT

Natural populations often face multiple mortality sources. Adaptive responses to one mortality source might be also beneficial with respect to other sources of mortality, resulting in “reinforcing adaptations”; or they might be detrimental with respect to other sources of mortality, resulting in “conflicting adaptations”. We explored these possibilities by testing experimentally if the responses of guppies (Poecilia reticulata) to the monogenean ectoparasitic worm Gyrodactylus differed between populations adapted to different predation regimes. In experimental stream channels designed to replicate the natural environment, we exposed eight guppy populations (high-predation and low- predation populations from each of four separate rivers) to either their local Gyrodactylus parasites (infection treatment) or to the absence of those parasites (control). We found that infection dynamics varied dramatically among populations in a repeatable fashion, but that this variation was not related to the predation regime of origin. Consistent with previous work, high-predation guppy females gained more mass, had lower reproductive investment, and had more but smaller embryos than did low-predation guppy females.

Relative to control (no parasite) channels, guppies from treatment (infected) channels gained less mass but produced similar numbers and sizes of embryos - and thus had a higher reproductive effort. However, no interaction was evident between infection treatment and predation regime. We conclude that parasitism by Gyrodactylus and predation are both likely selective forces for guppies, but that adaptation to predation does not have an obvious deterministic effect on host-parasite dynamics or on life-history traits of female guppies.

49! ! INTRODUCTION

Natural populations experience a wide variety of extrinsic sources of mortality that can have important consequences for the fitness of individuals and the dynamics of populations: e.g., predation (Stoks and McPeak 2003), parasitism (Mouritsen and Poulin

2010), resource limitation (Holland and DeAngelis 2009), and pollution (Billoir et al.

2007). Populations tend to adapt to these local mortality sources by evolving traits that reduce mortality and increase the chances of reproduction (Stearns 1977; Roff 1992;

Stearns 1992). Most studies of these responses tend to focus on overall mortality rate or on mortality from one specific cause, such as predation. The natural situation, however, is much more complex because populations typically face multiple mortality sources that might simultaneously include a variety of different predators, parasites, competitors, and abiotic stressors. Under such conditions, predicting evolutionary responses is difficult because one needs to account not only for multiple sources of mortality but also the ways in which they interact.

Interactions between different sources of mortality could be reinforcing or conflicting with respect to adaptation. As an example of reinforcing effects, life history theory predicts that elevated mortality at a given stage will generally select for particular adaptations of age at maturity, reproductive allocation and body growth, irrespective of the specific source of that mortality (Roff 1992; Stearns 1992). As an example of conflicting effects, induced morphological changes in Daphnia can decrease mortality due to fish predation, while increasing the risk of parasitism (Yin et al. 2011). Whether or not adaptations to different mortality sources have reinforcing or conflicting effects has

50! ! important evolutionary implications. In particular, it would seem easier for a population to adapt to reinforcing selective pressures, because the same adaptive trait changes will decrease mortality from multiple sources. In contrast, conflicting influences dictate that trait changes that are adaptive with respect to one mortality factor could well be maladaptive with respect to other factors.

We are specifically interested in the interaction between two common sources of mortality: predation and parasitism. On the one hand, reinforcing effects could be possible because both predators and parasites can cause high mortality, which should generally favor the evolution of earlier age at maturity, higher reproductive investment, and more but smaller offspring (Law 1979; Roff 1992; Stearns 1992; Reznick et al.

1996). Thus, adaptation to parasitism could aid adaptation to predation. On the other hand, conflicting effects are also possible because predation can select for behaviors, such as increased grouping (Seghers 1974; Seghers and Magurran 1995), that protect against predation but might increase exposure to directly transmitted parasites (Hatcher et al. 2006). Also, some aspects of adaptation to parasites, such as the immune response

(Dybdahl and Storfer 2003), may have little influence on adaptation to predators. We have commenced a research effort in Trinidadian guppy (Poecilia reticulata) populations to consider these various possibilities.

Guppies, predators, and parasites

The Trinidadian guppy is a sexually dimorphic live-bearing freshwater fish endemic to north-eastern South America. This species is frequently used in evolutionary studies due

51! ! to its dramatic and rapid adaptation to its local environment (reviews: Endler 1995;

Houde 1997; Magurran 2005). The best-studied environmental contrast is predation intensity, commonly classified as “high” (many dangerous predatory fishes that have major effects on guppy demography) versus “low” (fewer and less dangerous predatory fishes that have only minor effects on guppy demography). This contrast is clearly an important evolutionary force given the ample evidence that high-predation and low- predation guppy populations exhibit genetically-based adaptive divergence in a great diversity of traits (reviews: Endler 1995; Houde 1997; Magurran 2005). Furthermore, these differences evolve rapidly when high-predation guppies are introduced into low- predation environments, or when dangerous predators are introduced into formerly low- predation environments (Endler 1980; Reznick and Bryga 1987; Reznick et al. 1996;

Gordon et al. 2009).

Despite this focus on the effects of predation regime, parasites could also be very important (van Oosterhout 2006; Cable and van Oosterhout 2007a; van Oosterhout et al.

2007; Fraser and Neff 2010; Fraser et al. 2010). Most of the parasite work conducted thus far has focused on Gyrodactylus, a genus of ubiquitous host-specific viviparous monogenean ectoparasites of freshwater and marine fishes (Kearn 1994; Harris et al

2004). Natural guppy populations in Trinidad are infected by up to two species of

Gyrodactylus: G. bullatarudis and G. turnbulli (Harris and Lyles 1992; Martin and

Johnsen 2007). Both species can be found in sympatry on the same individual guppy, with G. bullatarudis commonly found on the anterior end of the fish, particularly on the

52! ! head and gills, and G. turnbulli commonly found on the posterior end, particularly on the caudal fin, peduncle, and anal fin (Harris and Lyles 1992).

After infecting a guppy, each individual G. turnbulli gives birth to up to three individuals that are released onto the skin of the fish (Scott 1982). The first daughter is a clone of the mother that is born approximately 24 h after the mother matures, and it carries its own developing embryo inside. The second and third daughters start developing immediately after this first birth, and are themselves born 2.5 and 4.5 days later (Scott 1982). This life cycle characteristic of gyrodactylids allows populations to grow rapidly on an individual host. Transmission to other hosts occurs primarily when the parasites on one host “jump” to another nearby host. This direct transmission, together with rapid population growth on individual fish, leads to epidemic spread of infection through fish populations (Scott and Anderson 1984). Moreover, infections by

Gyrodactylus can cause high guppy mortality in the laboratory (Scott and Anderson

1984; Cable and van Oosterhout 2007a; Cable and van Oosterhout 2007b) and in nature

(van Oosterhout et al. 2007).

Gyrodactylus is thus a good candidate for the study of whether adaptation to predation influences adaptation to parasitism. Some previous work has considered this possibility. In particular, wild-caught guppies from a high-predation site on the Aripo

River in Trinidad had lower parasite loads and shorter infection durations (when tested in the laboratory) than did guppies from a low-predation site in the same river (van

Oosterhout et al. 2003; Cable and van Oosterhout 2007a; Cable and van Oosterhout

53! ! 2007b). However, the generality of this apparent high-predation versus low-predation effect is uncertain given that no other populations have been tested in a similar manner, and that guppy traits differ dramatically among populations within a given predation regime (Endler 1978; Reznick et al. 1996; Millar and Hendry 2011).

Our experiment

Our goal was to explore how wild-caught guppies from high- versus low-predation environments differ in (1) the infection dynamics of their sympatric Gyrodactylus parasite, and (2) in their life-history responses to Gyrodactylus infection. We therefore collected young guppies from high-predation and low-predation populations in four different rivers and placed them into experimental stream channels with and without their local Gyrodactylus parasites. Local parasites were used (i.e., each guppy population with its own parasite strain), as opposed to cross-infections, because we were interested in host-parasite interactions that actually occur in nature.

MATERIALS AND METHODS

Experimental design

We conducted two experiments using guppies collected from four different rivers in northern Trinidad: Quare, Aripo, Marianne, and El Cedro (Fig. 1). In each case, a guppy population was divided into a parasite treatment (with Gyrodactylus) and a parasite-free control, each in a different experimental stream channel. Although here the real experimental manipulation relative to the natural population involves the removal of

54! ! Gyrodactylus, our goal is to infer whether the effect of the presence of Gyrodactylus differs among populations. We thus follow this logic and refer to the with parasite condition as our treatment and the without parasite condition as our control. Experiment

1 involved fish collected only from high predation sites in the four rivers, resulting in two infection treatments crossed with four populations. Experiment 2 involved fish from the same high-predation sites, as in experiment 1, paired with low-predation sites from the same four rivers, resulting in two infection treatments crossed with four rivers and two predation regimes in each river. Experiment 1 served as an initial trial testing the feasibility and effect size of our infection treatment in the experimental stream channels.

Also, because the experimental procedures were identical for the two experiments, data from experiment 1 allowed us to consider whether results were repeatable for the four populations that were included in both experiments. Experiment 2 then served as our formal test for effects of predation and parasitism.

The experimental stream channels were the same as those used in previous work on guppies (Palkovacs et al. 2009; Bassar et al. 2010) we used eight channels for experiment 1 and 16 for experiment 2. These channels are 0.5 m wide by 3 m long by 0.2 m deep, and throughout the experiments receive flowing water from a tributary to the

Arima River. Seven days prior to the introduction of fish, the channels were thoroughly cleaned and gravel was added, making them available for natural colonization by the invertebrates and algae that provide food for the guppies. Predators were not included because we were not interested in the direct effects of predation risk on Gyrodactylus-

55! ! guppy interactions, but rather the ecological implications of evolving under different predation environments.

It is important to note that our use of water and invertebrates/algae from the

Arima River provides a useful standardization in the sense that none of the other populations experienced these specific conditions in nature – because they all came from other rivers. It is certainly possible that the experimental results might have differed in another “common garden” (water from another river), but it is at least known that the specific channels and conditions we used have been effective in revealing differences between guppy populations from other rivers (Palkovacs et al. 2009; Bassar et al. 2010).

We are therefore confident that our results reveal host-parasite mediated differences between populations.

Guppy populations and collections

During September (experiment 1) and November (experiment 2) of 2009, we collected guppies from four rivers: Quare, Aripo, Marianne, and El Cedro (Fig. 1). Guppies in the first three rivers represent different guppy lineages, likely separated for millions of years

(Fajen and Breden 1992; Willing et al. 2010). Guppies in the El Cedro river represent the same lineage as those in the Aripo river, but are nevertheless geographically separate and genetically differentiated (Willing et al. 2010). Within the El Cedro, the high-predation guppies are natural (as they are in the three other rivers), whereas the low-predation guppies are descendants of the downstream high-predation guppies that were introduced upstream of a waterfall in 1981 into a previously guppy free environment (Reznick and

56! ! Bryga 1987). Predatory fauna across the south-slope rivers (El Cedro, Aripo and Quare) consists primarily of cichlids (Crenicichla alta and Aequiedens pucher:), whereas the

Marianne site is dominated by eleotrids (Eleotris pisonis and Gobiomorus dormitor:

Reznick and Bryga 1996).

The specific high-predation and low-predation collection sites within each river

(Fig. 1) were chosen based on previous studies establishing the local predation regime by

(1) the presence/absence of dangerous predators (Endler 1978; Reznick et al. 1996; Rodd and Reznick 1997; Magurran and Phillip 2001), and (2) direct measurements of mortality rate (Reznick et al. 1996; Bryant and Reznick 2004; Weese et al. 2010). As far as is known from previous studies, the basic predation regime (high or low) is consistent through time at each of these sites, although predation intensity might well vary. Thus, our inferences relate to the basic predation rather than the specific predation intensity.

At each collection site (four for experiment 1, eight for experiment 2), 30 mature males and 30 juvenile females were collected and transported in individual 8 oz Whirl-Pak bags

(Fisher Canada) to our laboratory in Trinidad. Juvenile females were selected (based on their small size and poorly developed “gravid spot”) to maximize the chance that they were virgins. On arrival at the laboratory, the fish were anaesthetized in 0.02% Tricaine

Methanesulfonate (Finquel MS222 from Fisher Canada) (1: 8000) buffered to a neutral pH using NaHCO3. The parasite load of each individual was then determined by scanning its entire surface using a dissecting microscope under illumination by a cold light source.

Each fish was also measured (standard length to the nearest mm), weighed (mass to the nearest mg), and given an individual intra-dermal mark with nontoxic Visible Implant

57! ! Elastomer dye (Northwest Marine Technology Inc.) This marking methodology has been used in numerous guppy studies and has no apparent influence on guppy survival

(Reznick et al. 1996; Bryant and Reznick 2004; Gordon et al. 2009; Weese et al. 2010).

After the above processing, Gyrodactylus infections were eliminated from the fish by treating them with an aqueous solution of N-cyclopropyl- 1, 3, 5- triazine- 2, 4, 6- triamine (cyromazine), (Lice And Anchor Worm Treatment Ecological Laboratories Inc.), according to manufacturer’s recommendations. During and after the parasite removal period, the guppies were held in gender- and population-specific recovery tanks, where they were scanned for parasites (as above) over five consecutive days. Individuals found to still carry parasites on any of these days were immediately transferred to individual 1 L aquaria, and again treated as above. By the end of the five-day period, no parasites were observed on any of the fish. All fish were then kept in recovery aquaria for a period of four weeks, which allowed them to recover full susceptibility to future Gyrodactylus infections (Scott 1985). During this period no mature males and no fry were found in the female tanks – confirming that they were indeed virgins.

Experimental infection

After the four-week recovery period, all fish were measured and weighed in the same manner as described above. Fish from each population were then randomly divided into two experimental groups: one to be infected and one to serve as an uninfected control.

Ten males and ten females were then introduced into each of the experimental channels.

The resulting densities were below those normally seen in the wild (Rodd and Reznick

1997; Rodd et al. 2001) in hopes that life history traits would not be influenced by

58! ! variation in density that resulted from any parasite-induced mortality during the experiment.

To initiate an epidemic outbreak in the infection treatment channels, a few additional male guppies were collected from each of the field sites. One naturally- infected male with two to four Gyrodactylus attached to his caudal peduncle or caudal fin was measured and weighed as above, and then introduced to the infection treatment channels two days after introduction of the other fish. Live Gyrodactylus on experimental guppies cannot be identified to species level, but the above location of attachment on the fish is consistent with G. turnbulli (Harris and Lyles 1992). The same procedure was implemented by introducing an uninfected male (confirmed by scanning three times) into the no-infection control channels. The day of introduction of the infected or uninfected male was considered to be experimental day 0, after which fish survival and infection dynamics were tracked for four weeks by counting the individual parasite loads on each fish on every second day (using the above methodology). This process was replicated for the control channels to ensure that guppies in the control treatments were indeed uninfected and that they experienced similar handling conditions to guppies in the infected treatments.

Two small modifications were required to the above procedure for experiment 2.

In the El Cedro low-predation infection treatment, the introduced infected male shed its parasites on the first day. We therefore re-infected this male using two parasites from another recently caught male, and re-introduced him into the channel. Similarly, in the

59! ! Aripo high-predation infection treatment, the introduced infected male shed its parasites twice and so was re-infected twice. In both cases, the day of the final re-introduction of the infected male was considered as day 0.

After four weeks, all fish were captured from the channels, measured (standard length to the nearest 0.1 mm), weighed (to the nearest 0.1 mg), and euthanized with an overdose of MS222. The embryos in each female were then counted and their stage of development determined according to Haynes (1995) identification keys. In addition, the total fresh mass of all embryos was measured for each female. The remaining guppies that were not used in the experiment were euthanized with an overdose of MS-222. All the procedures in the experiments were in accordance with ethical norms and approved by the McGill University Animal Use Committee in the protocol No. 5759.

Data evaluation and statistical analysis

Data were categorized at individual- and population-level parameters. The former relate the phenotypic response of each individual fish, and the latter to each individual experimental stream channel. Parasite-related parameters recorded at the population level included maximum prevalence of infection (highest percentage of the population infected on a given day), peak of infection (day with the highest total parasite abundance), time to peak of infection, time to exclusion (time until all parasites in the fish population had disappeared), maximum duration of infection (longest individual infection), maximum number of parasites (highest total parasite abundance) and maximum parasite load per individual (highest parasite load on an individual fish at any given day). In three

60! ! populations (Table 1) the parasite population was still growing at day 27. We acknowledge the possibility that parasite levels might have reached much higher levels, however, for the purpose of our analysis, differences observed during the length of the experiment represent plausible differences in host-parasite dynamics between populations.

The individual phenotypic responses (growth, reproductive effort, number of embryos, embryo mass and embryo development) of female guppies to Gyrodactylus infection were analyzed using general linear models in R version 2.12.2011-02-22 (R

Development Core Team 2010). In these models, predation regime of origin (high or low: for the second experiment only), experimental treatment (infected or control) and their interaction were treated as fixed factors, whereas river of origin was considered as a random factor, mainly because our conclusions about river effects are not intended to be specific to those rivers, but to the overall difference between high- and low-predation environments. The individual-based covariates included in the analysis were initial body mass at introduction into the experimental channels and the stage of embryo development. Parasite load-related measurements were entered as covariates to test the regression relationship between parasite load and the strength of female guppy response.

However, these proved non-significant and were therefore dropped from all analyses. In the particular case of female growth, neither initial mass nor embryo development was statistically significant, so these covariates were dropped from the model. Response variables, each considered in a separate analysis, were a) female growth over the 27 days of the experimental infection (males were excluded from this analysis because they stop

61! ! growing at sexual maturity: Reznick 1982), b) female reproductive effort (proportion of female growth due to embryonic fresh mass), c) number of embryos per female and d) total mass of embryos per female. In addition to the individual phenotypic response to infection, duration of infection was calculated for each fish as the number of consecutive days it was observed infected during the entire length of the experiment. A few fish became infected, then eliminated their parasites, and then became re-infected. For these individuals, the duration of infection was calculated as the total number of days when parasites were observed for at least two consecutive sampling periods. Fish that were infected on only one sampling date were assigned an infection duration of 1 day. Mean parasite load was calculated for each fish as the mean number of parasites observed during the length of its infection.

Normality was evaluated using Shapiro-Wilk tests and homoscedasticity was examined using Bartlett’s tests. Only for duration of infection and mean parasite load was the assumption of normality violated (W= 0.968 P=0.012; W= 0.505 P=<0.0002, respectively). Transformations did not improve normality. These variables were then analyzed using Kruskal-Wallis nonparametric test with predation as a main effect and rivers pooled together per predation regime. Males and females were pooled together for each river.

Several additional details are pertinent. First, fish that died during the experiment were excluded from analyses of phenotypic responses. Second, the infection did not establish or spread in several of the experimentally infected groups in the second

62! ! experiment (Table 1). These populations, including their respective controls, were therefore removed from the analysis of guppy phenotypic responses. This resulted in an unbalanced design with three low-predation populations (Quare, Aripo, and Marianne) and two high-predation populations (El Cedro and Marianne). The level of significance for all analyses was set at P < 0.05.

RESULTS

Infection dynamics

Infection levels recorded on the fish immediately after their collection from the wild varied dramatically among rivers, both between and within predation regimes (Table 1).

Within each river, however, high-predation sites had a higher Gyrodactylus prevalence than those from low-predation sites (t = 3.851; df = 3; P= 0.030).

In the experimental stream channels, the introduction of a Gyrodactylus-infected male led to remarkably similar infection dynamics for the four (high-predation) guppy populations that were replicated between experiments 1 and 2. In both cases, parasites spread and established infections in the Marianne and El Cedro guppy populations (Fig.

2a, c) but did not spread to new hosts and were soon eliminated in the Quare and Aripo guppy populations (data not shown). Also in both experiments, the total Gyrodactylus population peaked at higher levels in Marianne guppies than in El Cedro guppies: 351 versus 120 respectively in experiment 1 and 111 versus 72 respectively in experiment 2

(Fig. 2b, d). These comparisons indicate that infection dynamics are a repeatable property

63! ! of host-parasite interactions for a given population, thus validating our use of only a single population per river (experiment 2) for comparing high-predation and low- predation guppy-Gyrodactylus interactions.

In experiment 2, infection dynamics differed among populations but did not show a consistent association with the predation regime of origin (Fig. 3) – as two comparisons reveal. First, the infection did not spread in both a low-predation population (El Cedro) and two high-predation populations (Aripo and Quare). Second, for populations where the infection did spread, predation regime did not significantly affect mean parasite load

(Kruskal-Wallis test: H= 1.413; df=1; P= 0.234) or the mean duration of infection

(Kruskal-Wallis test: H= 0.529; df=1; P= 0.466).

Guppy responses

In experiment 2, guppy mortality was similar between the infected and control channels

(Table 1): average mortality in infected channels was 22.0% compared to 22.6% for control channels. Among the infected channels, El Cedro and Quare low-predation channels had the highest mortality (33% and 28%, respectively). Among the control channels, highest mortality was observed in El Cedro and Marianne low-predation, 48% and 28% respectively.

Female growth was reduced in the infection treatment channels compared to the control channels, and it was higher in guppies from high-predation populations than from

64! ! low-predation populations (Fig. 4a; Table 2). The interaction between predation and infection did not have a significant effect (Table 2).

When controlling for initial mass, females from high-predation environments had more embryos, greater total embryo mass, and higher reproductive effort than did those from low-predation environments (Fig. 4c,d; Table 2). Parasite infection (i.e., infected versus control channels) had no detectable influence on the number of embryos or total embryo mass (Table 2) but it did influence our measure of reproductive effort. In particular, females from infection treatment channels had an approximately 4% greater reproductive effort than did females from the uninfected channels (Fig. 4b; Table 2). This result occurred because females in the infection treatment channels attained a similar embryo mass despite their reduced growth. However, no interaction was evident between predation regime and infection treatment for any of the variables. Interestingly, females in low-predation channels exposed to parasites had a reproductive effort that was close to that of high-predation females not exposed to parasites (Figure 4b).

DISCUSSION

We explored Gyrodactylus infection dynamics and guppy traits for four replicate instances (i.e., different rivers) of guppy population divergence between high-predation and low-predation environments. In a first experiment, we found that infection dynamics in experimental stream channels differed in a repeatable way among guppy populations, but a second experiment showed that this variation was not consistently associated with predation regime. We also found that guppy traits differed consistently between high-

65! ! predation and low-predation guppies and were influenced by whether or not parasites were present in the channels, but that the effects of predation regime of origin and parasite presence did not interact. In the following sections, we first discuss in more detail the nature of infection dynamics and then further consider effects on guppy traits.

We close with a discussion of how our data motivate the logical next experimental steps in examining the interactions between predation and parasitism in this system.

Impact of predation regime on infection dynamics

The lack of a consistent association between predation regime of origin for the guppies and infection dynamics of their local parasites was evident in several comparisons. First, in both experiments, high-predation guppies from the Quare and Aripo rivers were hard to infect with Gyrodactylus whereas high-predation guppies from the Marianne and El

Cedro rivers were not. Second, of the four populations in which infections persisted for the full duration of experiment 2, two were high-predation and two were low-predation.

Finally, none of our measures of dynamics differed between high and low predation sites, despite higher prevalence of infection in wild fish from high-predation sites at the time of collection from the field. This is in contrast to the results of previous studies using only the Aripo river, where experimentally-infected individual guppies from a high-predation site (Lower Aripo) had lower parasite loads and shorter duration of infection than those from a low-predation site (Upper Aripo) (van Oosterhout et al. 2003). It is feasible that the observed differences between our results and those previously reported are due to particularities of the rivers or experimental designs. For instance, this is the first time that

Gyrodactylus dynamics and guppy response of eight different populations from four

66! ! rivers are compared. Moreover, here we report Gyrodactylus-guppy dynamics in a setting that closely mimics wild populations in a natural environment, allowing free parasite transmission or parasite avoidance, rather than individual infection trials in a laboratory setting. In addition, it is possible that in populations where the cost of parasitism is high, guppies have evolved a strong innate immune response that prevents, or at least severely limits, the establishment and spread of the parasite. Such could be the case for this particular river. In a later study, Cable and van Oosterhout (2007b) compared parasite dynamics of highly virulent Gyrodactylus strain in fish from a high- and a low-predation population from the Aripo River. They found that parasites had a much lower initial establishment rate on fish from the high-predation population. This fits with our results.

By contrast, when they re-infected the fish 53 days later, low-predation Aripo guppies had lower parasite loads and were infected for fewer days than were guppies from the high predation population. It would therefore be valuable to repeat our experiment to examine other aspects of the immune response.

The fact that we did not detect a consistent effect of predation regime of origin on infection dynamics could reflect limitations due to lack of replication. However, infection dynamics showed consistency between the four high-predation populations used in experiments 1 and 2. This consistency increases our confidence in concluding that predation regime was not a crucial driver of infection dynamics, at least in our experimental stream channels. At the same time, the variation among populations suggests that other properties of guppies, their parasites, or the local environments where

67! ! they co-evolved do indeed influence infection dynamics. Examining these factors will require further experimentation.

Guppy traits

A first important point is that our results closely parallel the previously-documented differences in growth and life history traits between high-predation and low-predation guppies. Most obviously, guppies from high-predation sites have more embryos, greater total embryo mass, and higher reproductive effort (Reznick and Endler 1982; Reznick and Bryga 1987; Reznick 1989; Reznick and Bryga 1996; Gordon et al. 2009; Figure 4).

A second important point is that, regardless of predation regime, female guppies in infected channels grew more slowly and had higher reproductive effort (proportion of growth composed of embryonic fresh weight) than did female guppies in the control channels. A third important point is that no interaction was evident between predation regime and parasite treatment: that is, divergence owing to evolutionary history of predation regime did not influence responses to local Gyrodactylus parasites.

The reduction of female growth in the infection treatment channels might be explained either by (1) a negative impact of Gyrodactylus infection on host health or (2) the activation and maintenance of an energetically costly immune response. The first possibility implies a strictly pathological consequence of the infection, whereas the second possibility could represent an adaptive host response. The parasite loads in our experiments are significantly lower than those typically observed in the laboratory (Scott

1982; Scott and Anderson 1984; Scott 1985) or in field populations (Harris and Lyles

68! ! 1992; van Oosterhout et al. 2003; van Oosterhout et al. 2006; but see Faria et al. 2010).

