Parasites and Cleaning Behaviour in Damselfishes Derek

Home , Amblyglyphidodon curacao, Amphiprion akindynos, Canthigaster papua, Chaetodon aureofasciatus, Chaetodon rainfordi, Chlorurus sordidus

Parasites and cleaning behaviour in damselfishes

Derek Sun

BMarSt, Honours I

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2015 School of Biological Sciences

Abstract

Pomacentrids (damselfishes) are one of the most common and diverse group of marine fishes found on coral reefs. However, their digenean fauna and cleaning interactions with the bluestreak cleaner wrasse, Labroides dimidiatus, are poorly studied. This thesis explores the digenean trematode fauna in damselfishes from Lizard Island, Great Barrier Reef (GBR), Australia and examines several aspects of the role of L. dimidiatus in the recruitment of young damselfishes. My first study aimed to expand our current knowledge of the digenean trematode fauna of damselfishes by examining this group of fishes from Lizard Island on the northern GBR. In a comprehensive study of the digenean trematodes of damselfishes, 358 individuals from 32 species of damselfishes were examined. I found 19 species of digeneans, 54 host/parasite combinations, 18 were new host records, and three were new species (Fellodistomidae n. sp., Gyliauchenidae n. sp. and Pseudobacciger cheneyae). Combined molecular and morphological analyses show that Hysterolecitha nahaensis, the single most common trematode, comprises a complex of cryptic species rather than just one species. This work highlights the importance of using both techniques in conjunction in order to identify digenean species. The host-specificity of digeneans within this group of fishes is relatively low. Most of the species possess either euryxenic (infecting multiple related species) or stenoxenic (infecting a diverse range of hosts) specificity, with only a handful of species being convincingly oioxenic (only found in one host species). Overall, the trematode richness in pomacentrid fishes is relatively low compared with that of many other coral reef fish families. The richness is probably best explained by the small size of most pomacentrids. In terms of beta diversity, of the total 23 trematode species having now been reported for the GBR in pomacentrids, just 14 are shared between Heron and Lizard Island. Five species have been recorded only from Lizard Island and four only from Heron Island. A total of 14 species reported from the two sites can be classified as core pomacentrid-specific parasites. I then examined whether the abundance and diversity of damselfish recruits differed on reefs relative to reefs from which cleaner wrasse had been removed for 12 years. Past studies have shown that cleaner wrasse enhance abundance and diversity of adult resident damselfishes (species that are site-attached), however, it is not known whether this effect occurs at recruitment or is driven by post-settlement events such as migration or differential mortality. I characterised the abundance of damselfish recruits on reefs with and without cleaner wrasse for five days in each of three lunar cycles, beginning four days after the new moon (November, December and January). The total abundance of damselfish recruits and specifically two species, Chrysiptera rollandi and P. amboinensis, was higher on reefs with cleaner wrasse than on reefs without. However, overall

species diversity of damselfish recruits did not differ. This study shows that the effects of the presence of cleaner wrasse begin at recruitment. Although the presence of cleaner wrasse increases the abundance of recruiting damselfish, it is unclear what mechanism(s) are involved in this process. I conducted aquarium experiments to examine whether young P. amboinensis preferentially select a microhabitat that is adjacent to a cleaner wrasse. The study showed that both settlement-stage and juvenile P. amboinensis spend a greater proportion of time in a microhabitat adjacent to a cleaner wrasse than in one adjacent to a control wrasse (a non-cleaner wrasse, Halichoeres melanurus) or where no fish was present. More time was spent next to juvenile than adult cleaner wrasse. I then conducted field observations to determine if recruits that are in a 1 m radius of a juvenile cleaner wrasse receive any direct benefits (cleaning) and whether juvenile cleaner wrasse preferentially cleaned larger individuals. The size and identity of all fish that were cleaned within a 1 m radius of a cleaner were recorded, along with nearby non-cleaned conspecifics. I showed that only 2 % of the clients that a juvenile cleaner wrasse clean per 20 min observation were small recruits (< 20 mm, TL) whereas larger damselfishes comprised 58 % of clients. Juvenile cleaner wrasse also preferred to clean larger individuals than the median size of the surrounding nearby non-cleaned conspecifics. These results show that cleaner wrasse serve as a positive cue during microhabitat selection but that direct benefits to a small recruit are not obvious. The possibilities of indirect rather than direct (cleaning) benefits are discussed. I then used sentinel traps to investigate the cumulative number of gnathiid isopods infecting the ambon damselfish, Pomacentrus amboinensis, on patch reefs with and without cleaner wrasse. Previous studies may have under surveyed the number of gnathiids on a host related to the presence or absence of cleaner fish, due to the sampling techniques used (wired cages or handnets). As gnathiids are temporary parasites, they can drop off a host and return to the benthos before the fish is sampled; sentinel traps are able to retain such individuals. Nocturnal collections produced almost three times more gnathiids than diurnal collections. These results are consistent with previous studies. There was no difference in the number of gnathiids found in sentinel traps placed on reefs with and without a cleaner wrasse. These results appear initially to be at variance with the many advantages to fishes reported as being associated with the presence of L. dimidiatus. These observations lead to the novel hypothesis that the advantage to clients from cleaners is in the reduction in the length of parasitism rather than the number of infections being reduced. This research has provided important information regarding the parasites of damselfishes, from the starting point of the characterisation of the parasite fauna to how cleaning behaviour affects the ecology of recruiting fish, thus increasing our understanding of recruitment processes on coral reefs. My study has demonstrated that the richness of damselfishes exceeds that of their iii

trematode parasites on the GBR although the parasite diversity (11 families) is remarkable. The work has contributed insight into both host range and species richness, having uncovered both novel trematode species, and novel hosts for known species. Regarding cleaning interactions, I have shown that 1) the abundance of damselfish recruits is higher on reefs with cleaner wrasse than on those without; 2) damselfish recruits preferentially select a microhabitat that is adjacent to a cleaner wrasse; and 3) that the benefits for young fish to be near cleaners are likely to be indirect rather than direct. These aspects of my study demonstrates the importance of cleaning mutualism not only for adult fishes, but for juvenile fishes during the recruitment period.

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

Publications during candidature

Sun D, Bray RA, Yong R, QY, Cutmore SC, Cribb TH (2014) Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia. Systematic Parasitology 88:141-152

Sun D, Blomberg SP, Cribb TH, McCormick MI, Grutter AS (2012) The effects of parasites on the early life stages of a damselfish. Coral Reefs 31:1065-1075

Publications included in this thesis

Incorporated as part of Chapter 2: Sun D, Bray RA, Yong R, QY, Cutmore SC, Cribb TH (2014) Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia. Systematic parasitology 88:141-152

TH Cribb, RA Bray and SC Cutmore provided feedback and comments on the manuscript. R. Q-Y Yong provided editorial advice and assistance with specimen measurements. SC Cutmore assisted with phylogenetic analyses. D Sun collected and analysed the data and wrote the manuscript.

Contributions by others to the thesis

Chapter 1: TH Cribb provided feedback and comments. D Sun was responsible for the remainder of the work.

Chapter 2: TH Cribb, RA Bray and R. Q-Y Yong provided feedback and comments on the manuscript. SC Cutmore assisted with phylogenetic analyses. D Sun collected and analysed the data and wrote the manuscript.

Part of Chapter 2: TH Cribb, RA Bray and SC Cutmore provided feedback and comments on the manuscript. R. Q-Y Yong provided editorial advice and assistance with specimen measurements. SC Cutmore assisted with phylogenetic analyses. D Sun collected and analysed the data and wrote the manuscript.

Chapter 3: KL Cheney and AS Grutter provided advice on experiment design. KL Cheney, J Werminghausen, TH Cribb and AS Grutter provided feedback and comments on the manuscript. D Sun designed and conducted the study, collected and analysed the data and wrote the manuscript.

Chapter 4: KL Cheney and AS Grutter provided advice on experimental design and statistical procedures. J Werminghausen assisted with data collection. KL Cheney, MI McCormick, MG Meekan, TH Cribb and AS Grutter provided with feedback and comments on the manuscript. D Sun designed and conducted the study, collected and analysed the data and wrote the manuscript.

Chapter 5: KL Cheney and AS Grutter provided advice on experimental design and statistical procedures. J Werminghausen and EC McClure assisted with field data collection. KL Cheney, J. Werminghausen, EC McClure, MI McCormick, MG Meekan, TH Cribb and AS Grutter provided editorial and feedback on the manuscript. D Sun designed and conducted the study, collected and analysed the data and wrote the manuscript.

Chapter 6: TH Cribb provided feedback and comments. D Sun was responsible for the remainder of the work.

Statement of parts of the thesis submitted to qualify for the award of another degree

None

vii

Acknowledgements

There are many people who need be thanked for their contribution to this project. However, first I would like to thank my two supervisors, Tom Cribb and Karen Cheney. Tom, your enthusiasm and knowledge in regards to parasites is remarkable. Throughout the years, including my honours years, you have guided and supported me and for that I am thankful. Karen, thank you for being supportive, sharing your knowledge and providing me with invaluable advice. I particularly want to acknowledge the support from Alexandra Grutter. Your encouragement, passion and enthusiasm for anything that I did, whether it be small or large made my candidature enjoyable. I’ve learnt so much and I can’t thank you enough for all of your guidance and advice. Words cannot express my gratitude to you for everything you have done for me. I would also like to thank both Mark McCormick and Mark Meekan for reading drafts, providing me with advice and for allowing me to use their light traps. Also a thank you to both Simon Blomberg and Jeff Hanson for helping me with all of my questions about statistics. I am truly grateful for the staff at the Lizard Island Research Station. I thank you for the support and the friendship that you have all showed me throughout the years. To Lyle, Anne, Marianne, Lance, Tania and Bob I can’t thank you all enough for helping me with this research, you have all gone way beyond what is required. Thank you all for the memorable events that we have all shared. They will never be forgotten. I am deeply thankful to all the people who have helped me throughout my candidature; Johanna, Eva, Fabio, Genevieve, Maarten, Russell, Scott, Anne and Andy. I would especially like to thank Johanna, Eva and Fabio for assisting me in the field, all without ever once complaining for having to get up at ungodly hours to help me with the light traps or the sentinel traps. Your assistance is one of the main reasons that I am able to complete this thesis. Genevieve and Maarten thank you for all of your help in the field, which includes helping me get out of that nally-bin on the boat. A special thank you goes to Russell, who throughout this PhD has read countless drafts and has taught me so much about marine parasitology. Scott for helping me with molecular analysis, as I’m utterly hopeless at it; Matt for teaching me how to put out the light traps and Anne and Andy for keeping me sane with our regular need for coffee and for your helpful comments and support. I am also extremely thankful to the members of the Marine Parasitology Lab, Coral Reef Ecology Lab and the people who occupy room 113 of Goddard throughout my candidature. I would like to thank the organisations that have provided me with financial support for the research conducted throughout my candidature. These include the Australian Research Council Discovery Grant (ARC grant to A. S. G. et al. and K. L. C. and T. H. C. et al.) and Sea World Research and Rescue Foundation, Australia (SWRRFI) for project funding. I would also like to viii

thank the International Society for Behavioural Ecology for the student travel award that has allowed me the opportunity to present this research at the conference in New York. Finally, I would like to thank my family and friends for always being supportive. Mum and Dad, thank you for all of your love and financial support. I can’t thank you enough. To my brothers and sister, William, Henry and Michelle, thank you for always being there for me, even though you thought that all I did was swim with fishes. I am also grateful to my close friends Karina, Glen, Justin and Simon for their constant support and kindness. I could always count on you guys to put a smile on my face. Love all of you guys. Thanks.

Keywords gnathiid isopods, cleaning mutualism, recruitment, trematodes, damselfish, ectoparasites, host- specificity, Great Barrier Reef, beta diversity

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 060201, Behavioural Ecology, 50% ANZSRC code: 060205, Marine and Estuarine Ecology, 20% ANZSRC code: 060301, Animal Systematics and Taxonomy, 30%

Fields of Research (FoR) Classification

FoR code: 0602, Ecology, 70% FoR code: 0699, Other Biological Sciences, 30%

Table of Contents

Abstract ...... ii Declaration by author ...... v Publications during candidature ...... vi Publications included in this thesis ...... vi Contributions by others to the thesis...... vii Statement of parts of the thesis submitted to qualify for the award of another degree ...... vii Acknowledgements...... viii Keywords ...... x Australian and New Zealand Standard Research Classifications (ANZSRC) ...... x Fields of Research (FoR) Classification ...... x Table of Contents ...... xi List of Figures ...... xiii List of Tables ...... xiv Abbreviations ...... xv Chapter 1: General Introduction...... 1 1.1 Preamble 2 1.2 Pomacentrids 2 1.3 Parasitism on coral reefs 2 1.4 Parasite diversity on coral reefs 3 1.5 Ecological importance of parasites 5 1.6 Cleaning mutualism 8 1.7 Aims and Thesis structure 11 Chapter 2: Trematodes of northern Great Barrier Reef Pomacentridae: richness, host- specificity and beta diversity ...... 14 Abstract 15 Introduction 15 Materials and methods 17 Results 19 Discussion 34 Acknowledgments 41 Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia ...... 42 Abstract 43 Introduction 43 Materials and methods 44 Gymnophalloidea Odhner, 1905 incertae familiae 47 Discussion 51 Acknowledgements 56 Chapter 3: Presence of cleaner wrasse increases the recruitment of fishes to coral reefs ...... 57 Abstract 58 Introduction 58 Material and methods 59 Results 60 Discussion 64 Acknowledgements 65

Chapter 4: Cleaner wrasse influence habitat selection of young damselfish ...... 66 Abstract 67 Introduction 68 Materials and methods 69 Results 74 Discussion 78 Acknowledgements 82 Chapter 5: Diel patterns of infection by gnathiid isopods on the ambon damsel, Pomacentrus amboinensis, and the influence of the presence of cleaner wrasse ...... 83 Abstract 84 Introduction 84 Materials and methods 88 Results 93 Discussion 95 Acknowledgements 98 Chapter 6: General Discussion ...... 99 References ...... 103 Appendices ...... 113

xii

List of Figures

Chapter 2: Trematodes of northern Great Barrier Reef Pomacentridae: richness, host- specificity and beta diversity

Figure 1 Body length and ventral sucker length of the three morphological types of Hysterolecitha nahaensis 24

Figure 2 Neighbour joining tree based on ITS2 rDNA sequences of Hysterolecitha nahaensis complex from Lizard Island. Samples are identified by host and dissection number. 25

Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia

Figures 12 Pseudobacciger cheneyae n. sp. ex Chromis weberi from off Lizard Island 49

Chapter 3: Presence of cleaner wrasse increases the recruitment of fishes to coral reefs

Figure 1 Total abundance of common damselfishes on experimental reefs. 61

Figure 2 Mean (± SE) abundance of recruiting damselfish species. 62

Figure 3 Diversity of all damselfish recruits per reef. 63

Chapter 4: Cleaner wrasse influence habitat selection of young damselfish

Figure 1 Front view of the laboratory experimental tank design. 71

Figure 2 The percentage of time spent on microhabitat adjacent to fish treatment by damselfish client Pomacentrus amboinensis. 75

Figure 3 Cleaning interactions of juvenile cleaner wrasse, Labroides dimidiatus and damselfishes. 77 Figure 4 The probability of an individual recruit (< 20 mm) being cleaned by a juvenile cleaner wrasse, Labroides dimidiatus, in relation to the abundance of a) other fish species and b) abundance of recruits per 20 min observations. 78

Chapter 5: Diel patterns of infection by gnathiid isopods on the ambon damsel, Pomacentrus amboinensis, and the influence of the presence of cleaner wrasse

Figure 1 Photograph of the sentinel trap used in the experiment. 91

Figure 2 Individual subcages used to hold Pomacentrus amboinensis in the sentinel traps. 92 xiii

List of Tables

Chapter 2: Trematodes of northern Great Barrier Reef Pomacentridae: richness, host- specificity and beta diversity

Table 1 Host-parasite associations for pomacentrid fishes and digenean trematodes from Lizard Island. 32

Table 2 Comparison between data collected from Heron Island (Barker et al. 1994) and Lizard Island (present study) for digeneans from pomacentrid fishes. 38

Table 3 Core pomacentrid specific digeneans from Heron and Lizard Island. 40

Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia

Table 1 Collection data and GenBank accession numbers for taxa sequenced during this study 46

Table 2 28S rDNA sequence data from GenBank included in phylogenetic analyses 50

Chapter 4: Cleaner wrasse influence habitat selection of young damselfish

Table 1 Ontogenetic stage and size of fish used in each part for laboratory experiment. Size is mean ± SE, TL. 71

Chapter 5: Diel patterns of infection by gnathiid isopods on the ambon damsel, Pomacentrus amboinensis, and the influence of the presence of cleaner wrasse

Table 1 Number of traps (with and with no fish) sampled per sampling day, site, reef pair cleaner presence treatment, and day and night. Total n per host = 382 traps. 90

Table 2 Number of sentinel traps with gnathiids, day and night trap samples and the number of gnathiids per trap collected from the sentinel trap experiment on Lizard Island 94

xiv

Abbreviations

GBR: Great Barrier Reef

PCO: Principle coordinate analysis

TK-HSD: Tukey-Kramer HSD

GLM: General linear model

LMER: Linear mixed-effect models

GLMER: Generalized linear mixed effect model

LRT: Likelihood ratio test

REML: Restricted maximum likelihood

PERMANOVA: Permutational multivariate analysis of variance

NJ: Neighbour joining

PVC: Polyvinyl chloride

SE: Standard error

TL: Total length

ITS2: Internal transcribed spacer 2 rDNA: Ribosomal Deoxyribonucleic acid mm: Millimetre h: Hours s: Seconds min: Minute n: Number m: Metre

Chapter 1: General Introduction

1.1 Preamble This study explores the complex interactions between three key faunal components on the Great Barrier Reef (GBR): the parasites (both endo- and ectoparasitics) that infect reef fishes, cleaner wrasse (Labroides dimidiatus) that feed on, and hence remove parasites from fishes, and pomacentrids fishes (damselfishes), a large and abundant reef fish family. In these interactions, pomacentrids are contextualised as both hosts to parasites, and clients to the cleaner wrasse. The fish and their parasites are abundant and diverse; the parasites remain poorly known. The effects of the parasites are subtle – we never see a pomacentrid dying because of its parasites – but real. The cleaner wrasse are keystone species whose parasite-eating behaviour has remarkable implications for the ecosystems overall. Here I review these components and conclude by outlining the studies that I have completed to further our understanding of them.

1.2 Pomacentrids The family Pomacentridae (damselfishes) is one of the most abundant and diverse group of marine fishes that inhabit coral reefs (Randall et al. 1997). Currently, there are 387 species of pomacentrids recognised globally of which more than 89 are reported for the GBR (Grutter, in review). They have been placed into five subfamilies: Abudefdufinae, Chrominae, Lepidozyginae (monospecific), Pomacentrinae and Stegastinae (Cooper et al. 2009). Damselfishes are known to occupy a wide range of habitats (e.g. rubble, coral and algae substrates) (Medeiros et al. 2010) and collectively have a rather broad diet, including planktivory, herbivory and omnivory (Meekan et al. 1995). These traits most likely have contributed to the success of this family on coral reefs. Damselfishes are thus an important family of fishes that are a major part of and directly influence the structure of benthic communities (Hixon and Brostoff 1983; Ceccarelli 2007).

1.3 Parasitism on coral reefs Parasitism is a non-mutualistic symbiotic relationship and is considered one of the most successful modes of life given the number of parasite species that exist (Poulin and Morand 2000). All groups of plants and animals are susceptible to parasitism and many groups have adopted parasitism as a biological strategy. The abundance of parasitism in the marine environment is demonstrated by the estimate of Rohde (2002) that there are 100,000 species of protistan and metazoan parasites of marine fishes globally.

Coral reefs are the most diverse and complex marine ecosystems and an important component of the diversity is comprised of parasites. Multiple parasite groups are abundant on coral reefs including Protista (multiple phyla), Myxozoa, Monogenea, Digenea, Cestoda, Nematoda, Acanthocephala, Copepoda, and Isopoda (Justine et al. 2010). Justine et al. (2010) predicted that the 1,700 fish species of New Caledonia harbour 17,000 parasite species in total. For trematode worms, Cribb et al. (2014b) predicted a total of 1,100 ̶ 1,800 species for the GBR and for monogeneans Whittington (1998) predicted there is at least one species for each GBR fish species. Despite these high estimates of richness, it is thought, by all estimates, that only a small fraction of species have been discovered or described (Cribb et al. 2014b). The parasite fauna of many vertebrate groups (e.g. many fish families) has not been surveyed. In addition, little is known about the ecology or ecological role of parasites on coral reefs. Relatively few ecological studies have focused on coral reef parasites relative to the volume of terrestrial and freshwater studies.

1.4 Parasite diversity on coral reefs Parasitological studies have demonstrated that parasites are highly abundant and diverse in the organisms that inhabit coral reefs, especially among coral reef fishes (Justine 2010). Here I focus on one group, the digenean trematodes. These parasites comprise one of the largest groups of internal metazoan parasites, comprising approximately 18,000 nominal species worldwide (Cribb et al. 2001). So far, 326 species have been described from the GBR, with a total of 1,100•1,800 species predicted for the region (Cribb et al. 2014b), making digeneans a dominant group of internal parasites of fishes on the GBR. Although studies examining the trematode fauna of GBR fishes span more than a century, it has only been in the last three decades that the group has been accorded detailed attention; the present total of 326 described species known now grew from only 30 species known in 1988 (Cribb et al. 2014b). The majority of known species of trematodes infect the gut, with a minority specialized in other organs, e.g. Aporocotylidae in the circulatory system (Smith 2002), Gorgoderidae in the urinary bladder (Ho et al. 2014) and Transversotrematidae under the scales (Hunter and Cribb 2012). They have complex life cycles that require multiple hosts (Poulin and Cribb 2002). In the most common life cycle form, infection of a molluscan first intermediate host by a miracidial larva leads to the asexual production of cercariae, which emerge from the mollusc and infect either the final host directly (Aprorocotylidae and Transversotrematidae), encyst in the open (e.g. Haploporidae and Gyliauchenidae), or infect a second intermediate host (e.g. Cryptogonimidae and Bucephalidae). In the second intermediate host the cercaria develops into a metacercaria, which has the biological goal of being consumed, with its host by the final host. This mode of infection is called trophic transmission (Brown et al. 2001).

The digenean fauna of GBR fishes has been examined systematically, with many of the major families of fish having been extensively examined. Understanding the host specificity of digeneans is a critical aspect in understanding transmission in this group. Host specificity among parasites has been classified into three categories; oioxenicity, where parasites infect a single host species (e.g. Lepotrema adlardi known only from Abudefduf bengalensis); stenoxenicity, where species of parasites infect related host species (e.g. Derogenes pearsoni from damselfishes) and euryxenicity where parasites infect a wide range of hosts (e.g. Aponurus laguncula known from 67 species from 29 fish families) (Miller et al. 2011b). Most of the survey effort for digeneans of GBR fishes has been concentrated at two locations, Heron Island on the southern end of the GBR, and Lizard Island in the north, with the reefs in between little surveyed. The explanation for this is the presence of dedicated research stations at the two locations, allowing for systematic collection and processing of samples. The first fish digeneans to be described from the GBR were two species from teleost fishes (Tetraodontidae) and two from elasmobranchs (Dasyatidae) by Johnston (1913). Of the 326 currently-known species, just 66 (20 %) species have been reported from both Heron and Lizard Island. However, many fish families have been examined more thoroughly at one location than the other. An important example is the Pomacentridae, which have been extensively surveyed at Heron Island, but not at Lizard Island. Studies on pomacentrids of the southern GBR (Cribb et al. 1992; Bray et al. 1993b; Bray et al. 1993a; Barker et al. 1994; Cribb et al. 1994a; Bray and Cribb 1998) demonstrated a fauna of seven families and 18 species of adult trematodes. These findings were based on a survey at Heron Island, comprising 312 individuals from 39 species. For Lizard Island on the northern GBR, where the present studies were based, nine species of trematodes from five families have been reported (Bray et al. 1993b; Grutter 1994; Cribb et al. 1998; Grutter et al. 2010; Sun et al. 2012), two of which (Lepotrema oyabitcha and Transversotrema damsella) are not known from Heron Island (Bray et al. 1993b; Hunter and Cribb 2012). Despite the recording of nine species at Lizard Island, the fauna there has not been studied systematically and the limited data available suggested that there are significant differences between the northern and southern GBR. The extent to which the parasite fauna of the northern and southern GBR is shared by and between fish species is poorly understood. There has been no systematic study of matched sets of statistically informative samples of the same species of fishes between the two sites.

1.5 Ecological importance of parasites The importance of parasites to fish health is a complex issue. Although a high proportion of individual fish everywhere are infected with parasites, pathogenic effects are rarely immediately evident and are only rarely studied in wild fish. It has been speculated that many parasitic infections are relatively benign (Candolin and Voigt 2001; Carrassón and Cribb 2014). When closely studied, however, fish parasites have often been shown to have a variety of negative effects on their host. They have been shown to affect critical life factors such as host growth rate (Adlard and Lester 1994; Jones and Grutter 2008), body condition (Lemly and Esch 1984), reproduction (Fogelman et al. 2009; Mensink et al. 2014), population dynamics (Scott and Dobson 1989), and behaviour (Barber et al. 2000; Roche et al. 2013b). For coral reef fishes, most parasite related research has been heavily biased towards diversity studies of the parasites of adult coral reef fishes (Cribb et al. 1994b; Grutter 1994; Morand et al. 2000) and little is known about the ecological role of most parasites. Few studies have examined the ecology of internal parasites on coral reef fish (Muñoz et al. 2006; Muñoz et al. 2007). Studies by Grutter et al. (2010) and Sun et al. (2012) showed that there was a difference in parasite (mostly worms) prevalence with age group in the young damselfishes. These studies also demonstrated that the presence of internal parasites may have an effect on the growth rate of an individual at or during settlement. One possible explanation for the limited amount of study of the effects of internal parasite on the ecology and behaviour of a host is that they are hidden within a host. There has been more work done on the ecology of ectoparasites, including copepods (e.g. Finley and Forrester 2003) and especially cymothoids and gnathiids. Adlard and Lester (1994) showed that barrier reef chromis, Chromis nitida, that were infected with the cymothoid isopod Anilocra pomacentri had lower growth rate and higher mortality than non-parasitised fish. Mortality was caused by several factors, including erratic swimming behaviour upon parasitic attachment leading to increased susceptibility to predation, and gross pathology leading to deterioration in body condition, particularly in juvenile fishes. Östlund-Nilsson et al. (2005) recorded higher oxygen consumption, an increase in pectoral fin beat frequency and reduced swimming speed for an apogonid fish parasitised with Anilocra apogonae. Fogelman and Grutter (2008) showed that three apogonid species infected by A. apogonae had reduced growth and survival. Gnathiid isopods are arguably the most ecologically important and certainly the most carefully studied of the ectoparasites of coral reef fishes. They are abundant in both temperate and tropical waters (Sikkel et al. 2009) and are common on coral reef fishes (Grutter 1994). They do not possess a pelagic phase like most other marine invertebrates (Smit and Davies 2004). The life cycle incorporates three juvenile (larval) stages - three zuphea stages (unfed phases), three praniza stages (fed phases) and the non-feeding adult (Smit and Davies 2004). These mobile ectoparasites emerge 5

from the benthos as the zuphea phase and locate a host fish. Once engorged on blood and body fluids, as the praniza phase, they return to the benthos and moult into the next larval phase. Gnathiids thus exhibit protelian parasitism - parasitic as juveniles and free-living as adults (Smit and Davies 2004). Unlike other ectoparasites such as copepods and isopods which may be permanent parasites, gnathiids isopods are temporary parasites and are also often referred to as micropredators (Lafferty and Kuris 2002). The mouthparts of gnathiids are highly modified enabling them to attach and feed on the surfaces of fishes (Monod 1926). Gnathiids attach themselves to the body, fins, eye rims, gill chambers, nares, lips and the buccal cavity of their fish host (Monod 1926). In feeding on the blood and tissue fluids from their hosts they can cause either tissue damage or, when present in large numbers, death (Marino et al. 2004; Hayes et al. 2011). The amount of time required for gnathiids to become fully engorged varies among both species and developmental stages, with feeding period being longer for larger species and later juvenile stages (Grutter 2003; Smit et al. 2003; Smit and Davies 2004). Gnathiids can become engorged in less than 10 minutes (Grutter 2003) but can attach for up to 10 h (Grutter 2003; Smit et al. 2003; Grutter et al. 2011). Previous studies on the ecology of gnathiids have demonstrated their importance on coral reefs. They can have a wide range of effects on adult fish. They can cause tissue damage (Heupel and Bennett 1999) and gill mucus shedding (Hayes et al. 2011), reduce haematocrit value (Jones and Grutter 2005), are probable vectors of haemogregarine blood parasites (Davies and Smit 2001; Davies et al. 2004; Curtis et al. 2013) and viral diseases (Davies et al. 2009), and can even cause mortality due to blood loss when numbers are high (Mugridge and Stallybrass 1983; Hayes et al. 2011). Infection by gnathiids also influences the behavioural interactions between cleaners and their clients (Grutter 2001; Sikkel et al. 2004). In contrast to the relatively detailed knowledge of the role/importance of gnathiids in adult fish biology, the nature of the interaction between gnathiids and young fish is poorly known. What is already known about gnathiids on adult fishes (i.e. prevalence, and feeding ecology) (Grutter 1995a,1996a,2003; Grutter 2008) suggests that they may also significantly affect young reef fish. However, only a handful of studies have investigated their effects on young reef fish. Jones and Grutter (2008) found that settling Dischistodus perspicillatus damselfish infected with three gnathiids each night for one week had slower growth, measured by a percent increase in size, than those infected with one or none. Grutter et al. (2008) recorded mortality (12 % and 16 %) for settling Neopomacentrus azysron damselfish infected with one or three gnathiids, respectively. Penfold et al. (2008) found that in the laboratory, gnathiid isopods killed 6.3 % of juvenile Acanthochromis polyacanthus damselfish that were < 10 mm SL compared with 0 % for uninfected individuals. Grutter et al. (2011) showed that settling P. amboinensis previously infected with a 6

gnathiid swam more slowly, had higher oxygen consumption, and had reduced survival relative to uninfected individuals. In the wild, 3.5 % of a population of juvenile P. amboinensis caught at dawn were infected with a gnathiid and in the laboratory the infection was found to last for up to 10 h (Grutter et al. 2011). These studies highlight the potential importance of gnathiids in the early life stages of coral reef fishes and how they may affect successful recruitment to the adult population. Although numerous studies have examined the emergence of gnathiids in both the Caribbean and on the GBR in regards to diurnal variation (Jacoby and Greenwood 1988; Grutter and Hendrikz 1999; Sikkel et al. 2006; Jones and Grutter 2007), little is known as to whether the presence of cleaner wrasse has a significant effect on the number of gnathiids on reef fish. Estimating the number of gnathiids on a fish is complicated. As gnathiids are temporary parasites, they are able to detach themselves from the host when it is disturbed (e.g. host is captured) or can be eaten by a cleaner fish before the host has been examined for gnathiids. Therefore, studies that have examined the number of gnathiids in relation to cleaner fish presence may have underestimated the number of gnathiids on a host. Previous studies that have examined the number of gnathiids on reef fishes have used sampling methods that may also have overlooked the possible interaction of cleaner fish and the behaviour of gnathiids. Grutter (1999a) showed that individual caged thicklip wrasse, Hemigymnus melapterus, had a 4.5-fold reduction in the number of gnathiid isopods on individuals when on reefs with a cleaner wrasse relative to those on reefs without. Cheney and Côté (2003) showed that longfin damselfish, Stegastes diencaeus, that were captured from territories that had access to a cleaning station had lower numbers of gnathiids than individuals that did not. Although these studies demonstrated that the presence of cleaner fish can affect the number of gnathiid isopods on reef fish, the findings only provide information in regards to the gnathiid load on the fish at a given time (time of host captured) and do not take into consideration the issues raised above in regards to the potential loss of gnathiids. Sikkel et al. (2011) reported a trap design that is able to quantify the cumulative number of gnathiids on a host over time. These ‘sentinel’ traps use live ‘bait’ fish to attract gnathiids that are seeking a host. However, unlike wire cages which have been used in the past (e.g. Grutter 1999a; Grutter 1999b) from which gnathiids can escape if disturbed or fed, these traps retain gnathiids once they have entered the trap. The traps also prevent any interaction between cleaner fish and the live fish inside the traps, thus preventing loss of gnathiids due to removal by cleaner fish. These traps thus have the capacity to provide an accurate representation of the cumulative number of gnathiids attracted to a fish over a period of time.

1.6 Cleaning mutualism Positive interspecific interactions are encounters between organisms that occur where at least one individual benefits while the other is not harmed (Bronstein 1994). Such interactions have been classified as either mutualism or facilitation/commensalism, with the outcome of the interaction often creating a more favourable environment, either directly or indirectly, for one or multiple species (Stachowicz 2001; Bronstein et al. 2004). Mutualistic interactions have been shown to be of great ecological importance within terrestrial (Bascompte and Jordano 2007), freshwater (Brown et al. 2012) and marine ecosystems (Knowlton and Rohwer 2003). Some mutualistic relationships can extend beyond the two species involved in the initial interaction to other species within the community, and sometimes the overall structure and function of ecosystems (Stachowicz 2001). Some organisms within mutualistic relationships can have a disproportionate effect on the success of other species and are known as “keystone” mutualists (Bond 1994). The presence of keystone mutualists may be pivotal in shaping the structure of communities. Keystone mutualists are known in both terrestrial (Harms et al. 2000) and aquatic environments (Caley et al. 1996). This is seen, for example, in the relationship between frugivores (notably mammals and birds) and plants via seed dispersal, which can lead to enhanced biodiversity of plants (Jordano 2000). On coral reefs, the bluestreak cleaner wrasse, Labroides dimidiatus, is considered a key stone mutualist because their presence affects coral reef fish biodiversity (Bshary 2003; Grutter et al. 2003; Waldie et al. 2011) through cleaning symbioses. Cleaning symbiosis is the removal and subsequent ingestion of ectoparasites, and diseased and injured tissue by cleaning organisms (Losey 1972; Côté 2000); it occurs in both the terrestrial and aquatic environment (Poulin and Grutter 1996; Weeks 2000; Cheney and Côté 2001; Becker et al. 2005; Biani et al. 2009). It has long been held that cleaning symbioses in the marine environment are mutualistic interactions in which cleaners benefit from the removal of parasites and tissue from their fish clients (Grutter 1996a; Arnal and Côté 2000; Arnal and Morand 2001) and clients benefit from a reduced ectoparasite load (Losey 1979; Gorlick et al. 1987; Grutter 1999b). In the marine environment cleaning behaviour is performed by narrow range of fish and crustaceans (see list in Poulin and Grutter 1996). Typically cleaners have defined territories, ‘cleaning stations’, which are visited by organisms, the ‘clients’ (Côté 2000). There is strong evidence to show that clients do obtain net benefits from the interaction (Bshary et al. 2007; Clague et al. 2011; Ros et al. 2011; Soares et al. 2011). Several studies have shown that the presence of cleaner fish affects the size and abundance of crustacean parasites on fish (Gorlick et al. 1987; Grutter 1999a; Cheney and Côté 2001; Grutter and Lester 2002; Clague et al. 2011), increases the size distribution of fish (Clague et al. 2011; Waldie et al. 2011), and affects the composition of client fish communities (Bshary 2003; Grutter et al. 2003; Waldie et al. 2011). 8