Interestingly, Kolluru et al. (2009) report that infected first-generation male descendants from the Quare and Madamas low-predation populations increased their foraging behavior when experimentally infected with Gyrodactylus. If this is true in wild populations, such compensatory foraging behavior will increase carotenoid intake for its allocation to immune response (Kolluru et al. 2006), and support the hypothesis that the activation and maintenance of an immune response against Gyrodactylus has a great energetic cost. Whether or not these responses then influence susceptibility to predation is not known, although the pathological clamping of fins that typically occurs as a consequence of infection is likely to increase susceptibility to predation due to reduced mobility. This warrants further investigation.

Despite having a lower growth rate, females in infection treatment channels did not have lower total embryo mass, which thus means that their reproductive effort increased relative to those in control channels (Figure 4b). This response is consistent with expectations for a host infected with parasites (like Gyrodactylus) that have low initial population growth followed by rapid proliferation. The basic idea is that infections that will be potentially debilitating in the future should provoke the transfer of resources from growth to reproduction (Forbes 1993). Interestingly, this particular plastic response is likely to interact with adaptation to high predation, which also favors increased reproductive effort in guppies (Endler and Reznick 1982; Reznick and Bryga 1987;

Reznick et al. 1996) and other taxa (Stibor 1992; Gliwicz 2007). Our own data confirm this idea because the relative increase in reproductive investment associated with either

69! ! high predation or parasite infection was almost identical (Fig. 4b), and because

Gyrodactylus infection levels were higher in wild high-predation than low-predation guppy populations, both in our study and in previous studies (Martin and Johnsen 2007;

Fraser and Neff 2010). Thus, our results on guppy life-history traits suggest that responses to parasites and predators are reinforcing rather than conflicting.

The way forward

Our study shows that Gyrodactylus infection dynamics differ greatly among guppy populations from different rivers, and that infection of a guppy population influences guppy growth and life history. Although we found no interaction between predation regime of the guppy populations and the dynamics or effects of parasites, our study provides evidence for reinforcing responses to these factors. In essence, guppy response to Gyrodactylus was in the same direction as previous studies have reported for predation

(Reznick and Endler 1982; Reznick and Bryga 1987; Reznick 1989; Reznick and Bryga

1996; Gordon et al. 2009). However, further experimentation is needed to reveal the exact nature of the relationship between adaptations to predation and parasitism. First, it would be useful to perform the same experiment under the risk of predation. For instance, predator chemical cues have been known to alter a number of traits and behaviors in guppies (Dzikowski et al. 2004; Gosline and Rodd 2008). Perhaps these alterations would also modify guppy-Gyrodactylus interactions. Second, our experimental design was intended to mimic dynamics that might occur in nature – and so we used each guppy population’s local parasites. This means, however, that we cannot separate guppy characteristics from parasite characteristics, which are likely to be associated through co-

70! ! evolution (van Valen 1973; Price 1980; Ebert and Hamilton 1996; Lively 1999; Kawecki and Ebert 2004). Future work could use the same parasite strain across multiple guppy populations, which would better reveal whether the guppies themselves show responses to parasites that are modified by their predation regime of origin. It would also be useful to cross host genotypes (guppy populations) with parasite genotypes (parasite populations) with environments (presence or absence of predator cues) to see if hosts or parasites are better adapted to each other, and if this depends on the environmental context (Nuismer and Gandon 2008; Gandon and Nuismer 2009). Third, we targeted infection with G. turnbulli when G. bullatarudis might well have different effects (Cable and van Oosterhout 2007a). Fourth, we used wild-caught fish, whose prior experience could have influenced the results. Although many high-predation versus low-predation guppy differences are genetically based (Reznick 1982; Arendt and Reznick 2005; Karim et al. 2007), some plastic effects are also evident (Rodd et al. 1997; Bashey 2006; Gosline and Rodd 2008). An important extension could therefore be to use lab-reared fish to disentangle genetic effects from plastic influences.

Conclusions

We found that guppies exposed to a natural epidemic cycle of infection with their local

Gyrodactylus parasites showed decreased body growth rate and increased reproductive effort, independent of predation regime of origin. We therefore do not have clear evidence that parasite population dynamics are consistently different between guppies from high-predation environments versus those from low-predation environments.

However, the parasite-induced decrease in growth rate and increase in reproductive

71! ! allocation would presumably be less disadvantageous, and even might be advantageous, for high-predation populations that are selected by predators in the same direction. These results suggest that, in this system, adaptation to one agent of mortality (parasitism) could reinforce adaptation to another agent of mortality (predation), although we have here considered only plastic effects of parasitism. Whether genetic divergence in response to parasitism parallels genetic divergence with regard to predation remains to be seen.

ACKNOWLEDGMENTS

We would like to thank Alfredo Marquez for his help with fish collection and parasite sampling. We also thank David Reznick and the FIBR Program (From Genes to

Ecosystems) for facilitating use of the experimental channels, valuable suggestions, and overall support. Felipe Dargent and Etienne Low-Decarie provided valuable comments and discussions. Funding was provided by the Natural Sciences and Engineering

Research Council of Canada (NSERC) in the form of a Special Research Opportunity

Grant to G. Fussmann, A. Hendry, P. Bentzen and M. Scott, Discovery Grants to M.

Scott, and PhD scholarship from the Consejo Nacional de Ciencia y Tecnología (Mexico) to F. Pérez-Jvostov. Research at the Institute of Parasitology is supported by a regroupement stratégique from Fonds Québecois de la recherche sur la nature et les technologies.

72! ! REFERENCES

1.! Arendt JD, Reznick DN (2005) Evolution of juvenile growth rates in female guppies

(Poecilia reticulata): predator regime or resource level? Proc R Soc Biol 272:333-

337

2.! Bashey F (2006) Cross-generational environmental effects and the evolution of

offspring size in the Trinidadian guppy Poecilia reticulata. Evolution 60:348–361

3.! Bassar RD, Marshall MC, Lopez-Sepulcre A, Zandona E, Travis J, Pringle C.M,

Flecker AS, Thomas SA, Fraser DF, Reznick DN (2010) Local adaptation in

Trinidadian guppies alters ecosystem processes. Proc Natl Acad Sci 107:8:3616-

3621

4.! Billoir E, Péry ARR, Charles S (2007) Integrating the lethal and sublethal effects of

toxic compounds into the population dynamics of Daphnia magna: A combination of

the DEBtox and matrix population models. Ecol Model 203:204-214

5.! Bryant MJ, Reznick DN (2004) Comparative studies of senescence in natural

populations of guppies. Am Nat 163:55-68

6.! Cable J, van Oosterhout C (2007a) The impact of parasites on the life history

evolution of guppies (Poecilia reticulata): The effects of host size on parasite

virulence. Int J Parasitol 37: 1449-1458

7.! Cable J, van Oosterhout C (2007b) The role of innate and acquired resistance in two

natural populations of guppies (Poecilia reticulata) infected with the ectoparasite

Gyrodactylus turnbulli. Biol J Linn Soc 90:647-655

8.! Dybdahl MF, Storfer A (2003) Parasite local adaptation: Red Queen versus Suicide

King. Trends Ecol Evol 18:3:523-530

73! ! 9.! Dzikowski R, Hulata G, Harpaz S, Karplus I (2004) Inducible reproductive plasticity

of the guppy Poecilia reticulata in response to predation cues. J Exp Zool 301A:776-

782

10.!Ebert D, Hamilton WD (1996) Sex against virulence: the coevolution of parasitic

diseases. Trends Ecol Evol 11:79-82

11.!Endler JA (1978) A predatorʼs view of animal color patterns. Evol Biol 11:319-364

12.!Endler JA (1980) Natural selection on color patterns in Poecilia reticulata. Evolution

34:1:76-91

13.!Endler JA (1995) Multiple-trait coevolution and environmental gradients in guppies.

Trends Ecol Evol 10:1:22-29

14.!Fajen A, Breden F (1992) Mitochondrial DNA sequence variation among natural

populations of the Trinidad guppy, Poecilia reticulata. Evolution 46:1457-1465

15.!Faria PJ, van Oosterhout C, Cable J (2010) Optimal release strategies for captive-

bred animals in reintroduction programs: Experimental infections using the guppy as

a model organism. Conserv Biol 143:35-45

16.!Forbes M (1993) Parasitism and host reproductive effort. Oikos 67:3:444-450

17.!Fraser B, Neff BD (2010) Parasite mediated homogenizing selection at the MHC in

guppies. Genetica 138:273-278

18.!Fraser B, Ramnarine IW, Neff BD (2010) Temporal variation at the MHC class IIB

in wild populations of the guppy (Poecilia Reticulata). Evolution 64:7:2086-2096

19.!Gandon S, Nuismer SL (2009) Interactions between genetic drift, gene flow, and

selection mosaics drive parasite local adaptation. Am Nat 173:212-224.

74! ! 20.!Gliwicz J (2007) Increased reproductive effort as a life history response of Microtus

to predation 14:3:314-317

21.!Gordon SP, Reznick ND, Kinnison MT, Bryant MJ, Weese DJ, Rasanen K, Millar

NP, Hendry AP (2009) Adaptive changes in life history and survival following a new

guppy introduction. Am Nat 174:1:34-45

22.!Gosline AK, Rodd FH (2008) Predator-induced plasticity in guppy (Poecilia

reticulata) life history traits. Aquat Ecol 42:693-699

23.!Harris PD, Lyles AM (1992) Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. J Parasitol

78:912-914

24.!Harris PD, Shinn AP, Cable J, Bakke TA (2004) Nominal species of the genus

Gyrodactylus von Nordmann 1832 (Monogenea: Gyrodactylidae), with a list of

principal host species. Systematic Parasitol 59:1:1-27

25.!Hatcher MJ, Dick JTA, Dunn A (2006) How parasites affect interactions between

competitors and predators. Ecol Lett 9:1253-1271

26.!Haynes JL (1995) Standardized classification of poeciliid development for life-

history studies. Copeia 1995:147-154

27.!Holland JN, DeAngelis DL (2009) Consumer-resource theory predicts dynamic

transitions between outcomes of interspecific interactions. Ecol Lett 12:1357-1366

28.!Houde AE (1997) Sex, color, and mate choice in guppies. Princeton University

Press, Princeton, NJ, USA.

75! ! 29.!Karim N, Gordon SP, Schwartz AK, Hendry AP (2007) This is not déjà vu all over

again: male guppy colour in a new experimental introduction. J Eur Soc Evol Biol 1-

12

30.!Kawecki TJ, Ebert D (2004) Conceptual issues in local adaptation. Ecol Lett 7:1225-

1241

31.!Kearn GC (1994) Evolutionary expansion of the monogenea. Int J Parasitol 24:1227-

1271

32.!Kolluru GR, Grether GF, Dunlop E, South S (2009) Food availability and parasite

infection influence mating tactics in guppies (Poecilia reticulata). Behav Ecol

20:1:131-137

33.!Kolluru GR, Ruiz NC, del Cid N, Dunlop E, Grether GF (2006) The effects of

carotenoid and food intake on caudal fin regeneration in male guppies. J Fish Biol

68: 1002-1012

34.!Law R (1979) Optimal life histories under age-specific predation. Am Nat 114:399-

417

35.!Lively CM (1999) Migration, virulence, and the geographic mosaic of adaptation by

parasites. Am Nat 153:S34-S47

36.!Magurran AE (2005) Evolutionary ecology the Trinidadian guppy. Oxford Univ.

Press, Oxford, UK

37.!Magurran AE, Phillip DAT (2001) Evolutionary implications of large-scale patterns

in the ecology of Trinidadian guppies, Poecilia reticulata. Biol J Linn Soc 731-9

38.!Martin CH, Johnsen S (2007) A field test of the Hamilton-Zuk hypothesis in the

Trinidadian guppy (Poecilia reticulata). Behav Ecol Sociobiol, 61:1897-1909

76! ! 39.!Millar NP, Hendry AP (In Press) Population divergence of private and non-private

signals in wild guppies. Environ Biol Fish

40.!Mouritsen KN, Poulin R (2010) Parasitism as a determinant of community structure

on intertidal flats. Mar Biol 157:1:201-213

41.!Nuismer SL, Gandon S (2008) Moving beyond common-garden and transplant

designs: insight into the causes of local adaptation in species interactions. Am Nat

171:658-668

42.!Palkovacs EP, Marshall MC, Lamphere BA, Lynch BR, Weese DJ, Fraser DF,

Reznick DN, Pringle CM, Kinnison MT (2009) Experimental evaluation of evolution

and coevolution as agents of ecosystem change in Trinidadian streams. Philos Trans

R Soc B 364:1617-1628

43.!Price PW (1980) Evolutionary biology of parasites. Princeton University Press,

Princeton, NJ, USA

44.!R Core Development Team (2008) R: a language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna.

45.!Reznick DN (1982) The impact of predation on life history evolution in Trinidadian

guppies: genetic basis of observed life history patterns. Evolution 36:6:1236-1250

46.!Reznick DN (1989) Life-history evolution in guppies: 2. Repeatability of field

observations and the effects of season on life histories. Evolution 43:6:1285-1297

47.!Reznick DN, Bryga H (1987) Life-history evolution in guppies (Poecilia reticulata):

1. Phenotypic and genetic changes in an introduction experiment. Evolution

41:6:1370-1385

77! ! 48.!Reznick DN, Bryga H (1996) Life-History evolution in guppies (Poecilia reticulata:

Poeciliidae). V. Genetic basis of parallelism in life histories. Am Nat 147:3:339-359

49.!Reznick DN, Butler MJ IV, Rodd FH, Ross P (1996) Life-history evolution in

guppies (Poecilia reticulata) 6. Differential mortality as a mechanism for natural

selection. Evolution 50:4:1651-1660

50.!Reznick DN, Endler J (1982) The impact of predation on life history evolution in

Trinidadian guppies (Poecilia reticulata). Evolution 36:1:160-177

51.!Rodd FH, Reznick DN (1997) Variation in the demography of guppy populations:

the importance of predation and life histories. Ecology 78:2:405-418

52.!Roff DA (1992) The evolution of life histories. Chapman and Hall, NY.

53.!Scott ME (1982) Reproductive potential of Gyrodactylus bullatarudis (Monogenea)

on guppies (Poecilia reticulata). Parasitology 85:217-236

54.!Scott ME (1985) Dynamics of challenge infections of Gyrodactulus bullatarudis

Turbull (Monogenea) on guppies, Poecilia reticulata (Peters). J Fish Dis 8:495-503

55.!Scott ME, Anderson RM (1984) The population dynamics of Gyrodactylus

bullatarudis (Monogenea) within laboratory populations of the fish host Poecilia

reticulata. Parasitology 89:159-194

56.!Seghers BH (1974) Schooling behavior in the guppy (Poecilia reticulata): an

evolutionary response to predation. Evolution 28:486-489

57.!Seghers BH, Magurran AE (1995) Population differences in the schooling behavior

of the Trinidadian guppy, Poecilia reticulata—adaptation or constraint. Can J Zool

73:1100-1105

78! ! 58.!Stearns SC (1977) The evolution of life history traits: A critique to the theory and a

review of the data. Ann Rev Ecol Syst 8:145-171

59.!Stearns SC (1992) The evolution of life histories. Oxford Univ. Press, Oxford.

60.!Stibor H (1992) Predator induced life-history shifts in a freshwater cladoceran.

Oceologia 92:162-165

61.!Stoks R, McPeek MA (2003) Predators and life histories shape Lestes damselfly

assemblages along a freshwater habitat gradient. Ecology 84:6:1576-1587

62.!van Oosterhout C, Domino AJ, Cummings SM, Blais J, Barson NJ, Ramnarine IW,

Mohammed RS, Persad N, Cable J (2006) Balancing genetic drift, and genetic

variation at the Major Histocompatibility Complex in two wild populations of

guppies (Poecilia reticulata). Evolution 60:12:2562-2574

63.!van Oosterhout C, Harris PD, Cable J (2003) Marked variation in parasite resistance

between two wild populations of the Trinidadian guppy, Poecilia reticulata (Pisces:

Poeciliidae). Biol J Linn Soc 79: 4, 645-651.

64.!van Oosterhout C, Mohammed RS, Hansen H, Archard GA, McMullan M, Weese

DJ, Cable J (2007) Selection by parasites in spate conditions in wild Trinidadian

guppies (Poecilia reticulata). Int J Parasitol 37:7:805-812

65.!van Valen L (1973) A new evolutionary law. Evol Theory 1:1-30

66.!Weese DJ, Gordon SP, Hendry AP, Kinnison MT (2010) Spatiotemporal variation

in linear natural selection on body color in wild guppies (Poecilia reticulata).

Evolution 1802-1815

79! ! 67.!Willing EM, Bentzen P, van Oosterhout C, Hoffmann M, Cable J, Breden F, Weigel

D and Dreyer C (2010) Genome-wide single nucleotide polymorphisms reveal

population history and adaptive divergence in wild guppies. Mol Ecol 19:968-984

68.!Yin M, Laforsch C, Lohr JN, Wollinska J (2011) Predator-induced defense makes

Daphnia more vulnerable to parasites. Evolution 65:6:1482-1488

80! ! TABLES

Table 1. Characterization of Gyrodactylus infection dynamics in the eight experimental stream channels from experiment 2.

Predation Max. Max. Max. Max. Time to regime of Prevalenc prevalence Time to duration number parasite peak of River wild e in the during exclusio of of load / Mortality infection populatio wild experimen n [days] infection parasite individua [days] n t [days] s l

Quare Low 17.5% 88.9% 27 - 20 305 151 28.5%

High 35.8% 19% 11 11 12 17 13 33.3%

El Cedro Low 7.1% 4.74% 1 5 6 2 2 33.3%

High 24.2% 73.68% 27 - 18 72 17 19.0%

Marianne Low 7.5% 70% 27 - 18 40 6 23.8%

High 28.5% 95% 25 - 27 111 21 4.7%

Aripo Low 38.2% 52.38% 19 26 18 22 10 14.2%

High 43.2% 9.52% 5 11 11 5 4 19.0%

81! ! Table 2. Statistical analyses of female life-history traits in experiment 2. All traits were analyzed using general linear models with “infection” and “predation” as fixed effects.

“River” was entered as a blocking factor. “Initial mass” and “Embryo development” were removed from female growth analysis due to non-significance.

Variable Female Growth Number of embryos Embryo mass Reproductive effort F (D.F.) P F (D.F.) P F (D.F.) P F (D.F.) P Fixed factors: 5.118 0.076 Infection 0.026 0.810 1.465 (1,72) 0.230 6.020 (1,72) 0.016 (1,74) (1,72) 20.546 29.683 Predation <0.001 <0.001 22.799 (1,72) <0.001 8.240 (1,72) 0.005 (1,74) (1,72) Predation 0.865 0.126 x 0.355 0.753 0.0108 (1,72) 0.917 0.317 (1,72) 0.575 (1,74) (1,72) Infection 15.056 Initial mass - - 0.009 60.762 (1,72) <0.001 58.864 (1,72) <0.001 (1,72) 22.821 47.501 32.744 Embryo Development - - <0.001 <0.001 <0.001 (1,72) (1,72) (1,72) Random factors: Standard Deviation Standard Deviation Standard Deviation Standard Deviation Intercept Residual Intercept Residual Intercept Residual Intercept Residual River 0.011 0.032 0.000 2.819 0.000 0.006 0.000 6.354

82! ! FIGURES

Caribbean Sea N

0 5 10 Kilometers

A

E

G H F B C D

Figure 1. Geographical location of Arima tributary mesocosms used for the experiment and the eight sites from which guppies were collected: mesocosms (star), Marianne high- predation (A), El Cedro high-predation (B), Aripo high-predation (C), Quare high- predation (D), Marianne low-predation (E), El Cedro low-predation (F), Aripo low- predation (G) and Quare low-predation (H).

83! ! Marianne high-predation Marianne high-predation

100 a) 350 b) 90 200 Experiment 1 80 100 Experiment 2 70 50 60

50 20

Prevalence (%) 40 10

30 Total number of parasites +1 5

20

10 2

0 1

1 3 5 7 9 11 13 15 17 19 21 23 25 27 1 3 5 7 9 11 13 15 17 19 21 23 25 27

El Cedro high-predation El Cedro high-predation

100 c) 100 d)

90 50 80

70 20 60

50 10

Prevalence (%) 40 5 30 Total number of parasites +1

20 2 10

0 1

1 3 5 7 9 11 13 15 17 19 21 23 25 27 1 3 5 7 9 11 13 15 17 19 21 23 25 27

Days Days

Figure 2. Epidemic dynamics in El Cedro and Marianne high-predation guppies in experiment 1 and experiment 2.a) Parasite prevalence in Marianne high-predation guppies, b) Gyrodactylus population dynamics in Marianne high-predation guppies, c)

Parasite prevalence in El Cedro high-predation guppies, and d) Gyrodactylus population dynamics in El Cedro high-predation guppies.

84! ! Aripo El Cedro

a) 120 b) 120

High predation 50 50

Low predation 20 20 10 10 5 5 Total number of parasites +1 1 1

1 3 5 7 9 11 13 15 17 19 21 23 25 27 1 3 5 7 9 11 13 15 17 19 21 23 25 27

Marianne Quare 350 120 c) d) 200 50 100 50 20 20 10 10 5 Total number of parasites +1 5 1 1

1 3 5 7 9 11 13 15 17 19 21 23 25 27 1 3 5 7 9 11 13 15 17 19 21 23 25 27 Days Days

Figure 3. Gyrodactylus population dynamics in the eight experimental guppy populations from four rivers used in experiment 2: a) Aripo, b) El Cedro, c) Marianne and d) Quare.

Parasite dynamics on high-predation guppies are represented by solid lines, dashed line for low-predation guppies.

85! ! a) b) 0.13 Control 7 Infected 0.12 6 0.11 5 0.10 Growth (g) 4 0.09 Number of embryos 0.08 3 0.07 2 0.06

c) d) 18

0.020 16 0.018 0.016 14 0.014 12 0.012 Embryo weight (g) 10 0.010 8 % of gained mass devoted to reproduction 0.008 6 0.006 High Low High Low Predation Predation

Figure 4. Differences in phenotypic traits between females from infected and control channels, and high-predation and low-predation experimental populations from experiment 2: a) growth. b) reproductive effort (proportion of female growth due to embryonic fresh mass ), c) embryo mass, and d) number of embryos. Only those populations in which the infection did establish and their respective controls were analyzed. Dashed line-open triangle, control treatment; solid line-closed square, infected treatment. Error bars represent ± 1 standard error.

86! ! CONNECTING STATEMENT NO. 2

In the previous chapter I explored how guppy-Gyrodactylus adaptation can be influenced by host adaptation to predation. The results suggest that both predation and Gyrodactylus parasitism induce similar phenotypic responses in guppies (i.e. reduction in growth and higher reproductive allocation). It thus seems feasible that adaptations to predation could reinforce adaptations to Gyrodactylus parasites, and vice versa. There was also strong and repeated variation in guppy-Gyrodactylus dynamics across the different rivers, which was independent of predation.

An obvious first question that rises from the previous chapter is whether host- parasite adaptations are dependent on the environment or on host lineage. If adaptation is dependent on predation environment (i.e. high predation, high virulence; low predation, low virulence), one may predict that Gyrodactylus from high-predation environments will be well adapted to their guppy host. If this adaptation is largely dependent on host lineage

(i.e. some rivers have highly virulent Gyrodactylus, and other do not), then one may predict that guppy-Gyrodactylus adaptation is independent of predation and will vary greatly among rivers – as suggested by the results in the previous chapter. In the next chapter I, thus, investigate if similar predation environments results in parallel guppy-

Gyrodactylus local adaptation, or whether diverged lineages (i.e. rivers) undergo independent coevolutionary trajectories.

87! !

CHAPTER 4

Testing for host-parasite local adaptation: an experiment with Gyrodactylus

ectoparasites and guppy hosts

Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and Scott, M. E. Int J Parasitol.45:

409-417.

… the adjustment between parasites and hosts is very delicate, and very small changes in the environment in the host […] with some variability on the part of individual parasites to adapt themselves to variations in environment, may determine whether many or few or no parasites are able to develop in a particular host animal. – ASA C. CHANDLER Speciation and host relationship of parasites, 1923

88! ! ABSTRACT

Hosts and parasites are in a perpetual co-evolutionary ‘‘arms race’’. Due to their short generation time and large reproductive output, parasites are commonly believed to be ahead in this race, although increasing evidence exists that parasites are not always ahead in the arms race – in part owing to evolutionary lineage and recent ecological history. We assess local adaptation of hosts and parasites, and determine whether adaptation was influenced by ecological or evolutionary history, using full reciprocal cross-infections of four Gyrodactylus ectoparasite populations and their four guppy (Poecilia reticulata) host populations in Trinidad. To consider effects of evolutionary lineage and recent ecology, these four populations were collected from two different river drainages (Marianne and

Aripo) and two different predation environments (high and low). The highest infection levels were obtained when parasites from the Aripo lineage infected guppies from the

Marianne lineage, indicating a higher infectivity, virulence and/or reproductive success of the Aripo parasites. Aripo lineage guppies were also better able to limit Gyrodactylus population growth than guppies from the Marianne River, indicating their strong

‘‘resistance’’ to Gyrodactylus regardless of the source of the parasite. Predation environment had no detectable influence on host–parasite population dynamics of sympatric or allopatric combinations. The much stronger effect of evolutionary lineage

(i.e., river) than recent ecological history (i.e., predation) emphasizes its importance in driving co-evolutionary dynamics, and should be explored further in future studies on local host–parasite adaptation.

89! ! INTRODUCTION

Adaptation in host–parasite systems is a dynamic ‘‘arms race’’ in which adaptive peaks for the host and the parasite continuously shift in response to evolution of the opposing party (Ebert, 1994; Kaltz and Shykoff, 1998; Gandon and Michalakis, 2002; Kawecki and Ebert, 2004). Parasites are generally considered to be ahead in this arms race due to their shorter generation times which should increase their evolutionary speed (see Lively,

1999; Gandon and Michalakis, 2002; Greischar and Koskella, 2007; Hoeksema and

Forde, 2008), and because hosts are usually exposed to many parasite species which makes adaptation to any one species more difficult (Kawecki and Ebert, 2004).

Consistent with this, many studies have found that parasites show stronger signals of local adaptation to their hosts than hosts do to their parasites (reviewed in: Greischar and

Koskella, 2007; Hoeksema and Forde, 2008) as evidenced by higher infection levels for a given parasite population on sympatric hosts than on allopatric hosts (Ebert, 1994;

Saarinen and Taskinen, 2005). However, other studies have not found evidence of local parasite adaptation, or have found apparent local maladaptation of parasites: e.g., infection levels are higher on allopatric than sympatric hosts (Lemoine et al., 2012; Roth et al., 2012; Konijnendijk et al., 2013; Sternberg et al., 2013).

One set of potential reasons for these varied results is methodological. First, many studies have measured parasite fitness (e.g., infection levels) without also measuring host fitness (e.g., survival or growth) which means that local adaptation cannot be considered independently for both host and parasite. Second, many studies have been conducted in the laboratory whereas very different results might be obtained in the natural environment

90! ! (Brockhurst and Koskella, 2013). Third, many studies have not per- formed full reciprocal cross-infection experiments which makes it difficult to separate the confounding influences of virulence and resistance co-evolution (Greischar and Koskella,

2007).

Another set of potential reasons for varied results in local host–parasite adaptation studies is untested interactions with other factors related to evolutionary history or ecological context (Thompson, 1994, 1999; Morgan et al., 2005). Evolutionarily, different host lineages and their co-evolved parasites could have had different histories of selection, genetic bottlenecks, drift and founder events which might have strongly shaped co-evolutionary trajectories. Ecologically, recent ecological history whereby different host–parasite populations have experienced different biotic or abiotic conditions could have imposed selection that directly or indirectly influenced co-evolutionary trajectories

(Thompson, 1999). As one example, environments with high predation-induced host mortality are likely to select both for parasites that reproduce more quickly (and thus might be more virulent), and for hosts that invest less in parasite defense (Lively, 1999;

Gandon and Michalakis, 2002). Of course, the inverse might occur if parasitism increases susceptibility to other sources of mortality (Choo et al., 2003).

The Trinidadian guppy is frequently used in evolutionary studies due to its capacity for rapid and repeatable adaptation to different ecological environments (see reviews: Endler, 1995; Houde, 1997; Magurran, 2005; Dargent et al., 2013). The ecological force that has received the most attention is predation intensity, with guppy

91! ! populations commonly classified as either high predation (HP), with many dangerous predatory fishes that have major effects on guppy survival, or low predation (LP), with fewer and less dangerous predatory fishes that have only minor effects on guppy survival

(Reznick et al., 1996a; Gordon et al., 2009; Weese et al., 2010). In response to these different mortality regimes, HP and LP guppies have evolved a number of behavioural, life history and morphological differences (see reviews: Endler, 1995; Houde, 1997;

Magurran, 2005). As one example, HP guppies show earlier maturation and increased reproductive investment, with more frequent reproductive events and many but smaller embryos (Reznick, 1982; Reznick and Endler, 1982). Moreover, this evolution occurs rapidly following experimental introductions in nature (Reznick and Bryga, 1987;

Reznick et al., 1990, 1997; Gordon et al., 2009) and is repeatable across watersheds colonised by very diver- gent guppy lineages and with different predator faunas (Reznick and Bryga, 1996; Reznick et al., 1996b).

Guppies are commonly infected by the monogenean worm Gyrodactylus, a genus of ubiquitous host-specific ectoparasites on fishes (Harris and Lyles, 1992; Kearn, 1994;

Harris et al., 2004). Gyrodactylus are viviparous and reproduce directly on the host, exhibiting hyperviviparity: a mature female has in its uterus a fully developed embryo that in turn has a developing embryo within its uterus (Kearn, 1994). Transmission between hosts occurs through contact when the parasite ‘jumps’ to a new host. These characteristics result in a rapid increase in parasite numbers on an individual host and epidemic spread of infection through fish populations (Scott and Anderson, 1984).

Infections by Gyrodactylus can cause high guppy mortality in the laboratory (Scott and

92! ! Anderson, 1984; van Oosterhout et al., 2003; Cable and van Oosterhout, 2007a,b) and in nature (van Oosterhout et al., 2007). Not surprisingly, then, some evidence exists that guppy populations have evolved in response to Gyrodactylus, particularly through variation in the immune response (van Oosterhout et al., 2003) and at loci of the Major

Histocompatibility Complex (MHC) (Fraser and Neff, 2009; Fraser et al., 2010).

In a previous study (Pérez-Jvostov et al., 2012), we used experimental infections in semi-natural mesocosms to test whether adaptation to different predation environments

(HP versus LP) influenced Gyrodactylus–guppy interactions. We found strong and repeatable differences in Gyrodactylus infection dynamics between host–parasite assemblages taken from different field locations, but we found that the differences were not related to predation regime. However, because each guppy population was infected only with its own local parasite population, we were unable to disentangle the confounding effects between highly resistant hosts and highly virulent parasites, and those from low-resistance hosts and low-virulence parasites, which restricted any potential inferences on local adaptation.

The objective of this study was to assess local adaptation of hosts and parasites, and to determine whether adaptation was influenced by ecological or evolutionary history, using the well-studied ectoparasite Gyrodactylus infecting the Trinidadian guppy

(Poecilia reticulata). Our design allowed us to circumvent methodological limitations

(Hoeksema and Forde, 2008) by (i) generating separate measures of parasite and host fitness, (ii) conducting experiments in reasonably natural (mesocosm) environments, and

93! ! (iii) conducting a full reciprocal cross-infection experiment with four Gyrodactylus– guppy populations to disentangle local adaptation from effects of host–parasite co- evolution. We specifically tested whether parasites or hosts showed evidence of local adaptation (higher performance of parasites with sympatric than with allopatric hosts, or higher performance of hosts with sympatric than with allopatric parasites), and whether any local mal- adaptation was related to drainage of origin (evolutionary lineage) or predation regime (ecological differences).

MATERIALS AND METHODS

Fish collection and treatment

Immature guppies were collected from an HP population and an LP population within each of two rivers in the northern mountain range of Trinidad: the Marianne River (HP,

N10°460 30.52500 , E- 61°180 25.86100 ; LP, N10°440 51.8500 , E-61°170 30.615) on the northern slope and the Aripo River (HP, N10°390 25.83200 , E-61°130 39.39500 ;

LP, N10°410 15.49600 , E-61°140 4.45500 ) on the southern slope. These two rivers represent different guppy lineages (and probably separate colonisation events) as genetic distances between them are very large (see Suk and Neff, 2009; Willing et al., 2010). The

Gyrodactylus populations in these different drainages are probably also distinct (given their host specificity for guppies), but this has not yet been confirmed.

At each site, the fish were collected with butterfly nets and immediately placed in individual 8 oz. whirl-pak bags (Spectrum Nasco, U.S.A.) to prevent movement of

94! ! parasites among fish. After transfer to our laboratory in Trinidad, all fish were anaesthetized with MS-222 (Finquel MS222 from Fisher Canada; 1:8000 dilution and buffered to a neutral pH using NaHCO3) and then immediately scanned for

Gyrodactylus, using a dissecting microscope. Infected fish were isolated in individual containers to prevent the spread of infection.

All fish, regardless of whether or not they were initially infected, were treated with N-cyclopropyl-1,3,5-triazine-2,4,6-tri- amine (cyromazine; Lice And Anchor Worm

Treatment, Ecological Laboratories Inc., U.S.A.) which effectively eliminates

Gyrodactylus (Pérez-Jvostov et al., 2012). When no Gyrodactylus were seen on a fish over three consecutive days of visual inspection (as above), the fish was considered parasite-free. Elastomer dyes (Northwest Marine Technology Inc., U.S.A.) were then injected to give each fish a distinct intra-dermic mark, a procedure used effectively in many previous guppy studies (Bassar et al., 2010; Weese et al., 2010; Pérez-Jvostov et al., 2012). The elastomer marks were no longer than 2 mm and no marked fish showed signs of reduced mobility or altered behaviour. Guppies were then held in population- and sex-specific aquaria. No fry were observed in the recovery aquaria, confirming that females had been virgin prior the experiment.

Mesocosms

The mesocosms were 0.5 m wide by 3 m long by 0.2 m deep, and received continuous flowing water from a tributary adjacent to the Arima River without guppies, thus also preventing any potential introduction of Gyrodactylus into the mesocosms. This natural

95! ! flow allowed colonisation of the mesocosms by algae and invertebrates, including natural foods for guppies, but excluded any non-experimental guppies. These specific mesocosms have been used in a number guppy studies and are a good mimic of natural conditions (for technical specifications see Palkovacs et al., 2009; Bassar et al., 2010;

Pérez-Jvostov et al., 2012).

Experimental design

Our experiment used a full reciprocal cross-infection design for the four host-parasite populations (Fig. 1). Each of the four guppy populations was tested with each of the four

Gyrodactylus populations. This design led to four sympatric pairs (hosts and parasites from the same locations) and 12 allopatric pairs (hosts and parasites from different locations). Due to a limited number of mesocosms (16 channels), we were unable to perform replicates for the particular guppy-Gyrodactylus combinations.

Experimental protocol

Four weeks after parasite removal and marking (see Section 2.1), each fish was weighed

(to the nearest 0.1 mg), measured for standard length (to the nearest 1 mm), and scanned for Gyrodactylus. No parasites were found, confirming that parasite treatment had eliminated Gyrodactylus from all experimental fish. Guppies were then separated into 16 experimental groups (four for each population) each with eight females and eight males.

The 16 groups were then introduced into 16 mesocosms – one group per mesocosm.

96! ! Gyrodactylus for the experiment came from an infected “donor” fish collected immediately prior to the experiment from each of the four natural populations. To initiate a Gyrodactylus epidemic, we first transferred two to four parasites from the caudal fin of each of the four infected “donor” fish onto a male guppy selected from each of the four populations from the above-described recovery tanks. This transfer was done using a dissecting microscope by individually moving Gyrodactylus from a donor fish onto a naïve male. The experimentally infected males were kept overnight in individual

1 L containers and parasite establishment was confirmed the following day by visual inspection using a microscope. One infected male guppy was then introduced into each mesocosm to generate every possible combination of hosts and parasite sources.

Gyrodactylus epidemics in each mesocosm were monitored every second day over a period of 23 days. All fish were captured individually, anaesthetised (see Section

2.1), identified and inspected using a dissecting microscope to count parasites. After each inspection, the fish were released back into their mesocosm. At the end of the experiment, the weight (to the nearest 0.1 mg) and length (to the nearest 1 mm) were recorded for all fish. All females were euthanized with MS-222 and then dissected to count their embryos. Reproductive allocation was calculated as the percentage of gained weight devoted to embryo weight.

All procedures in the experiments were in accordance with ethical practices and approved by the McGill University, Canada, Animal Use Committee (Protocol No.

5759).

97! !

Statistical analysis

Our analyses focus on two aspects of local host–parasite adaptation: (i) Gyrodactylus performance on different guppy populations, and (ii) guppy performance when exposed to different Gyrodactylus populations. Gyrodactylus performance was evaluated in two separate types of model, and guppy performance was evaluated in a third type of model.

All analyses were performed in R version 2.14.1 (R Core Development Team 2011), and

P values were obtained using a Satterthwaite approximation for degrees of freedom with the package lmerTest, with levels of significance set at P < 0.05.

Models of Gyrodactylus performance

As a first step in evaluating Gyrodactylus performance, two linear mixed effects (LME) models were constructed with different response variables: (i) mean abundance of infection (average number of parasites observed on all guppies throughout the experiment), and (ii) duration of infection (number of consecutive days each guppy was infected throughout the experiment). Mean abundance of infection was log-transformed to meet the assumptions of normality and homoscedasticity of residuals. In both models, the random factor was guppy population and the fixed factors were host–parasite combination (sympatric versus allopatric – see Fig. 1), predation regime (HP versus LP) of Gyrodactylus, and drainage (Marianne versus Aripo Rivers) source of Gyrodactylus.

Simplified alternative models did not have lower Akaike Information Criterion (AIC) values (not shown), thus we only pre- sent results for the full model including all interactions. In this analysis, local Gyrodactylus adaptation would be inferred if parasite

98! ! performance was higher on sympatric than allopatric hosts, taking into account the predation regime and drainage of origin of Gyrodactylus.

As a second step in evaluating Gyrodactylus performance, the 16 host–parasite combinations were categorised according to the ‘‘degree of similarity’’ between hosts and parasites. We generated a new fixed factor with four levels representing hosts and parasites from (i) the same drainage and same predation regime (‘‘sympatric’’ as described above), (ii) the same drainage but different predation regimes, (iii) different drainages but the same predation regime, or (iv) different drainages and different predation regimes. We then fitted a general linear model with the response variable being mean intensity of infection (log transformed) and the explanatory variables being the degree of similarity, predation regime (high versus low) of Gyrodactylus, and drainage

(Marianne versus Aripo Rivers) source of Gyrodactylus. In this analysis we purposely ranked the degree of similarity based on drainage source rather than predation, based on the assumption that host genetic makeup would be more important than the predation environ- ment, but the opposite could also have been explored. Local Gyrodactylus adaptation would be inferred relative to ecological difference (is Gyrodactylus performance higher on guppies from the same predation regime?) and phylogenetic distance (is Gyrodactylus performance higher on guppies from the same drainage source?).

99! ! Models of guppy performance

To evaluate guppy performance, three LME models were constructed with different response variables: (i) change in female body mass (final weight–initial weight), (ii) reproductive allocation (proportion of body mass devoted to embryonic mass), and (iii) number of embryos. In these models, the random factor was Gyrodactylus population and the fixed factors were host–parasite combination (sympatric versus allopatric – see Fig.

1), predation regime (high versus low) of the guppy population, and drainage source

(Marianne versus Aripo Rivers) of the guppy population. Initial female mass was also added as a covariate. Based on AIC comparisons of alternative models (not shown), we present a reduced model that excluded the three-way interactions, second order interactions with host–parasite combination (sympatric versus allopatric) and the initial mass covariate.

RESULTS

Gyrodactylus performance

Gyrodactylus infections established and spread through the experimental guppy population in all of the guppy–Gyrodactylus combinations (Fig. 1). Nonetheless, parasite performance on allopatric pairs varied greatly whereas parasite performance on sympatric hosts, measured as mean abundance, was similar across all Gyrodactylus populations

(Fig. 2A). We first describe results based on the two response variables for sympatric- allopatric comparisons and then results based on host–parasite ‘‘degree of similarity’’.

100! ! Sympatric-allopatric analyses based on mean intensity of infection showed that

Gyrodactylus from the Aripo River were maladapted to their sympatric hosts in that they achieved higher intensities on allopatric Marianne River guppies (Fig. 2A; Table 1). This pattern held, regardless of the predation regime of the hosts or parasites, suggesting that maladaptation is best explained at the drainage source level. This was best exemplified in the Marianne River LP Gyrodactylus, which was the only parasite population showing a higher intensity on its sympatric host than on allopatric hosts – even though the infection intensity of Marianne River HP Gyrodactylus was similar in sympatric and allopatric comparisons (Figs. 2A, B and 3A, B).

Sympatric-allopatric analyses based on the duration of infection on individual fish yielded results similar to those described above for the mean intensity of infection. In particular, Aripo River Gyrodactylus (both LP and HP) were maladapted in that infections were 6–8days shorter on sympatric than allopatric hosts; Marianne River LP

Gryodactylus were locally adapted in that infections were up to 6 days longer on sympatric than allopatric hosts; and Marianne River HP Gryodactylus maintained similar infection durations on sympatric and allopatric guppies (Fig. 2B; Table 1).

“Degree of similarity’’ analyses showed that Aripo River Gyrodactylus performance was the highest on allopatric hosts that shared the same predation regime, especially for HP Gyrodactylus (Fig. 2C; Table 2). By contrast, Marianne River

Gyrodactylus performed similarly on all allopatric hosts, regardless on the degree of similarity in the environment (HP versus LP) and phylogenetics (drainage source).

101! !

Guppy performance

Female guppy growth was higher when fish sympatric than allopatric parasites for three of the four guppy populations (Fig. 4A; Table 3). Overall, Aripo River HP females showed the highest growth rate, particularly when infected with their sympatric parasite, whereas Marianne River HP females had the lowest growth rate regardless of parasite origin. The number of embryos per female was similar among populations when females were exposed to allopatric Gyrodactylus (Fig. 4B; Table 3). However, when infected with sympatric Gyrodactylus, Aripo River LP and Marianne River HP guppies produced fewer embryos, whereas sympatric infection of Aripo River HP and Marianne River LP guppies resulted in a higher number of embryos. The analysis of reproductive allocation did not reveal any significant effects of predation regime, drainage source or sympatric/allopatric association (Table 3).

DISCUSSION

Many previous studies of host–parasite interactions have not been designed in a way that allows clear insights into local adaptation and co-evolution (Hoeksema and Forde, 2008).

In an effort to reduce a number of these limitations, we tracked separate measures of parasite and host fitness in a fully reciprocal cross-infection design conducted in stream mesocosms using four populations that differed in recent ecological history and evolutionary history. Our first key finding was that parasites were not – contrary to typical expectations – locally adapted to their hosts. This raises questions about the

102! ! conditions under which para- sites or hosts are more likely to lead the evolutionary ‘‘arms race’’. Our second key finding was that patterns of local parasite mal- adaptation were strongly influenced by evolutionary lineage (drainage source and therefore host, and perhaps parasite, lineage) but were not influenced by recent ecological history (predation regime). In the following sections we expand on the potential explanations for local

Gyrodactylus maladaptation, as well as the relative importance of ecological history and evolutionary lineage as drivers of host–parasite co-evolution.

Although parasites are generally expected to have an evolutionary advantage over hosts (Ebert, 1994; Saarinen and Taskinen, 2005) due to their short generation time and potentially high host specificity, parasites have not always shown signatures of local adaptation to their sympatric hosts (Kaltz et al., 1999; Oppliger et al., 1999; Koskela et al., 2000; Lemoine et al., 2012; Roth et al., 2012; Konijnendijk et al., 2013; Sternberg et al., 2013). Our study reduced the methodological limitations of many previous studies, and our data indicate that Gyrodactylus does not generally show strong local adaptation to their sympatric guppy hosts. As an example, Aripo River Gyrodactylus achieved the highest and the longest infection intensities when exposed to guppies from the Marianne

River drainage, yet relatively low infection levels when infecting guppies from their shared drainage – regardless of their predation environment (Figs. 2A, B and 3).

One possible reason for this apparent lack of local adaptation by Gyrodactylus – or even maladaptation in the case of Aripo River parasites – is a strong divergence in resistance between Marianne River and Aripo River guppies. Such differences, coupled

103! ! with strong divergence in Gyrodactylus virulence, could result in an apparent lack of strong local parasite adaptation, or even mal- adaptation, when highly virulent parasites interact with low resistance hosts. Some evidence for this possibility exists in our study as Aripo River guppies were least affected by their sympatric para- sites, with highest growth rates (and sometimes more embryos) when faced with sympatric, relative to allopatric, parasites, while Aripo River Gyrodactylus performed best when infecting the less resistant Marianne River guppies (Fig. 4A, B). Yet another possible explanation is that guppies are leading the ‘‘arms race’’ and have become locally adapted to their sympatric parasites. Previous work has suggested that guppies from the Paria River, with common MHC alleles, have lower Gyrodactylus infection levels in the laboratory (Fraser and Neff, 2009), with similar observations in the wild (Fraser et al., 2010), suggesting that, indeed, guppies could be locally adapted to their sympatric Gyrodactylus.

Why might guppies be leading the ‘‘arms race’’? First, it should be noted that, although Gyrodactylus have much shorter generation times than guppies, guppies have relatively short generation times (approximately four per year) compared with hosts in other frequently studied host–parasite associations. Furthermore, it is conceivable that the reproductive system of Gyrodactylus (i.e., hyperviviparity and parthenogenesis), although allowing for very fast population growth, can also result in a reduction of genetic diversity in the population. Indeed, a recent laboratory study reported that sexual reproduction accounted for only 3.7% to 10.9% of population diversity, suggesting that the vast majority of the individuals are clones (Schelkle et al., 2012). Such a reduction in

Gyrodactylus genetic diversity could allow outbred guppy populations to quickly adapt to

104! ! the most common parasite genotype and thus put them at a lesser evolutionary disadvantage than other hosts to their parasites. Second, Gyrodactylus can cause high guppy mortality in the field and, especially, in the laboratory (Scott and Anderson, 1984; van Oosterhout et al., 2003; Cable and van Oosterhout, 2007a,b). It may be conceivable that very high virulence in Gyrodactylus might be maladaptive if it increases host mortality in such way that the basic Gyrodactylus reproductive rate is reduced (Dybdahl and Storfer, 2003; Alizon et al., 2008). In this way, lower virulence in Gyrodactylus could evolve in response of the lower resistance of their guppy host (Altizer, 2001;

Sternberg et al., 2013) – although the opposite effect has also been reported (Hoeksema and Forde, 2008). Some evidence for this exists in our study in that the least resistant hosts (Marianne River HP and LP) also had the parasites with the lowest performance. Of course, cause and effect could be reversed here in that Marianne River Gyrodactylus do not have to evolve high infectivity because their hosts show such low resistance.

Regardless of the specific reasons, these results should be taken as another challenge to the established paradigm that parasites generally are ahead in the ‘‘arms race’’ with hosts

(Hoeksema and Forde, 2008).

Beyond this basic result, considerable variance was present between our measures of both parasite fitness (number of para- sites and duration of infection) and host fitness

(growth and reproduction). We now explore the causes of this variation by considering different evolutionary histories (different guppy, and likely Gyrodactylus, lineages) and differences in recent ecological histories (HP versus LP).

105! ! How evolutionary history has shaped host–parasite interactions remains to be determined. One possibility is that patterns of guppy genetic variation differ dramatically between the two lineages, which could then cause different evolutionary trajectories even in the case of similar selection. For instance, recent studies have shown considerable variation between northern and southern slope guppies in the genetic basis for adaptive traits, potentially due to founder effects (Willing et al., 2010). Another possibility is that genetic differentiation between HP and LP populations in a given drainage is relatively low – thus making Gyrodactylus adaptation to particular genotypes easier. Indeed, low microsatellite genetic differentiation between predation environments has been shown for both Marianne River and Aripo River drainages (Suk and Neff, 2009), but is still unknown for genes of the immune system (e.g., MHC). Irrespective of the particular mechanisms by which evolutionary history could affect the evolution of host resistance and parasite infectivity, our study indicates that these effects have major implications on co-evolutionary dynamics. Taking evolutionary history into account is thus important in studies of host–parasite dynamics.

Theoretical models often predict that increased parasite transmission (and therefore perhaps increased virulence) should evolve under increased host mortality

(Anderson and May, 1982; Gandon and Michalakis, 2002). Given that higher parasite reproductive and transmission rates can increase local parasite adaptation (Alizon et al.,

2008), Gyrodactylus from HP environments would be expected to perform better (e.g., parasite growth rate, infectivity, prevalence, mean abundance) on their sympatric compared with allopatric hosts (higher performance leads to higher fitness) – although

106! ! this of course would also select for higher resistance in HP guppies. This expectation has not been demonstrated in our study as we found no relationship between predation regime and local maladaptation by parasites. For instance, of the two parasite populations showing the highest maladaptation (i.e., lowest growth rate on their sympatric host), one was from an HP site and the other was from an LP site.

One possible reason for our inability to detect an effect of recent ecological history is that our experimental conditions were not realistic enough – for instance, we excluded predators from the mesocosms. Predators and predator cues are certainly known to have very strong plastic effects on guppy behaviour and growth (Rodd and Reznick,

1997; Evans et al., 2007; Gosline and Rodd, 2007; Brown et al., 2013), and these guppy traits are known to influence infection (Johnson et al., 2011; Richards et al., 2010). Thus, perhaps we would have found very different results had we exposed guppies to predators or predator cues.

Beyond parasites, we might also expect different predation regimes to influence guppy resistance to parasites. For instance, high mortality rates in HP environments might reduce the benefits of investing in resistance (or tolerance) to parasites (Dargent et al., 2013). Alternatively, heavily infected individuals might experience higher predation risk, which would thus select for increased resistance by the hosts (Packer et al., 2003).

However, our study failed to find an association between the predation regime of guppy hosts and their response to parasites. Perhaps the two effects described above (high parasite-independent predation and effects of infection on predation) offset each other,

107! ! leading to no net effect of predation regime. This possibility might be interesting to investigate in future experiments.

Although the precise mechanisms will have to be further established, our study adds to the growing body of work that suggests that parasites and predation do not seem to have strong interactive effects on guppies (Gotanda et al., 2013; Pérez-Jvostov et al.,

2012; Dargent et al., 2013). We understand the limitations of our lack of replication for any given guppy–Gyrodactylus combination, yet we are confident of our results given that our sympatric infection dynamics closely resemble those of a previous study where strong drainage source effects were also observed (Pérez-Jvostov et al., 2012). Predation is only one ecological factor that might influence host–parasite co-evolution, thus it might seem tenuous to use our results to conclude that ecological context is not important. However, predation is thought to be the strongest ecological context shaping the evolution of guppy traits (Endler, 1995; Houde, 1997; Magurran, 2005) – thus it was a reasonable place to start.

Our observation that parasites were not locally adapted to their hosts may stimulate further work on the conditions under which parasites or hosts are more likely to be leading the evolutionary ‘‘arms race’’. Our finding that patterns of local host and parasite maladaptation were not influenced by the predation regime but were strongly influenced by the drainage source and therefore host (and perhaps parasite) lineage runs counter to the idea that natural selection owing to ecological differences leads to deterministic patterns of parallel (or convergent) evolution (Endler, 1986; Schluter,

108! ! 2000), and to evidence from a number of guppy traits for deterministic responses to predation (Reznick and Endler, 1982; Reznick et al., 1990; Rodd and Reznick, 1991;

Endler, 1995, but see Torres-Dowdall et al., 2012). However, recent studies are increasingly emphasising the fact that evolution in similar environments is often not very similar (i.e., non-parallel or non-convergent) which suggests a considerable role for historical contingency (Kaeuffer et al., 2012; Fitzpatrick et al., 2013). Our study provides direct support for this contingency by showing that patterns of local host–parasite maladaptation are predictable by drainage source (and likely lineage) rather than the

(otherwise) most important ecological context for guppies (predation regime). Overall, our study thus provides additional support for the importance of considering non- deterministic aspects of evolution and the causes thereof.