Labroides dimidiatus is the most common cleaner fish species on the GBR (Randall et al. 1997). At Lizard Island, some individual client fish visit a cleaning station up to 144 times a day (Grutter 1995a). Individual cleaner wrasses have been shown to consume an average of 1,218 ± 118 parasites, mostly gnathiids, from 2,297 ± 83 fish inspected each day (Grutter 1996a). These studies suggest that cleaner wrasse presence plays an important role in reducing the number of parasites, which may ultimately lead to fish choosing to be near a cleaner fish. On patch reefs around Lizard Island, GBR, the species richness and abundance of visitors (fish species that move between patch reefs) were reduced when L. dimidiatus individuals were removed over an 18-month period (Grutter et al. 2003). After 8.5 years of maintaining these same reefs free of L. dimidiatus, Waldie et al. (2011) found a 23 % and 37 % reduction in species richness and abundance of all resident fishes (species that are site-attached) relative to reefs that had cleaners. Furthermore, there was a 66 % reduction in the abundance of juvenile (< 70 mm TL) visitor fishes on reefs without L. dimidiatus. This reduction in juvenile abundance could explain the 23 % lower abundance and 33 % lower species richness of visitor fishes on these reefs. Although the benefits of L. dimidiatus are thus well documented for adults and/or visiting reef fishes, its importance in settlement processes is largely unknown. The ecological determinants of settlement choice for pelagic coral reef fish larvae are complex, but dramatically affect population dynamics and the size of future adult populations (Hoey and McCormick 2004; Heinlein et al. 2010). Selecting an optimum habitat can maximise the growth and survival of coral reef fish during and after settlement (McCormick and Hoey 2004; Lecchini et al. 2007a). Factors affecting microhabitat choice include water temperature (Grorud- Colvert and Sponaugle 2011; Lienart et al. 2014), habitat structure (Feary et al. 2007; Coker et al. 2012b), coral type (Jones 1988; Danilowicz 1996), position on the reef (Srinivasan 2003; Jordan et al. 2012), and presence of conspecifics (Ohman et al. 1998; Adam 2010). There has been only limited research on the significance of the presence of heterospecifics in influencing the microhabitat choice of young reef fish. A few studies have demonstrated a positive or negative correlation between presence of heterospecifics and fish settlement. The abundance of some wrasses (Labridae) is highest in the algal territories of herbivorous damselfishes, possibly due to increased algal cover providing shelter and food resources at settlement (Green 1994; Green 1998), suggesting correlation rather cause. Resident piscivorous groupers and moray eels inhibit levels of settlement of damselfish and acanthurids, but increase the settlement of a labrid, Thalassoma bifasciatum (Almany 2003). As mortality from predation is often low for cleaners (Côté 2000; Grutter 2004b), it was hypothesised that juvenile T. bifasciatum may be immune to piscivory due to the species being a cleaner (Almany 2003). In contrast, the presence of two resident damselfishes in the absence of a resident piscivore 9

increases the abundance of acanthurid settlers (Almany 2003). Given the high fish diversity of coral reefs (Allen 2008), the potential for complex interactions that influence settlement processes is considerable. The importance of gnathiids and other external parasites to fish health has also been demonstrated indirectly by studies such as that of Grutter et al. (2003), which showed a positive association between cleaner wrasse presence and fish richness and abundance. This relationship suggested that mobile reef fish actively seek cleaner wrasses in order to have gnathiids and other parasites removed. Clague et al. (2011) showed that on reefs with cleaner fish, older P. moluccensis were larger at a certain age than those from reefs with no cleaner fish. They also found that copepod abundance was higher on fish from reefs with no cleaner fish, but only for larger host fish. Gnathiids have also been demonstrated to be of importance during settlement, in that even a single gnathiid infection can decrease the settlement success of P. amboinensis larvae (Grutter et al. 2011). Thus, if gnathiids are not removed from fish by cleaner fish or gnathiid population densities are higher on reefs without cleaner fish, there could be a reduction in juvenile fish abundance. It would thus be an advantage for young reef fish to settle near cleaning stations. However, the role of cleaning behaviour with respect to influencing the settlement behaviour of juvenile fish remains relatively unstudied. As cleaner fish are known to reduce their clients’ ectoparasite loads, proximity to a cleaner station should be advantageous for resident fish species. Waldie et al. (2011) found a 66 % increase in juveniles of visiting fish species on reefs associated with a cleaner fish relative to those without them. However, no study has examined the effect of cleaner wrasse presence on the juveniles of resident fish species, most of which consist of damselfishes. Previous studies have shown that the ability to locate suitable habitats during settlement can have significant advantages in terms of enhancing growth and/or increasing survival of young reef fish (Bonin et al. 2011; Vail and McCormick 2011; Rankin and Sponaugle 2014). Young reef fish have highly developed visual (Wenger et al. 2011; Lecchini et al. 2014), olfactory (McCormick et al. 2010; Paris et al. 2013) and auditory senses (Radford et al. 2011; Huijbers et al. 2012), which are known to play a critical role when selecting suitable habitats. The humbug damselfish Dascyllus aruanus has been demonstrated to use olfactory cues released by conspecifics in selecting reefs to settle on (Sweatman 1988). Vail and McCormick (2011) demonstrated that settling damselfishes use olfaction to recognise predators and thus avoid settling onto habitats where predators are present. At present, it is not known whether young reef fish that are found near a cleaning station have selected these reefs by using visual or other cues. It is known that naïve larvae of Hawaiian humbug Dascyllus albisella, raised in captivity innately recognise the Hawaiian cleaner wrasse, Labroides phthirophagus (Losey et al. 1995). That study demonstrated that experience with a 10

cleaner fish is not required for D. albisella to recognise a cleaner fish and to adopt the client ‘pose’. Although this study showed that young reef fish recognise cleaner fish, no studies have tested whether cleaner fish presence plays a role in larval habitat choice. A similar mechanism may occur with young reef fish that are approaching or are making the transition to the benthos, in that they may partly select reefs based on cues related, directly or indirectly, to cleaner fish presence. As most coral reef fish (especially damselfishes) are relatively site attached and will not leave their territory to seek a cleaner fish (Randall et al. 1997), any selection would have to occur while they are in the process of settling onto the reef or soon after as post-settlement movement (Lewis 1997; Simpson et al. 2008). There are other potential benefits besides a reduction in parasite loads for fish that settle in the vicinity of a cleaner wrasse. Cheney et al. (2008) showed that the presence of a cleaner wrasse reduces the frequency of aggressive chases by a piscivore towards potential prey species. This reduction in aggression by predators at cleaning stations may indirectly benefit prey species, by reducing the risk of being consumed and influencing how long they may remain at a cleaning station when a predator is present. No doubt such an indirect effect of cleaner fish presence would benefit such vulnerable young reef fish during settlement, as predation is the major cause of mortality at that period.

1.7 Aims and thesis structure The broad aim of this study was to improve understanding of a range of parasite-related issues from the richness of digenean trematodes to the importance of cleaning behaviour in damselfishes. The thesis thus explores a series of interrelated topics relating to the parasites of damselfishes and their ecological implications: (1) the richness and prevalence of digenean trematodes in damselfishes from the northern GBR, Lizard Island; (2) the effects of cleaner wrasse presence on the abundance of recruiting damselfishes; (3) the effect of cleaner wrasse presence in a microhabitat choice of a damselfish; and (4) the cumulative number of gnathiid isopods infecting a species of damselfish relative to cleaner wrasse presence. These topics are explored through both field and laboratory- based experiments. The thesis is presented as a collection of chapters written in a style that is suitable for publication. The thesis consists of six chapters of which four (2 ̶ 5) are in a manuscript format.

Chapter 2. Trematodes of northern Great Barrier Reef Pomacentridae: richness, host- specificity and beta diversity

There are 326 species of trematodes reported so far for the GBR (Cribb et al. 2014b) including 18 species known from damselfishes. Although damselfishes (Pomacentridae) are a major component of the coral reef fish assemblage, and much is known in regards to their ecology and biology, the trematode fauna of damselfishes on the northern GBR, has not been examined systematically. Current knowledge regarding the trematode fauna of damselfishes is mainly from systematic surveys conducted on Heron Island, southern GBR (Cribb et al. 1992; Bray et al. 1993a; Bray et al. 1993b; Barker et al. 1994). In contrast to damselfishes, other families of fishes have been extensively surveyed for trematodes at both Heron and Lizard Island - chaetodontids (Bray et al. 1994; McNamara et al. 2012; Diaz et al. 2013), lutjanids (Miller and Cribb 2007; Miller and Cribb 2008) and serranids (Bott and Cribb 2009; Bott et al. 2013). An extensive survey of damselfish trematodes from Lizard Island, far north GBR formed the first component of my PhD research. This study led to the description of one new species, Pseudobacciger cheneyae, published as separate standalone paper (incorporated in chapter 2).

Chapter 3. Presence of cleaner wrasse increases the recruitment of fishes to coral reefs. Although the direct and indirect effects of cleaner fish have been demonstrated for adult visitor reef fish, they have yet to be examined for resident larval reef fishes. Clague et al. (2011) showed that, on reefs with cleaner fish, older Pomacentrus moluccensis were larger at a certain age than those from reefs without cleaner fish. Gnathiids have been demonstrated to be important during settlement, in that a single gnathiid can decrease the success of settlement of P. amboinensis larvae significantly (Grutter et al. 2011). Therefore, if gnathiids are not removed from fish by cleaner fish (direct affect) or gnathiid population densities are higher on reefs without cleaner fish (indirect affect), there could be a reduction in juvenile fish abundance. Cleaner presence could also affect larval fish growth or survival through modification of the behaviour of predatory fish and thus the risk of predation, as there is evidence that predators reduce aggression towards bystanders in the presence of a cleaner wrasse (Cheney et al. 2008). I examined the effects of cleaner fish presence on juvenile density using patch reefs that have been experimentally manipulated for 12 years by being kept free of cleaner wrasses (Clague et al. 2011; Waldie et al. 2011).

Chapter 4. Cleaner wrasse influence habitat selection of young damselfish. The ability to locate suitable habitats during settlement can confer significant advantages on young fish by enhancing growth and/or increasing survival (McCormick and Hoey 2004). Young reef fish have highly developed visual and olfactory senses (Lecchini et al. 2007b) and these play a critical role in selecting suitable habitats. Naïve captive-raised larval Hawaiian humbug, Dascyllus albisella, innately recognise Hawaiian cleaner wrasse, Labroides phthirophagus (Losey et al. 1995). 12

Although that study showed that reef fish larvae recognise cleaner fish, no studies have demonstrated a link between cleaner fish presence and larval microhabitat choice. Parasites such as gnathiids have already been shown to decrease settlement success of a fish, thus I hypothesise that it would be an advantage for young reef fish to settle near cleaning stations. I tested whether young reef fish preferentially choose a microhabitat that is adjacent to a cleaner fish and whether young fish found at cleaning stations receive direct benefits in the form of cleaning services. If young fish actively select a microhabitat that is near a cleaner fish, this selection may help increase their survival after settlement.

Chapter 5. Diel patterns of infection by gnathiid isopods on the ambon damsel, Pomacentrus amboinensis, and the influence of the presence of cleaner wrasse. Gnathiid isopods are the most common temporary ectoparasites of coral reef fishes (Grutter and Poulin 1998), and are the key food of cleaner wrasse (Grutter 1997). The presence of cleaner fish is known to reduce the number of gnathiids on reef fish (Grutter 1999a; Cheney and Côté 2001) which could explain the higher abundances of fishes on reefs where cleaner wrasse are present (Grutter et al. 2003; Waldie et al. 2011) and also the abundance of recruiting fishes. I hypothesise that predation of gnathiids by cleaner wrasse leads to a reduction in the overall population of gnathiids on reefs with cleaners relative to those without. Although previous studies have examined the number of gnathiids on reef fishes (Grutter 1994; Grutter 1999a; Cheney and Côté 2001; Grutter et al. 2011), these studies only characterise the number of gnathiids on an individual at a specific time point. As a result, the rate at which gnathiids infect reef fish, and whether that rate changes over the course of a day, is not understood. Fish might be under surveyed due to gnathiids either dropping off due to being fully fed or having been eaten by a cleaner fish. In this study I used sentinel traps that retain gnathiids once they drop off and prevent physical interaction between cleaner wrasse and client. It is possible that individuals that are vulnerable to gnathiid infection such as young fish are more commonly infected with gnathiids than has been shown previously. By collecting samples throughout the day and night (24 hours), I explored the overall cumulative number of gnathiids and also examined of how it varies with respect to cleaner wrasse presence.

Chapter 6. General Discussion This chapter integrates the major findings of each study and makes recommendations for future research.

Chapter 2: Trematodes of northern Great Barrier Reef Pomacentridae: richness, host-specificity and beta diversity

Derek Sun, Rodney A. Bray, Russell Q-Y. Yong, Scott C. Cutmore & Thomas H. Cribb

Abstract Species richness and host specificity of the digenean fauna of the family Pomacentridae was studied by examining 358 individuals of 32 species of pomacentrids from Lizard Island, northern Great Barrier Reef (GBR). In total, 19 clear morphotype species of digeneans, 54 host/parasite combinations, 19 new host distributions, 18 new host records and three new species, Fellodistomidae n. sp., Gyliauchenidae n. sp., and Pseudobacciger cheneyae Sun, Bray, Yong, Cutmore & Cribb, 2014, were recorded from this large and important group of coral reef fishes. Combined morphological and molecular analysis of specimens initially identified as Hysterolecitha nahaensis suggested that on the basis of morphology it is possible to recognise three species and from molecular analysis as many as seven. The two sources of data could not be completely reconciled and the system requires further work. For the purpose of this analysis all such forms are identified as “H. nahaensis complex”. Further examination of the host-parasite relationship showed that host specificity is relatively low; only three species are apparently oioxenic whereas the majority possessed either stenoxenic or euryxenic host specificity. Overall, a total of 23 species of digenean trematodes have now been reported in pomacentrids from the GBR of which 13 are present at both Lizard and Heron Island. Five and four species have been reported only from Lizard and Heron Island, respectively. Lastly, 14 of the 23 species have been reported from only pomacentrids. The richness that was recorded is similar to that reported previously at Heron Island. This study suggests that the diversity of pomacentrids species far exceeds that of their digenean parasites at Lizard Island.

Introduction Damselfishes (Pomacentridae) are one of the most diverse families of marine fishes occurring on coral reefs (Randall et al. 1997). Currently, there are 387 species of pomacentrids recognised globally in both marine and brackish systems (Froese and Pauly 2012). The family is divided into

five subfamilies: Abudefdufinae, Chrominae, Lepidozyginae (monospecific), Pomacentrinae and Stegastinae (Cooper et al. 2009); all five subfamilies are represented on the GBR and more than 80 species are known in the region (Randall et al. 1997). Habitat selection and a wide diversity of diet preferences of damselfishes may have aided diversification of this family on coral reefs. The large diversity within this group has allowed damselfishes to occupy and exploit a range of microhabitats such as living corals (Wilson et al. 2008), rubble/dead coral (Bay et al. 2001), and algae-dominated areas (Medeiros et al. 2010), as well as exploiting both benthic and pelagic niches (Randall et al. 1997). Damselfishes can be divided into three major trophic guilds: planktivores, herbivores and omnivores (Cooper et al. 2009); the scope for exploitation by trophically transmitted parasites, including trematodes, in this family is thus substantial. Host specificity is an important aspect of the biology of parasites (Adamson and Caira 1994). Host specificity of parasites can be classified into three broad categories: oioxenicity, where an individual parasite species is known to only infect a single host species; stenoxenicity, where the parasite infects multiple phylogenetically related species; and euryxenicity, where a species infects a diverse range of hosts united by ecology and physiology rather than taxonomy (Miller et al. 2011b). Miller et al. (2011b) showed that 60 % of the trematodes then known from the GBR were oioxenic, 32 % were stenoxenic and just 8 % were euryxenic. For the Pomacentridae specifically, Barker et al. (1994) found that, of the 18 species of digeneans known from pomacentrids at Heron Island, six [Bivesicula unexpecta Cribb, Bray & Barker, 1994, Haplosplanchnidae sp., Lepotrema adlardi (Bray, Cribb & Barker, 1993), Lepotrema sp. 1, Lepotrema sp. 2 and Steganoderma gibsoni Cribb, Bray &Barker, 1992] were oioxenic, three (Derogenes pearsoni Bray, Cribb & Barker, 1993, D. pharyngicola Bray, Cribb & Barker, 1993 and Hysterolecitha heronensis Bray, Cribb & Barker, 1993) were strictly stenoxenic, another two [Hysterolecitha nahaensis Yamaguti, 1942 and Schikhobalotrema pomacentri (Manter, 1937)] were predominantly stenoxenic (i.e. primarily found in one group of closely-related hosts, but occasionally found in hosts from unrelated families) and seven [Aponurus laguncula Loos, 1907, Lepotrema clavatum (Ozaki, 1932), Lecithaster stellatus Looss, 1907, Macvicaria heronensis Bray & Cribb, 1989, Preptetos cannoni Barker, Bray & Cribb, 1993, P. xesuri (Yamaguti, 1940) Pritchard, 1960 and Thulinia microrchis (Yamaguti, 1934) Bray, Cribb & Barker, 1993] were euryxenic. The host range thus varied greatly between species, with one, Hysterolecitha nahaensis, found in 24 pomacentrid species. Although damselfishes have been the focus of many ecological studies in reef fishes their parasitic fauna remains incompletely known. The first dedicated investigation into damselfish digenean fauna was conducted at Heron Island in the early 1990s. A total of 39 pomacentrid species was examined, resulting in the discovery of 18 species of trematodes and the description of 8 new 16

species (Cribb et al. 1992; Bray et al. 1993b; Bray et al. 1993a; Bray and Cribb 1998). Relatively few studies have been conducted elsewhere on the GBR; of these, work was primarily conducted at Lizard Island. However, unlike at Heron Island, these studies focused on individual species of damselfishes or a particular family of digenean trematode (Grutter et al. 2010; Hunter and Cribb 2010; Sun et al. 2012; Sun et al. 2014). Thus, the trematode fauna of damselfishes at other locations on the GBR, including Lizard Island, can be considered to be under-studied. Studies conducted elsewhere in the world (e.g. Yeo and Spieler 1980; Dyer et al. 1985) have demonstrated a relatively depauperate digenean fauna in damselfishes. This is the first study to have systematically examined the diversity and host specificity of digenean trematodes in the family Pomacentridae from Lizard Island in the northern Great Barrier Reef and compared its nature relative to the existing known fauna at Heron Island reef 1,184 km to the south.

Materials and methods Specimen collection Between 2011 and 2012, pomacentrids were collected from Lizard Island (1440S, 14528E), Great Barrier Reef, Australia. These fish were collected using a barrier net, spear gun or clove oil, and were returned to the research station and euthanized prior to dissection via cranial pithing. As transversotrematids live beneath the scales of fish the body was soaked in vertebrate saline for a minimum of 10 minutes. This allowed any transversotrematids to fall to the bottom of the container. The heart was removed from the body and examined carefully for aporocotylids. The contents of the gut cavity were removed and examined under a dissecting microscope using fine scissors and forceps. A “gut wash” was conducted in an attempt to dislodge parasites attached to the gut wall (Cribb and Bray 2010). Specimens were killed and heat-fixed by pipetting the worms into 0.85 % saline just off the boil and transferred to Eppendorf tubes containing 70 % absolute ethanol (Cribb and Bray 2010).

Morphological analysis Preserved worms were stained using Mayer’s haematoxylin, destained in 1 % hydrochloric acid, neutralised with 1 % ammonia solution, dehydrated through a graded series of ethanols (50, 70, 80, 90 and 100 %), cleared in methyl salicylate and then mounted on slides in Canada balsam. All measurements were made using a digital camera and the programme SPOT Advanced (version 4.6). Voucher specimens are lodged within the Queensland Museum.