ACKNOWLEDGMENTS

We would like to thank David Reznick and the FIBR Program (From Genes to

Ecosystems) for facilitating use of the experimental channels, valuable suggestions, and overall support. Funding was provided by the Natural Sciences and Engineering Research

Council of Canada (NSERC) in the form of a Special Research Opportunity Grant to G.

Fussmann, A. Hendry, P. Bentzen and M. Scott, Discovery Grants to M. Scott, and PhD scholarship from the Consejo Nacional de Ciencia y Tecnología (Mexico) to F. Pérez-

Jvostov. Research at the Institute of Parasitology is supported by a regroupement stratégique from Fonds Québecois de la recherche sur la nature et les technologies.

109! ! REFERENCES

1.! Alizon, S., Hurford, A., Mideo, N., van Baalen, M., 2008. Virulence evolution and the

trade-off hypothesis: history, current state of affairs and the future. J Evol. Biol. 22, 2:

245-259.

2.! Altizer, S. M., 2001. Migratory behaviour and host-parasite co-evolution in natural

populations of monarch butterflies infected with a protozoan parasite. Evol. Ecol. Res.

3, 611–632.

3.! Anderson, R.M., May, R.M., 1982. Coevolution of hosts and parasites. Parasitology

85, 411–426.

4.! Bassar R.D., Marshall, M.C., Lopez-Sepulcre, A., Zandona, E., Travis, J., Pringle,

C.M., Flecker, A.S., Thomas S,A., Fraser, D.F., Reznick, D.N., 2010. Local adaptation

in Trinidadian guppies alters ecosystem processes. Proc. Natl. Acad. Sci. 107, 3616-

3621.

5.! Brockhurst, M.A., Koskella, B., 2013. Experimental Coevolution of Species

Interactions. Trends Ecol. Evol. 28, 367–75. (DOI:10.1016/j.tree.2013.02.009.)

6.! Brown, G. E., Ferrari, M. C. O., Elvidge, C. K., Ramnarine, I. W., Chivers, D. P.,

2013. Phenotypically Plastic Neophobia : A Response to Variable Predation Risk.

Proc. R. Soc. B. 280: 20122712.

7.! Cable J., van Oosterhout, C., 2007a. The impact of parasites on the life history

evolution of guppies (Poecilia reticulata): The effects of host size on parasite

virulence. Int. J. Parasitol. 37, 1449-1458

110! ! 8.! Cable J., van Oosterhout, C., 2007b. The role of innate and acquired resistance in two

natural populations of guppies (Poecilia reticulata) infected with the ectoparasite

Gyrodactylus turnbulli. Biol. J. Linn. Soc. 90, 647-655.

9.! Choo, K., Williams, P.D., Day, T., 2003. Host Mortality, Predation and the Evolution

of Parasite Virulence. Ecol. Lett. 6, 310–315. (DOI:10.1046/j.1461-

0248.2003.00425.x.)

10.! Crispo, E., Bentzen, P., Reznick, D. N., Kinnison, M. T., Hendry, A. P., 2006. The

Relative Influence of Natural Selection and Geography on Gene Flow in Guppies.

Mol. Ecol. 15, 49–62. (DOI:10.1111/j.1365-294X.2005.02764.x.)

11.! Dargent F., Scott, M.E., Hendry, A.P., Fussmann, G.F., 2013. Experimental

elimination of parasites in nature leads to the evolution of increased resistance in

hosts. Proc. R. Soc. B. 280 (doi:10.1098/rspb.2013.2371).

12.! Ebert, D., 1994. Virulence and local adaptation of a horizontally transmitted parasite.

Science 265, 1084–1086.

13.! Ebert, D., Weisser, W.W., 1997. Obligate killing for obligate killers: the evolution of

life histories and virulence of semelparous parasites. Proc. R. Soc. Lond. Ser. B. 264,

965–991.

14.! Endler, J. A., 1986. Natural selection in the wild. Princeton Univ. Press, Princeton,

NJ.

15.! Endler, J. A., 1995. Multiple-trait coevolution and environmental gradients in

guppies. Trends Ecol. Evol. 10, 22-29.

111! ! 16.! Evans, J. P., Gasparini, C., Pilastro, A., 2007. Female Guppies Shorten Brood

Retention in Response to Predator Cues. Behav. Ecol. Sociobiol. 61, 719–727.

(DOI:10.1007/s00265-006-0302-1.)

17.! Fitzpatrick, S.W.,Torres-Dowdall, J., Reznick, D.N., Ghalambor, C.K., Funk, W.C.,

2013. Parallelism isn’t perfect: could disease and flooding drive a life history anomaly

in Trinidadian guppies? Am. Nat. 183, 290-300

18.! Fraser B., Neff, B.D., 2009. Parasite mediated homogenizing selection at the MHC

in guppies. Genetica 138, 273-278.

19.! Fraser B., Ramnarine, I.W., Neff, B.D, 2010.Temporal variation at the MHC class

IIB in wild populations of the guppy (Poecilia reticulata). Evolution 64, 2086-2096.

20.! Gandon, S., Michalakis, Y., 2002. Local adaptation, evolutionary potential and host-

parasite coevolution: interactions between migration, population size and generation

time. J. Evol. Biol. 15, 451–462.

21.! Gordon, S.P., Reznick, N.D., Kinnison, M.T., Bryant, M.J., Weese, D.J., Rasanen,

K., Millar, N. P., Hendry, A.P., 2009. Adaptive changes in life history and survival

following a new guppy introduction. Am. Nat. 174, 34-45.

22.! Weese, D.J., Gordon, S.P., Hendry, A.P., Kinnison, M.T., 2010. Spatiotemporal

variation in linear natural selection on body color in wild guppies (Poecilia reticulata).

Evolution 64, 1802-1815.

23.! Gosline, A.K., Rodd, F. H., 2007. Predator-Induced Plasticity in Guppy (Poecilia

reticulata) Life History Traits. Aquatic Ecol. 42, 693–699. (DOI:10.1007/s10452-007-

9138-7.)

112! ! 24.! Gotanda, K. M, Delaire, L. C., Raeymaekers, J. A. M., Pérez-Jvostov, F., Dargent,

F., Bentzen, P., Scott, M.E., Fussmann, G.F., Hendry, A.P., 2013. Adding Parasites to

the Guppy-Predation Story: Insights from Field Surveys. Oecologia 172, 155–66.

(DOI:10.1007/s00442-012-2485-7.)

25.! Greischar, M. A., Koskella, B., 2007. A Synthesis of Experimental Work on Parasite

Local Adaptation. Ecol. Lett. 10, 418–434. (DOI:10.1111/j.1461-0248.2007.01028.x.)

26.! Harris, P. D., Lyles A.M., 1992. Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. J. Parasitol. 78,

912-914.

27.! Hoeksema, J. D., Forde, S. E., 2008. A meta-analysis of factors affecting local

adaptation between interacting species. Am. Nat. 171, 275-290. (DOI:

10.1086/527496)

28.! Houde, A.E., 1997. Sex, color, and mate choice in guppies. Princeton University

Press, Princeton, NJ, USA.

29.! Johnson, M. B., Lafferty, K. D., van Oosterhout, C., Cable, J.C., 2011. Parasite

Transmission in Social Interacting Hosts: Monogenean Epidemics in Guppies. PloS

One 6, e22634. (DOI:10.1371/journal.pone.0022634.)

30.! Kaeuffer, R., Peichel, C. L., Bolnick, D. I. and Hendry, A. P., 2012 parallel and

nonparallel aspects of ecological, phenotypic, and genetic divergence across replicate

population pairs of lake and stream stickleback. Evolution, 66: 402–418. (DOI:

10.1111/j.1558-5646.2011.01440.x)

31.! Kaltz, O., Shykoff, J., 1998 Local Adaptation in Host–parasite Systems. Heredity 81,

361–370. doi:10.1046/j.1365-2540.1998.00435.x.

113! ! 32.! Kawecki, T. D., Ebert, D., 2004 Conceptual issues in local adaptation. Ecol. Lett. 7,

1225–1241.

33.! Kearn G.C., 1994 Evolutionary expansion of the monogenea. Int. J. Parasitol. 24,

1227-1271.

34.! Konijnendijk, N., Raeymaekers J. A. M., Vandeuren, S., Jacquemin L.,Volckaert, F.

A. M., 2013 Testing for Local Adaptation in the Gasterosteus–Gyrodactylus Host–

parasite System. Evol. Ecol. Res. 3, 489–502.

35.! Koskela, T., Salonen, V., Mutikainen, P., 2000 Local adaptation of a holoparasitic

plant, Cuscuta europaea: variation among populations. J. Evol. Biol. 13: 749–755.

36.! Lemoine, M., Doligez, B., Richner, H., 2012 On the Equivalence of Host Local

Adaptation and Parasite Maladaptation: An Experimental Test. Am. Nat. 179, 270–81.

(DOI:10.1086/663699.)

37.! Lively, C. M., 1999 Migration , Virulence , and the Geographic Mosaic of

Adaptation by Parasites. Am. Nat. 153, S34-S47.

38.! Magurran, A.E. 2005. Evolutionary ecology the Trinidadian guppy. Oxford Univ.

Press, Oxford, UK

39.! Millar, N.P., Hendry, A. P.,2012 Population divergence of private and non-private

signals in wild guppies. Environ. Biol. Fishes 94, 513-525.

40.! Morgan, A. D., Gandon, S., Buckling, A., 2005 The Effect of Migration on Local

Adaptation in a Coevolving Host-Parasite System. Nature 437, 253–256.

(DOI:10.1038/nature03913.)

114! ! 41.! Oppliger, A., R. Vernet, Baez, M., 1999 Parasite local maladaptationin the Canarian

lizard Gallotia galloti (Reptilia: Lacertigae) parasitized by haemogregarian blood

parasite. J. Evol. Biol. 12, 951–955.

42.! Packer, C., Holt, R. D., Hudson, P. J., Lafferty, K. D., Dobson, A. P., 2003. Keeping

the herds healthy and alert: implications of predator control for infectious disease.

Ecol. Lett. 6, 797–802.

43.! Palkovacs E. P., Marshall M.C., Lamphere, B.A., Lynch, B.R., Weese, D.J., Fraser,

D.F., Reznick, D.N., Pringle, C.M., Kinnison, M.T., 2009. Experimental evaluation of

evolution and coevolution as agents of ecosystem change in Trinidadian streams.

Philos. Trans. R. Soc. B. 364:1617-1628.

44.! Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F., Scott, M. E., 2012. Are host–

parasite interactions influenced by adaptation to predators#? A test with guppies and

Gyrodactylus in experimental stream channels. Oecologia, 170, 77-88.

(DOI:10.1007/s00442-012-2289-9)

45.! R Core Development Team (2011) R: a language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna.

46.! Reznick, D. N., Bryga, H., 1987. Life-history evolution in guppies (Poecilia

reticulata): 1. Phenotypic and genetic changes in an introduction experiment.

Evolution 41, 1370-1385.

47.! Reznick, D. N., Endler, J.A., 1982 The Impact of Predation on Life History

Evolution in Trinidadian Guppies (Poecilia Reticulata). Evolution 36, 160-177.

48.! Reznick, D. N. 1982. The impact of predation on life history evolution in Trinidadian

guppies: genetic basis of observed life history patterns. Evolution 36:6:1236-1250.

115! ! 49.! Reznick, D. N., Rodd, F. H., Cardenas, M. 1996b. Life-History Evolution in Guppies

(Poecilia reticulata : Poeciliidae). IV . Parallelism in Life- History Phenotypes. Am.

Nat. 147,319–338.

50.! Reznick, D. N., Shaw, F. H., Rodd, F.H, Shaw, R. 1997a Evaluation of the Rate of

Evolution in Natural Populations of Guppies (Poecilia Reticulata). Science 275,

1934–1937.

51.! Reznick, D.N., Bryga, H.A., Endler, J.A., 1990. Experimetally Induced Life-History

Evolution in a Natural Population. Nature 346, 357–359.

52.! Reznick, D.N., Butler, M.J. IV, Rodd, F.H., Ross, P., 1996a. Life-history evolution

in guppies (Poecilia reticulata) 6. Differential mortality as a mechanism for natural

selection. Evolution 50,1651-1660.

53.! Richards E. L., van Oosterhout C., Cable, J., 2010. Sex-specific differences in

shoaling affect parasite transmission in guppies. PLOS One 5, e13285.

(DOI:10.1371/journal.pone.0013285.)

54.! Rodd F. H., Reznick D. N., 1997. Variation in the demography of guppy populations:

the importance of predation and life histories. Ecology 78:2:405-418

55.! Rodd, F. H., Reznick, D. N., 1991. Life History Evolution in Guppies: III. The

Impact of Prawn Predation on Guppy Life Histories. Oikos 62, 13

(DOI:10.2307/3545440.)

56.! Roth, O., Keller, I., Landis, S. H., Salzburger, W., Reusch, T. B. H., 2012. Hosts are

ahead in a marine host–parasite coevolutionary arms race: Innate immune system

adaptation in pipefish Syngnathus typhie against Vibrio phylotypes. Evolution 66,

2528–2539. (DOI:10.5061/dryad.4h1q3qb2.)

116! ! 57.! Saarinen, M., Taskinen, J., 2005. Local adaptation in a crustacean parasite–

molluscan host interaction: a field experiment. Evol. Ecol. Res. 7, 1191–1199.

58.! Schelkle, B., Faria, P., Johnson, M.B., van Oosterhout, C., Cable, J., 2012. Mixed

infections and hybridisation in monogenean parasites. PLoS ONE 7 7: e39506.

59.! Schluter, D. 2000. The ecology of adaptive radiation. Oxford Univ. Press, Oxford,

UK.

60.! Scott, M.E., Anderson, R. M., 1984. The population dynamics of Gyrodactylus

bullatarudis (Monogenea) within laboratory populations of the fish host Poecilia

reticulata. Parasitology 89, 159-194.

61.! Sternberg, E. D., Li, H., Wang, R., Gowler, C., de Roode, J. C., 2013. Patterns of

Host-Parasite Adaptation in Three Populations of Monarch Butterflies Infected with a

Naturally Occurring Protozoan Disease: Virulence, Resistance, and Tolerance. Am

Nat 182, E235–E248. (DOI:10.1086/673442.)

62.! Suk, H.Y., Neff, B. D., 2009. Microsatellite Genetic Differentiation among

Populations of the Trinidadian Guppy. Heredity 102, 425–34.

(DOI:10.1038/hdy.2009.7.)

63.! Thompson, J. N., 1999. The evolution of species interactions. Science 284, 2116–

2118.

64.! Thompson, J. N., 1994. Specific hypotheses on the geographic mosaic of

coevolution. Am. Nat. 153, S1-S14.

65.! Torres-Dowdall, J., Handelsman., C.A., 2012. Local Adaptation and the Evolution

of Phenotypic Plasticity in Trinidadian Guppies (Poecilia Reticulata). Evolution 66,

3432–3443. (DOI:10.5061/dryad.84gf5.)

117! ! 66.! Torres-Mejia, R. M., 2011. Ecomorphology of Body Shape and Life History in two

Poeciliids (Gambusia spp. and Poecilia reticulata). UC Riverside: Biology

67.! van Oosterhout, C., Domino, A.J., Cummings, S.M., Blais, J., Barson, N.J.,

Ramnarine, I.W., Mohammed, R.S., Persad, N., Cable, J., 2006. Balancing genetic

drift, and genetic variation at the Major Histocompatibility Complex in two wild

populations of guppies (Poecilia reticulata). Evolution 60, 2562-2574.

68.! van Oosterhout, C., Harris, P.D., Cable, J., 2003. Marked variation in parasite

resistance between two wild populations of the Trinidadian guppy, Poecilia reticulata

(Pisces: Poeciliidae). Biol. J. Linn. Soc. 79, 645-651.

69.! van Oosterhout, C., Mohammed, R.S., Hansen, H., Archard, G.A., McMullan, M.,

Weese, D.J., Cable, J., 2007. Selection by parasites in spate conditions in wild

Trinidadian guppies (Poecilia reticulata). Int. J. Parasitol. 37, 805-812.

70.! Willing, E.M., Bentzen, P., van Oosterhout, c., Hoffmann, M., Cable, J., Breden, F.,

Weigel, D., Dreyer, C., 2010. Genome-Wide Single Nucleotide Polymorphisms

Reveal Population History and Adaptive Divergence in Wild Guppies. Mol. Ecol. 19,

968–84. (DOI:10.1111/j.1365-294X.2010.04528.x.)

118! ! FIGURES

Figure 1. A schematic of the experimental design and Gyrodactylus population dynamics when infectin each guppy population. Horizontal labels indicate the source of guppies, vertical labels indicate the source of the parasites used in the experiment. White squares represent allopatric combinations (guppies and Gyrodactylus from different locations) and gray squares represent sympatric combinations (guppies and Gyrodactylus from the same locations).

GUPPIES HP LP HP LP ARIPO ARIPO MARIANNE MARIANNE

5

3 HP ARIPO 1

5

3 LP ;abundance) 1 ARIPO

5

3 Gyrodactylus HP Gyrodactylus 1 MARIANNE

5 SQRT;(

3 LP 1 MARIANNE

2 8 16 24 2 8 16 24 2 8 16 24 2 8 16 24 Days

119! ! Figure 2. Gyrodactylus population dynamics on the four guppy host populations. Each figure represents the mean number of parasites on individual fish. Symbols represent guppy populations: Marianne HP (filled squares), Marianne LP (empty squares), Aripo

HP (filled circles) and Aripo LP (empty circles). Error bars represent standard errors.

Parasite infection dynamics on all guppy populations

Marianne LP Gyrodactylus Marianne HP Gyrodactylus a) b)

5 Marianne HP Aripo HP Aripo Low 2 Marianne LP

0 Mean intensity of infection Aripo LP Gyrodactylus Aripo HP Gyrodactylus c) d)

30

15 10

5 0

Mean intensity of infection 1 8 16 24 1 8 16 24 Days Days

120! ! Figure 3. LS Means for Gyrodactylus performance. a) mean number of parasites / fish / day when infecting the sympatric host vs. all allopatric hosts, b) mean duration of infection on sympatric vs all allopatric hosts c) mean number of parasites / fish / day according to the degree of similarity between the parasite strain and the guppy strain in the mesocosms. Marianne HP Gyro (filled squares), Marianne LP Gyro (empty squares),

Aripo HP Gyro (filled circles) and Aripo LP Gyro (empty circles). Error bars represent standard errors.

Gyrodactylus performance on sympatric vs. allopatric hosts a) b)

10 10 8 5 6 4 2 2 Length of infection (days) Mean intensity of infection Allopatric Sympatric Allopatric Sympatric Host-parasite combination Host-parasite combination Gyrodactylus performance on all hosts c) 25 Marianne.LP 15 Marianne.HP Aripo.HP 5 Aripo.LP

1

Mean intensity of infection Dif. Predation/ Same Predation/ Dif. Predation/ Same Predation/ Dif. Drainage Dif. Drainage Same Drainage Same Drainage Degree of similarity

121! ! Figure 4. LS Means for guppy performance when infected with sympatric vs. allopatric hosts. A) female guppy growth during the length of the experiment, b) number of embryos per female. Symbols represent guppy populations: Marianne HP (filled squares),

Marianne LP (empty squares), Aripo HP (filled circles) and Aripo LP (empty circles).

Error bars represent standard errors.

Guppy performance under allopatric vs. sympatric Gyrodactylus a) b) 6 0.08

0.06 4

0.04 2 (per female) Female growth (g) Number of embryos Allopatric Sympatric Allopatric Sympatric

Host-parasite combination Host-parasite combination

122! ! TABLES

Table 1. Statistical analysis for Gyrodactylus performance on sympatric vs. allopatric guppies populations. Analyses were performed using linear mixed effects model. P values and denominator degrees of freedom were obtained using Satterthwaite approximation for degrees of freedom.

Mean+intensity+of+ Duration+of+infection Explanatory+variables infection

Fixed+effects F!(D.F.) P F!(D.F) P

Gyrodactylus’!drainage!of!origin 49.065 <0.001 11.75 <0.001

!(1,!263.20) (1,!262.67)

Gyrodactylus’!predation!regime!of! 1.358! 0.244 3.451 0.064

origin (1,!263.20) (1,!262.67)

HostDparasite!combination 15.086! <0.001 5.008 0.026

(1,!260.96) (1,!260.99)

Gyrodactylus’!drainage!of!origin 4.063 0.044 1.980 0.160 X Gyrodactylus’!predation!regime!of! !(1,!263.20) (1,!262.67)

origin

Gyrodactylus’!drainage!of!origin 4.945! 0.027 12.702 <0.001 X HostDparasite!combination (1,!263.16) (1,!219.90)

Gyrodactylus’!predation!regime!of! 7.925! 0.005 0.708 0.400 origin X (1,!263.16) (1,!219.90) HostDparasite!combination

Gyrodactylus’!drainage!of!origin 3.06! 0.081 2.615 0.107 X Gyrodactylus’!predation!regime!of! (1,!263.16) (1,!219.90) origin X HostDparasite!combination

Random+effects Variance St.+Dev. Variance St.+Dev.

Guppy!population!(intercept) 0.1708 0.4133 0.6564 0.8102

Residual 0.4144 0.6435 5.9542 2.4401

123! !

Table 2. Statistical analysis for Gyrodactylus performance on guppy populations, according to their degree of similarity (Same drainage and same predation environment, same drainage and different predation environment, different drainage and same predation environment, different drainage and different predation environment). Analysis were performed using general linear model.

Variables Mean+intensity+of+infection

F!(D.F.) P

Gyrodactylus’!drainage!of!origin!(Marianne! 137.167 <0.001

vs!Aripo) (1,!256)

Gyrodactylus’!predation!of!origin! 15.473 <0.001

(1,!256)

Degree!of!similarity! 24.672 <0.001

(3,!256)

Gyrodactylus’!drainage!of!origin 2.045 0.153 X Gyrodactylus’!predation!regime!of!origin (1,!256)

Gyrodactylus’!drainage!of!origin 57.439 <0.001 X Degree!of!similarity (3,!256)

Gyrodactylus’!predation!regime!of!origin 4.289 0.005 X Degree!of!similarity (3,!256)

Gyrodactylus’!drainage!of!origin 4.723 0.0031 X Gyrodactylus’!predation!regime!of!origin (3,!256) X Degree!of!similarity

124! ! Table 3. Statistical analysis for guppy performance when infected with sympatric vs. allopatric Gyrodactylus. Analyses were performed using linear mixed effects models. P values and denominator degrees of freedom were obtained using Satterthwaite approximation for degrees of freedom.

Fixed+effects Female+growth Number+of+embryos Reproductive+allocation

F!(D.F) P F!(D.F.) P F!(D.F.) P

Guppy!drainage!of!origin 7.546 0.003 0.033 0.855 3.123 0.080

(1,!97.56) (1,!87.99) (1,!87.95)

Guppy!predation!of!origin 0.214 0.644 0.291 0.590 1.052 0.307

(1,!97.56) (1,!87.99) (1,!87.95)

HostDparasite! 4.043 0.047 2.574 0.111 0.380 0.539 combination (1,!97.56) (1,!87.99) (1,!87.95)

Guppy!drainage!of!origin 10.349 0.001 8.183 0.005 1.573 0.213

X (1,!97.56) (1,!87.99) (1,!87.95)

Guppy!predation!of!origin

Random+effects Variance St.+Dev. Variance St.+Dev. Variance St.+Dev.

Gyrodactylus!population 0.001 0.000 0.001 0.000 0.000 0.000

Residual 0.005 0.022 5.865 2.422 0.032 0.179

125! ! CONNECTING STATEMENT NO. 3

In the previous two chapters I investigated how guppy-Gyrodactylus interactions can be shaped by various ecological and evolutionary factors. This, of course, is not uni- directional. Parasites can often induce strong phenotypic and behaviorual changes in their hosts that can in turn influence other interacting species such as predators and competitors. This is also the case for Gyrodactylus.

As I explored in chapter 2 – and subsequently demonstrated in chapter 3 and 4,

Gyrodactylus-infected guppies, more often than not, show a strong reduction in growth.

Similarly, previous studies have demonstrated that behavioral traits like feeding rate and aggressiveness can be higher in infected individuals, compared to healthy ones. It is, thus, conceivable that Gyrodactylus could influence guppy interspecific interactions. However, how and to what extent? In the next chapter I answer this question by assessing the interaction between guppies and Rivulus hartii – a well known predator and competitor of guppies in Trinidad – in the presence and absence of Gyrodactylus.

126! !

CHAPTER 5

An experimental test of antagonistic effects of competition and parasitism on host performance in semi-natural mesocosms

Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and Scott, M. E. CONDITIONAL ACCEPTANCE, Oikos

We do not at present know what maintains the state of equilibrium between the different genera actually found in the natural communities analysed, but must postulate that there is some ecological condition that buffers or cuts down the effectiveness of competition between species separated by generic characters.

– CHARLES ELTON J Anim Ecol. 1946

127! ! ABSTRACT

The mechanisms by which parasites can mediate the interactions between species have received increased interest in recent years. Nonetheless, most research has focused on the role of shared parasites as mediators of interspecific competition. Here, we explore the relative effects of Gyrodactylus specialist ectoparasites of Trinidadian guppies (Poecilia reticulata) on competition between their host and juveniles of the killyfish Rivulus hartii.

In mesocosms that replicate natural streams, we exposed guppies to only competitors, to only parasites, to both parasites and competitors, or the absence of both. Consistent with previous studies, we found that female guppies grew significantly less where only

Gyrodactylus were present, and this was regardless of infection status or parasite load.

Surprisingly, this effect of Gyrodactylus on the growth of female guppies was greatly reduced when both parasites and competitors were present in the mesocosms. We conclude that guppies can mediate the effects of Gyrodactylus on competition with

Rivulus, by adaptively fine-tuning their phenotype when simultaneously facing multiple enemies.

128! !

INTRODUCTION

Every species coexists with many other species, and thus inevitably faces competition, predation, and parasitism (Hatcher et al., 2006). To date, however, ecological research examining interactions between these effects has focused mainly on predation and competition (Chase et al., 2002). Although increasing attention is being given to the influence of parasitism on predator-prey interactions (Lafferty and Morris 1996; Raffel et al., 2010; Marino and Werner 2013; Marino et al., 2014), the interaction between parasitism and competition has been largely neglected (Hatcher et al. 2006; Pedersen and

Fenton 2007). An important exception is “apparent competition” where two species may have deleterious effects on each other through shared parasites (Holt 1977; Hudson and

Greenman 1998). In the present paper, we will consider a different type of interaction between parasitism and competition: how a non-shared parasite can influence competition through its effects on the host.

Specialist parasites, evidently, can only modify competitive interactions through their effects on their particular host (Hatcher et al., 2006), but are also likely to be strongly influenced by the effect of the competitor on its host, as they greatly depend on host density. Theoretical and empirical studies have shown that the presence of competitors can reduce parasite transmission by reducing the host population density

(Keesing et al., 2006; Brunner et al., 2008; LoGiudice et al., 2008; Hall et al., 2009) or can increase transmission by enhancing host-parasite encounters owing to differential use of space, increased activity, and/or foraging behavior (Hatcher et al., 2006). In such

129! ! cases, a host’s response to competition might reduce its capacity to deal with parasites, but at the same time, parasite-induced changes in host traits can negatively impact the ability of the host to deal with competitors. Depending on the risk and fitness costs associated with each of these threats (i.e. relative effects), hosts should show a stronger response to the more costly interaction (Reylea, 2004; Raffel et al., 2010). These reciprocal effects of competition influencing parasitism, and parasitism influencing competition, can modify the net effect of both interactions (i.e. interactive effects), yet it has been largely under-explored in experimental research (Raffel et al., 2010).

Understanding the relative importance of parasites, in the realm of the interactions to which the host is exposed, is of utmost importance (Price et al. 1986; Minchella and

Scott, 1991; Hatcher et al., 2006; Dunn et al., 2012), particularly because both density- and trait-mediated effects can affect competitive and consumer-resource relationships, and propagate through trophic levels within the community (Dunn et al., 2012).

In this paper, we explore the relative and interactive effects of parasitism and competition by comparing fish growth in the presence of specialist parasites and/or competitors. We do so in artificial streams that replicate natural streams, and we take advantage of a well-known study system: the Trinidadian guppy (Poecilia reticulata), its competitor Rivulus hartii, and a guppy-specific ectoparasites of the genus Gyrodactylus.

Empirical system

The upstream reaches of rivers in Trinidad are commonly inhabited by only guppies and

Rivulus – because upstream migration of most fishes is prevented by a series of

130! ! waterfalls. Adult Rivulus (maximum total length ~100 mm) are much larger than adult guppies (maximum total length ~ 45 mm) and are strict predators, foraging mainly on invertebrates and small fish, including juvenile guppies (Gilliam et al., 1992; Mattingly and Butler 1994; Fraser et al., 1999). Juvenile Rivulus, on the other hand, are of similar size as guppies, and directly compete with guppies for shelter and food (i.e. aquatic invertebrates) (Dussault and Kramer, 1981; Gilliam et al. 1993; Palkovacs et al., 2009).

In addition to these effects of Rivulus on guppies, reciprocal effects also occur. For example, the presence of guppies decreases the growth rate of juvenile Rivulus (Gilliam et al., 1993) – through resource competition – but dramatically increases the growth rate of adult Rivulus (Walsh et al., 2011) through guppy predation on Rivulus young, and the release of adult Rivulus from intra-specific competition (Walsh et al., 2011; Fraser and

Lamphere 2013). An indirect effect, the importance of which will become clear later, is that both guppies and Rivulus impose strong selection on each other for rapid juvenile growth so as to avoid inter-specific predation (Seghers 1973; Seghers 1974; McKellar and Hendry, 2011; Gosline and Rodd 2007).

The above guppy-Rivulus interactions might be strongly influenced by the specialist monogenean ectoparasites Gyrodactylus, that complete their life cycle on guppy hosts (Kearn 1994; Cable and Harris, 2002; Harris et al., 2004). In Trinidad, three species of Gyrodactylus are known to infect guppies: G. poeciliae, G. turnbulli, and G. bullatarudis (Harris and Lyles 1992; van Oosterhout et al., 2003; Xavier et al., 2015).

The genus is characterized by an extreme progenesis and hyper-viviparity: adults give birth to fully-grown offspring that in turn contain a developing embryo in utero (Cohen

131! ! 1977). Newborn Gyrodactylus are fully-grown and directly attach to the host, and feed on host mucus and epithelial cells in a manner similar to other surface-browsing monogeneans (Kearn 1999). Transmission between guppy hosts occurs through direct contact and infections can cause high guppy mortality in the laboratory (Scott and

Anderson 1984; Cable and van Oosterhout 2007) and in nature (van Oosterhout et al.,

2007). Sub-lethal effects also occur: guppies exposed to Gyrodactylus show reduced foraging behavior (van Oosterhout et al., 2003), and reduced growth perhaps due to a potential reallocation of resources to immune responses (Pérez-Jvostov et al., 2012).

For the above reasons, the guppy-Gyrodactylus-Rivulus system is well suited for testing the role of non-shared parasites in mediating interactions between competing species, a subject that has been largely unexplored experimentally. We predict that

Gyrodactylus will modify guppy-Rivulus competitive interaction in favor of Rivulus through its detrimental effects on guppy growth and behavior. Thus, our objective is to quantify the relative and interactive effects of parasitism and guppy-Rivulus competitive interactions on the performance (i.e. growth) of both fish species. Although differences in virulence have been reported for G. turnbulli and G. bullatarudis in one river (van

Oosterhout et al., 2003), guppy immune response seems not to be species-specific

(Richards and Chubb, 1996).

132! ! MATERIALS AND METHODS

Fish collection and treatment

In July 2013, we collected mature guppies (17 – 40 mm) and juvenile Rivulus (20 – 45 mm) from the Paria River in Trinidad (P7 in Millar et al., 2006) and transported them in 2

L containers to the laboratory in Trinidad where they were scanned for Gyrodactylus infections using a dissecting microscope. Infected and uninfected fish were then separated into species, sex and size specific groups (small, medium and large). All guppies, regardless of infection status, were then treated for Gyrodactylus infections with an application of N-cyclopropyl-1,3,5-triazine-2,4,6-triamine (cyromazine; Lice And

Anchor Worm Treatment, Ecological Laboratories Inc.) – some infected guppies were set aside for use as a source of infection for the mesocosm experiments. The successful elimination of Gyrodactylus was verified four days later by scanning all guppies using a dissecting microscope. Gyrodactylus infections have been reported to survive for ~ 5 hrs on Rivulus (King and Cable 2007; King et al., 2009; Cable et al., 2013), so all Rivulus were also treated.

Two days prior to the beginning of the experiment, all collected Rivulus and guppies were anaesthetized using 0.02% Tricaine Methanesulfonate (Finquel MS-222 from Fisher Canada) (1: 8000) buffered to a neutral pH using NaHCO3, weighed (nearest mg), measured (standard length to the nearest mm), and given a distinct intra-dermic mark using an elastomer dye (Northwest Marine Technology Inc.). This marking procedure is standard for guppies and has been used in many studies within minimal mortality (Weese et al., 2010; Bassar et al., 2010; Pérez-Jvostov et al., 2012).

133! ! The experiment

The experiment was performed in experimental stream channels (mesocosms) that have been used in previous work on guppies (e.g. Palkovacs et al., 2009; Bassar et al., 2010;

Pérez-Jvostov et al., 2012). These channels (0.5 m wide by 3 m long by 0.2 m deep) received flowing water from a tributary to the Arima River that had neither guppies nor

Rivulus, and they were covered with netting to prevent bird predation. Two weeks prior to introduction of fish, river gravel was added to the channels making them available for natural colonization by the invertebrates and algae that provide food for both guppies and

Rivulus.

The experiment consisted of three replicates of each of five experimental treatments: guppies only (GO); guppies and Gyrodactylus (GG); guppies, Gyrodactylus, and Rivulus (GGR); guppies and Rivulus (GR); and Rivulus only (RO). Single species treatments (RO, GO, GG) consisted of ten randomly selected Rivulus (six small, two medium, and two large Rivulus: RO) or twelve randomly selected guppies (four large males, four medium females, and four large females: GO and GG). The largest Rivulus used in the experiment measured 45 mm and the largest female guppy measured 39 mm.

Mixed species treatments (GR and GGR) consisted of five Rivulus and two male and four female guppies in the same size distribution as in single species treatments and typical in guppy-Rivulus experiments (Palkovacs et al., 2009). In addition, to minimize any potential familiarity between fish, we specifically ensured that fish that were kept together prior to the experiment were introduced into different mesocosms.

134! ! In the parasite treatments (GG and GGR), one of the male guppies and one of the female guppies used for each mesocosm was selected from the group of fish collected from the field but not treated (see above). Each of these guppies had only 2-3

Gyrodactylus located only on the caudal fin. This increased the likelihood that only one species of Gyrodactylus (presumably G. turnbulli based on location) was used in the experiment (Harris, 1989; Harris and Lyles, 1992).

The mesocosms were checked daily for dead fish which were immediately removed and identified based on their elastomer marks. After 20 days, all remaining fish were collected, identified, and weighed, and all Gyrodactylus were counted by scanning guppies using a dissecting microscope.

Statistical analysis

Growth was calculated as the difference in mass as a proxy for performance, and was analyzed with Generalized Linear Mixed Models (GLMMs) with post-hoc Tukey’s honestly significant difference (HSD) tests to examine pairwise differences between treatments. The model included treatment, sex and their interaction as fixed factors; replicate (1 to 3) was nested as a random factor within treatment. Juvenile Rivulus cannot be differentiated based on sex, and so this factor was removed from the Rivulus model.

Starting mass of individual fish was included in all initial models but was later removed owing to non-significance. These analyses were used to address three key questions.

First, we evaluated the effect of competition by comparing the growth (change in mass) of guppies and of Rivulus in the competition treatment (GR) with growth in the respective

135! ! single species treatments (GO and RO). Second, we examined the effect of Gyrodacytlus infection on guppies in the absence of Rivulus by comparing the growth of guppies between treatments with and without parasites (GG vs. GO). Third, we examined the interactive effects of parasitism and competition by comparing growth of guppies and

Rivulus between treatments with and without Gyrodactylus (GR vs. GGR; GG vs. GGR).

Analyses were conducted in R version 2.14.1 (R Core Development Team 2011) using the nlme package, and the multcomp package for paired-wise comparisons in

Generalized Linear Mixed Effects Models (GLMMs). All the levels of significance were set at P < 0.05.

RESULTS

Overall mortality of guppies (7% ‒ all dead guppies were found) and Rivulus (10% of all

Rivulus were found dead and an additional 6.6% were missing and presumed dead at the end of the experiment) was low. Gyrodactylus infections persisted to the end of the experiment in all GG and GGR mesocosms, reaching 33.5 % prevalence in the GGR treatments and 25% in the GG treatments (Table 1).

Guppy performance (i.e. growth) was the highest in the absence of Rivulus and

Gyrodactylus (Fig. 1). We detected no evidence that interspecific competition affected growth of either Rivulus (RO = GR) (F2,4= 0.703, P= 0.547) or guppies (GO = GR) (Z=

1.922, P= 0.219), although a trend for less growth of guppies was evident (Fig. 1; Table

3). However, growth of guppies was negatively influenced by Gyrodactylus (Fig. 1).

136! ! This parasite-induced depression in growth was evident for female guppies in the presence of Rivulus (GGR 32% less than GO), and most dramatic in the absence of

Rivulus (GG 75% less than GO) (Fig. 1; Table 3). In contrast to guppy females, male guppy growth did not differ among any treatments (Fig. 2; Table 3).

DISCUSSION

Our goal was to quantify the effects of parasitism and guppy-Rivulus interactions on the growth of both fish species – with particular attention being paid to the relative importance of parasitism and competition and the nature of any interactions between them. Although no effects were detected for Rivulus or for male guppies, the presence of

Gyrodactylus parasites decreased female guppy growth, and this effect was much stronger than the effect of Rivulus (Fig. 1). We also found a very strong interaction:

Gyrodactylus reduced female guppy growth in the absence but not the presence of

Rivulus. In short, the relative effect of Gyrodactylus on the growth of female guppies was much greater than that of a competitor (and potential predator), but the two effects were strongly interactive. These results generate several important insights into the nature of guppy-Gyrodactylus-Rivulus interactions and, more generally, food web interactions.

The coexistence between guppies and Rivulus has been commonly viewed as a balance between predation and competition, with guppies being the better competitors

(Gilliam et al., 1993), but large adult Rivulus actively preying upon juvenile guppies

(Fraser and Lamphere, 2013). So why was Rivulus unaffected by the better competitor in this experiment? Although we did find a trend for a decrease in growth of Rivulus in the

137! ! presence of size-matched guppies, this was not significantly different from the Rivulus- only control, and similar results have been reported in a previous mesocosms experiments

(Palkovacs et al., 2009), where changes in biomass did not differ between mixed and single species treatments. It is possible that under these experimental conditions competition is lessened due to relatively low fish density per mesocosms; however, an alternative possibility is that Rivulus grow larger than guppies and shift their diet towards terrestrial prey that are too large for guppies to eat, releasing them from resource competition (Fraser and Lamphere, 2013). Indeed, at the end of the experiment Rivulus in the mixed-species treatment were almost three times larger than female guppies, despite being of similar size as the largest females in the mesocosms at the beginning of the experiment. Our results, thus, support the notion that guppies compete with size-matched

Rivulus until these are large enough to feed on alternative prey types, and avoid resource competition.

Why might the growth of female guppies be lower in infected compared with uninfected mesocosms? It is important to point out that the parasite-induced reduction in female guppy growth observed here was consistent with our previous mesocosms experiment of eight guppy populations (Pérez-Jvostov et al., 2012). In that experiment, the reduction in female growth was not influenced by parasite load or infection status, but rather by the presence of Gyrodactylus. We suggest that guppies exposed to Gyrodactylus face an energetic trade-off between growth and the activation and the maintenance of the immune response – as has been frequently documented in birds and mammals (Rauw

2012). Crucially, the reduction in guppy growth is not due to any pathological effects of

138! ! the infection, but rather an immunoprophylactic response in which guppies invest in resistance at the expense of growth. This would also help explain the strong reduction in growth despite the small infection levels observed at the end of the experiment in the GG treatment (Table 1), and reinforce the notion that it is not the pathology commonly associated with high Gyrodactylus infection levels, but rather a phenotypic response of the host.

How did the interactive effect arise, whereby female guppies exposed to

Gyrodactylus grew better in the presence of a competitor? We suggest two potential mechanisms. Gyrodactylus transmission is positively density-dependent, so the per-capita risk of infection should be higher at higher guppy densities (Anderson and May, 1981).

Thus, it is predicted that guppies will increase their investment in Gyrodactylus resistance mechanisms as population density increases. Under this scenario, differences in guppy density between treatments could explain the strong reduction in guppy growth in the GG

(12 guppies) compared to the GGR treatment (6 guppies), as the risk of infection is larger in the higher-density GG treatment. This density-dependent-prophylaxis has been extensively reported in insects (Wilson and Cotter, 2009), but whether it plays an important role in vertebrates is still largely unknown (Sadd and Schmid-Hempel, 2009).

Alternatively, intraguild predation provides a more feasible explanation for the observed differences in female growth between treatments. As an apparently adaptive response to reduce Rivulus predation on juvenile guppies, these increase their growth rate when exposed to chemical cues from adult Rivulus (Gosline and Rodd 2008). Guppies might thus show a phenotypic response to Rivulus as a potential predator: although the Rivulus

139! ! in our experiment were not large enough to eat the guppies, the presence of small Rivulus is presumably a reliable cue of the likely presence of larger Rivulus. If guppies increased their growth in response to chemical cues signaling the presence of Rivulus, this would have partially counteracted the negative effects of parasitism on guppy growth, consistent with our observation that guppy growth in the presence of Gyrodactylus and Rivulus was intermediate between guppy-only and guppy-Gyrodactylus treatments.

An interesting result worth exploring is the observed difference in the effect of

Gyrodactylus between male and female guppies, where only females showed a strong decrease in growth in the presence of Gyrodactylus. This could simply be related to behavioral differences that could indirectly reduce the cost of infection. For example, in the wild females tend to school more and invest more time in feeding behavior, which may increase their susceptibility to Gyrodactylus infections (Richards et al., 2010).

Indeed, female guppies in the wild tend to be found more commonly infected than males, with the largest females usually being the most infected ones (Gotanda et al., 2013).

Although behavioral differences would help to explain differences in infection levels in the wild, they do not explain the strong reduction in growth in females but not males in our experiment. It has been recently shown that females in low predation localities – such as the one used here – have 1.6 times longer life span than do males, independently of any extrinsic source of mortality (Arendt et al., 2014). This short male-life span may favor the evolution of higher resistance – given that they have little to no time to waste fighting Gyrodactylus infections. Indeed, Dargent et al. (2014) have recently shown that male guppies are overall more resistant to Gyrodactylus, than are females. Thus, the

140! ! observed reduction of growth rate in females, but not in males, seems to be directly related to the intrinsic cost of infection and their reduced capacity to fight Gyrodactylus infections (i.e. reduced tolerance and resistance).

A natural question that rises from our results is whether Gyrodactylus could change the dynamics and structure of the broader community. Natural communities are subject to different environmental conditions that might exacerbate Gyrodactylus effects. For example, during the rainy season in Trinidad, heavy flooding could make it more difficult for guppies to find the necessary resources to fight an infection. During these periods infected guppies are more easily swept downstream (van Oosterhout et al., 2007), and this may drastically reduce guppy population density. It is thus possible that the

Gyrodactylus effects reported here are lower than those during strong flooding events. In addition, if differential foraging, or even differential dietary preference of guppies for higher quality items, to fight the infection, Gyrodactylus could have strong top-down effects, and potentially influence the structure and composition of lower trophic levels

(i.e. invertebrates and algae). Finally, we used guppies and Rivulus that have coevolved in the presence of Gyrodactylus; it would be interesting to test how Gyrodactylus can influence guppy-Rivulus competition and the broader community when neither of the species has previously encountered Gyrodactylus, and it would provide to be useful to do so in the wild.

Our results have demonstrated that host phenotypic plasticity plays an important role in mediating the interaction between hosts, parasites and competitors/ predators, and

141! ! contrast with those of previous studies where parasites were the main drivers of interspecific interactions through their effects on host phenotype (Werner and Peacor

2003; Hatcher et al., 2006). Nonetheless, in most of these studies parasites had synergistic effects with predators and competitors, whereas in our experiment parasitism and competition had antagonistic effects: increased growth in the presence of Rivulus, and decreased growth in the presence of the parasite. Such antagonistic effects have also been reported for amphibians (Reviewed in Reylea 2007). For example, tadpoles of Bufo americanus delayed development in the presence of an echinostome-infected snail, but accelerated their development in response to a caged newt predator (Notophthalmus viridescens) (Raffel et al., 2010). However, to our knowledge, this is the first time that this has been reported in a fish system. Our results thus add to the growing evidence that host phenotypic plasticity in response to multiple enemies is fine-tuned to balance the opposing phenotypic optima. If we ought to better understand how host respond to multiple threats simultaneously, it is critical that we appreciate the natural complexity in which organisms live and how phenotypically plastic responses to one source of mortality may be highly contingent on another.

142! ! REFERENCES

1.! Anderson, R. M. and May, R. M. 1981. The population dynamics of

microparasites and their invertebrate hosts. – Phil. Trans. R. Soc. B: Biol. Sci.

291: 1054, 451-524.

2.! Arendt, J. D. et al. 2014. Replicated origin of female-biased adult sex ratio in

introduced populations of the Trinidadian guppy (Poecilia reticulata). – Evolution

68: 2343–2356. doi: 10.1111/evo.12445

3.! Bassar, R. D. et al. 2010. Local adaptation in Trinidadian guppies alters

ecosystem processes. – Proc. Nat. Acad. Sci. 107: 8, 3616-3621.

4.! Brunner, J. L. and Ostfeld, R. S. 2008. Multiple causes of variable tick burdens on

small-mammal hosts. – Ecology 89: 8, 2259.

5.! Cable, J. and Harris, P. D. 2002. Gyrodactylid developmental biology: historical

review, current status and future trends. – Int. J. Parasitol. 32, 255-280.

6.! Cable J. and van Oosterhout, C. 2007. The role of innate and acquired resistance

in two natural populations of guppies (Poecilia reticulata) infected with the

ectoparasite Gyrodactylus turnbulli. – Biol. J. Linn. Soc. 90, 647-655.

7.! Cable, J. et al. 2013. Can parasites use predators to spread between primary hosts?

– Parasitology 140, 1138-1143. Doi:10.1017/S003118201300067X.

8.! Chase, J. M. et al. 2002. The interaction between predation and competition: a

review and synthesis. – Ecol. Lett. 5: 302–315. doi: 10.1046/j.1461-

0248.2002.00315.x

143! ! 9.! Dargent, F. et al. 2013. Experimental elimination of parasites in nature leads to

the evolution of increased resistance in hosts. – Proc. R. Soc: Biol. Sci. 280: 1773,

20132371. DOI: 10.1098/rspb.2013.2371

10.!Dunn, A. M. et al. 2012. Indirect effects of parasites in invasions. – Func. Ecol.

26: 1262–1274. doi: 10.1111/j.1365-2435.2012.02041.x

11.!Dussault, G. V. and Kramer, D. L. 1981. Food and feeding behavior of the guppy,

Poecilia reticulata (Pisces: Poeciliidae). – Can. J. Zool. 59:684-701.

12.!Fraser, D. F. and Lamphere, B. A. 2013. Experimental evaluation of predation as

a facilitator of invasion success in a stream fish. – Ecology 94: 640–649.

13.!Fraser, D. F. et al. 1999. Habitat quality in a hostile river corridor. – Ecology 80,

597–607.

14.!Gilliam J. F. et al. 1993. Structure of a tropical stream fish community: a role for

biotic interactions. – Ecology 74, 1856–1870. doi:10.2307/1939943.

15.!Gosline, A. K. and Rodd, F. H. 2007. Predator-induced plasticity in guppy

(Poecilia reticulata) life history traits. – Aquatic Ecol. 42: 4, 693–699.

doi:10.1007/s10452-007-9138-7.

16.!Gotanda, K. M. et al. 2013. Adding parasites to the guppy-predation story:

insights from field surveys. – Oecologia 172: 1, 155-166.

17.!Hall, S. R. 2009. Friendly competition: evidence for a dilution effect among

competitors in a planktonic host–parasite system. – Ecology 90, 791–801.

18.!Harris, P. D. 1989. Interactions between population growth and sexual

reproduction in the viviparous monogenean Gyrodactylus turnbulli Harris, 1986

from the guppy, Poecilia reticulata Peters. – Parasitology 98: 02, 245-251.

144! ! 19.!Harris, P. D. and Lyles, A. M. 1992. Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. – Int. J.

Parasitol. 78: 5, 912-4.

20.!Harris, P. D. et al. 2004. Nominal species of the genus Gyrodactylus von

Nordmann 1832 (Monogenea: Gyrodactylidae), with a list of principal host

species. – Systematic Parasitol. 59:1, 1-27.

21.!Hatcher, M. J. et al. 2006. How parasites affect interactions between competitors

and predators. – Ecol. Lett. 9, 1253–1271. doi: 10.1111/j.1461-

0248.2006.00964.x

22.!Holt, R. D. 1977. Predation, apparent competition, and the structure of prey

communities. – Theo. Pop. Biol. 12, 197-229.

23.!Hudson, P. and Greenman, J. 1998. Competition mediated by parasites: biological

and theoretical progress. – Trends Ecol. Evol. 13:10, 387–390.

24.!Kearn G. C. 1994. Evolutionary expansion of the monogenea. – Int. J.Parasitol.

24, 1227-1271.

25.!Keesing, F. 2006. Effects of species diversity on disease risk. – Ecol. Lett. 9: 4,

485-498.

26.!King, T. A. and Cable, J. 2007. Experimental infections of the monogenean

Gyrodactylus turnbulli indicate that it is not a strict specialist. – Int. J. Parasitol.

37, 663–672.

27.!King, T. A. et al. 2009. Experimental infections with the tropical monogenean,

Gyrodactylus bullatarudis: potential invader or experimental fluke? – Int. J.

Parasitol. 58, 249–254.

145! ! 28.!Kolluru et al. 2009. Food availability and parasite infection influence mating

tactics in guppies (Poecilia reticulata). – Behav. Ecol. 20: 1, 131–137.

29.!Lafferty, K. D. and Morris, A. K. 1996. Altered behavior of parasitized killifish

increases susceptibility to predation by bird final hosts. – Ecology 77, 1390-1397.

30.!LoGiudice, K. et al. 2008. Impact of host community composition on lyme

disease risk. –Ecology 89: 10, 2841.

31.!Marino, J. A. 2014. Predators and trematode parasites jointly affect larval anuran

functional traits and corticosterone levels. – Oikos 123: 4, 451–460.

32.!Marino, J. A. and Werner, J. E. E. 2013. Synergistic effects of predators and

trematode parasites on larval green frog (Rana clamitans) survival. – Ecology

94:12, 2697-2708.

33.!Mattingly, H. T. and Buttler, M. J. IV. 1994. Laboratory predation on the

Trinidadian guppy: implications for the size-selective hypothesis and guppy life-

history evolution. – Oikos 69: 54–64.

34.!McKellar, A. E. and Hendry, A. P. 2011. Environmental factors influencing adult

sex ratio in Poecilia reticulata: laboratory experiments. – J. Fish Biol. 79: 4, 937–

53. Doi:10.1111/j.1095-8649.2011.03065.x

35.!Millar, P. et al. 2006. Disentangling the selective factors that act on male colour in

wild guppies. – Oikos 113: 1–12.

36.!Minchella, D. J. and Scott, M. E. 1991. Parasitism: A cryptic determinant of

animal community structure. – Trends Ecol. Evol. 6, 8: 250–4. Doi:10.1016/0169-

5347(91)90071-5

146! ! 37.!Palkovacs, E. P. et al. 2009. Experimental evaluation of evolution and coevolution

as agents of ecosystem change in Trinidadian streams. – Philos. Trans. R. Soc. B.

364:1617-1628.

38.!Pedersen, A.B. and Fenton, A. 2007. Emphasizing the ecology in parasite

community ecology. –Trends Ecol. Evol. 22: 3, 133–9.