Molecular analysis All voucher specimens are paragenophores (sensu Pleijel et al. 2008). All vouchers are lodged in the Queensland Museum, Brisbane, Australia. Total genomic DNA was extracted using standard phenol/chloroform extraction techniques (Sambrook and Russell 2001). The D1-D3 regions of 28S nuclear ribosomal DNA was amplified using the primers LSU5 (5'-TAG GTC GAC CCG CTG AAY TTA AGC-3') (Littlewood et al. 2000) and ECD2 (5'-CTT GGT CCG TGT TTC AAG ACG GG-3') (Littlewood et al. 2000) and the ITS2 region using the primers 3S (5'-GGT ACC GGT GGA TCA CGT GGC TAG TG-3') (Bowles et al. 1993) and ITS2.2 (5'-CCT GGT TAG TTT CTT TTC CTC CGC-3' ) (Cribb et al. 1998) PCR for both the 28S and ITS2 regions was performed with a total volume of 20 μl consisting of approximately 10 ng of DNA, 5μl of 5x MyTaq Reaction Buffer (Bioline), 0.75 μl of each primer (10 pmols) and 0.25 μl of Taq DNA polymerase (BiolineMyTaq™ DNA Polymerase), made up to 20 μl with Invitrogen™ ultraPURE™ distilled water. Amplification was carried out on a MJ Research PTC-150 thermocycler. The following profile was used to amplify the 28S region: an initial 95°C denaturation for 4 min, followed by 30 cycles of 95°C denaturation for 1 min, 56°C annealing for 1 min, 72°C extension for 2 min, followed by a single cycle of 95°C denaturation for 1 min, 55°C annealing for 45 sec and a final 72°C extension for 4 min. The following profile was used to amplify the ITS2 region: an initial single cycle of 95°C denaturation for 3 min, 45°C annealing for 2 min, 72°C extension for 90 sec, followed by 4 cycles of 95°C denaturation for 45 sec, 50°C annealing for 45 sec, 72°C extension for 90 sec, followed by 30 cycles of 95°C denaturation for 20 sec, 52°C annealing for 20 sec, 72°C extension for 90 sec, followed by a final 72°C extension for 5 min. Amplified DNA was purified using a BiolineISOLATE II PCR and Gel Kit, according to the manufacturer’s protocol. Cycle sequencing of purified DNA was carried out for the ITS2 rDNA region using the forward primer GA1 (5'-AGA ACA TCG ACA TCT TGA AC-3') (Anderson and Barker 1998) and the reverse primer ITS2.2 and for the 28S rDNA region using the same primers used for PCR amplification as well as the additional primer 300F (5'-CAA GTA CCG TGA GGG AAA GTT-3')(Littlewood et al. 2000). Cycle sequencing was carried out at the Australian Genome Research Facility using an AB3730xl capillary sequencer. Sequencher™ version 4.5 (GeneCodes Corp.) was used to assemble and edit contiguous sequences. Sequences were aligned using Muscle implemented with MEGA 5.2. Differences between sequences were explored by creating distance matrices and neighbour joining (NJ) trees.

Results A total of 358 individual fish were examined from 32 species of pomacentrids from Lizard Island. In total, 19 clear morphotype species of digeneans, 54 host/parasite combinations, 18 new host records, 19 new host parasite/locality records and three new species were recorded (see Table 1).

Below, the details of observations relating to the 19 species are summarised. Bivesiculidae Yamaguti, 1934 Bivesicula unexpecta Cribb, Bray & Barker, 1994 Host: Acanthochromis polyacanthus (Bleeker, 1855). Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: This species was described from Acanthochromis polyacanthus by Cribb et al. (1994a) at Heron Island. Sequence data for this species from the same host species was subsequently reported for both Heron and Lizard Islands (Cribb et al. 1998). The present specimens are entirely consistent with the original description of this species. Bivesicula unexpecta remains the only bivesiculid species to have been reported as an adult from a pomacentrid and A. polyacanthus remains the only known host species. Immature bivesiculids were found in several other pomacentrids in this study, but these were all consistent with Bivesicula claviformis Yamaguti, 1934, a species that matures in serranid fishes and has immature stages in the intestine of a range of fishes preyed on by serranids (Cribb et al. 1998).

Derogenidae Nicoll, 1910 Derogenes pearsoni Bray, Cribb & Barker, 1993 Hosts: Chrysiptera cyanea (Quoy & Gaimard, 1825), Pomacentrus adelus Allen, 1991, Stegastes apicalis (De Vis, 1885). Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Our specimens of this species match those reported by Bray et al. (1993a) from the damselfishes Amblyglyphidodon curacao (Bloch, 1787), Amphiprion akindynos Allen, 1972, Plectroglyphidodon lacrymatus (Quoy & Gaimard, 1985), Pomacentrus chrysurus Cuvier, 1830, P. moluccensis Bleeker, 1853, P. tripunctatus Cuvier, 1830, and an unidentified species of Pomacentrus sp. from Heron Island, in the southern GBR. Sun et al. (2012) reported D. pearsoni from Pomacentrus amboinensis Bleeker, 1868 from Lizard Island. The three hosts reported here represent new host records. Derogenes pearsoni differs from the other species of Derogenes found in GBR pomacentrids, D. pharyngicola Bray, Cribb & Barker, 1993, in having shorter caeca, the 19

structure of the sinus-sac, the location of the posterior genital pore and the development of the uterus in the forebody (Bray et al. 1993a). Derogenes pharyngicola was not detected in the present study.

Faustulidae Poche, 1926 Pseudobacciger cheneyae Sun, Bray, Yong, Cutmore & Cribb, 2014 Host: Chromis weberi Fowler & Bean, 1928. Locality: Lizard Island Voucher specimens: QM G 234368G 234387 Remarks: This species was described from Chromis weberi by Sun et al. (2014) (as part of this study, published paper appended to this chapter) from Lizard Island. Pseudobacciger cheneyae is the first faustulid to have been reported from a pomacentrid, with other members of the genus known primarily from clupeids, but also from pomacanthids and sparids (Parukhin 1976b; Machida 1984; Dimitrov et al. 1999). The systematics of this species, its family and higher classification are problematic. Although the genus Pseudobacciger is classified within the family Faustulidae, phylogenetic analyses show that P. cheneyae is more closely related to members of the superfamily Gymnophalloidea, which includes, amongst others, the Fellodistomidae and Tandanicolidae, to which P. cheneyae is sister (Sun et al. 2014). Currently, P. cheneyae displays oioxenic host-specificity in that it has only been recorded from one host species, despite the ecology and behaviour of C. weberi resembling that of other planktivorous damselfishes.

Fellodistomidae Nicoll, 1909 Fellodistomidae n. sp. Host: Chrysiptera rollandi (Whitley, 1961). Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: This species was represented by just a single specimen from 1 of 9 Chrysiptera rollandi examined. The specimen agrees with the family Fellodistomidae in its smooth tegument, restricted vitellarian and Y-shaped excretory vesicle but not with any presently recognised genus. The fact that only a single specimen was found casts doubt on its status as a genuine parasite of pomacentrids, but no such form has been seen in any of thousands of other fishes examined at Lizard Island. The status of this form remains unclear.

Gyliauchenidae Fukui, 1929 Gyliauchenidae n. sp. Host: Stegastes apicalis (De Vis, 1885). Locality: Lizard Island Deposited specimens: QM XXXXX Remarks: The Gyliauchenidae is a small family of trematodes in the superfamily Lepocreadioidea, whose members primarily infect herbivorous reef fishes (Bray et al. 2009). Only one species, Gyliauchen pomacentri Nahhas & Wetzel, 1995, has been previously reported from a pomacentrid fish, Pomacentrus philippinus from off Suva, Fiji (Nahhas and Wetzel 1995). A description of this new species is being prepared in collaboration with Dr Kathryn Hall of the Queensland Museum.

Haplosplanchnidae Poche, 1926 Schikhobalotrema pomacentri Skrjabin & Guschanskaja, 1955 Hosts: Amblyglyphidodon curacao (Bloch, 1787), Chromis viridis (Cuvier, 1830), Pomacentrus moluccensis Bleeker, 1853. Voucher specimen: QM XXXXX Remarks: The family Haplosplanchnidae is known from a wide range of marine teleosts, with three species from the genus Schikhobalotrema Skrjabin & Guschanskaja, 1955, reported from pomacentrids. Schikhobalotrema crassum Pritchard & Manter, 1961, was described from Stegastes fasciolatus (Ogilby, 1889) from Hawaii, S. robustum Pritchard & Manter, 1961 was described from S. fasciolatus as well as two acanthurid and one chaetodontid species, also from Hawaii (Pritchard and Manter 1961), and S. pomacentri (Manter, 1937) Skrjabin & Guschanskaja, 1955, has been reported from 17 species of pomacentrid (plus one labrid species), from both Pacific and Atlantic coasts of North, Central and South America, as well as Heron Island (Cribb et al. 1994). The present specimens were consistent with S. pomacentri. The taxonomy of this genus has not been considered in any detail for specimens from the GBR and we think it is likely that parallel molecular studies will be needed to resolve it. The identification of S. pomacentri should be considered provisional.

Lecithasteridae Odhner, 1905 Aponurus laguncula Looss, 1907 Host: Abudefduf whitleyi Allen & Robertson, 1974 Locality: Lizard Island Voucher specimens: QM XXXXX

Remarks: This species is reported as being cosmopolitan, being found in various locations worldwide and infecting over 60 different host species from 28 families (Bray & Mackenzie, 1990). Bray et al. (1993a) first reported and described this species from the Pomacentridae, having discovered it in Pomacentrus moluccensis from Heron Island, while Grutter et al. (2010) and Sun et al. (2012) reported it from P. moluccensis and P. amboinensis from Lizard Island, respectively. Our discovery of A. laguncula from the Whitley’s sergeant, Abudefduf whitleyi, represents a new host record for this species. This species differs from the closely-related A. rhinoplagusiae Yamaguti, 1934, which also infects pomacentrids, in having a smaller seminal receptacle (Bray et al. 1993a).

Hysterolecitha heronensis Bray, Cribb & Barker, 1993 Host: Pomacentrus amboinensis Bleeker, 1868 Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Our specimens of this species match the original description by Bray et al. (1993a) from five species of Pomacentridae from Heron Island, Pomacentrus amboinensis, P. moluccensis, P. nigromarginatus Allen, 1973, P. philippinus Evermann & Searle, 1907, and an unidentified Pomacentrus sp. It was later reported from P. amboinensis from Lizard Island (Sun et al. 2012). Hysterolecitha heronensis is similar to H. nahaensis Yamaguti, 1942, with which it may have been confused historically (Bray et al. 1993a). It is distinguished from H. nahaensis by possessing notably digitiform vitelline lobes, a larger sinus-sac, a greater average sucker ratio and more distinct genital atrium. It is noticeable that at Lizard Island this species has been found only in species of one pomacentrid genus, Pomacentrus.

Hysterolecitha nahaensis Yamaguti, 1942 complex Host: Abudefduf bengalensis (Bloch, 1787), Abudefduf sexfasciatus (Lacépède, 1801), Acanthochromis polyacanthus (Bleeker, 1855), Amblyglyphidodon curacao (Bloch, 1787), Chromis atripectoralis Welander & Schultz, 1951, Dascyllus reticulatus (Linnaeus, 1758), Dischistodus melanotus (Bleeker, 1858), Neopomacentrus melas (Cuvier, 1830), Pomacentrus adelus Allen, 1991, Pomacentrus amboinensis Bleeker 1868, Pomacentrus chrysurus Cuvier, 1830, Pomacentrus moluccensis Bleeker, 1853, Pomacentrus nagasakiensis Tanaka, 1917, Pomacentrus pavo (Bloch, 1787), Pomacentrus wardi Whitley, 1927. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Specimens recorded here as “Hysterolecitha nahaensis complex” created a special identification problem. This species was described by Yamaguti (1942) from two species of 22

Pomacentridae and one of Scaridae with Dascyllus aruanus (Pomacentridae) as the type-host. From the same general region it was subsequently re-reported from Japanese waters (Ichihara 1974; Dyer et al. 1988), south Vietnam (King 1964), Indonesia (Sulawesi) (Yamaguti 1953), and the Great Barrier Reef (Lester and Sewell 1990; Bray et al. 1993a; Barker et al. 1994; Sun et al. 2012). It has also been reported from “Southern Seas” by Parukhin (1989) and from the Mediterranean by Rizk et al. (1996). In Asia and waters from the Great Barrier Reef, records of this species have been overwhelmingly from pomacentrids although there are single records from an acanthurid, a lobotid and a scarid. Bray et al. (1993a) reported 23 species of Pomacentridae (and no other families of fishes) as hosts for this species on the GBR. In that study a multivariate analysis of 62 specimens identified some variation in which specimens from Dascyllus reticulatus showed the greatest tendency to clumping, but the variation was considered inconclusive and all the specimens were identified as a single species. Bray et al. (1993a) considered the record of Parukhin (1989) from “Southern Seas” to be unconvincing. In the present study H. nahaensis was by far the most common infection. However, morphological examination of 105 specimens initially identified as H. nahaensis showed evidence of representing three clearly distinct species as follows:

Type 1. Multiple specimens from Dascyllus reticulatus, two from Pomacentrus moluccensis, and single specimens from Acanthochromis polyacanthus and Stegastes apicalis were distinguishable from all other samples on the basis of their ventral suckers being proportionally distinctly smaller than those of comparable-sized individuals from all other fishes (Figure 1).

Type 2. Eight morphological specimens (1 ex N. melas, 3 ex P. amboinensis, 2 ex P. moluccensis and 2 ex P. wardi) differed from all other specimens in the sample in the ventral sucker being proportionally far larger than those of comparable-sized specimens (Figure 1). The sucker length ratio in these specimens ranged from 4.31-5.90. In addition, in these specimens the testes are unusually close to the ventral sucker, the vitellarium is proportionally large, and the oral sucker is proportionally small.

Type 3. This was by far the most common type (71 measured specimens) and it was found in 14 pomacentrid species Abudefduf sexfasciatus, Acanthochromis polyacanthus, Amblyglyphidodon curacao, Dascyllus reticulatus, Dischistodus melanotus, Neoglyphidodon melas, Pomacentrus adelus, P. amboinensis, P. chrysurus, P. moluccensis, P. nagasakiensis, P. pavo and P. wardi. The size of the ventral sucker of these specimens was clearly intermediate between those of

Morphotypes 1 and 2 (Figure 1). Examination of specimens and comparison of basic measurements suggested no consistent basis for further subdivision of this form.

Figure 1 Body length and ventral sucker length of the three morphological types of Hysterolecitha nahaensis.

In addition to the morphological examination of specimens, ITS2 rDNA sequences were generated for 18 specimens from 7 pomacentrid species. These formed seven distinct genotypes (Genotypes A-G) which each differed from each other by at least 7 bp. Three of the genotypes were represented by a single sequence; the others had 2, 2, 5 and 6 replicates. In numerous studies of trematodes of marine fishes consistent differences in ITS2 rDNA sequences at the level of 7 bp have been found consistent with the recognition of separate species (e.g. McNamara et al. 2014). In the replicated sequences there was no variation except for a single bp difference in one specimen of the genotype with 5 replicates. Distinctions between these taxa are illustrated in an NJ tree (Figure 2). We are able to make the following observations about these genotypes:

Figure 2 Neighbour joining tree based on ITS2 rDNA sequences of Hysterolecitha nahaensis complex from Lizard Island. Samples are identified by host and dissection numbers.

Genotype A. 5 sequences ex Dascyllus reticulatus and 1 ex Stegastes apicalis. This genotype corresponds with Morphotype 1 in that all sequences from D. reticulatus were of this genotype and the great majority of morphological specimens were of Morphotype 1. The single sequence and morphological specimen from S. apicalis corresponded in the same way.

Genotype B. Two identical sequences ex Chromis atripectoralis. No complete morphological specimens corresponding to this genotype are available; partial paragenophore specimens are consistent with H. nahaensis but are unusually large. One paragenophore specimen with part of the posterior body removed measures 1,816 µm to the vitellarium, far longer than any other specimen in the complete collection for which the maximum measurement is 1,505 µm. The combination of genetic and morphological distinction are presumably consistent with the presence of a distinct species.

Genotype C. A single completely distinct sequence ex Pomacentrus moluccensis. This is one of three genotypes reported from this species; it is the repeated sharing of hosts by multiple genotypes

that makes delineation in this group especially problematic. The single morphological specimen from the dissection corresponding to the sequenced specimen is a standard and unremarkable Morphotype 3. It is not presently possible to relate this genotype to a distinct morphotype.

Genotype D. A single completely distinct sequence ex Pomacentrus adelus. This host is also represented in Genotype E. There is no corresponding morphological specimen from the fish from which the sequenced specimen was obtained. It is not presently possible to relate this genotype to a distinct morphotype.

Genotype E. Five sequences from Chromis viridis, Amblyglyphidodon curacao, Pomacentrus adelus and P. moluccensis (4 x identical, the C. viridis sample differs at 1 bp). The great bulk of the specimens from these hosts relate to Morphotype 3 including two sequenced from the same dissections as are represented among the morphological material. This combination presumably reflects a distinct species and Genotype C and Morphotype 3 are here provisionally identified as the same species. However, this identification is compromised by the fact that other genotypes have been found in two of the fish species (P. adelus and P. moluccensis).

Genotype F. A single completely distinct sequence ex Abudefduf whitleyi. No morphological specimens at all are available from this host. The few specimens from Abudefduf sexfasciatus (for which there are no sequences) are slightly distinctive in having unusually large ventral suckers for the specimens in Morphotype 3. It is not presently possible to relate this genotype to a distinct morphotype, but it seems possible that it represents a distinct “Abudefduf” species.

Genotype G. Two identical sequences ex P. moluccensis from two separate individual fish. One of those fish had infections of both Morphotype 2 and 3, the other only Morphotype 3 among the examined morphological specimens. It is not presently possible to relate this genotype to a distinct morphotype. There are no ITS2 rDNA sequences for specimens of Hysterolecitha nahaensis available for comparison from any other locality; indeed, there are none for the genus except for an 18S rDNA sequence of H. nahaensis (see Blair et al. 1998). Perusal of specimens identified as H. nahaensis by Bray et al. (1993a) from pomacentrids from Heron Island suggests that there is almost certainly more than one species represented. We conclude that it is impossible to be definitive about the richness of Hysterolecitha in GBR pomacentrids. The subject requires further intensive study, especially the sequencing of more paragenophore specimens. It is noteworthy that the type-host of H. nahaensis, Dascyllus aruanus, 26

was surveyed thoroughly (n = 20) but had no infections of any sort at Lizard Island. For the purposes of the present analysis we have chosen to record all the forms discussed above as “H. nahaensis complex”. This is for two reasons. First, although some of the putative taxa within the complex appear to be clearly distinct, there are many grey areas in which the differences between putative taxa are completely unresolved; we are unable to propose a convincing hypothesis for the number of species present although we conclude that there are at least four, but potentially as many as seven. Given that three genotypes were represented by only single sequences and that specimens from many pomacentrid species have not been sequenced at all, it seems possible that even more diversity remains undetected in this system. Second, a key aim of this study was to compare the trematode fauna of pomacentrids of the northern GBR with that reported previously for the southern GBR. We conclude that the identifications of the H. nahaensis complex for the southern GBR are also confounded by the presence of a complex of species and reiterate that no molecular data at all are available from there. We thus propose to identify all these forms as “H. nahaensis complex” with a view to future resolution; determining the extent to which the fauna of the northern and southern GBR is comparable for this complex is thus beyond the scope of the present study.