Doi:10.1016/j.tree.2006.11.005

39.!Pérez-Jvostov, F. et al. 2012. Are host – parasite interactions influenced by

adaptation to predators ? A test with guppies and Gyrodactylus in experimental

stream channels. – Oecologia 170: 1, 77-88. Doi:10.1007/s00442-012-2289-9.

40.!Pérez-Jvostov, F. et al. 2015. Local adaptation in the context of a mosaic of

coevolution: A test with guppies and Gyrodactylus. – Int. J. Parasitol.

41.!Price, P. et al. 1986. Parasite mediation in ecological interactions. – Ann. Rev.

Ecol. System 1: 7, 487-505.

42.!Raffel, T. R. et al. 2010. Parasitism in a community context: trait-mediated

interactions with competition and predation. – Ecology 91: 7, 1900–7.

43.!Rauw, W. M. 2012. Immune response from a resource allocation perspective. –

Front. Genet. 3, 267. Doi:10.3389/fgene.2012.00267.

44.!Reylea, R. 2002. Competitor-induced plasticity in tadpoles: consequences, cues,

and connections to predator-induced plasticity. – Ecol. Monograph. 72: 4, 523–

540.

45.!Reylea, R. 2007. Getting out alive: how predators affect the decision to

metamorphose. – Oecologia 152: 389-400.

46.!Reznick, D. N. 1982. The impact of predation on life history evolution in

147! ! Trinidadian guppies: genetic basis of observed life history patterns. – Evolution

36: 6, 1236–1250.

47.!Richards, E. L. et al. 2010. Sex-specific differences in shoaling affect parasite

transmission in guppies. – PLoS One 5:10, e13285.

48.!Richards, G. R. and Chubb, J. C. 1996. Host response to initial and challenge

infections, following treatment, of Gyrodactylus bullatarudis and G. turnbulli

(Monogenea) on the guppy (Poecilia reticulata). – Parasitol. Res. 82: 3, 242–247.

49.!Scott, M. E. and Anderson, R. M. 1984. The population dynamics of

Gyrodactylus bullatarudis (Monogenea) within laboratory populations of the fish

host Poecilia reticulata. – Parasitology 89: 159–194.

50.!Seghers, B. H. 1973. An analysis of geographic variation in the antipredator

adaptations of the guppy, Poecilia reticulata. Ph.D. thesis, University of British

Columbia, Vancouver.

51.!Seghers, B. H. 1974. Schooling behavior in the guppy (Poecilia reticulata): an

evolutionary response to predation. – Evolution 28, 486–489.

52.!van Oosterhout, C. et al. 2003. Marked variation in parasite resistance between

two wild populations of the Trinidadian guppy, Poecilia reticulata (Pisces:

Poeciliidae). – Biol. J. Linnean. Soc. 79, 64

53.!van Oosterhout, C. et al. 2007. Selection by parasites in spate conditions in wild

Trinidadian guppies (Poecilia reticulata). – Int. J. Parasitol. 37:7, 805-812.

54.!Walsh, M. R. et al. 2011. The direct and indirect effects of guppies: implications

for life-history evolution in Rivulus hartii. – Func. Ecol. 25:1, 227–237.

Doi:10.1111/j.1365-2435.2010.01786.x.

148! ! 55.!Weese, D. J. et al. 2010. Spatiotemporal variation in linear natural selection on

body color in wild guppies (Poecilia reticulata). – Evolution 64, 1802-181.

56.!Werner, E. E. and Peacor, S. D. 2003. A review of trait-mediated indirect

interactions in ecological communities. – Ecology 84:1083–1100.

57.!Wilson, K. and Cotter, S.C. 2009. Density-dependent prophylaxis in insects. In

Ananthakrishnan, T.N., Whitman, T.W., Editors, Phenotypic plasticity of insects:

mechanisms and consequences. – Enfield: Science Publ Inc. p. 381-420.

149! ! TABLES

Table 1. Descriptive statistics. Treatment abbreviations are as follows: GO, guppy-only;

GG, guppy-Gyrodactylus; GR, guppy-Rivulus; GGR, guppy-Gyrodactylus-Rivulus; RO,

Rivulus-only. Mixed species treatments consisted of 4 guppy females, 2 guppy males and

5 Rivulus. The number of fish was doubled in the single species treatments. Initial and final parasite prevalence (percentage of infected individuals in the mesocosm) and abundance (total number of Gyrodactylus per mesocosms) were determined on days 1 and 20 of the experiment, respectively.

Initial Final Initial Final Initial Final Rivulus Guppy Treatment Replicate prevalence prevalence mean mean parasite parasite mortality mortality (%) (%) intensity intensity abundance abundance GO 1 ------0/12

(12 guppies) 2 ------0/12

3 ------0/12

GG 1 16.66 25 1.5 2 3 6 - 0/12

(12 guppies) 2 16.66 8.3 2 3 4 3 - 4/12

3 16.66 25 1 1.66 2 5 - 0/12

GGR 1 33.33 33.3 1.5 3 3 6 1/5 0/6 (6 guppies, 5 2 33.33 16.66 2 5 4 5 0/5 1/6 Rivulus) 3 33.33 33.3 1.5 1.5 3 3 2/5 0/6

GR 1 ------0/5 1/6 (6 guppies, 5 2 ------0/5 0/6 Rivulus) 3 ------2/5 1/6

RO 1 ------2/10 -

(10 Rivulus) 2 ------1/10 -

3 ------2/10 -

150! ! Table 2. Statistical analysis for guppy growth across the different treatments (GO, guppy- only; GG, guppy-Gyrodactylus; GGR, guppy-Gyrodactylus-Rivulus; GR, guppy-Rivulus;

RO, Rivulus-only). Analyses were performed using linear mixed effects models with replicate nested in treatment.

F P Explanatory variable D.F.

<0.001 Gender 1, 85 47.122 0.005 Treatment 3, 6 12.434 0.002 Gender X Treatment 3, 85 5.040

151! ! Table 3. Specific Tukey’s HSD Contrasts for female and male guppies. GO, guppy-only;

GG, guppy-Gyrodactylus; GGR, guppy-Gyrodactylus-Rivulus; GR, guppy-Rivulus; RO,

Rivulus-only.

Females P Contrast Estimate S.E. Z <0.001 GG vs. GO -0.162 0.023 -7.125 <0.001 GG vs. GR -0.108 0.028 -3.799 0.005 GG vs. GGR -0.092 0.028 -3.341

GO vs. GR 0.054 0.028 1.922 0.219 0.052 GO vs. GGR 0.070 0.027 2.556

GR vs. GGR 0.016 0.032 0.488 0.962

Males P Contrast Estimate S.E. Z

GG vs. GO -0.007 0.033 -0.209 0.997

GG vs. GR -0.035 0.042 -0.836 0.837

GG vs. GGR 0.002 0.042 0.040 0.993

GO vs. GR -0.028 0.041 -0.692 0.900

GO vs. GGR 0.009 0.041 0.210 0.997

GR vs. GGR 0.037 0.049 0.759 0.873

152! ! FIGURES

A

AB B

C

Figure 1. Least Square means in growth for a) female and male guppies, and b) Rivulus pooled across genders. The five treatments as shown are: GO, guppy-only; GG, guppy-

Gyrodactylus; GR, guppy-Rivulus; GGR, guppy-Gyrodactylus-Rivulus; RO, Rivulus- only. Error bars represent standard errors.

153! ! CONNECTING STATEMENT NO. 4

The functional importance of parasites in community structure has received increased interest in the last two decades, and it is now commonly acknowledged that parasites can dramatically alter community dynamics through their effects on their host ecology. In the previous chapter, I approached this experimentally, by comparing the outcome of competition between Rivulus and guppies in the presence and absence of Gyrodactylus.

The results were consistent with those from Chapter 3, where the sole exposure to

Gyrodactylus resulted in a strong decrease in female growth. Interestingly, this effect was smaller when both Gyrodactylus and Rivulus were present. However, when comparing both effects independently, I found that parasitism had the strongest negative effect on guppy growth.

The following chapter is a natural extension of Chapter 5. The main limitation of experimental studies lies in that they rarely fully replicate natural conditions. Although experimental approaches like in the previous chapters are important in understanding the mechanisms behind particular processes, their extrapolation to natural communities is, many times, unfeasible. In Chapter 6, thus, I explore the functional importance of

Gyrodactylus in community structure directly in the wild, by introducing Gyrodactylus into two previously Gyrodactylus-free guppy populations. I quantify the effects of

Gyrodactylus on guppy and Rivulus demography, as well as assess whether these effects are predominantly density- or trait- mediated.

154! !

CHAPTER 6

A direct assessment of the ecological importance of Gyrodactylus ectoparasites in two distinct Trinidadian rivers

Pérez-Jvostov, F., Hendry, A. P., Fussmann, G. F. and Scott, M. E. Functional Ecology. TO BE SUBMITTED

[The] trips were successful; for the regions we explored were wilderness wonderlands, – full of beauty, bounding in the romance which ever enhances wild creatures and wild men

– CHARLES WILLIAM BEEBE In Our Search For a Wilderness, 1910

155! ! ABSTRACT

1. Parasites can have important implications for the structure and composition of natural biological communities that may be directly associated with reduction in host fecundity and survival (i.e. density-mediated effects), or indirectly associated through changes in host traits such as behaviour, life-history, morphology and physiology (i.e. trait-mediated effects). Nonetheless, few studies have tested whether parasites modify interspecific interactions through density- and/or trait- mediated effects on their hosts in the wild.

2. We directly tested the density- and trait-mediated effects of parasites in natural communities. We translocated Gyrodactylus ectoparasites of the Trinidadian guppy into two previously Gyrodactylus-free guppy populations within two distinct rivers in the northern mountain range in Trinidad – the Marianne and the Paria rivers – and measured phenotypic and demographic changes in the guppy host and its competitor, Rivulus hartii.

3. We found strong phenotypic differences between guppies from the two rivers. In the

Marianne where guppies invested more in reproduction (i.e. size and number of embryos), the introduction of the ectoparasite resulted in a strong reduction in guppy survival, but this was not reflected in any reduction in guppy density, nor phenotypic differences. In contrast, in the Paria where guppies invested less in reproductive effort, the introduction of Gyrodactylus resulted in a strong reduction in female growth, but did not influence survival nor density. In neither river did the introduction of Gyrodactylus influence Rivulus phenotype or demography.

156! ! 4. Our data emphasize that parasite effects can greatly vary between host populations

(i.e., context specific), and reinforce the importance of cohesive and integrative studies to fully understand how parasites can influence the ecology of the local community.

157! ! INTRODUCTION

Given that hosts are embedded in a community of interacting species, any changes in host phenotype or host population associated with parasite effects would be expected to have cascading effects on predators and competitors (Minchella and Scott, 1993; Hatcher et al., 2006). At the level of the host population, effects such as parasite-induced mortality are likely to have strong implications for community dynamics through reductions in population density and abundance (i.e., density-mediated effects) (Holt, 1977). At the phenotypic level, parasites can influence community dynamics through host’s resource reallocation in response to energetically costly immune responses (Lochmiller and

Deerenberg, 2000) at the expense of reproduction or growth (Sheldon and Verhulst,

1996; Bonneaud et al., 2003; Schmid-Hempel and Ebert, 2003; Schmid-Hempel, 2011) and through changes in host health (i.e., trait-mediated effects). For example, infected individuals often show lethargy and erratic behaviour that can reduce their competitive ability (Barber et al., 2000; Hatcher et al., 2006) or increase their likelihood of being consumed (Lozano, 1991; Marcogliese, 2004; Barber, 2007; Mikheev, 2009). Thus, the extent and manner by which parasites influence non-host species is likely to depend not only on their effects on host reproduction and survival, but also on the capacity of the host to respond to the presence of parasites and reduce the cost of infection (Hudson et al., 2006).

Although the mechanisms by which parasites can influence community structure have been laid out in theory, many of the empirical studies have involved experimental work in the laboratory where the influence of extrinsic factors is removed (Mouritsen and

158! ! Poulin, 2004; Wood et al., 2007), or comparative studies in the field where host population history (i.e. parasitism and genetic bottlenecks) is generally unknown (Poulin,

1999; Werner and Peacor, 2003; Hatcher et al., 2006). In the present study we have overcome these limitations by experimentally transplanting a parasite (Gyrodactylus spp.) into two wild host populations (guppy, Poecilia reticulata) that had no previous history of infection with this parasite, and we measured the associated changes at the level of the host population (density, biomass) and at the level of the individual phenotype (growth, number and size of embryos). We did this both in the host species and in a sympatric non-host competitor (killifish: Rivulus hartii).

Study system

The geological history of Trinidad has resulted in series of waterfalls along rivers of the northern mountain range. These waterfalls serve as natural barriers to upstream migration of fish, with more diverse fish communities in the lower reaches, and a step-wise decrease in community complexity in the upstream direction (Guilliam et al., 1993;

Magurran and Phillips, 2001). Upstream sites are often inhabited by only two fish: guppies and Rivulus. Similar size classes compete for food (aquatic invertebrates) and shelter (Grether et al., 2001; Zandona et al., 2011). Larger size classes, however, can prey on smaller ones: large guppies on juvenile Rivulus, and large Rivulus on all but female guppies (Seghers, 1967; Gilliam et al., 1993; Mattingly and Butler, 1994; Fraser and

Lamphere, 2011; Hughes et al., 2013). The presence of waterfalls also results in contrasting parasite exposure, as downstream guppy populations always have the

159! ! ectoparasite Gyrodactylus spp. whereas some upstream populations do not (Gotanda et al., 2013).

Gyrodactylids complete their life cycle on their hosts (Kearn, 1994; Harris et al.,

2004) and transmission occurs through direct contact (Cable and Harris, 2002). Mature

Gyrodactylus carry a fully developed embryo that already has a second-generation embryo developing inside (Kearn, 1994). This direct life cycle with extreme progenesis allows Gyrodactylus to achieve large population sizes in very short periods of time

(Scott, 1982; Scott and Anderson, 1984). Gyrodactylus can have effects at the host population level through increased guppy mortality (Scott, 1982; Scott and Anderson,

1984; Cable and van Oosterhout, 2007). The parasite can similarly influence host phenotype. Heavily-infected guppies show increased lethargy (van Oosterhout et al.,

2003), erratic swimming (Hockley et al., 2014), reduced feeding behaviour (Lopez, 1998; van Oosterhout et al., 2003), and reduced reproductive success (Kennedy et al., 1987,

Kolluru et al., 2009, van Oosterhout et al., 2007). In addition, guppies exposed to

Gyrodactylus show reduced growth, suggesting an energetic trade-off between immune response and growth (Pérez-Jvostov et al., 2012; Pérez-Jvostov et al., 2015). Considering all these effects, Gyrodactylus infections would seem likely to influence the structure of the local community through their effects on guppy hosts.

Our objectives were to a) quantify the effects of the introduction of Gyrodactylus on guppies and Rivulus in a natural field setting, and b) assess whether these effects were predominantly at the population level (i.e., density-mediated effects) or at the level of the

160! ! individual phenotype (i.e., trait-mediated effects). Two initially Gyrodactylus-free tributaries were each divided into two adjacent sections: a Gyrodactylus introduction section directly above a waterfall and a control section further upstream separated by an artificial barrier to prevent upstream movement of infected guppies from the introduction section. Prior to the parasite introduction (pre-introduction phase), the demography and life history of guppies and Rivulus was assessed by mark-recapture. Eight months after the parasite introduction (post-introduction phase), a second mark-recapture experiment was performed. This replicated BACI (before-after: control-impact) design allowed us to assess, for both Rivulus and guppies, the changes at the population and phenotypic level associated with the Gyrodactylus introduction.

MATERIALS AND METHODS

Experimental design

We identified two low-predation Gyrodactylus-free rivers on the northern slope of the northern mountain range in Trinidad. One river was in the Paria watershed and the other was in the Marianne watershed, with the two watersheds having genetically distinct low- predation guppy populations that represented true evolutionary replicates (Crispo et al.,

2006; Fraser et al., 2009; Xavier et al., 2015). In both rivers, a waterfall in the lower reach served as a barrier preventing the upstream migration of Gyrodactylus-infected guppies from the downstream population.

161! ! To confirm the presence of Gyrodactylus below the waterfall and the absence of

Gyrodactylus above the waterfall, 100 guppies were collected upstream and downstream of the waterfall in both rivers. Each fish was anaesthetized using a buffered solution of

0.02% tricaine methanesulfonate (MS-222) and the surface of the fish was scanned for parasites using a dissecting microscope. This is the standard method for establishing

Gyrodactylus prevalence (Scott, 1982), surveys of 100 fish are considered reliable for estimating Gyrodactylus presence/absence and abundance (Martin and Johnsen, 2007;

Gotanda et al., 2013; Xavier et al., 2015), and results from one time period have been shown to be highly repeatable at other time periods (Fraser et al., 2009; Gotanda et al.,

2013). Infection prevalence immediately below the waterfalls was 75% (mean intensity =

5.1 + 3.5 S.D. per fish) in the Marianne and 32% (mean intensity= 1.4 + 1.0 S.D.) in the

Paria. By contrast, no Gyrodactylus-infected fish were found above the waterfall in either river.

Experimental protocol

Each reach above the waterfall was divided into a downstream “experimental” section of five experimental pools where we would later introduce Gyrodactylus and an upstream

“control” section of five pools that would remain Gyrodactylus-free. The control section was separated from the experimental section by creating a 50 cm high barrier made with large rocks, gravel and fallen tree trunks to prevent any potential upstream spread of the ectoparasites. For each pool at each capture period (details below), we estimated individual pool volume using: ! = #×%, where ! is pool volume, # is pool area and % is the average pool depth.

162! ! Pre-introduction mark-recapture methodology was applied to each of the 5 control and 5 experimental pools in each river. Previous studies have shown that intensive effort in small pools can be effective at collecting all guppies and Rivulus (Reznick et al., 1996;

Reznick et al., 2001). All fish from each pool were collected, fish were separated by pool and species, and held overnight in the field prior to transportation to the laboratory the next morning. In the laboratory, the fish were held in pool/species/sex-specific aquaria.

Within two days, each fish was scanned for parasites (as above), weighed (nearest 0.001 g), and individually-marked using Visible Implant Elastomer dyes (Northwest Marine

Technology Inc.). These procedures have minimal effects on guppy survival as (Gordon et al., 2009; Weese et al., 2010; Pérez-Jvostov et al., 2012, 2015). None of the collected guppies were infected with Gyrodactylus parasites, which further confirmed that these populations were Gyrodactylus-free. All fish were then returned to their specific pool of origin. One month later, we again collected all fish, identified them using their marks, scanned them for parasites (again none were found), and weighed them as described above. In addition, we quantified the life-history traits of female guppies (mass and number of embryos) by dissecting 25 randomly selected recaptured adult females from each of the introduction and the control sections.

To introduce parasites into the experimental section of each river, we collected fish from five pools immediately below the waterfall and scanned them for parasites.

After identifying infected fish, we manually transferred three parasites onto each of three randomly-selected marked fish from each of the five pools in the experimental section.

After confirming (24 h later) that the infection had successfully established on each fish, they were returned to their home pools in the experimental section.

163! !

Eight months after introduction, we confirmed that Gyrodactylus was still present in the experimental pools in the Paria and that Gryodactylus was not present in the control section. However, in the Marianne river, Gyrodactylus had spread from the experimental into the control pools, despite our constructed barrier. We therefore explored further upstream to delineate how far Gyrodactylus had spread. Approximately

250 m upstream of our initial control section, a large fallen tree formed a natural barrier that had prevented further upstream guppy (and therefore Gyrodactylus) movement, which we confirmed by inspecting captured fish. A mark-recapture experiment, identical to that described for the pre-introduction phase, was then conducted in both rivers in order to obtain comparable data on demography and phenotypic traits over a one-month post-introduction interval. In addition to including the 10 pools from each river, we also conducted the mark-recapture experiment in two large pools approximately 250 m upstream in the Marianne, where Gyrodactylus had not yet spread.

Statistical analyses

Our analyses focused on three potential mechanisms by which Gyrodactylus could influence community structure: a) changes at the population level in guppy or Rivulus survival, biomass and density (i.e. density-mediated effects), and c) changes at the individual level in guppy or Rivulus size and life history (i.e. trait-mediated effects).

Paria comparisons used the originally designed experimental and control sections for the pre- and post-introduction phases. However, given that the exact time

164! ! Gyrodactylus spread into the control section in the Marianne river is unknown, Marianne comparisons used the originally designed experimental and control sections for the pre- introduction phase, but removed the initial five control pools, and used the two farther- upstream pools as the control for the post-introduction phase.

Population-level survival estimates were calculated using Wilson Score Intervals for binomial counts and proportions (number of fish recaptured/ total number of fish at first capture). However, since we only have one recapture episode we cannot independently differentiate between survival and recapture probability. Survival was independently calculated for each control and experimental section, in both pre- and post- introduction phases using the binom package in R version 3.1.2. GUI (R Core

Development Team 2011). P values for within-river pairwise comparison of proportions were adjusted using the Holm method for multiple comparisons. Additional population- level differences were analyzed with generalized linear mixed effects models (GLMMs) in the nlme package in R. The response variables, each considered in a separate analysis, were guppy density (number of guppies/ pool volume) and guppy biomass (weight of all guppies / pool volume). Explanatory variables included the fixed effects of river

(Marianne versus Paria), experimental treatment (experimental versus control), phase

(pre- versus post-introduction), all two-way interactions, and the three-way interaction.

Pool was a random factor nested within treatment. We performed a similar set of analyses with the same model structure for Rivulus density and biomass.

165! ! Individual-level differences were analyzed with separate GLMMs for a) guppy mass (weight in grams at first capture), b) the change in individual guppy mass over the

30-day period between capture and recapture, c) female reproductive effort (proportion of female growth due to embryonic fresh mass at the time of recapture), d) the number of embryos per female at recapture, and e) the total embryonic mass per female at recapture.

Fixed effects were river of origin (Marianne versus Paria), experimental treatment

(experimental or control), phase (pre- or post-introduction), and their interactions. Sex was included as a fixed effect for analyses of mass and change in mass. Pool of origin was included as a random factor nested in treatment. Mean parasite load (i.e. individual parasite load averaged between capture and recapture) and body mass at capture were entered as covariates, but mean parasite load was removed owing to non-significance. We fit a similar model for Rivulus growth with the exception that sex was not included as a main effect because the sex of Rivulus was often unknown.

RESULTS

Overall, we collected 1579 guppies (981 in the Marianne and 598 in the Paria) and 513

Rivulus (285 in the Marianne and 228 in the Paria) (Table 1). Almost complete turnover of individuals occurred between the pre- and post-introduction phases in both rivers, with the exception being one female in the Paria. Gyrodactylus successfully established in both the Paria (post-introduction capture prevalence = 25.2% and mean intensity= 3.1 worms/infected guppy; post-introduction recapture prevalence = 18.6% and mean intensity = 2.7 worms/infected guppy) and the Marianne River (post-introduction capture

166! ! prevalence = 27.0% and mean intensity = 2.3 worms/infected guppy; post-introduction recapture prevalence = 16.9%, and mean intensity = 0.5 worms/infected guppy). The maximum parasite load on a fish was 12 in the Paria and 19 in the Marianne.

Population level effects

Changes in guppy survival were observed in both rivers. In the Paria, although no differences between treatments were evident in either pre- (P=0.98) or post-introduction phase (P=0.97), survival of fish in the both the control and experimental sections post- introduction was higher than that of the experimental treatment in the pre-introduction phase (P= 0.053 and P=0.051; Fig. 1a). In the Marianne river, the introduction of

Gyrodactylus resulted in a strong reduction in guppy survival (Fig. 1b). Fish from the experimental treatment in the post-introduction phase showed a reduction in survival relative both to the pre-introduction phase (P<0.001), and to the post-introduction control

(P<0.001). No detectable differences were observed in the survival of guppies in the post-introduction control compared to the pre-introduction phase (P= 0.961 and P=

0.713). Additionally, our analyses revealed significant two-way interactions between river (Paria vs. Marianne) and treatment (experimental vs. control), and river and phase

(pre- vs. post-introduction) for both guppy density and biomass, but no significant main effects were detected (Table 2). In the Paria river, guppy density in both control and experimental treatments was 46% lower at the post-introduction phase compared with the pre-introduction phase (Fig. 2a). In contrast, in the Marianne, guppy density remained constant between the pre- and post-introduction phases, regardless of treatment (Fig. 1a).

167! ! These density patterns were mirrored by biomass, with a reduction of ca. 0.4g/m3 in guppy biomass in the Paria post-introduction, and no change in the the Marianne (Fig.

2b).

Individual level effects

Analysis at the individual level for guppy mass revealed a significant interaction between river (Paria versus Marianne) and treatment (experimental versus control), significant interactions between river and phase and significant main effects of sex (Table 2). In the

Paria, female mass was greater at the post-introduction phase than the pre-introduction phase regardless of treatment (Table 2; Fig. 3a). On the other hand, both females and males in the Marianne river had lower mass during the post-introduction phase than during the pre-introduction phase.

Significant interactions between river and phase (Table 2) and significant main effects of sex, river and treatment were detected for change in mass (Table 2). In the

Paria, females in the experimental treatment grew 34% less during the post-introduction phase than during the pre-introduction phase whereas females from the control treatment grew twice as much during the post-introduction phase than during the pre-introduction stage. Considering only the post-introduction phase, Paria females in the experimental treatment grew 55% less than those in the control treatment. In the Marianne, on the other hand, female growth did not differ between treatments either at the pre- nor post- secondary phases (Fig. 3c; Table 3).

168! ! Total embryonic mass per female was not affected by river, treatment or phase

(Table 3). However, reproductive effort and number of embryos were both affected by river while controlling for initial female weight (Table 3). Females from the Marianne river had more and larger embryos than females from the Paria regardless of treatment or phase.

Rivulus population and individual-level effects

None of the investigated explanatory variables had any detectable effect on

Rivulus survival, density or biomass (data not shown). No significant effects of introduction of Gyrodactylus spp were found for Rivulus growth or average size in either river (data not shown).