Lecithaster stellatus Looss, 1906 Hosts: Chrysiptera rollandi (Whitley, 1961), Pomacentrus amboinensis Bleeker, 1868, Pomacentrus moluccensis Bleeker, 1853. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Our specimens of this species match the description by Bray et al. (1993a). This species has a large distribution range, having been recorded throughout the Indo-west and west-central Pacific, as well as from the Mediterranean Basin, in both tropical and temperate waters (Bray et al. 1993a). It has been recorded from 17 fish families so far (reviewed by Bray et al. 1993a), including at least one record from fresh water in anadromous fishes (Salmonidae) (Bykhovskaya-Pavlovskaya et al. 1962). Bray et al. (1993a) reported L. stellatus from 11 species of Pomacentridae from Heron Island, while Sun et al. (2012) reported it from Pomacentrus amboinensis from Lizard Island. Our discovery of this species in Chrysiptera rollandi and Pomacentrus moluccensis represent new host records. The species has also been reported from Lizard Island, from the labrid Coris batuensis (Bleeker, 1856) by Muñoz and Cribb (2006).

Thulinia microrchis Yamaguti, 1934 Hosts: Abudefduf septemfasciatus (Cuvier, 1830), Acanthochromis polyacanthus (Bleeker, 1855), Dischistodus melanotus (Bleeker, 1858), Pomacentrus amboinensis Bleeker, 1868. 27

Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Our specimens match the description of Bray et al. (1993a). This species has a widespread distribution and is reported to be highly euryxenic (Bray et al. 1993a; Miller et al. 2011b). It has been reported from Heron Island in 15 species of Pomacentridae as well as in species of Acanthuridae, Chaetodontidae, Kyphosidae, Lethrinidae, Pomacanthidae, Serranidae and Siganidae (Yamaguti 1938; Manter 1969a; Ichihara 1986; Lester and Sewell 1990; Bray et al. 1993a; Barker et al. 1994). Despite the disparity in dietary preferences of these fishes, unpublished molecular evidence indicates the conspecificity of this species, at least across several host families (Miller et al. 2011b). Our finding of T. microrchis in Abudefduf septemfasciatus represents a new host record; T. microrchis was previously reported from three other species of Abudefduf, A. bengalensis, A. sexfasciatus and A. whitleyi (see Bray et al. 1993a).

Lepocreadiidae Odhner, 1905 Lepocreadium oyabitcha Machida, 1984 Hosts: Abudefduf bengalensis (Bloch, 1787), Abudefduf sexfasciatus (Lacépède, 1801), Abudefduf whitleyi Allen & Robertson, 1974. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: This species was described by Machida (1984) from Abudefduf vaigiensis (Quoy & Gaimard, 1825) from Okinawa, Japan. A single specimen was reported from A. whitleyi from Lizard Island by Bray and Cribb (1998), who noted several differences between their specimens and those of Machida (1984), including in body shape, having a tri-lobed (instead of ovoid) ovary, and in infection site (the gut, as opposed to the gall-bladder). Despite this, Bray and Cribb (1998) regard the two as being conspecific, pointing out the distinctive “ramifying moniliform-vitellarium” as an important unifying feature. Our findings agree with those of Bray and Cribb (1998), and expand the host range of this species by two (Abudefduf bengalensis and A. sexfasciatus). It seems likely that this species is restricted to fishes of the genus Abudefduf.

Lepotrema adlardi (Bray, Cribb & Barker, 1993) Host: Abudefduf bengalensis (Bloch, 1787). Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: This species was described from Abudefduf bengalensis from Heron Island by Bray et al. (1993b) (as Lepocreadium adlardi). Revision of the Lepocreadiidae by Bray and Cribb (1996) led 28

to the transfer of this species to Lepotrema. All observed features in our material are consistent with those of L. adlardi. The species differs from other species of Lepotrema in the length of its forebody, the extent of its vitelline follicles and having a relatively long prepharynx. Abudefduf bengalensis remains the only known host of L. adlardi, despite the examination of several other species of Abudefduf at both Heron and Lizard Island.

Lepotrema clavatum Ozaki, 1932 Hosts: Acanthochromis polyacanthus (Bleeker, 1855), Pomacentrus amboinensis Bleeker, 1868. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: This species was described by Ozaki (1932) from the monocanthid Stephanolepis cirrhifer (Temminck & Schlegel, 1850) from Japan. It has since been reported in four fish families from across the Indo-Pacific (Balistidae, Bothidae, Chaetodontidae and Pomacentridae) (see Bray and Cribb 1996 for a full list of infected fish species). This species has been recorded in two species of pomacentrids (Acanthochromis polyacanthus and Parma polylepis Günther, 1862) at Heron Island on the GBR. It has also been recorded from another pomacentrid species in Hawaii, Dascyllus albisella Gill, 1862, (Pritchard 1963). The specimens we collected are consistent with those reported by Bray et al. (1993b) from Heron Island (as Lepocreadium clavatum), with Pomacentrus amboinensis representing a new host record for this species. Lepotrema clavatum is easily distinguished from the other two species of Lepotrema which infect pomacentrids. From L. adlardi it differs in the extent of the vitelline follicles, the length of the prepharynx and length of the forebody, and from L. monile Bray & Cribb, 1998 it differs in having a tri-lobed ovary and lacking a sphincter around the distal metraterm. The host-specificity of L. clavatum is unusual for this genus, being present in several fish families. Research presently underway suggests that L. clavatum will prove to be a species complex comprising species with narrower specificity than currently reported. This form from pomacentrids may ultimately became known under a separate name. Records for the Great Barrier Reef show that this species is abundant in A. polyacanthus (This study; Bray et al. 1993b) and at best uncommon in all other pomacentrids.

Lepotrema monile Bray & Cribb, 1998 Hosts: Abudefduf septemfasciatus (Cuvier, 1830), Pomacentrus amboinensis Bleeker, 1868, Pomacentrus chrysurus Cuvier, 1830, Pomacentrus wardi Whitley, 1927. Locality: Lizard Island Voucher specimens: QM XXXXX

Remarks: This species was first reported as Lepocreadium sp. from Pomacentrus wardi from Heron Island (Bray et al. 1993b). Bray and Cribb (1996) revived the genus Lepotrema and included five species within it, defined by having a combination of a dorsally-positioned excretory pore, a tri- lobed ovary and a “folded muscular pad” on the distal surface of the metraterm. Bray and Cribb (1998) proposed L. monile for this form, noting that while it lacks a tri-lobed ovary and the folded muscular pad, it does possess a dorsal excretory pore. This species was subsequently reported from Pomacentrus amboinensis from Lizard Island by Sun et al. (2012). Our specimens agree closely with the original description of this species. Abudefduf septemfasciatus and P. chrysurus represent new hosts for this species.

Preptetos cannoni Barker, Bray and Cribb, 1993 Hosts: Abudefduf septemfasciatus (Cuvier, 1830). Locality: Lizard Island Deposited specimen: QM XXXXX Remarks: This species was described by Barker et al. (1993) from three species of Siganidae from Heron Island. It has also been reported from Pomacentrus bankanensis (see Bray et al. 1993b). Species of the genus Preptetos resemble Lepocreadium Stossich, 1903, but can be distinguished by the excretory vesicle passing between the testes. The specimen from A. septemfasciatus agrees with the original descriptions by Barker et al. (1993) and represents a new host record for this species. In this study, only a single individual of P. cannoni was found. The low prevalence and intensity detected is consistent with the findings of Bray et al. (1993b) who also reported just a single individual pomacentrid (Pomacentrus bankanensis) being infected. It seems that the species overwhelmingly infects siganids, and occurs in pomacentrids only incidentally.

Monorchiidae Odhner, 1911 Hurleytrematoides morandi McNamara & Cribb, 2011 Host: Pomacentrus wardi Whitley, 1927. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: Species of Hurleytrematoides have been reported from several locations including Florida (Manter 1942), China (Wang and Wang 1993), Indonesia (Machida 2005) and Australia (McNamara et al. 2012). They overwhelmingly infect fishes from the family Chaetodontidae, with some records from Acanthuridae, Pomacanthidae and Tetraodontidae (McNamara and Cribb 2011). This species can be distinguished from other species of Hurleytrematoides by its distinctive body shape and by its distinctively spined cirrus-sac (McNamara and Cribb 2011). This record, based on 30

the discovery of a single specimen in Pomacentrus wardi, is the first of this family from the Pomacentridae. This species has previously been found only (and repeatedly) in species of Chaetodon. This individual infection is best considered a straggler or ‘accidental’.

Transversotrematidae Witenberg, 1944 Transversotrema damsella Hunter & Cribb, 2012 Hosts: Abudefduf septemfasciatus (Cuvier, 1830), Abudefduf sexfasciatus (Lacépède, 1801). Locality: Lizard Island Deposited specimens: QM XXXXX Remarks: This species was described from Abudefduf bengalensis from Lizard Island by Hunter and Cribb (2012) and has also been recorded from several other pomacentrid species: Abudefduf septemfasciatus, A. sexfasciatus, Acanthochromis polyacanthus and Amblyglyphidodon curacao, all from Lizard Island. Further examination of the host records of T. damsella suggests that this species predominantly infects damselfishes of the genus Abudefduf (see Hunter and Cribb 2012). Interestingly, T. damsella has not been found at Heron Island despite numerous damselfish species having been examined there (Barker et al. 1994). Transversotrema damsella is distinguished from other Transversotrema species by the ovary being larger than the testes and also has a relatively distinctive overall shape.

Zoogonidae Odhner, 1902 Steganoderma gibsoni Cribb, Bray & Barker, 1992 Hosts: Abudefduf septemfasciatus (Cuvier, 1830), Abudefduf sexfasciatus (Lacépède, 1801), Abudefduf whitleyi Allen & Robertson, 1974. Locality: Lizard Island Voucher specimens: QM XXXXX Remarks: The present specimens agree well with S. gibsoni as described originally from A. whitleyi from the southern GBR (Cribb et al. 1992) and it was found in two further Abudefduf species. This is the only zoogonid known from a GBR pomacentrid. The records reported here suggest that this species is probably restricted to species of Abudefduf. There have been several other records of zoogonids from pomacentrids; Deretrema fusillus Linton, 1910 from Abudefduf saxatilis (Linnaeus, 1758) from Florida (Linton 1910; Manter 1934,1947), Deretrema sp. from A. sordidus (Forsskål, 1775) from Saipan in Micronesia (Toman and Kamegai 1974) and Diphterostomum americanum Manter, 1947 from Stegastes leucostictus (Müller & Troschel, 1848) from Florida (Overstreet 1969).

Table 1 Host-parasite associations for pomacentrid fishes and digenean trematodes from Lizard Island. Numbers indicate number of worms: numbers in parentheses indicate number of hosts infected. Key to parasites. Derogenidae: A Derogenes pearsoni. Lecithasteridae: B, Aponurus languncula; C, Hysterolecitha heronensis; D, Hysterolecitha nahaensis; E, Lecithaster stellatus; F, Thulinia microrchis. Lepocreadiidae: G, Lepocreadium oyabitcha; H, Lepotrema adlardi; I, Lepotrema clavatum; J, Lepotrema monile; K, Preptetos cannoni. Bivesiculidae: L, Bivesicula unexpecta. Faustulidae: M, Pseudobacciger cheneyae. Haplosplanchnidae: N, Schikhobalotrema sp. Zoogonidae: O, Steganoderma gibsoni. Transversotrematidae: P, Transversotrema damsella. Monorchiidae: Q, Hurleytrematoides morandi. Fellodistomidae: R, Fellodistomidae n. sp. Gyliauchenidae: S, Gyliauchenidae n. sp.

Parasite species Hosts # Host A B C D E F G H I J K L M N O P Q R S

Abudefduf bengalensis 3 4 (1) 5 (1) 2 (2) septemfasciatus 1 3 (1) 1 (1) 1 (1) 2 (1) sexfasciatus 15 6 (3) 10 (3) 2 (2) 3 (1) vaigensis 4 whitleyi 8 1 (1) 5 (2) 1 (1) Acanthochromis polyacanthus 20 3 (3) 8 (4) 18 (10) 23 (9) Amblyglyphidodon curacao 20 8 (5) 1 (1) Chromis atripectoralis 20 2 (2) weberi 15 96 (9) viridis 4 1 (1) Chrysiptera cyanea 9 7 (2) rex 3 rollandi 9 3 (2) 1 (1) taupoa 1 1 (1) Dascyllus aruanus 20 reticulatus 22 25 (11) trimaculatus 4 32

Dischistodus melanotus 4 3 (1) 2 (2) pseudochrysopecilus 4

Neoglyphidodon melas 20 48 (6) Neopomacentrus azysron 21 1 (1) Plectroglyphidodon lacrymatus 2 Pomacentrus adelus 17 7 (1) 2 (1) 10 (5) amboinensis 23 1 (1) 10 (6) 3 (3) 1 (1) chrysurus 22 7 (4) 1 (1) 1 (1) coelestis 12 1 (1) moluccensis 27 18 (6) 2 (2) 1 (1) 2 (1) nagasakiensis 7 3 (2) pavo 5 1 (1) wardi 8 4 (3) 1 (1) Stegastes apicalis 7 8 (2) 1 (1) 7 (3) fasciolatus 1

Discussion Identification of the fauna From a total of 358 individuals of 32 species of pomacentrids examined we found 19 species of digenean trematodes and 54 host/parasite combinations. A total of 19 host distributions were new, 18 were new host records, and three were new species. The characterisation of this fauna proved in part straightforward and in part complex. It was straightforward in that the great majority of the species were easily identified as forms previously reported from pomacentrids or, in a few cases, other fishes. The three new species, a faustulid, a fellodistomid and a gyliauchenid, were all readily recognisable as distinct and were the sole representatives of their respective families encountered. The complexities of identification relate almost entirely to the delineation of the Hysterolecitha nahaensis complex. As outlined above, this species was by far the most common infection encountered but combined molecular and morphological evidence suggests that it comprises at least four species, there is limited evidence for as many as seven species, and it seems entirely possible that there could be more than that. On the basis of the present evidence it is possible to distinguish only three taxa by morphology of which the most common, Morphotype 3, almost certainly is a complex of multiple species. In addition to the problem relative to H. nahaensis, it seems likely that some of the other species identified (especially Aponurus languncula, Lecithaster stellatus and Lepotrema clavatum) will ultimately become recognised under different names, but their distinction as single taxa in pomacentrids does not appear to be in doubt. The second complexity in the characterisation of this fauna is in the size of the task. The study incorporated the examination of 372 individuals of 32 pomacentrid species. Despite the fact that these numbers represent a substantial number of individual fishes for field and processing effort, the incompleteness of the survey of the fauna is demonstrated by two features of the data. First, there remain a further 58 pomacentrid species known from Lizard Island that have never been examined for parasites; undoubtedly these will harbour many new host-parasite combinations. The second feature of the fauna is the rarity of many infections. Of 54 host/parasite combinations, 27 (exactly half) were detected in only a single individual fish. Such a level of rarity implies that a great many more host/parasite combinations remain undetected for the hosts that have been examined. Some of the infections found only once are evidently stragglers (e.g. Preptetos cannoni and Hurleytrematoides morandi) and are thus insignificant in the understanding of the fauna. In contrast, the distribution of oioxenic (single host) species is difficult to predict and there are likely to be a number of those that remain completely undetected so far. Below we consider the nature of host-specificity, overall richness patterns and beta diversity for this fauna with the understanding that the data set remains incomplete.

Richness Despite the Pomacentridae being a large and diverse group of coral reef fishes, the mean species richness of digeneans in species within this family was low at 1.69. Results show that 20 of 32 pomacentrid species examined from Lizard Island were infected with only 1 ̶ 3 species of digeneans; just five species harboured four species of digenean; seven of the 32 species of pomacentrids had no digeneans at all. For many pomacentrid species the samples were low, but it is striking that the 20 Dascyllus aruanus produced no infections at all. It is well-established that parasite richness within host species is not random, but is broadly related to host species traits, especially size, density and distribution (Guégan et al. 1992; Kamiya et al. 2014). Host size probably explains the relatively low richness of adult digeneans among pomacentrid species relative to larger coral reef fishes such as chaetodontids, lutjanids and serranids as well as among the pomacentrid species. Notably the larger species in the study, species of Abudefduf, were infected with most digenean species. Generally, host size is a strong predictor for parasite richness because larger-bodied individuals provide greater space and resources for parasites than smaller-bodied individuals (Sasal et al. 1997; Kamiya et al. 2014). However, at the infracommunity level, parasite richness is more variable when examining the effects of increasing host size. Two medium-sized pomacentrid species, Pomacentrus amboinensis and P. moluccensis, were also infected with 4 species of digeneans; notably, however these two species had the two largest samples examined in this study indicating the importance of sample size. Overall, although host size is a good predictor of parasite richness it will not necessarily over-rule the effects of different habitats, densities and diet (Lo et al. 1998; Torres et al. 2006; Morand 2015). The reported digenean fauna of pomacentrids from both the northern and southern GBR suggests that there has been little radiation of families of trematodes species within this group; the total of 23 species are drawn from 11 trematode families. Apart from the presence of three species of Lepotrema the clear exception is the H. nahaensis complex; however, interpretation of the nature of the H. nahaensis complex must await its better characterisation. The limited radiation of individual genera is in contrast to many other GBR fish family/trematode family complexes Bucephalidae Poche, 1907 in serranids (Bott and Cribb 2009), Paradiscogaster Yamaguti, 1934 (Diaz et al. 2013) and Hurleytrematoides Yamaguti,1953 (McNamara and Cribb 2011) in chaetodontids, Opecoelidae Ozaki, 1925 in mullids (Rohner and Cribb 2013) and Cryptogonimidae Ward, 1917 in lutjanids (Miller and Cribb 2007). In all these combinations there is much greater parasite richness than seen for any of the families represented in pomacentrids. As found by Barker et al. (1994) at Heron Island, the diversity of pomacentrid species is far greater than the diversity of digenean trematodes at Lizard Island. Other studies from other locations have also shown that the diversity of digeneans in pomacentrids is relatively low (Yeo and

Spieler 1980; Dyer et al. 1985). However, it should be noted that this study only examined 32 species of pomacentrids from a possible 89 species that have been recorded at Lizard Island (Grutter et al. in review). Thus, there is the likelihood that other species of digenean worms will be found, as well as extending the host records for already known species.