DISCUSSION

The importance of parasites in community ecology has received increasing interest in recent years. Parasites can alter community dynamics through a reduction in host population density (i.e., density-mediated direct effects) and/or through changes in host phenotype (i.e., trait-mediated indirect effects), which can individually or jointly have strong repercussions for coexisting species as changes in density can alter phenotypes and vice versa. We explored these potential parasite induced effects in the wild by directly transplanting Gyrodactylus ectoparasites of Trinidadian guppies into previously

Gyrodactylus-free guppy populations. First, we found strong phenotypic differences

169! ! between guppies from both rivers: guppies in the Marianne invested more in reproduction

(i.e., size and number of embryos) than guppies from the Paria which grew faster but had smaller and less embryos. Eight months after parasite introduction, we found infection levels similar to those of other wild guppy populations (Martin and Johnsen, 2007;

Gotanda et al., 2013), both as prevalence (Marianne: 22%; Paria: 21%) and parasite loads

(Marianne: 1.3 + 0.1 S.D.; Paria: 2.9 + 0.5). Despite such low infection levels, a strong reduction in guppy survival was observed in the river with the higher reproductive effort, the Marianne (Fig. 1b), although no other changes in density or phenotype were associated with the introduction of Gyrodactylus. In the river with lower reproductive effort, the Paria, despite a reduction in guppy density of over 50%, a strong reduction in female guppy growth was associated with the introduction of Gyrodactylus, but no changes in survival, density or biomass were evident. In addition, we were unable to detect any effects of parasitism on guppies for Rivulus demography or growth. We now provide potential explanations for the observed results, while expanding our consideration of the relative importance that Gyrodactylus might have on the ecology of local communities.

Population-level effects

Guppies heavily infected with Gyrodactylus show high mortality in the laboratory (Scott

1982; Scott and Anderson, 1984; Cable and van Oosterhout, 2003), but only one study has reported such effects in nature (van Oosterout et al., 2007). Our study introduced

Gyrodactylus into two populations that were Gyrodactylus-free, and might well have been so for a long time. This introduction, surprisingly, did not result in a decrease in

170! ! guppy density, despite a strong reduction in guppy survival in the experimental treatment in the Marianne river (Table 2). What might prevent Gyrodactylus from inducing negative population-level effects, and why might the inconsistency across tributaries arise?

It is possible that a fast turnover of resident guppies in the focal populations could reduce the negative effects Gyrodactylus could have on guppy density, by continuously moving infecting individuals from the population and preventing the establishment of epidemics. Our results suggest that this could certainly be possible, as new immigrants accounted for large proportions of both populations. Alternatively, guppies, particularly in the Marianne, might simply migrate more to escape the risk of Gyrodactylus infection, and new migrants compensate for the loss in density and biomass, and thus removing any population-level effects of Gyrodactylus. Other mechanisms might influence the disparity between rivers. First, guppy densities in the wild might have been too low (average of 10 guppies/ m3) to allow the explosive Gyrodactylus epidemics that are so detriment to lab- reared guppies. Second, Gyrodactylus might evolve low virulence in some populations and, indeed, variation in virulence among Gyrodactylus strains has been documented

(Cable and van Ooosterhout 2003; Pérez-Jvostov et al., 2015). Third, many other factors influence guppy mortality to the point that Gyrodactylus might have little additional effect. For instance, Paria and Marianne guppy populations similar to ours experience predation from Rivulus and Macrobrachium prawns (Rodd and Reznick, 1991) and kingfishers (Templeton and Shriner, 2004), mortality from high flows during flooding

(Weese et al., 2011), and density-dependence due to very high competition (Grether et

171! ! al., 2001). Fourth, under density limitation, increased mortality owing to Gyrodactylus might simply increase the survival of the remaining guppies, leading to no net effect on demography.

Elucidating the specific mechanisms driving the apparent lack of Gyrodactylus demographic effects on one population, while resulting in positive effects in another will require further work. However, our results highlight the importance of ecological contingencies (e.g., host life-history, parasite virulence, population structure, river, etc.), and their potential in influencing the strength and direction of parasite density-mediated effects.

Individual-level effects

Phenotypic changes in the host induced directly by parasitic infection or indirectly through a reduction in the fitness costs of infection can strongly modify interspecific interactions even on short time scales (Holt 1977). We found only one such effect and only in one river. Specifically, female guppies in the Gyrodactylus treatment in the Paria river had reduced growth rates. Although similar changes were not observed in the

Marianne, we have previously shown in mescocosm experiments that female guppies exposed to Gyrodactylus do reduce their growth compared to unexposed fish (Pérez-

Jvostov et al., 2012; Pérez-Jvostov et al., 2015). This reduction in growth was not related to the individual infection status, but rather the sole presence of Gyrodactylus in the mesocosm resulted in a reduction in growth, suggesting a potential trade-off with a prophylactic immune response (Sadd and Schmid-Hempel, 2009). Moreover, the

172! ! reduction in growth observed in the present experiment occurred in spite of a reduction in guppy density. Surprisingly, we found no differences between treatments in average female mass or reproductive effort. These results suggest that guppies can invest in immune response at the expense of growth, but without compromising reproductive success.

One would expect a priori – as we did – that the introduction of a known deleterious parasite into a naïve population would have strong negative effects for the hosts in the wild, given that this has been the case in other systems (Simberloff, 2000;

Turchin et al., 2003). Yet our results instead suggest that the immediate effects of parasites are few and context-specific, for the demography and traits of hosts and the species within which it interacts.

What might we have missed for guppies? We looked for effects only eight months after introducing Gyrodactylus, and it remains possible that stronger effects are yet to come. Yet Gyrodactylus life histories enable extremely rapid rates of increase and guppies also turn over very quickly. Natural communities are subject to intermittently environmental conditions that might influence Gyrodactylus effects. For example, our study was conducted during the wet season, whereas guppy densities are higher in the dry season (Reznick and Endler, 1982; Rodd and Reznick, 1997; Reznick et al., 2001;

Grether et al., 2001), which might enhance Gyrodactylus epidemics.

173! ! What might we have missed for Rivulus? It is likely that given the absence of strong population-level effects of Gyrodactylus on guppies, we did not detect any changes in Rivulus. Alternatively, Rivulus might have reduced the effects of

Gyrodactylus on guppies. Indeed, we have recently used mesocosm experiments to show that guppies exposed to both Gyrodactylus and Rivulus show intermediate growth between guppies exposed to only the parasite or to only the competitor, without any additional changes in density or biomass (Pérez-Jvostov et al., 2015). This pattern is likely due to intra-guild predation, as large Rivulus can actively prey on small guppies, thus, intermediate guppy growth speeds reaching a safe size while still allowing for allocation into immune response.

CONCLUSIONS

The introduction of Gyrodactylus ectoparasites into two previously naïve guppy populations in Trinidad resulted in strong but different effects in each host population. In the population with the higher reproductive effort, the Marianne, the introduction of the ectoparasite resulted in a ~ 50% reduction in guppy survival, whereas in the population with higher growth rate, the Paria, the introduction of the ectoparasite resulted in a ~50% reduction in female growth – a potential trade-off with immune response. Despite these strong individual-level effects, Gyrodactylus had no consequences at the population level, neither at guppy density nor biomass. An important message to take from our findings is that trait- and density-mediated effects of parasites appear to be context specific, in this case differing between “replicate” rivers. We do not know the specific reason for such differences but low-predation tributaries have been shown to differ in resource levels,

174! ! predator densities, flow rates, and primary productivity (Grether et al. 2001; Millar et al.

2006; McKellar et al. 2009). Any of these factors could – in principal – influence the costs of being parasitized or of mounting and deploying defenses against predators. These realizations further highlight the complexity behind host-parasite interactions in the wild, and the importance of local processes in influencing such dynamics. We hope our results may stimulate future research to understand how parasites and competitors jointly affect host and community ecology in natural populations, as host response to parasites can also be mediated by the presence of interacting species.

ACKNOWLEDGMENTS

We would like to thank Maryse Boisjoly, Christianne Aikins, Elizabeth Turner and Jack Torresdall for their assistance with fieldwork. This work was supported by

NSERC-Special Research Opportunity grant to APH, MES and GFF, and a Ph.D.

Fellowship from the Consejo Nacional de Ciencia y Tecnología (CONACYT) to FPJ.

175! ! REFERENCES

1.! Anderson, R.M., & May, R. M. (1978) Regulation and stability of host-parasite

population interactions: I. Regulatory processes. The Journal of Animal Ecology,

47: 219-247.

2.! Barber, I. (2007) Parasites, behaviour and welfare in fish. Applied Animal

Behaviour Science, 104: 251-264.

3.! Barber, I., Hoare, D., & Krause, J. (2000) Effects of parasites on fish behaviour: a

review and evolutionary perspective. Reviews in Fish Biology and Fisheries, 10:

131-165.

4.! Bonneaud, C., Mazuc, J., Gonzalez, G., Haussy, C., Chastel, O., Faivre, B., &

Sorci, G. (2003) Assessing the cost of mounting an immune response. The

American Naturalist, 161: 367-379.

5.! Cable, J., & Van Oosterhout, C. (2007) The role of innate and acquired resistance

in two natural populations of guppies (Poecilia reticulata) infected with the

ectoparasite Gyrodactylus turnbulli. Biological Journal of the Linnean Society,

90: 647-655.

6.! Cable, J., Scott, E. C.G., Tinsley, R.C., & Harris, P.D. (2002) Behavior favoring

transmission in the viviparous monogenean Gyrodactylus turnbulli. Journal of

Parasitology, 88: 183-184.

7.! Dargent, F., Scott, M.E., Hendry, A.P., & Fussmann, G.F. (2014) Experimental

evolution of parasite resistance in wild guppies: natural and multifarious

selection. Proceedings of the Royal Society B: Biological Sciences, 281:

20141820.

176! ! 8.! Diekmann O., & Heesterbeek J.A.P. (2000) Mathematical epidemiology of

infectious diseases: model building, analysis and interpretation. Wiley,

Chichester, UK.

9.! Gilliam, J.F., Fraser, D.F., & Alkins-Koo, M. (1993) Structure of a tropical stream

fish community: a role for biotic interactions. Ecology, 74:1856–1870.

10.!Grether, G.F., Millie, D.F., Bryant, M.J., Reznick, D.N., & Mayea, W. (2001)

Rain forest canopy cover, resource availability, and life history evolution in

guppies. Ecology, 82: 1546-1559.

11.!Harris, P.D., & Lyles, A.M. (1992) Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. The Journal

of Parasitology, 78: 912-914.

12.!Harris, P.D., Shinn, A.P., Cable, J., & Bakke, T.A. (2004) Nominal species of the

genus Gyrodactylus von Nordmann 1832 (Monogenea: Gyrodactylidae), with a

list of principal host species. Systematic Parasitology, 59: 1-27.

13.!Hedrick, S.M. (2004). The acquired immune system: A vantage from beneath.

Immunity, 21: 607-615.

14.!Hockley, F.A., Wilson, C.A.M.E., Brew, A., & Cable, J. (2014) Fish responses to

flow velocity and turbulence in relation to size, sex and parasite load. Journal of

The Royal Society Interface, 11: 20130814.

15.!Hockley, F.A., Wilson, C.A.M.E., Graham, N., & Cable, J. (2014) Combined

effects of flow condition and parasitism on shoaling behaviour of female guppies

Poecilia reticulata. Behavioral Ecology and Sociobiology, 68: 1513-1520.

177! ! 16.!Holt, R.D. (1977) Predation, apparent competition, and the structure of prey

communities. Theoretical Population Biology, 12: 197-229.

17.!Houde, A.E., & Torio, A.J. (1992) Effect of parasitic infection on male color

pattern and female choice in guppies. Behavioral Ecology, 3: 346-351.

18.!Hudson, P.J., Dobson, A.P., & Lafferty, K.D. (2006) Is a healthy ecosystem one

that is rich in parasites? Trends in Ecology & Evolution, 21: 381-385.

19.!Kearn, G.C. (1994) Evolutionary expansion of the Monogenea. International

Journal for Parasitology, 24: 1227-1271.

20.!Kennedy, C.E.J., Endler, J.A., Poynton, S.L., & McMinn, H. (1987) Parasite load

predicts mate choice in guppies. Behavioral Ecology and Sociobiology, 21: 291-

295.

21.!Kennedy, C.E.J., Endler, J.A., Poynton, S.L., & McMinn, H. (1987) Parasite load

predicts mate choice in guppies. Behavioral Ecology and Sociobiology, 21: 291-

295.

22.!Kolluru, G.R., Grether, G.F., Dunlop, E., & South, S.H. (2009) Food availability

and parasite infection influence mating tactics in guppies (Poecilia reticulata).

Behavioral Ecology, 20: 131-137.

23.!Kolluru, G.R., Grether, G.F., South, S.H., Dunlop, E., Cardinali, A., Liu, L., &

Carapiet, A. (2006) The effects of carotenoid and food availability on resistance

to a naturally occurring parasite (Gyrodactylus turnbulli) in guppies (Poecilia

reticulata). Biological Journal of the Linnean Society, 89: 301-309.

178! ! 24.!Lafferty, K.D., Allesina, S., Arim, M., Briggs, C.J., De Leo, G., Dobson, A.P., &

Thieltges, D.W. (2008) Parasites in food webs: the ultimate missing links.

Ecology Letters, 11: 533-546.

25.!Lochmiller, R.L. & Deerenberg, C. (2000) Trade-offs in evolutionary

immunology: just what is the cost of immunity? Oikos, 88: 87-98.

26.!López, S. (1998) Acquired resistance affects male sexual display and female

choice in guppies. Proceedings of the Royal Society of London. Series B:

Biological Sciences, 265: 717-723.

27.!López, S. (1999) Parasitized female guppies do not prefer showy males. Animal

Behaviour, 57: 1129-1134.

28.!Lozano, G. A. (1991) Optimal foraging theory: a possible role for parasites.

Oikos, 60: 391-395.

29.!Magurran, A.E. & Phillip, D.A. (2001). Implications of species loss in freshwater

fish assemblages. Ecography, 24: 645-650.

30.!Marcogliese, D.J. (2004) Parasites: small players with crucial roles in the

ecological theater. EcoHealth, 1: 151-164.

31.!Mattingly, H.T., & Butler IV, M.J. (1994) Laboratory predation on the

Trinidadian guppy: implications for the size-selective predation hypothesis and

guppy life history evolution. Oikos, 69: 54-64.

32.!May, R.M., & Anderson, R.M. (1978) Regulation and stability of host-parasite

population interactions: II. Destabilizing processes. The Journal of Animal

Ecology, 47: 249-267.

179! ! 33.!Mikheev, V.N. (2009) Combined effects of predators and parasites on shoaling

behavior of fishes. Journal of Ichthyology, 49: 1032-1041.

34.!Pérez-Jvostov, F., Hendry, A.P., Fussmann, G.F., & Scott, M.E. (2012) Are host–

parasite interactions influenced by adaptation to predators? A test with guppies

and Gyrodactylus in experimental stream channels. Oecologia, 170: 77-88.

35.!Pérez-Jvostov, F., Hendry, A.P., Fussmann, G.F., & Scott, M.E. (2015) Testing

for local host–parasite adaptation: an experiment with Gyrodactylus ectoparasites

and guppy hosts. International Journal for Parasitology, 45: 409-17.

36.!Poulin, R. (1999) The functional importance of parasites in animal communities:

many roles at many levels? International Journal for Parasitology, 29: 903-914.

37.!Reznick, D., & Endler, J.A. (1982) The impact of predation on life history

evolution in Trinidadian guppies (Poecilia reticulata). Evolution, 36: 160-177.

38.!Reznick, D., Butler IV, M.J., & Rodd, H. (2001) Life-history evolution in

guppies. VII. The comparative ecology of high- and low-predation environments.

The American Naturalist, 157: 126-140.

39.!Rodd, F.H., & Reznick, D.N. (1997) Variation in the demography of guppy

populations: the importance of predation and life histories. Ecology, 78: 405-418.

40.!Schmid-Hempel, P. (2003) Variation in immune defense as a question of

evolutionary ecology. Proceedings of the Royal Society of London B: Biological

Sciences, 270: 357-366.

41.!Schmid-Hempel, P. (2011) Evolutionary parasitology: the integrated study of

infections, immunology, ecology, and genetics. Oxford University Press, New

York.

180! ! 42.!Schmid-Hempel, P., & Ebert, D. (2003) On the evolutionary ecology of specific

immune defense. Trends in Ecology & Evolution, 18: 27-32.

43.!Scott, M. E. (1982) Reproductive potential of Gyrodactylus bullatarudis

(Monogenea) on guppies (Poecilia reticulata). Parasitology, 85: 217-236.

44.!Scott, M.E., & Anderson, R.M. (1984) The population dynamics of Gyrodactylus

bullatarudis (Monogenea) within laboratory populations of the fish host Poecilia

reticulata. Parasitology, 89: 159-194.

45.!Seghers, B.H. (1967) An analysis of geographic variation in the antipredator

adaptations of the guppy, Poecilia reticulata (Doctoral dissertation, University of

British Columbia).

46.!Sheldon, B.C., & Verhulst, S. (1996) Ecological immunology: costly parasite

defences and trade-offs in evolutionary ecology. Trends in Ecology & Evolution,

11: 317-321.

47.!Van Oosterhout, C., Harris, P.D., & Cable, J. (2003) Marked variation in parasite

resistance between two wild populations of the Trinidadian guppy, Poecilia

reticulata (Pisces: Poeciliidae). Biological Journal of the Linnean Society, 79:

645-651.

48.!Van Oosterhout, C., Mohammed, R.S., Hansen, H., Archard, G.A., McMullan,

M., Weese, D.J., & Cable, J. (2007) Selection by parasites in spate conditions in

wild Trinidadian guppies (Poecilia reticulata). International Journal for

Parasitology, 37: 805-812.

49.!Zandona, E., Auer, S.K., Kilham, S.S., Howard, J.L., López-Sepulcre, A.,

O’Connor, M.P. & Reznick, D.N. (2011) Diet quality and prey selectivity

181! ! correlate with life histories and predation regime in Trinidadian guppies.

Functional Ecology, 25: 964-973.

182! ! FIGURES

Paria Marianne a) b)

0.8 A A A 0.6 A A AB B 0.4 B

0.2 Estimated recapture probablity

Pre-Intro Pre-Intro Post-Intro Post-Intro Pre-Intro Pre-Intro Post-Intro Post-Intro Control Experimental Control Experimental Control Experimental Control Experimental

Figure 1. Normal approximations for 95% binomial confidence intervals on recapture probability using the Wilson Score Interval.

183! !

Paria:Control a) )

3 15

10

5 Density (guppies/m

b) ) 3

1.5

1.0

0.5 Guppy biomass (g/m Pre-Introduction Post-Introduction

Figure 2. . LS Means for guppy density (a) and biomassStage (b). Error bars represent standard errors. Symbols represent Paria Control (), Paria Experimental (), Marianne Control

(),Marianne experimental ().

184! ! Female guppies Male guppies 0.30 a) 0.13 b) 0.25 0.11 0.20

0.15 0.09

Average weight (g) 0.10

0.05 0.07

0.03 c) d) 0.06 0.02 0.04 0.01 0.02 0.00 Change in weight (g) 0.00 -0.01 Pre-Introduction Post-Introduction Pre-Introduction Post-Introduction

Paria:Control Marianne:Control Figure 3. LS Means for guppy weight and growth.Marianne:Gyro Error bars represent standard errors. Paria:Gyro Symbols represent Paria Control (), Paria Experimental (), Marianne Control (),

Marianne experimental ().

185! ! TABLES

Guppies

Tributary Treatment Stage Total number Total number of Number of Percentage Number of new of fish at first fish at second recaptures recaptured (%) collections capture capture

Marianne Control Pre-Introduction 89 75 43 48.31 32

Gyrodactylus Pre-Introduction 130 156 74 56.92 82 Control Post-Introduction 217 183 108 49.77 75 Gyrodactylus Post-Introduction 157 113 40 25.4 73

Paria Control Pre-Introduction 89 71 32 35.96 39 Gyrodactylus Pre-Introduction 187 132 63 33.69 69 Control Post-Introduction 32 39 19 59.38 20 Gyrodactylus Post-Introduction 99 113 50 50.51 63

Rivulus

Tributary Treatment Stage Total fish at Total fish at Number of Percentage Number of new first capture second capture recaptures recaptured (%) collections

Marianne Control Pre-Introduction 32 33 10 31.25 23 Gyrodactylus Pre-Introduction 27 34 10 37.04 24 Control Post-Introduction 37 59 14 37.84 45 Gyrodactylus Post-Introduction 54 57 15 27.78 42

Paria Control Pre-Introduction 24 25 11 45.83 14 Gyrodactylus Pre-Introduction 43 35 9 20.93 26 Control Post-Introduction 31 45 9 29.03 36 Gyrodactylus Post-Introduction 26 33 5 19.23 28

Table 1. Descriptive statistics of collection parameters during the pre- and post-

introduction stages in the Gyrodactylus and control treatments.

186! ! Guppy density Guppy biomass Explanatory variable F (D.F.) P F (D.F.) P

River 3.363 (1,41) 0.073 1.249 (1,41) 0.270 Phase 2.459 (1,41) 0.124 1.028 (1,41) 0.316 Treatment 0.040 (1,41) 0.844 0.036 (1,41) 0.852 River X Phase 6.374 (1,41) 0.015 3.631 (1,41) 0.063 River X Treatment 9.141 (1,41) 0.004 11.579 (1,41) 0.001 Phase X Treatment 0.064 (1,41) 0.801 0.082 (1,41) 0.775 River X Phase X Treatment 0.457 (1,41) 0.502 0.076 (1,41) 0.783

Table 2. Analyses for density mediated effects of Gyrodactylus introduction. Density and biomass was calculated for each pool during the mark and the recapture collections.

Significant p values highlighted in bold.

187! !

Total embryonic Reproductive Number of Guppy mass Change in mass Explanatory mass per female effort embryos variable F (D.F.) P F (D.F) P F (D.F) P F (D.F) P F (D.F) P 9.48 91.224 2.146 46.788 8.72 River 0.022 <0.001 0.145 <0.001 0.004 (1,441) (1,440) (1,133) (1,133) (1,133) 5.019 19.828 0.174 0.134 1.035 Phase 0.025 <0.001 0.677 0.715 0.311 (1,441) (1,440) (1,133) (1,133) (1,133) 0.259 5.008 0.019 0.862 0.565 Treatment 0.611 0.026 0.89 0.355 0.453 (1,441) (1,440) (1,133) (1,133) (1,133) 64.861 4.928 Sex <0.001 0.008 ------(1,441) (1,440) 74.669 74.669 47.656 5.804 25.213 Guppy weight - <0.001 <0.001 0.017 <0.001 (1,441) (1,440) (1,133) (1,133) (1,133) River 6.502 0.576 1.047 0.292 1.428 X 0.011 0.448 0.308 0.59 0.234 (1,441) (1,440) (1,133) (1,133) (1,133) Treatment River 20.557 5.499 0.175 0.059 1.18 X <0.001 0.02 0.677 0.809 0.279 (1,441) (1,440) (1,133) (1,133) (1,133) Phase Treatment 0.259 0.242 1.225 1.459 0.076 X 0.61 0.623 0.27 0.229 0.784 (1,441) (1,440) (1,133) (1,133) (1,133) Phase River X 0.184 6.317 0.454 1.115 1.93 Treatment (1,441) 0.667 (1,440) 0.012 (1,133) 0.502 (1,133) 0.293 (1,133) 0.167

X Phase Table 3. Analyses on changes in mean trait values of guppies using general linear mixed effects models.

188! !

CHAPTER 7

General discussion

Perseverance is the hard work that you do after you get tired of doing the hard work that you already did. – UNKNOWN Trini phrase found on a tarp in the woods of the Paria River, Trinidad, 2013

189! ! For decades, parasitologists have appealed for the incorporation of parasitism into contemporary ecological research (Minchella and Scott, 1991; Marcogliese and Cone,

1997; Hatcher et al., 2006; Lafferty and Morrison, 1996; Hudson et al., 2006; Lafferty et al., 2006; Lefèvre et al., 2008), yet despite important progress in our understanding of how parasites can influence a variety of ecological processes, several issues remain

(Hatcher et al., 2006; Hatcher et al., 2015). Model organisms provide a solution for this, as the effects of other sources of selection on their ecology and evolution are well documented – thus facilitating the incorporation of parasitism into an integrated ecological and evolutionary theory. The work presented in this thesis provides a cohesive examination of the relative importance of Gyrodactylus ectoparasites in the Trinidadian guppy. Each of the chapters examined one aspect of the interaction between guppies,

Gyrodactylus, and the environment where both coevolve (Fig. 1). In particular, I was able to explore how Gyrodactylus can influence the evolutionary ecology of their guppy host in the context of environmental heterogeneity (e.g. predation), and how this environmental heterogeneity influences guppy-Gyrodactylus convolution, and how guppy-Gyrodactylus interactions can 'spill over' and modify the ecology of non-host species (Rivulus hartii). In the following sections I expand on each of these questions and discuss the importance of my work, its’ major implications and caveats.

190! ! Predation

Gyrodactylus Chapters 3 and 4

Chapters 5 and 6

Guppies Rivulus

Figure 1. Schematic of this thesis. Chapters 3 and 4 explore how predation shape guppy-

Gyrodactylus adaptation (dark blue), chapters 5 and 6 explore how guppy-Gyrodactylus interactions are influenced by Rivulus, and how Rivulus is influenced by guppy-

Gyrodactylus interactions (light blue).

191! ! In chapter 2, I review various mechanisms by which specialist parasites could drive community dynamics through their direct and indirect effects on their host, and I differentiate the mechanisms with those of generalist parasites. In chapters 3 to 6, I experimentally tested some of these mechanisms. In particular, in chapter 3 I presented results of an experiment designed to test how adaptation to predation, can in turn influence hosts’ adaptation to parasites. I tested whether guppies from high-predation environments can maintain shorter and lower Gyrodactylus infections, compared with those from environments where predation is low. I found that Gyrodactylus infection dynamics strongly differ between different rivers, but were not influenced by predation.

Moreover, both high- and low- predation female guppies exposed to Gyrodactylus grew less than the controls, suggesting an energetic cost associated with activation and maintenance of an immune response. My results provide new insights on how multiple sources of selection can jointly act together, and reinforce the notion that adaptation to predation, does not necessary imply adaptation to parasitism – despite being reinforcing sources of selection. Chapter 4 follows to explore host and parasite local adaptation, and whether this adaptation is dependant on the predation environment where both coevolve.

Parasites have generally been expected to have an evolutionary advantage over their hosts

(Ebert, 1994; Saarinen and Taskinen, 2005) due to short generation time and potentially high host specificity, yet they have not always shown signatures of local adaptation to their sympatric hosts (Kaltz et al., 1999; Oppliger et al., 1999; Koskela et al., 2000;

Lemoine et al., 2012; Roth et al., 2012; Konijnendijk et al., 2013; Sternberg et al., 2013).

Studies on how additional sources of mortality could influence host-parasite local adaptation, and whether they prevent or facilitate host-parasite adaptation, are still largely

192! ! missing (Hoeksema and Forde, 2008) – despite that they would help to better understand the mechanisms of local adaptation. I addressed this using fully reciprocal cross- infections with guppies from two rivers, each with a high- and a low-predation population. I found strong differences in the capacity of guppies from different rivers to limit Gyrodactylus population growth, and this was regardless of their predation environment of origin. These results suggest population evolutionary history (i.e. genetic makeup, founder effects, genetic bottle necks, etc.) can strongly influence the intensity and direction of host-parasite coevolution.

The importance of parasites in predator-prey have been has been extensively studied in the literature. However, the importance of parasites in interspecific competition is still lagging behind (Hatcher et al., 2006). Moreover, the extent to which non-shared parasites (i.e., specialist) can influence interspecific competition, to my knowledge, has been addressed in only one theoretical studies (Fenton and Brockhurst, 2008) – and it certainly deserves more attention. In chapter 5 I directly tested how competition can be influenced by the presence of no-shared parasites. Using mesocosms channels, I analyzed the effects of Gyrodactylus ectoparasites on the competition between guppies and their natural enemy, Rivulus hartii, by contrasting competition in the presence and absence of

Gyrodactylus. I found no strong evidence for parasite-mediated competition. Even though

Gyrodactylus strongly reduced the growth of their guppy host, this effect was strongly reduced in the presence of the non-host competitor, Rivulus. Chapter 6 presents the results of a two-year long mark-recapture experiment where I studied how Gyrodactylus can influence community dynamics. I directly compared changes in community dynamics

193! ! before and after introducing Gyrodactylus into two previously Gyrodactylus-free wild guppy populations. This chapter is based solely on work performed in the natural community, compared to the experimental setting of the previous three chapters. I found that effects of Gyrodactylus are largely context dependent. I was unable to detect any demographic changes in guppy populations. Nonetheless, a strong reduction in guppy growth rates in one population, while a strong decrease in survival in another, were associated with the introduction of Gyrodactylus. Interestingly, adding a parasite into these two freshwater fish communities did not result in any evident changes in Rivulus ecology. The results in chapters 5 and 6 suggest that hosts can mediate the effects of parasites on competition with, by adaptively fine-tuning their phenotype when simultaneously facing multiple enemies.

Many questions, of course, still remain. For example, is there differential mortality between infected and healthy individuals in high-predation environments? An argument that has been commonly used is that infected individuals suffer higher mortality from predation, and thus natural selection should favour the evolution of higher resistance in guppies (Cable and van Oosterhout, 2003), and thus higher virulence in

Gyrodactylus (Williams and Day, 2001; Choo and Day, 2003). Nonetheless, this argument has no empirical support, and testing it would be certainly useful. Another relevant question relates to the interaction between mate-choice, predation and

Gyrodactylus. There is extensive evidence that females prefer brightly coloured males

(Kennedy et al., 1987; Houde and Torio, 1992), and that Gyrodactylus infections can reduced male colouration due to a trade-off between resource allocation to carotenoid

194! ! colouration and immune response (Houde and Torio, 1992). Thus, bright colouration should be an honest signal for resistance genes, as those males that can invest in colouration without jeopardizing immune response should be the most resistant

(Hamilton and Zuk, 1982). How does predation fit into this picture? It is plausible that predators will prey on brightly coloured males because they are easily detected – as proved by Endler (1980) – but it is also plausible that predators will prey on infected individuals, as they are an easier prey. If females and predators select for the more brightly coloured and resistant males, female preference should switch towards less bright males (Houde, 1997). Here, again, empirical evidence is lacking. Understanding how parasitism, predation and mate choice shape guppy evolution is important if we ought to have an evolutionary theory that incorporates the interaction between multiple sources of selection in the wild.

MAJOR CONTRIBUTIONS

Through out this thesis I have used Gyrodactylus ectoparasites to better understand the ecological consequences of parasitism. This genus has a long and extensive history of research, mainly driven by the devastating epidemics of G. salaris in the farmed and wild populations of European Atlantic salmon (Salmo salar) (Johnsen and Jensen, 1986;

Bakke and MacKenzie, 1993; Heggberget et al., 1993; Appleby and Mo, 1997), but they also commonly infect model organisms – which makes them an even more attractive system to study host-parasite interactions in the context of the broader ecological and evolutionary theory. For example, across its’ geographic range in Europe and North

America, the three-spined stickleback (Gasterosteus aculeatus) is the primary host for at

195! ! least six species of Gyrodactylus (G. alexandrei, G. avaloniae, G. branchicus, G. canadensis, G. gasterastei and G. arcuatus: Bakke et al., 2002), and spatial divergence in resistance to infection has been reported for several populations in the U.K. (de Roji et al., 2011) and Belgium (Raeymaekers et al., 2011), as well as between lake and stream ecotypes in Northern Germany (Eizaguirre et al., 2011). Although the broader ecological implications of such infections are still largely unknown, there is evidence that particular

MHC haplotypes confer resistance to Gyrodactylus in one environment, while they increase the susceptibility in another (Eizaguirre et al., 2011), which could then facilitate speciation through MHC-mediated mate choice (Wegner et al., 2003; Wegner et al.,

2006; Milinski et al., 2006).

Ian Barber (2013) has advocated the benefits of using the stickleback system to better understand the genetic architecture behind host-parasite adaptations and their ecological implications in divergent environments (e.g. lake vs. stream) (Barber, 2013).

In this thesis I demonstrate that the guppy-Gyrodactylus system is equally suitable, as it also provides important advantages for the study of the interactions between hosts, parasites and their environment. In particular, the relative depauperate community of both of fish and parasites, and the heterogeneity of the environments where guppies naturally occur in Trinidad, make this system suitable for understanding the relative importance of parasitism in the evolutionary ecology of the community – a subject that is ever important in contemporary research (Poulin, 2011). I believe I have provided the first steps in understanding how Gyrodactylus can influence the evolution of their host in the wild, but also how Gyrodactylus can modify competitive interactions through their direct and

196! ! indirect effects on host ecology and evolution.

IMPLICATIONS

With the growing incidence of biological invasions, it has become increasingly important to understand the mechanisms that facilitate the expansion and establishment of non- native organisms. In addition to facilitate the invasion success of their host, parasites can be important invaders themselves, when they are transported with their hosts. In these cases, parasitic invasions generally result in the decimation of the native-naïve host population. One such example is the case of the Caspian Sea sturgeon (Huso huso) that was introduced into the Aral Sea with its monogenean ectoparasite (Nitzchia sturionis), and resulted in the local extinction of the native sturgeon (Acipenser nudiventris) (Prenter et al., 2004). Due to the global aquarium trade, guppies have also been transported around the world, with wild populations already reported in Australia (Dove and Ernst, 1998). In

Europe, guppies have been introduced to Italy and the U.K. (Harris, 1986), and an established population has been reported in the Netherlands (Magurran, 2005).

Gyrodactylus are likely to play an important role in guppy invasive success. With a direct life cycle and the absence of a specialized transmission stage Gyrodactylus have probably also been successfully translocated worldwide (Torchin et al., 2004) – although the way and to which extent they can facilitate invasion is still unknown. The work I have presented here using the Trinidadian guppy and Gyrodactylus, particularly in chapters 5 and 6, provided an important insight into how Gyrodactylus can modify competitive interactions through their direct and indirect effects on guppies. I believe this is the first

197! ! step in understanding the ecological implications that such invasions could have on native fauna, and it will be invaluable for future conservation efforts.

LIMITATIONS

Despite the important findings that have emerged from the work presented in this thesis, it is pertinent to acknowledge its’ caveats and limitations. Some of these drawbacks were the trade-off between choosing one experimental design over another – which made them subject of much debate – but others were simply due to the day-to-day complications of fieldwork. Below, I extend on the three main caveats of my work: guppy phenotypic plasticity, Gyrodactylus host range and Gyrodactylus cryptic species.

Phenotypic plasticity

One of the main problems associated with the use of wild-caught individuals to assess the presence of adaptive evolution is phenotypic plasticity. For example, if populations inhabit different environments, divergent selection can result in adaptive phenotypic differentiation in traits that increase local fitness (Kawecki and Ebert, 2004). However, such differentiation can be driven by local adaptation but also through adaptive phenotypic plasticity (Pigliucci, 2005). If gene flow between populations is low, divergent selection can result in the evolution of locally adapted ecotypes: resident genotypes will produce phenotypes with higher relative fitness than genotypes from other environments (Kawecki and Ebert, 2004). Conversely, adaptive phenotypic plasticity in response to environmental pressures can move phenotypes closer to the optimum without

198! ! genetic divergence (Yeh and Price, 2004). Second generation lab-reared offspring (F2), where maternal and phenotypic (i.e. environmental) effects are removed are thus better alternatives to wild-caught individuals to test if adaptation exists. In the case of guppy adaptation to the predation environment, there is extensive evidence that the phenotypic differences observed between high- and low-predation guppies (e.g. growth rate, size at maturity, reproductive allocation) do have a genetic basis (Reznick, 1982; Reznick et al.,

1990; Reznick and Bryga, 1996; Torres-Dowdall et al., 2012). Much less is known about the influence of plasticity on guppy response to parasitism.

Despite the potential effects of phenotypic plasticity, wild-caught individuals – for which prior history of exposure and infection with Gyrodactylus is unknown – can lose any acquired resistance after an infection-free refractory period (Scott and Robinson

1984; Scott, 1985; Richards and Chubb, 1996; van Oosterhout et al., 2003). This refractory period will potentially change from population to population, depending on the fitness costs of Gyrodactylus infections (see bellow). Scott and Robinson (1984) first documented that ornamental guppies retain some degree of resistance to challenge infections with G. turnbulli for at least two weeks after experimentally clearing the infection. Scott (1985), and Richards and Chubb (1996), further demonstrated that full- susceptibility can be regained in ornamental guppies by increasing the refractory period to four-six weeks, while van Ooosterhout et al. (2003) have recently shown that in wild- guppies from the Aripo river, this refractory period can last up to 53 days.

199! ! Evidently, the use of second-generation lab-reared guppies would be optimal to study the interaction between guppy adaptations to predation and adaptations to parasitism, as they would not have been previously exposed to neither predators nor parasites. However, given that guppy adaptations to predation are known to be the result of adaptive evolution, provided a long enough refractory period wild-caught guppies can also be used to accurately assess divergence in immune competence between populations, environments, and even rivers (van Oosterhout et al., 2007).

Gyrodactylus host range

Throughout this thesis I have referred to Gyrodactylus as specialist parasites or parasites with narrow host range. However, there is certainly variation in the degree of host specificity within this genus. In a study of the host range of parasites of Canadian fishes,

Poulin (1992) reported that, on average, Gyrodactylus infect four host species, but at least

60% of all Gyrodactylus investigated in that study were strict specialists (Poulin, 1992) – although in a more recent and extended review, Bakke et al. (2007) reported that the proportion of strictly specialist Gyrodactylus could be closer to 30%. Of course, this view of specialization can be regarded as relatively simplistic way to define what is a specialist versus a generalist parasite, as it only relies in the number of host species without accounting their phylogenetic relatedness. However, it does provide an important first step in understanding variation in host specificity within Gyrodactylus.

How specialist are Gyrodactylus of the Trinidadian guppy? Several studies have investigated this question, particularly in respect to G. turnbulli. King et al. (2007)

200! ! recently tested infection success of an isogenic line of G. turnbulli (Gt3) on ten different fish species with various degrees of relatedness to guppies. After experimentally transferring individual worms onto naive hosts, Gt3 was able to successfully infect and attain high infection levels on all the closely related Poeciliid fishes, but the infection success was greatly reduced when infecting unrelated hosts (<20%). King et al. (2009) further tested host switching in Gt3, under more natural conditions (i.e. mixed species groups in aquaria). Their findings present a slight contrast with those of their previous study. Here, although successful infections on non-optimal hosts, Poecilia shenops

(33.3%) and Xiphophorous hellerii (13.3%) were reported, these were much lower than infections on guppies (46% success). Similarly, the intensity of infections was lower on both non-optimal hosts (maximum of 23 and 25 worms respectively) compared to those on guppies (maximum of 146 worms on one individual). In addition, this host shift was achieved under very high worm densities (~100 Gyrodactylus), which is rarely observed in natural guppy populations (Martin and Jonhsen, 2007; Gotanda et al., 2013). This migration to non-optimal hosts could be adaptive, in the case Gt3 is indeed capable of host switching, but it could also be non-adaptive, if it is an artefact of overcrowding.

Bakke et al. (2007) suggested that Gyrodactylus transmission must be linked either to the development of a hostile microenvironment, such as increased host immunity or death, or to the eventual density-dependent mechanisms within the parasite infra-population.

Although, to-date, the specific factors triggering Gyrodactylus migration are unknown, it is possible that at high parasite densities individual Gyrodactylus will experience diminishing results, and thus switching to another host – even if is not the optimal – can increase migrant’s fitness (Charnov, 1976). This alternative strategy provides a better

201! ! explanation to the observed host switching of Gt3, and it suggests that Gt3, and potentially wild G. turnbulli, are specialist parasites capable of exploiting non-optimal hosts under adverse conditions.

Cryptic species of Gyrodactylus

The two species of Gyrodactylus traditionally identified infecting guppies in Trinidad very similar morphologically, although preserved individuals can be properly identified based on their hamuli, ventral bars and marginal hooks (Harris, 1986; Harris and Lyles

1992). Despite the strong morphological similarities, both species display marked site preferences which can facilitate their distinction in a non-invasive way. For example, G. bullatarudis is commonly found on the anterior end of guppies, especially on the head and mouth (Harris and Lyles, 1992), while G. turnbulli on the caudal regions (Harris,

1986). Throughout this thesis I employ the general name of Gyrodactylus to refer to both species, acknowledging that without a proper identification it could be either of them.

Nonetheless, in an attempt to minimize the probability to mix multiple Gyrodactylus species in the experiments, all the parasites used were located in the posterior end, suggesting G. turnbulli. This is, of course, an important problem, as both species have been shown to differ in virulence – at least in the Aripo river (Cable and van Oosterhout et al., 2003), and can also form mixed infections (Harris and Lyles, 1992). Thus, it is possible that the results that I have presented here are simply an artifact of not having a consistent use of one or the other species. This is certainly a possibility. However, there is also evidence suggesting that despite species differences, guppies respond equally to both. Richards and Chubb (1996) tested guppy response to new and challenge infections

202! ! with G. turnbulli and G. bullatarudis, and found that the immune response in guppies did not depend on Gyrodactylus species. I believe the results I have presented in the four experimental chapters are a valid representation of guppy-Gyrodactylus interactions in their natural environment, and although I acknowledge that proper identification would have allowed me to make better and stronger inferences, I would have not been able to do the lengthy and complex experiments I have presented here. It is a trade-off I am happy to live with.

REFERENCES

1.! Appleby, C., and Mo, T.A. (1997) Population dynamics of Gyrodactylus salaris

(Monogenea) infecting Atlantic salmon, Salmo salar, in the river Batnfjordselva,

Norway. The Journal of Parasitology, 83: 23-30.

2.! Bakke, T.A., and MacKenzie, K. (1993) Comparative susceptibility of native

Scottish and Norwegian stocks of Atlantic salmon, Salmo salar L., to

Gyrodactylus salaris Malmberg: laboratory experiments. Fisheries Research, 17:

69-85.

3.! Bakke, T.A., Cable, J., and Harris, P.D. (2007) The biology of gyrodactylid

monogeneans: the “Russian-doll killers”. Advances in Parasitology, 64: 161-460.

4.! Bakke, T.A., Harris, P.D., and Cable, J. (2002) Host specificity dynamics:

observations on gyrodactylid monogeneans. International Journal for

Parasitology, 32: 281-308.

203! ! 5.! Barber, I. (2013) Sticklebacks as model hosts in ecological and evolutionary

parasitology. Trends in Parasitology, 29: 556-566.

6.! Charnov, E.L. (1976). Optimal foraging: the marginal value theorem. Theoretical

Population Biology. 9: 129-136.

7.! Choo, K., Williams, P.D., and Day, T. (2003) Host mortality, predation and the

evolution of parasite virulence. Ecology Letters, 6: 310-315.

8.! de Roij, J., Harris, P.D., & MacColl, A.D. (2011) Divergent resistance to a

monogenean flatworm among three‐spined stickleback populations. Functional

Ecology, 25: 217-226.

9.! Dove, A.D., and Ernst, I. (1998) Concurrent invaders—four exotic species of

Monogenea now established on exotic freshwater fishes in Australia.

International Journal for Parasitology, 28: 1755-1764.

10.!Ebert, D. (1994) Virulence and local adaptation of a horizontally transmitted

parasite. Science, 265: 1084–1086.

11.!Eizaguirre, C., Lenz, T.L., Sommerfeld, R.D., Harrod, C., Kalbe, M., and

Milinski, M. (2011) Parasite diversity, patterns of MHC II variation and olfactory

based mate choice in diverging three-spined stickleback ecotypes. Evolutionary

Ecology, 25: 605-622.

12.!Endler, J.A. (1980) Natural selection on color patterns in Poecilia reticulata.

Evolution, 34: 76-91.

13.! Fenton, A., and Brockhurst, M.A. (2008) The role of specialist parasites in

structuring host communities. Ecological research, 23: 795-804.

204! ! 14.!Gotanda, K.M., Delaire, L.C., Raeymaekers, J.A., Pérez-Jvostov, F., Dargent, F.,

Bentzen, P. and Hendry, A.P. (2013) Adding parasites to the guppy-predation

story: insights from field surveys. Oecologia, 172: 155-166.

15.!Hamilton, W.D., and Zuk, M. (1982) Heritable true fitness and bright birds: a role

for parasites? Science, 218: 384-387.

16.!Harris, P.D. (1983) The morphology and life-cycle of the oviparous

Oögyrodactylus farlowellae.(Monogenea, Gyrodactylidea). Parasitology, 87:

405-420.

17.!Harris, P.D. (1986) Species of Gyrodactylus von Nordmann, 1832 (Monogenea

Gyrodactylidae) from poeciliid fishes, with a description of G. turnbulli sp. nov.

from the guppy, Poecilia reticulata Peters. Journal of Natural History, 20: 183-

191.

18.!Harris, P.D., and Lyles, A.M. (1992) Infections of Gyrodactylus bullatarudis and

Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. The Journal

of Parasitology, 78: 912-914.

19.!Heggberget, T.G., Johnsen, B.O., Hindar, K., Jonsson, B., Hansen, L.P., Hvidsten,

N.A., and Jensen, A.J. (1993) Interactions between wild and cultured Atlantic

salmon: a review of the Norwegian experience. Fisheries Research, 18: 123-146.

20.! Hoeksema, J.D., and Forde, S.E. (2008) A meta-analysis of factors affecting local

adaptation between interacting species. American Naturalist, 171: 275-290. (DOI:

10.1086/527496)

21.!Houde, A.E. (1997) Sex, color, and mate choice in guppies. Princeton University

Press.

205! ! 22.!Johnsen, B.O., and Jensen, A.J. (1986) Infestations of Atlantic salmon, Salmo

salar, by Gyrodactylus salaris in Norwegian rivers. Journal of Fish Biology, 29:

233-241.

23.! Kaltz, O., and Shykoff, J. (1998) Local Adaptation in Host–parasite Systems.

Heredity 81: 361–370. doi:10.1046/j.1365-2540.1998.00435.x.

24.!Kawecki, T.J., and Ebert, D. (2004) Conceptual issues in local adaptation.

Ecology Letters, 7: 1225-1241.

25.!King, T.A., and Cable, J. (2007) Experimental infections of the monogenean

Gyrodactylus turnbulli indicate that it is not a strict specialist. International

Journal for Parasitology, 37: 663-672.

26.!King, T.A., van Oosterhout, C., and Cable, J. (2009) Experimental infections with

the tropical monogenean, Gyrodactylus bullatarudis: Potential invader or

experimental fluke? Parasitology International, 58: 249-254.

27.! Konijnendijk, N., Raeymaekers J.A.M., Vandeuren, S., Jacquemin L.,and

Volckaert, F.A.M. (2013) Testing for Local Adaptation in the Gasterosteus–

Gyrodactylus Host–parasite System. Evolutionary Ecology Research, 3: 489–502.

28.! Koskela, T., Salonen, V., and Mutikainen, P. (2000) Local adaptation of a

holoparasitic plant, Cuscuta europaea: variation among populations. Journal of

Evolutionary Biology, 13: 749–755.

29.! Lemoine, M., Doligez, B., and Richner, H. (2012) On the Equivalence of Host

Local Adaptation and Parasite Maladaptation: An Experimental Test. American

Naturalist. 179: 270–81. (DOI:10.1086/663699.)

206! ! 30.!Magurran, A.E. (2005) Evolutionary ecology: the Trinidadian guppy. Oxford

University Press.

31.!Martin, C.H., and Johnsen, S. (2007) A field test of the Hamilton–Zuk hypothesis

in the Trinidadian guppy (Poecilia reticulata). Behavioral Ecology and

Sociobiology, 61: 897-1909.

32.!Milinski, M. (2006) The major histocompatibility complex, sexual selection, and

mate choice. Annual Review of Ecology, Evolution, and Systematics, 37: 159-186.

33.!Pigliucci, M. (2005) Evolution of phenotypic plasticity: where are we going now?

Trends in Ecology and Evolution, 20: 481-486.

34.!Poulin, R. (1992) Determinants of host-specificity in parasites of freshwater

fishes. International Journal for Parasitology, 22: 753-758.

35.!Poulin, R. (2011) Evolutionary ecology of parasites. Princeton University Press.

36.!Prenter, J., MacNeil, C., Dick, J.T., and Dunn, A. M. (2004) Roles of parasites in

animal invasions. Trends in Ecology and Evolution, 19: 385-390.

37.!Raeymaekers, J.A., Wegner, M., Huyse, T., and Volckaert, F.A. (2011) Infection

dynamics of the monogenean parasite Gyrodactylus gasterostei on sympatric and

allopatric populations of the three-spined stickleback Gasterosteus aculeatus.

Folia Parasitologica, 58: 27-34.

38.!Yeh, P.J., and Price, T.D. (2004) Adaptive phenotypic plasticity and the

successful colonization of a novel environment. American Naturalist, 164: 531-

542.

207! ! 39.!Reznck, D.N. (1982) The impact of predation on life his- tory evolution in

Trinidadian guppies: the genetic basis of observed life history patterns. Evolution,

30: 160-177.

40.!Reznick, D.A., Bryga, H., and Endler, J.A. (1990) Experimentally induced life-

history evolution in a natural population. Nature, 346: 357-359.

41.!Reznick, D.N., and Bryga, H.A. (1996) Life-history evolution in guppies

(Poecilia reticulata: Poeciliidae). V. Genetic basis of parallelism in life histories.

American Naturalist, 147: 339-359.

42.!Richards, G.R., and Chubb, J.C. (1996) Host response to initial and challenge

infections, following treatment, of Gyrodactylus bullatarudis and G. turnbulli

(Monogenea) on the guppy (Poecilia reticulata). Parasitology Research, 82: 242-

247.

43.!Scott, M.E. (1985) Dynamics of challenge infections of Gyrodactylus bullatarudis

Turnbull (Monogenea) on guppies, Poecilia reticulata (Peters). Journal of Fish

Diseases, 8: 495-503.

44.!Scott, M.E., and Robinson, M.A. (1984) Challenge infections of Gyrodactylus

bullatamdis (Monogenea) on guppies, Poecilia reticulata (Peters), following

treatment. Journal of Fish Biology, 24: 581-586.

45.! Saarinen, M., and Taskinen, J. (2005) Local adaptation in a crustacean parasite–

molluscan host interaction: a field experiment. Evolutionary Ecology Research, 7:

1191–1199.

46.! Sternberg, E.D., Li, H., Wang, R., Gowler, C., de Roode, J.C. (2013) Patterns of

Host-Parasite Adaptation in Three Populations of Monarch Butterflies Infected

208! ! with a Naturally Occurring Protozoan Disease: Virulence, Resistance, and

Tolerance. American Naturalist, 182: E235–E248. (DOI:10.1086/673442.)

47.!Torchin, M.E., and Mitchell, C.E. (2004) Parasites, pathogens, and invasions by

plants and animals. Frontiers in Ecology and the Environment, 2: 183-190.

48.!Torres‐Dowdall, J., Handelsman, C.A., Reznick, D.N., and Ghalambor, C.K.

(2012) Local adaptation and the evolution of phenotypic plasticity in Trinidadian

guppies (Poecilia reticulata). Evolution, 66: 3432-3443.

49.!Van Oosterhout, C., Harris, P.D., and Cable, J. (2003) Marked variation in

parasite resistance between two wild populations of the Trinidadian guppy,

Poecilia reticulata (Pisces: Poeciliidae). Biological Journal of the Linnean

Society, 79: 645-651.

50.!Van Oosterhout, C., Mohammed, R.S., Hansen, H., Archard, G.A., McMullan,

M., Weese, D.J., and Cable, J. (2007) Selection by parasites in spate conditions in

wild Trinidadian guppies (Poecilia reticulata). International Journal for

Parasitology, 37: 805-812.

51.!Wegner, K.M., Kalbe, M., Rauch, G., Kurtz, J., Schaschl, H., and Reusch, T.B.H.

(2006) Genetic variation in MHC class II expression and interactions with MHC

sequence polymorphism in three‐spined sticklebacks. Molecular Ecology, 15:

1153-1164.

52.!Wegner, K.M., Reusch, T.B.H., and Kalbe, M. (2003) Multiple parasites are

driving major histocompatibility complex polymorphism in the wild. Journal of

Evolutionary Biology, 16: 224-232.

209! ! 53.!Williams, P.D., and Day, T. (2001) Interactions between sources of mortality and

the evolution of parasite virulence. Proceedings of the Royal Society of London B:

Biological Sciences, 268: 2331-2337.

210! !