Host-specificity The host specificity of the digeneans of pomacentrids varies, with a tendency towards being relatively low. We found just three1 species to be convincingly oioxenic (i.e found repeatedly in only one host species) at Lizard Island: Bivesicula unexpecta was only found in Acanthochromis polyacanthus, Lepotrema adlardi only in Abudefduf bengalensis and Pseudobacciger cheneyae only in Chromis weberi. Due to only a single specimen being recorded, further examination is required to determine whether Fellodistomidae sp. from Chrysiptera rollandi is indeed oioxenic or in fact occurs in other pomacentrids. Our findings are in support of earlier studies regarding Bivesicula unexpecta and Lepotrema adlardi (Bray et al. 1993b; Cribb et al. 1994a). The majority of the digenean species reported in this study are apparently stenoxenic, i.e. infecting multiple species within the same family; Derogenes pearsoni, Hysterolecitha heronensis, H. nahaensis complex, Lepocreadium oyabitcha, Lepotrema monile, Steganoderma gibsoni (see Cribb et al. 1992; Bray et al. 1993a; Bray and Cribb 1998) and Transversotrema damsella (Hunter and Cribb 2012) are all in this category. Two of these stenoxenic species, H. heronensis and L. oyabitcha, can be further categorised as being restricted to multiple congeners of Pomacentrus and Abudefduf, respectively. The remaining species are classified as euryxenic, infecting a wide range of hosts from multiple families; Aponurus laguncula, Lecithaster stellatus, Lepotrema clavatum, Hurleytrematoides morandi, Preptetos cannoni and Thulinia microrchis. However, it should be noted that H. morandi and P. cannoni, which are common in chaetodontids and siganids respectively are technically euryxenic, but are effectively stenoxenic to the Chaetodontidae and Siganidae, infecting pomacentrids only incidentally. This could also be the case for A. languncula which has not been reported other than from three species of pomacentrids. Infections of these pomacentrid species are rare, suggesting that the infections may also be incidental. A likely explanation for the generally broad host range of many species are the eco- physiological similarities between pomacentrid species, especially their broad and overlapping diets, which may facilitate host-sharing in digeneans (Marcogliese 2002). Use of intermediate hosts by metacercariae facilitates transmission into the definitive host (Poulin and Cribb 2002). Broadly overlapping diets increase the possibility that multiple species of fish will encounter the same suite

1 In this study the gyliauchenid was oioxenic, but subsequent collections by colleagues (not reported here) has shown that it infects at least two species in addition to Stegastes apicalis. 36

of metacercariae. The fact that hemiuroid metacercariae are known to infect a range of zooplankton such as calanoid copepods and polychaetes (Marcogliese 1995) which in turn are consumed by many pomacentrids, may explain the high prevalence of digeneans from this superfamily infecting multiple pomacentrid species. Although diet may explain some of the variation in the infection of pomacentrids by certain digeneans species (e.g. Hysterolecitha nahaensis complex), it does not explain the high specificity observed in other species (e.g. Pseudobacciger cheneyae to just one species of Chromis, Lepotrema adlardi to just one species of Abudefduf). In such cases the explanation may be physiological compatibility (Adamson and Caira 1994). The physiological compatibility between a parasite and its potential hosts may play an important role in determining the suitability of those hosts, thus driving host specificity. However, the extent to which physiology influences host specificity has not been explored for most fish-infecting digeneans and is poorly understood (Miller et al. 2011b). It is notable that some of the digeneans we found were species which are interpreted as having cosmopolitan distributions and exhibiting euryxenicity, i.e. extremely wide host ranges. We encountered five species which could be classed as being euryxenic, including two (Aponurus languncula and Lecithaster stellatus) that are reported to be distributed circumglobally. Euryxenicity is observed in a small number of marine digenean species worldwide (Bartoli et al. 2005) and has been reported for only 23 of the 290 species of digeneans on the GBR (Miller et al. 2011b). It is possible, perhaps probable, that some of these euryxenic species may comprise complexes of cryptic species that have distinct geographical distributions. The two cosmopolitan species reported in this study are from the superfamily Hemiuroidea, which are characterised by many of its member taxa exhibiting low host specificity (Gibson and Bray 1979). Although A. laguncula has not been reported from any fish family on the GBR other than pomacentrids (reported from 3 species only), it has been recorded in 67 teleost species from 29 families, making it putatively one of the most widespread digeneans with respect to both geographical distribution and host range (Bray and MacKenzie 1990; Bray et al. 1993a). It is currently unclear whether all records of A. laguncula comprise one species or form a species complex. Recent molecular data have indicated the presence of a cryptic species of Aponurus within the A. laguncula complex, A. mulli Carreras-Aubets, Repulles-Albelda, Kostadinova & Carrasson, 2011, which infects red mullet, Mullus barbatus Linnaeus, 1758, from the Mediterranean (Carreras- Aubets et al. 2011). It is possible that further exploration will reveal that Aponurus laguncula is actually comprised of multiple cryptic species, and that A. laguncula sensu stricto may actually not be found in Australian waters. Conversely, it is also possible that molecular work may reveal that A. laguncula is indeed one species, despite its apparent cosmopolitan distribution, as has been previously demonstrated for the tuna-infecting aporocotylid fluke Cardicola forsteri Cribb, Daintith

& Munday, 2000 (Aiken et al. 2007). The same principles may also apply to the other reportedly cosmopolitan species, Lecithaster stellatus. Cryptic species may also occur within some of the stenoxenic species found on the GBR, in particular Hysterolecitha nahaensis. Currently, H. nahaensis has been recorded in a total of 26 pomacentrid species from Lizard and Heron Islands, making it the digenean with the widest host range. This species closely resembles H. heronensis; the two species may have been historically confused with each other due to their morphological similarities (Bray et al. 1993a). These species require comprehensive reconsideration with combined morphological or molecular analysis. It is apparent that a number of cryptic species are present within these species, as has been demonstrated for other digeneans in other hosts (Jousson et al. 2000; Nolan and Cribb 2005; Diaz et al. 2015).

Beta diversity These data create a rare opportunity to consider beta diversity in parasite faunal composition on the GBR. Both the depth of study and richness of digenean trematodes encountered in this study is comparable to that of Barker et al. (1994), who examined 312 individuals from 39 species of pomacentrids from Heron Island, on the southern GBR (see Table 2). Twenty-one pomacentrid species were examined at both sites. Of these, six had robust samples of at least 10 individuals at both sites. Although the overall trematode richness is almost identical for the two sites, it is striking that mean richness and prevalence were reported as much higher at Heron than at Lizard Island.

Table 2 Comparison between data collected from Heron Island (Barker et al. 1994) and Lizard Island (present study) for digeneans from pomacentrid fishes.

HI LI Fish species 39 32 Fish individuals 312 358 Parasite species 18 19 Host/parasite combinations 86 54 Mean richness 2.205 1.688 Total infections 209 140 Mean overall prevalence 0.67 0.39

The total digenean fauna for GBR pomacentrids now stands at 23 species in 50 pomacentrid species of which 14 are now known from both locations, five have been only recorded from Lizard Island (Lepocreadium oyabitcha, Pseudobacciger cheneyae, Transversotrema damsella, Fellodistomid sp and Gyliauchenid sp.) and four only from Heron Island (Derogenes pharyngicola, Lepocreadium sp., Preptetos xesuri and Macvicaria heronensis). Of the five species known only from Lizard Island, only three can be inferred to be absent from Heron Island. As only a single specimen of the fellodistomid digenean was recorded and given that no Chromis weberi have been examined at Heron Island for P. cheneyae, despite records indicating that C. weberi does occur there (Russell 1983), we cannot conclude that these species do not occur at Lizard Island; it is possible that with further examination, these species may prove to be present at Heron Island. In contrast, the evidence is good that three species (L. oyabitcha, T. damsella and Gyliauchenid n. sp.) do not occur at Heron Island. Of the four species known only from Heron Island, only D. pharyngicola is a regularly occurring parasite restricted to pomacentrids; Preptetos xesuri and Macvicaria heronensis are mainly parasites of other fishes and the Lepocreadium sp. from Amblyglyphidodon curacao has been collected only twice. We conclude that it is possible to identify a “core pomacentrid fauna” of species that are regular in and dependent on pomacentrids (Table 3). From this core fauna are excluded species that are rare to the point that their status is uncertain (e.g. Aponurus laguncula, Fellodistomid sp.) and those that are clearly principally parasites of other fishes (e.g. Preptetos cannoni and Hurleytrematoides morandi). With such exclusions we recognise 14 morphospecies in this category with the understanding that “H. nahaensis complex” certainly represents multiple species.

Table 3 Core pomacentrid specific digeneans from Heron and Lizard Island.

Family Genus Species HI LI Bivesiculidae Bivesicula unexpecta ● ● Derogenidae Derogenes pearsoni ● ● Derogenes pharyngicola ●

Faustulidae Pseudobacciger cheneyae ● Gyliauchenidae Gyliauchen sp. ● Haplosplanchnidae Schickhobalotrema pomacentri ● ● Lecithasteridae Hysterolecitha heronensis ● ● nahaensis Hysterolecitha ● ● (complex) Lepocreadiidae Lepocreadium oyabitcha ●

Lepotrema adlardi ● ● Lepotrema clavatum ● ● Lepotrema monile ● ● Transversotrematidae Transversotrema damsella ●

Zoogonidae Steganoderma gibsoni ● ● Count 10 13

The only previous detailed comparison of the parasite fauna of comparable fish from the southern and northern GBR was that of monorchiid trematodes of chaetodontid fishes (McNamara and Cribb 2011; McNamara et al. 2012). A total of 10 species in total was reported, of which all occurred at Heron Island but only seven were found at Lizard Island. This finding ran counter to the general pattern of greater biological richness on the northern than on the southern GBR (Russell 1983). Here the comparison is less straightforward because of the complex (and partly uncertain) host-specificity of the pomacentrid parasites (whereas the monorchiids of chaetodontids show reliably high specificity) and because a smaller number of the host species sampled at the two sites were in common. Of the 14 core pomacentrid species, 10 (71 %) have been found at Heron Island and 13 (93 %) at Lizard Island (Table 3). These numbers are consistent with the interpretation that there are significant differences in the parasite fauna at the two sites and that overall richness is slightly higher in the north. That both mean species richness and mean overall prevalence were reported as much higher at Heron Island by Barker et al. (1994) is puzzling. There are so many areas of uncertainty in these comparisons (from the true identity of several of the species to the robustness of the sample sizes) that these findings can only be a basis for future analyses.

Acknowledgments We thank E. C. McClure, F. Cortesi, K. L. Cheney, S. Martin for their help collecting specimens and the staff of the Lizard Island Research Station for their support. This work was funded by a Sea World Research and Rescue Foundation, Australia (SWRRFI) awarded to D. Sun and T. H. Cribb and an Australian Research Council Discovery Grant awarded to T. H. Cribb.

Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia

Derek Sun, Rodney A. Bray, Russell Q-Y. Yong, Scott C. Cutmore & Thomas H. Cribb This section is published as an original article in Systematic Parasitology (2014) 8:141-152

Abstract A new species of digenean, Pseudobacciger cheneyae n. sp., is described from the intestines of Weber’s chromis (Chromis weberi Fowler & Bean) from off Lizard Island, Great Barrier Reef, Australia. This species differs from the three described species of Pseudobacciger Nahhas & Cable, 1964 [P. cablei Madhavi, 1975, P. harengulae (Yamaguti, 1938) and P. manteri Nahhas & Cable, 1964], in combinations of the size of the suckers and the length of the caeca. The host of the present species is a perciform (Family Pomacentridae) which contrasts with previous records of the genus which are almost exclusively from clupeiform fishes. The genus Pseudobacciger is presently recognised within the family Faustulidae (Poche, 1926) but phylogenetic analyses of 28S and ITS2 rDNA show that the new species bears no relationship to species of four other faustulid genera (Antorchis Linton, 1911, Bacciger Nicoll, 1924, Paradiscogaster Yamaguti, 1934 and Trigonocryptus Martin, 1958) but that instead it is nested within the Gymnophalloidea (Odhner, 1905) as sister to the Tandanicolidae (Johnston, 1927). This result suggests that the Faustulidae is polyphyletic.

Introduction The Faustulidae Poche, 1926 is a small family of digenean trematodes found in a wide range of mainly marine teleost families (Bray 2008). There are 11 genera, most of which have few species. Only four genera, Antorchis Linton, 1911, Bacciger Nicoll, 1924, Faustula Poche, 1926 and Paradiscogaster Yamaguti, 1934, have five or more described species. In Australia, faustulids have been reported from several fish families in tropical, sub-tropical and temperate waters (Korotaeva 1969; Kurochkin 1970; Bray 1982; Bray et al. 1994; Cribb et al. 1999; Diaz and Cribb 2013; Diaz et al. 2013).

Pseudobacciger Nahhas & Cable, 1964, as reviewed by Dimitrov et al. (1999), is represented by three species, all infecting marine teleosts. The hosts are overwhelmingly clupeiforms (Clupeidae and Engraulidae), with single records from a sparid and a pomacanthid (Madhavi 1975; Parukhin 1976a; Gaevskaja and Naidenova 1996; Dimitrov et al. 1999; Machida and Uchida 2001). Species of Pseudobacciger primarily infect the intestine, but also occur occasionally in the pyloric caeca (Dimitrov et al. 1999). There are reports of species of Pseudobacciger from the temperate to subtropical Atlantic, Pacific and Indian Oceans (Bray 2008) but to date there are no records from Australian waters. Sporocysts have been reported in bivalves and metacercariae in shrimps from Japan and Korea (Chun et al. 1981; Kim and Chun 1984). Pseudobacciger was previously included in the subfamily Baccigerinae Yamaguti, 1958 (family Fellodistomidae Nicoll, 1909) but molecular analysis by Hall et al. (1999) revealed the Fellodistomidae to be polyphyletic, and the Baccigerinae was raised to family level as its senior synonym, Faustulidae. Pomacentrids have been extensively examined for trematodes in Australia (Manter 1969b; Cribb et al. 1992; Bray et al. 1993a; Bray et al. 1993b; Bray and Cribb 1998; Bray and Cribb 2004; Hunter and Cribb 2012). To date 20 fully identified species from eight families of trematodes have been reported, but no faustulids have been recorded from pomacentrids from Australia or elsewhere. Between July 2011 and August 2012, several specimens of a planktivorous damselfish, Weber’s chromis (Chromis weberi Fowler & Bean), were caught from off Lizard Island, on the northern GBR and examined for trematodes. In these we discovered a previously unreported trematode, of which morphological analysis suggests is a new species of Pseudobacciger. Molecular phylogenetic analysis suggests that the family-level affiliation of this new species is not clear.

Materials and methods Specimen collection Fifteen Chromis weberi were collected from off Lizard Island, Great Barrier Reef, Australia in 2011 and 2012. Fish were collected using a barrier net and were killed prior to dissection by cranial pithing. The contents of the body cavity were removed and examined under a dissecting microscope using fine scissors and forceps. The gastrointestinal tract was examined for parasites using the gut- wash approach as described by Cribb and Bray (2010). Specimens were fixed by pipetting them into near boiling saline followed by immediate preservation in 70 % ethanol for both morphological and molecular analysis. Worms for morphological analysis were stained using Mayer’s haematoxylin, destained in 1 % hydrochloric acid, neutralized with 0.5 % ammonia solution, dehydrated through a

graded series of ethanol, cleared in methyl salicylate, and mounted on slides in Canada balsam. Drawings were made with the aid of a camera lucida. Measurements were made using a SPOT Insight™ digital camera (Diagnostic Instruments, Inc.) mounted on an Olympus BH-2 compound microscope using SPOT™ imaging software. Measurements are in micrometres (µm) and are presented as the range, followed by the mean in parentheses. Type-specimens are lodged in the Queensland Museum (QM), Brisbane, Australia.

Molecular analysis All voucher specimens are lodged in the Queensland Museum, Brisbane, Australia and are paragenophores (sensu Pleijel et al. 2008). Total genomic DNA was extracted using standard phenol/chloroform extraction techniques (Sambrook and Russell 2001). Partial 28S nuclear ribosomal DNA was amplified using the primers LSU5 (5'-TAG GTC GAC CCG CTG AAY TTA AGC-3') (Littlewood et al. 2000) and ECD2 (5'-CTT GGT CCG TGT TTC AAG ACG GG-3') (Littlewood et al. 2000) and the ITS2 region using the primers 3S (5'-GGT ACC GGT GGA TCA CGT GGC TAG TG-3') (Bowles et al. 1993) and ITS2.2 (5'-CCT GGT TAG TTT CTT TTC CTC CGC-3' ) (Cribb et al. 1998). PCR for both the 28S and ITS2 regions was performed with a total volume of 20 μl consisting of approximately 10 ng of DNA, 5 μl of 5 MyTaq Reaction Buffer (Bioline), 0.75 μl of each primer (10 pmols) and 0.25 μl of Taq DNA polymerase (Bioline MyTaq™ DNA Polymerase), made up to 20 μl with Invitrogen™ ultraPURE™ distilled water. Amplification was carried out on a MJ Research PTC-150 thermocycler. The following profile was used to amplify the 28S region: an initial 95°C denaturation for 4 min, followed by 30 cycles of 95°C denaturation for 1 min, 56°C annealing for 1 min, 72°C extension for 2 min, followed by a single cycle of 95°C denaturation for 1 min, 55°C annealing for 45 sec and a final 72°C extension for 4 min. The following profile was used to amplify the ITS2 region: an initial single cycle of 95°C denaturation for 3 min, 45°C annealing for 2 min, 72°C extension for 90 sec, followed by 4 cycles of 95°C denaturation for 45 sec, 50°C annealing for 45 sec, 72°C extension for 90 sec, followed by 30 cycles of 95°C denaturation for 20 sec, 52°C annealing for 20 sec, 72°C extension for 90 sec, followed by a final 72°C extension for 5 min. Amplified DNA was purified using a Bioline ISOLATE II PCR and Gel Kit, according to the manufacturer’s protocol. Cycle sequencing of purified DNA was carried out for the 28S rDNA region using the same primers used for PCR amplification as well as the additional primer 300F (5'-CAA GTA CCG TGA GGG AAA GTT-3') (Littlewood et al. 2000) and for the ITS2 rDNA region using the forward primer GA1 (5'-AGA ACA TCG ACA TCT TGA AC-3') (Anderson and Barker 1998) and the reverse primer ITS2.2. Cycle sequencing was carried out at the Australian Genome Research Facility using an AB3730xl capillary sequencer. Sequencher™ version 4.5 (GeneCodes Corp.) was used to assemble and edit 45

contiguous sequences. GenBank accession numbers for the taxa sequenced in this study are shown in Table 1. In addition to samples of the species under direct consideration, we generated 28S and ITS2 rDNA sequences for an unknown gymnophallid metacercaria collected from the bivalve Paphies elongata (Reeve) (Mesodesmatidae) collected from Woorim Beach, Bribie Island, southeast Queensland (27°04'S, 153°12'E). Collection of this specimen was required as there are currently no 28S sequence available on GenBank for a gymnophallid. Voucher specimens of this species are lodged within the Queensland Museum under registration numbers QM G 234368 – G 234387. We also generated an ITS2 rDNA sequence from a specimen of Tandanicola bancrofti Johnston, 1927 consistent with the original description of this species (Johnston 1927).

Table 1 Collection data and GenBank accession numbers for taxa sequenced during this study

Taxon Host Locality ITS2 rDNA 28S rDNA (no. of (no. of replicates) replicates) Pseudobacciger Chromis weberi Off Lizard Island, Great cheneyae n. sp. Fowler & Bean Barrier Reef, KJ648920 (2) KJ648919 (2) Queensland (14°40′S, 145°27′E) Gymnophallidae Paphies elongata Woorim Beach, Bribie sp. (Reeve) Island, Queensland KJ648918 (2) KJ648917 (1) (27°04'S, 153°12'E) Tandanicola Tandanus tandanus Kholo Crossing, KJ648921 (2)  bancrofti (Mitchell) Brisbane River, (Johnston, Queensland (27°33'S, 1927) 152°44'E)

Phylogenetic analysis The ITS2 and partial 28S rDNA sequences generated during this study were aligned with species of Bucephalidae Poche, 1907, Faustulidae, Fellodistomidae and Tandanicolidae Johnston, 1927 available on GenBank using MUSCLE version 3.7 (Edgar 2004) with ClustalW sequence weighting and UPGMA clustering for iterations 1 and 2. The resultant alignments were refined by eye using MESQUITE (Maddison and Maddison 2009) and the ends of each fragment were trimmed to match the shortest sequence in the alignments. Bayesian inference and Maximum Likelihood analyses of both the 28S and ITS2 rDNA datasets were conducted to explore relationships among these taxa. Bayesian inference analyses were performed using MrBayes version 3.2.2 (Ronquist et al. 2012), run on the CIPRES portal

(Miller et al. 2010). Maximum Likelihood analyses were performed using RAxML version 7.2.8 (Stamatakis et al. 2008). The software jModelTest version 0.1.1 (Posada 2008) was used to estimate the best nucleotide substitution model for the datasets. Bayesian inference analyses were conducted on the 28S dataset using the GTR+G model and on the ITS2 dataset using the TVM+G model predicted as the best estimators by the Akaike Information Criterion (AIC) in jModelTest. The closest approximation of these models was implemented in the subsequent Maximum Likelihood analyses. Bayesian inference analyses were run over 10,000,000 generations (ngen = 10,000,000) with two runs each containing four simultaneous Markov Chain Monte Carlo (MCMC) chains (nchains = 4) and every 1,000th tree saved (samplefreq = 1,000). Bayesian analyses used the following parameters: nst = 6, rates = gamma, ngammacat = 4, and the priors parameters of the combined dataset were set to ratepr = variable. Samples of substitution model parameters, and tree and branch lengths were summarised using the parameters ‘sump burnin = 3000’ and ‘sumt burnin = 3,000’. These ‘burnin’ parameters were chosen because the log likelihood scores ‘stabilised’ well before 300,000 replicates in the Bayesian inference analyses. Nodal support in the Maximum Likelihood analyses was estimated by performing 100 bootstrap pseudoreplicates. Species of the Hemiuridae (Looss, 1899) were designated as functional outgroups in all analyses.

Gymnophalloidea Odhner, 1905 incertae familiae Pseudobacciger Nahhas & Cable, 1964

Pseudobacciger cheneyae n. sp.

Type-host: Chromis weberi Fowler & Bean (Perciformes: Pomacentridae), Weber’s chromis. Type-locality: Off Lizard Island, northern Great Barrier Reef, Queensland, Australia (14°40'S, 145°27'E). Site in host: Intestine. Material examined: A total of 100 specimens collected from C. weberi; 2 for molecular analysis, the remainder for morphological analysis. Prevalence: 73% (11 of 15 fish infected). Intensity: 1−44 individuals per fish with a mean ( SE) intensity (10  4.128) Type-material: Holotype QM G 234368 and 19 paratypes QM G 234369–G 234387 deposited in the Queensland Museum, Brisbane, Australia.

urn:lsid:zoobank.org:act: 369B1E27-B899-48CD-ACD6-0893B314C1C4 urn:lsid:zoobank.org:act: AB0A60B3-E5AC-4EE4-8990-8664A738A1E9

Molecular sequence data: 28S rDNA, 2 replicates (1 submitted; GenBank KJ648919); ITS2 rDNA, 2 replicates (1 submitted; GenBank KJ648920). Etymology: This species is named in honour of Dr Karen Cheney, The University of Queensland, in recognition of her contribution to marine parasitology.

Description (Figure 1, 2) [Measurements taken from 20 specimens.] Body pyriform, posterior half somewhat broader than anterior, 371453 (407)  230327 (281). Forebody, 98138 (119) long, occupies 2334 (28) % body length. Tegumental spines minute, <1 long, cover entire body evenly, including surfaces of oral and ventral suckers. Oral sucker well delineated, opening ventro-subterminally, muscular, 6891  4663 (79  57). Ventral sucker well developed, in anterior half of body, 80109  69101 (94  87). Oral sucker length to ventral sucker length ratio 1:1.211.67 (1.47). Oral sucker width to ventral sucker width ratio 1: 1.031.24 (1.13). Prepharynx absent. Pharynx muscular, immediately posterior to oral sucker, 3647  2534 (40  29). Oesophagus distinct, often sinuous or coiling once before intestinal bifurcation, 3355 (43) long. Intestine bifurcates dorsal to ventral sucker; caeca usually hidden by uterine coils, extend to anterior hindbody, close to posterior margins of testes, 115166 (133) long (Figure 2). Testes 2, entire, nearly round or somewhat irregular in outline, opposite in anterior hindbody, intercaecal, well-separated from lateral margins of body, 4781  5181 (65  70). Sperm ducts inconspicuous. Cirrus-sac absent. Seminal vesicle bipartite, median, posterior to caecal bifurcation, reaching anteriorly from level of posterior margin of ventral sucker, 2671  3270 (43  47). Pars prostatica and ejaculatory ducts not detected. Genital pore median, immediately anterior to margin of ventral sucker. Ovary trilobed, median, between testes, posterior margin usually hidden by uterine coils, extends to anterior hindbody, 4473  5596 (54  75). Oviduct and oötype inconspicuous. Laurer’s canal opening dorsally towards posterior end of body (usually obscured by eggs, seen clearly in only a single laterally mounted specimen). Seminal receptacle not detected. Uterine coils extensive in posterior half of body from posterior margin of vitelline follicles and vitelline duct, and overlapping testes, ovary and seminal vesicle. Vitelline follicles in 2 small lateral fields at level of ventral sucker, 64100  3373 (84  47). Vitelline ducts undulate, unite medially anterior to ovary and posterior to seminal vesicle, straight length 82115 (96), total length 89143 (115). Eggs elliptical, 1821  1215 (20  13.5). Excretory vesicle Y-shaped; arms extend to level of anterior margin of ventral sucker, 243326 (286) long. Excretory pore terminal.

Figure 1-2 Pseudobacciger cheneyae n. sp. ex Chromis weberi from off Lizard Island. 1, Holotype, whole mount, ventral view; 2, Holotype, whole mount, ventral view, showing excretory vesicle, caeca and seminal vesicle. Scale-bars: 100 μm

Molecular and phylogenetic analysis BLAST analysis of the 28S sequence identified the greatest similarity of Pseudobacciger cheneyae n. sp. as being to two tandanicolids, Tandanicola bancrofti Johnston, 1927 and Prosogonarium angelae Cribb & Bray, 1994, rather than to faustulids as had been expected given the generic level identification of the species. The sequence for Pseudobacciger cheneyae n. sp. was thus aligned with and analysed relative to a wide range of trematodes of the suborder Bucephalata La Rue, 1926 (Bucephaloidea Poche, 1907: Bucephalidae; Gymnophalloidea Odhner, 1905: Fellodistomidae, Gymnophallidae Odhner, 1905 and Tandanicolidae) and, from the suborder Xiphidiata Olson, Cribb, Tkach, Bray & Littlewood, 2003: Faustulidae (Table 2). Figure 3 shows relationships derived from Bayesian inference and Maximum Likelihood analyses of P. cheneyae n. sp. relative to these taxa. In this analysis P. cheneyae n. sp. forms a strongly supported clade with the two tandanicolids within the well-supported Gymnophalloidea and is distant from the faustulids.

BLAST analysis of the ITS2 sequence identified the greatest similarity of P. cheneyae n. sp. as being to a range of bucephalids. Given this, and the results of the 28S rDNA analyses, the sequence was aligned with and analysed relative to a wide range of trematodes of the suborder Bucephalata (Bucephaloidea: Bucephalidae; Gymnophalloidea: Fellodistomidae, Gymnophallidae and Tandanicolidae) and, as above, from the suborder Xiphidiata, representatives of Faustulidae. In all analyses P. cheneyae n. sp. formed clades with a range of taxa from the Bucephalata; it was never associated with the clade comprising faustulids. However, posterior probability and bootstrap support were typically low and these analyses are not therefore shown.

Table 2 28S rDNA sequence data from GenBank included in phylogenetic analyses

Species GenBank Reference accession number Family Fellodistomidae Coomera brayi Dove & Cribb, 1995 KJ425462 Cribb et al. (2014a) Fellodistomum agnotum Nicoll, 1909 AJ405289 Bray et al. (1999) Fellodistomum fellis (Olsson, 1868) AJ405290 Bray et al. (1999) Oceroma praecox Cribb, Miller, Bray & Cutmore, 2014 KJ425464 Cribb et al. (2014a) Olssonium turneri Bray & Gibson, 1980 AY222283 Olson et al. (2003) Steringophorus blackeri Bray, 1973 AJ405296 Bray et al. (1999) Steringophorus dorsolineatus (Reimer, 1985) AJ405291 Bray et al. (1999) Steringophorus furciger (Olsson, 1867) AJ405292 Bray et al. (1999) Steringophorus margolisi Bray, 1995 AY222281 Olson et al. (2003) Steringophorus pritchardae (Campbell, 1975) AJ405295 Bray et al. (1999) Steringophorus thulini Bray & Gibson, 1980 AJ405298 Bray et al. (1999) Tergestia sp. KJ425467 Cribb et al. (2014a) Family Tandanicolidae Prosogonarium angelae Cribb & Bray, 1994 AY222285 Olson et al. (2003) Tandanicola bancrofti Johnston, 1927 KJ425466 Cribb et al. (2014a) Family Bucephalidae Rhipidocotyle minima (Wagener, 1852)* AY222225 Olson et al. (2003) Family Faustulidae Antorchis pomacanthi (Hafeezullah & Siddiqi, 1970) AY222268 Olson et al. (2003) Bacciger lesteri Bray, 1982 AY222269 Olson et al. (2003) Trigonocryptus conus Martin, 1958 AY222270 Olson et al. (2003) Outgroup Taxa Superfamily Hemiuroidea Aponurus sp. DQ354368 Pankov et al. (2006) Bunocotyle progenetica Chabaud & Buttner, 1959 DQ354365 Pankov et al. (2006) Robinia aurata Pankov, Webster, Blasco-Costa, Gibson, DQ354367 Pankov et al. (2006) Littlewood & Kostadinova, 2006

*Bartoli et al. (2006) re-identified this material as R. minima (Wagener, 1852) not R. galeata (Rudolphi, 1819) as originally identified. R. galeata, the type-species of the genus, was poorly known until redescribed by Derbel et al. (2011), who showed that it is clearly distinct from R. minima.

Figure 3 Relationships between Pseudobacciger cheneyae n. sp. and other Gymnophalloidea, Bucephalidae and Faustulidae taxa based on Bayesian inference and Maximum Likelihood analyses of the 28S rDNA dataset. Bayesian inference posterior probabilities are shown above the nodes and Maximum Likelihood bootstrap support values below the nodes; values < 50 not shown. Abbreviations: Bu, Bucephalidae; Faustul, Faustulidae; Gy, Gymnophallidae; Tand, Tandanicolidae.

Discussion Taxonomy The present species possesses features that are fully consistent with species of the faustulid genus Pseudobacciger as recognised by Bray (2008). The present species has the same general organization and, in particular, possesses the naked two-chambered seminal vesicle reported for that genus. As reviewed by Dimitrov et al. (1999), Pseudobacciger presently is represented by just three species: Pseudobacciger cablei Madhavi, 1975, P. manteri Nahhas & Cable, 1964 and P. harengulae (Yamaguti, 1938). The present form is distinguished from two of these species as follows. Pseudobacciger cablei is reported as a distinctly larger worm (528704 μm long vs 371453 μm for the present material) in which the suckers (especially the oral sucker) are actually

smaller than in the present form (Madhavi 1975). Pseudobacciger cablei differs further from the new species in having intestinal caeca that terminate close to the anterior margins of the testes rather than towards the posterior margin, and in having vitelline follicles that reach relatively close to the anterior end of the body. Finally, P. cablei differs from the present form in being reported as having the genital pore well anterior to the ventral sucker as opposed to immediately anterior to it. Pseudobacciger manteri is reported at about the same overall size as the present material (373747 μmlong vs 371453 μm for the present material), but possesses a distinctly smaller oral sucker and caeca that terminate well posterior to the testes (Nahhas and Cable 1964). The present material is most similar to the type-species of Pseudobacciger, P. harengulae, as originally described by Yamaguti (1938) from the clupeid Sardinella zunasi (Bleeker) (as Harengula zunasi), and reported several times subsequently as reviewed by Dimitrov et al. (1999) [We are doubtful that the form reported from a pomacanthid by Machida and Uchida (2001) as P. harengulae can be conspecific with the species reported originally by Yamaguti (1938); see further comments below]. The overall size and shape of P. harengulae as reported in those studies and the present form are comparable as are the relative arrangements of their digestive systems and reproductive organs, but we detect a difference, however, in the larger size of the suckers of the present form: both oral and ventral suckers have size ranges either completely non-overlapping with those reported for P. harengulae (see Gaevskaja 1996; Dimitrov et al. 1999) or overlapping with a higher range (Yamaguti 1938). In particular, the ventral sucker of the present form has a length of 80109 μm, in contrast to 5885 μm in diameter as reported by Yamaguti (1938). Yamaguti’s specimens appear to have been slightly flattened and it might thus be expected that the size of the suckers would be inflated whereas they are distinctly smaller than reported here. Machida and Uchida (2001) identified specimens from the pomacanthid Genicanthus lamarck Lacépède as P. harengulae. These are much larger (9201,270  670890 μm) than previously described for this species (and indeed for the genus), with proportionally larger dimensions for most characters, including suckers, pharynx, oesophagus, gonads and eggs. In addition the forebody is relatively longer than that of the new species (3545 vs 2334% of body length). Machida and Uchida (2001) noted that the ‘variation with P. harengulae [as extended by their description] seems to make the status of each of the three species rather obscure’. We think it is unlikely that Machida and Uchida’s (2001) worms are conspecific with the tiny worm reported earlier from a clupeiform by Yamaguti (1938) Our conclusions regarding the identity of the present form are influenced by our understanding of host-specificity of these parasites. Miller et al. (2011a) analysed the reported host- specificity of 290 trematodes known from fishes of the Great Barrier Reef. Of these, just four had been reported from more than one order of fishes and for none of the four had the wide specificity 52

been tested by molecular data. The high prevalence and intensity reported in the present study suggest strongly that the infection of Chromis weberi is in no way accidental yet P. cheneyae n. sp. has not been seen in a wide range of other pomacentrid species examined on the GBR. The three presently-recognised species of Pseudobacciger are reported almost entirely from clupeiform fishes; the exceptions are reports of P. harengulae from a herbivorous sparid (Perciformes, the Bogue Boops boops Linnaeus) by Parukhin (1976a) and from a planktivorous pomacanthid (Perciformes, Lamarck’s Angelfish Genicanthus lamarck) by Machida and Uchida (2001). For the present form to occur widely in clupeiforms but in just one of many examined species of Pomacentridae (Perciformes) seems highly anomalous. We thus think that the apparent distinctions between the host of the present species and those reported previously for species of Pseudobacciger are a likely indication that the present form is specifically distinct. We consider taxonomic decision making in this genus to be unusually difficult. Species of Pseudobacciger are usually exceptionally small (of the more than 300 species of trematodes reported from fishes of the GBR, only the bivesiculid Paucivitellosus fragilis Coil, Reid & Kuntz, 1965 is as small as the new species (Pearson 1968), and important morphological features tend to be obscured by eggs. We note that the analysis of Dimitrov et al. (1999) implied that P. harengulae is an essentially cosmopolitan species given that there are records from Japan, Korea, the Atlantic and Indian Oceans and the Black Sea. Although this is certainly possible, as demonstrated by the molecular evidence that the aporocotylid Cardicola forsteri Cribb, Daintith & Munday, 2000 is cosmopolitan (Aiken et al. 2007), such widespread species appear to be exceptional. We think it possible, perhaps likely, that Pseudobacciger will prove to comprise a complex of significantly more species than are presently recognised. Such a conclusion will almost certainly depend on the generation of molecular data which, unfortunately, are available only for the present species so far. In the light of analysis of the morphology and host-specificity of the present form we conclude that it is best treated as a new species. However, the new name is proposed with a heightened sense that all such proposals are hypotheses open to testing.

Systematics The classification of the present form raises a systematic problem in that although its morphology makes it convincing as a species of Pseudobacciger, it has been shown to be phylogenetically remote from the representatives of four faustulid genera for which 28S or ITS2 rDNA sequence data are available. Instead, its position, as can best be judged on the basis of molecular data, is robustly within the Gymnophalloidea and as sister taxon to species of the Tandanicolidae. The Tandanicolidae is a small family with species distinctive especially in the form of the terminal genitalia and in usually infecting both marine and freshwater siluriforms (catfish) (Cribb and Bray

1994; Bray 2001). The present form certainly cannot be readily recognised in that family. The gymnophallid metacercaria sequenced for this study formed the sister taxon to Pseudobacciger + Tandanicolidae [We note that its identity as a gymnophallid was confirmed by its nesting among several other identified gymnophallids in the analyses of ITS2 rDNA sequence data]. Gymnophallids are essentially parasites of birds (Scholz 2002) so that, although the present form is very small, like most gymnophallids, it would have been a surprise if, of all the gymnophalloids, P. cheneyae had proven most closely related to that family. Our analyses incorporated six genera of Fellodistomidae but these form a well-supported clade sister to the Gymnophallidae + (Pseudobacciger + Tandanicolidae). In addition to the Fellodistomidae, Gymnophallidae and Tandanicolidae, Bray (2002) included the Botulisaccidae Yamaguti, 1971 and the Callodistomidae Odhner, 1910 in the Gymnophalloidea; the latter two were included “for convenience of identification”. No molecular data are available for the former family, but the sole member of this monotypical family does not resemble Pseudobacciger cheneyae n. sp. The Callodistomidae, as represented by the species Prosthenhystera obesa (Diesing, 1850) was found, utilising 28S rDNA, to be the sister group to the Allocreadiidae Looss, 1902 within the superfamily Gorgoderoidea Looss, 1901 by Curran et al. (2006). Phylogenetically, therefore, P. cheneyae n. sp. can thus be best considered as Gymnophalloidea incertae sedis with the possibility that it may ultimately require a family separate from those presently recognised. We speculate, thus, that the Faustulidae as presently conceived may require division into two families, one related to the Zoogonidae as presently recognised from the analysis of Olson et al. (2003) and a separate family within the Gymnophalloidea. Potential division of the Faustulidae has implications for the understanding of morphology in the group. We observe that there is a potentially relevant dichotomy in the anatomy of the genera presently recognised in the Faustulidae. Species of Allofellodistomum Yamaguti, 1971, Pronoprymna Poche, 1926, Pseudobacciger and Pseudosellacotyle Yamaguti, 1953 have cirrus- sacs or naked terminal genitalia with a very small pars prostatica whereas species of Antorchis, Bacciger (but we note some variation within this genus), Echinobreviceca Dronen, Blend & McEachran, 1994, Faustula, Paradiscogaster, Parayamagutia Machida, 1971, Trigonocryptus Martin, 1958 and Yamagutia Srivastava, 1937 all have a cirrus-sac enclosing a large, prominent pars prostatica. The three faustulid genera for which 28S rDNA sequences were previously available (Antorchis, Bacciger and Trigonocryptus) and the single genus for which ITS2 rDNA sequences were available (Paradiscogaster) are all of the latter type whereas P. cheneyae n. sp. is of the former type. This difference may form a morphological basis for the recognition of separate families. Notably, there are already two family group names, Pseudosellacotylinae Yamaguti, 1958 and Pentagramminae Yamaguti, 1958, which might be raised and applied to a distinct family within

the Gymnophalloidea. Although we consider the evidence too preliminary to justify a formal proposal relating to family names at this stage, we feel unable to continue to recognise Pseudobacciger in the Faustulidae. We recognise that the morphological similarity between P. cheneyae n. sp. and other Pseudobacciger spp. may be convergent, but the similarity is so striking that this seems unlikely. Potential division of the Faustulidae also has implications for the understanding of life- cycles in the group. A close relationship of the Faustulidae with the Zoogonidae has been previously intriguing in that all the many known zoogonid life-cycles involve gastropods as first intermediate hosts as recently reviewed by Barnett et al. (2014) whereas the handful of known faustulid cycles (species of Bacciger and Pseudobacciger) involve bivalves as first intermediate hosts (see Kim and Chun 1984; Bartoli and Gibson 2007). Should these species of Bacciger and Pseudobacciger prove to form a clade within the Gymnophalloidea then that relationship would be consistent with the use of bivalves by all known members of that clade, as well as the sister taxon Bucephalidae. Such an outcome would also lead to the prediction that the Faustulidae that are related to the Zoogonidae may prove to have gastropod rather than bivalve first intermediate hosts.

Ecology of damselfish and their parasites The Pomacentridae is one of the most prominent families of fishes on the GBR with 79 species known from the region (Randall et al. 1997). Pomacentrids are found mainly on shallow coral reefs, but a few species have been recorded at depths of 80 metres or more. Despite this family being diverse, occupying a range of habitats (live and dead coral, sand, coral rubble and the water- column) (Wilson et al. 2008), having varying dietary preferences (herbivorous, omnivorous, and planktivorous) (Randall et al. 1997), and a variety of social habits (social or solitary) (Fishelson 1998), the trematode fauna of this family is considered depauperate relative to some other fish families (Barker et al. 1994). Although, pomacentrids have been extensively examined for trematode parasites on the Great Barrier Reef (Manter 1969b; Cribb et al. 1992; Bray et al. 1993a; Bray et al. 1993b; Bray and Cribb 1998; Bray and Cribb 2004; Hunter and Cribb 2012), no gymnophalloid taxa have been recorded previously. Trematodes of pomacentrids in Australian waters demonstrate a wide spectrum of patterns of host specificity. Several species are evidently euryxenic (infecting taxa related only by ecology). In this category are Hysterolecitha nahaensis Yamaguti, 1942, Lecithocladium moretonense Bray & Cribb, 2004, Lepotrema clavatum Ozaki, 1932 and Thulinia microrchis (Yamaguti 1934) which have all been reported from non-pomacentrids in addition to pomacentrids (Bray et al. 1993a; Machida and Uchida 2001; Bray and Cribb 2002,2004) although the host range of these species have not been tested by molecular approaches. Several species are stenoxenic (infecting

phylogenetically related host species); the haplosplanchnid Schickhobalotrema pomacentri (Manter, 1937) and the transversotrematid Transversotrema damsella Hunter & Cribb, 2012 have been reported from multiple pomacentrid species (Cribb et al., 1994; Hunter & Cribb, 2012). Finally, two species of Lepotrema Ozaki, 1932 (Lepocreadiidae), one species of zoogonid, and one species of bivesiculid have each been reported from only a single pomacentrid species (Barker et al. 1994; Cribb et al. 1994a; Bray and Cribb 1998). The specificity of P. cheneyae n. sp. is consistent with this latter pattern of strict specificity. The ecology and behaviour of C. weberi is similar to that of other planktivorous damselfish species, such as Abudefduf whitleyi Allen & Robertson, Chromis ternatensis Bleeker and C. viridis Cuvier, all of which share the habitat of infected C. weberi. It is thus unclear why P. cheneyae n. sp. should infect only C. weberi. As C. weberi is classified as a non-territorial fish species (Ormond et al. 1996), it is unlikely that infection is the direct result of a narrow habitat choice. Host-specificity is often thought of as being generated by encounter and compatibility filters (Euzet and Combes 1980). In the present case it is not yet possible to determine which filter might be driving the apparently strict specificity seen here.

Acknowledgements We thank K. L. Cheney and E. C. McClure for their help collecting specimens and the staff of the Lizard Island Research Station for their support. This work was funded by an Australian Research Council Discovery grant awarded to T. H. Cribb and a Sea World Research and Rescue Foundation, Australia (SWRRFI) awarded to D. Sun and T. H. Cribb.

Chapter 3: Presence of cleaner wrasse increases the recruitment of fishes to coral reefs

Chapter 3: Presence of cleaner wrasse increases the recruitment of fishes to

coral reefs

Derek Sun, Karen L. Cheney, Johanna Werminghausen, Mark G. Meekan, Mark I. McCormick,

Thomas H. Cribb & Alexandra S. Grutter

Submitted to a journal for review

Abstract Mutualisms affect the biodiversity, distribution and abundance of biological communities. However, ecological processes that drive mutualism-related shifts in population structure are often unclear and must be examined to elucidate how complex, multi-species mutualistic networks are formed and structured. In this study, we investigated how the presence of key marine mutualistic partners can drive the organisation of local communities on coral reefs. The cleaner wrasse, Labroides dimidiatus, removes ectoparasites and reduces stress hormones for multiple reef fish species, and their presence on coral reefs increases fish abundance and diversity. Such changes in population structure could be driven by increased recruitment of larval fish at settlement, or by post- settlement processes such as modified levels of migration or predation. We conducted a controlled field experiment to examine the effect of cleaners on recruitment processes of a common group of reef fishes, and showed that small patch reefs (61–285 m2) with cleaner wrasse had higher abundances of damselfish recruits, than reefs from which cleaner wrasse had been removed over a 12-year period. However, the presence of cleaner wrasse did not affect species diversity of damselfish recruits. Our study provides evidence of the ecological processes that underpin changes in local population structure in the presence of a key mutualistic partner.

Introduction On coral reefs, cleaning symbioses are key mutualistic networks that increase abundance and promote species diversity in local fish communities (Grutter et al. 2003; Waldie et al. 2011). Cleaner organisms, such as the bluestreak cleaner wrasse, Labroides dimidiatus, interact with multiple species of fish, and provide fitness benefits to their clients by reducing stress hormones

(Soares et al. 2011) and ectoparasite loads such as gnathiid isopods (Grutter 1999a). The presence of cleaner wrasse also affects fish population structure, size-frequency distribution and growth (Grutter 1999a; Clague et al. 2011; Waldie et al. 2011). Abundance and diversity of fishes, both adult damselfishes (site-attached) (Bshary 2003; Waldie et al. 2011) and visitors (moving between patch reefs) (Grutter et al. 2003) increase, as does the abundance of juvenile visitor fishes (Waldie et al. 2011) on reefs with cleaner wrasse. Damselfishes (Pomacentridae) are one of the most abundant groups of fishes found on coral reefs. Damselfishes play a critical role in structuring coral reef benthic communities (Ceccarelli et al. 2005), with their presence indirectly increasing the survival of juvenile fishes (McCormick and Meekan 2007). Although the abundance and diversity of adult damselfishes are higher on reefs where cleaner wrasse are present (Bshary 2003; Waldie et al. 2011), it is not known if this difference occurs during recruitment, as damselfishes are relatively sedentary after settlement (when fish first join the reef assemblage after a pelagic larval stage). Whether differences in reef fish population structure in the presence of cleaning organisms involve regulatory processes such as migration or predation during settlement, or post-settlement during the juvenile and adult life stages, has not been investigated. Understanding recruitment, a critical demographic process, is important as it determines the structure of many marine and terrestrial communities (Caley et al. 1996). In this study, we examined whether the presence of cleaner wrasses influences recruitment patterns of larval coral reef fishes on patch reefs, which could explain differences in the abundance and diversity of local assemblages of adult fishes. As previous studies have shown a decrease in abundance and diversity of reef fishes such as damselfishes and juvenile visitor fishes on reefs where cleaner wrasse are absent (Bshary 2003; Grutter et al. 2003; Waldie et al. 2011), we predict a similar trend occurring for damselfish recruits.

Material and methods Our study was conducted on 16 small patch reefs in shallow (3−7 m depth) areas in lagoon and back-reef habitats near Lizard Island, GBR, Australia. Since 2000, ‘removal’ reefs (n = 7) were inspected at 3-month intervals and cleaner wrasse removed with hand and barrier nets; ‘control’ reefs (n = 9) were left undisturbed (see Clague et al. 2011 for a description of reefs; Waldie et al. 2011). The number of cleaner wrasse present on control reefs were recorded. To investigate whether the abundance and species richness of newly-settled damselfishes differed between removal and control reefs, we counted young recruits (< 20 mm TL) for five days in each of three lunar cycles, beginning four days after the new moon (18−22 November 2012, 17−21 December 2012 and 16−20 January 2013). These times were selected because settlement of

fishes to coral reefs peak around the time of the new moon during these summer months (Milicich and Doherty 1994). Counts alternated between randomly-selected control and removal reefs and fish were identified to species where possible. Fish were counted once per reef by the same observer (D.S.) on SCUBA for 60−120 min, depending on reef size. Fish were counted by the same observer to reduce variability. As damselfishes are most active from morning to midday (Bosiger and McCormick 2014), reefs were surveyed between 0900−1200 hr. The observer circled ~ 1 m above the reef counting an abundant, or several less-abundant, species (in the same order for each reef) on each pass. For analysis, fish were categorised either as common, wherein recruits of these species were found on 90% of reefs across all three months (of which there were five species: Pomacentrus adelus, P. amboinensis, P. moluccensis, P. nagasakiensis and Chrysiptera rollandi), or uncommon (all other damselfishes). We compared the abundance and three measures of diversity of damselfish recruits on control and removal reefs. We conducted generalized linear mixed effect models (GLMER) with a Poisson distribution error, to examine whether the total abundance of all the five common damselfishes combined (and also separately by species) and the total abundance of uncommon damselfishes differed relative to the presence of cleaner wrasse. The terms treatment (cleaner or no cleaner presence), site (Lagoon or Casuarina Beach) and time (November 2012, December 2012 and January 2013) were fixed factors; reef identity was a random factor and reef area was a covariate. A posteriori Tukey-Kramer HSD post-hoc test was used to interpret significant interactions. We also used full linear mixed- effect models (LMER) with a REML procedure to analyse species richness and Simpson’s indices of diversity (Krebs 1999) and used a permutational multivariate analysis of variance (PERMANOVA) and principle coordinates analysis (PCO) for the species composition. To determine the best model in all analyses, we compared the full model with models in which one of the explanatory terms was dropped using the “drop1” function (Chambers 1992). The term was dropped if the analysis of variance found that a dropped term had no significant effect on the model using the Chi-square distribution. A Bonferroni correction (α = 0.01) was applied to account for multiple comparisons. All statistical analyses were conducted using R version 3.0.1 (R Development Core Team 2013), except for the species composition which was analysed using PRIMER 6.0 (Anderson et al. 2008).

Results Thirty-one species of damselfish recruits were identified (see Appendix I). The total abundance of common damselfishes was lower on removal than on control reefs (p = 0.006, Figure 1). At the species level, the abundances of C. rollandi (p = 0.0001, December only) and P. amboinensis (p <

0.0001, Figure 2a-b) were significantly lower on reefs where cleaner wrasse had been removed than on control reefs. Although there were similar trends in the average abundance of the other common species, P. adelus, P. moluccensis and P. nagasakiensis, these differences were not significant (p = 0.0769, 0.0689 and 0.0573 respectively, Figure 2c-e). Total abundance of uncommon damselfishes was also lower on removal reefs (p = 0.0070, Figure 2f). We found no differences in measures of diversity between removal and control reefs (species richness: p = 0.3578; Simpson’s diversity index: p = 0.8418; species composition: PERMANOVA,

F (1 ̶ 36) = 1.83, p = 0.1029, Figure 3a-c).

Figure 1 Total abundance of common damselfishes on experimental reefs. Means (± SE) abundances are from reefs with and without cleaner wrasse. * Significant difference in cleaner wrasse treatment (p < 0.01).

Figure 2 Mean (± SE) abundance of recruiting damselfish species. Control and cleaner wrasse removal reefs: a-e) abundance of individual common species, f) abundance of uncommon species. * Significant difference in cleaner wrasse treatment (p < 0.01), n.s = p > 0.01.

Figure 3 Diversity of all damselfish recruits per reef. a) Species richness, b) Simpson’s diversity index, c) species composition on control and cleaner wrasse removal reef, represented by a principle coordinate analysis (PCO).

Discussion The combined abundances of the five common species, the individual abundances of two common species (C. rollandi and P. amboinensis), and the total abundance of uncommon recruited damselfishes were higher on reefs with cleaner wrasse present than on those without. There was also a strong (but not statistically significant) trend for abundances of the other common damselfishes (P. adelus, P. moluccensis and P. nagasakiensis) to be higher on reefs with cleaners. These results suggest that the presence of a key mutualistic species, creates changes in population structure around the time of settlement and recruitment. This in turn could explain the observed changes in abundance and distribution of coral reef fish populations in relation to cleaner wrasse presence (Waldie et al. 2011). The presence of cleaner wrasse could influence recruitment processes by a number of direct or indirect mechanisms. If young fish use visual or olfactory cues to locate potential microhabitats, then sensory cues may enable them to locate cleaners directly. Indeed, a recent study suggests recently-settled fish can recognise and preferentially choose a microhabitat near a cleaner (Sun et al. in review, Chapter 4). Alternatively, cleaner wrasse could influence recruitment processes indirectly by reducing the abundance of ectoparasites in the local environment. Higher abundances of parasites in a system have the potential to detrimentally affect the recruitment of juveniles, as demonstrated in lower recruitment success of juvenile cricket frogs (Acris crepitans) in ponds with high rates of parasitic infection (Beasley et al. 2005). Although both adults and recruit damselfishes are known to be infected with gnathiids, cleaner wrasse rarely clean recruits and tend to clean larger individuals (Sun et al. in review, Chapter 4). However, with gnathiids decreasing swimming performance and increasing mortality of settling fish (Grutter et al. 2011) the removal of even a single gnathiid can increase the survival of an individual. Cleaner wrasse may also indirectly promote fish recruitment by influencing the abundance of conspecifics, which have been shown to influence microhabitat selection by settling fish, as settling fish may use the presence of conspecifics as a measure of habitat quality (Coppock et al. 2013) or the behaviour of predators (Cheney et al. 2008). The abundance of all (adult and juvenile) conspecifics for P. amboinensis, P. moluccensis and P. nagasakiensis are also higher on reefs with cleaner wrasse than those without (D. Sun, unpublished data). Further investigation is required to disentangle the relative importance of these direct and indirect mechanisms. It should be noted that the gradual decrease in the abundance of damselfish recruits from November to January is due to the larval fish season ending. Interestingly, there was no difference in species richness, Simpson’s diversity index, or composition of damselfish recruits between control and removal reefs, suggesting that the presence of cleaner wrasse affects the abundance of individual species but not the overall community diversity of recruit damselfishes. Previous studies have shown an increase in the abundance and

species richness of resident adult fishes in the presence of cleaner wrasse (Bshary 2003; Waldie et al. 2011). Our results suggest that such patterns in diversity are not likely due to differential recruitment, but due to post-settlement events, including migration in response to cleaner wrasse presence or the effects of parasitism (Waldie et al. 2011). Parasitic infection by gnathiid isopods can make recruits more susceptible to the high rates of predation that occur immediately after settlement (typically 56 % within two days of settlement) (Almany and Webster 2006). Ultimately, this process could affect the diversity of fishes in the presence of cleaner wrasse. In summary, we have demonstrated for the first time that the presence of cleaner wrasse has the potential to influence the structure and demography of reef fish populations via effects on the process of recruitment. Thus, mutualism is an important mechanism that contributes to the already- complex recruitment processes of reef fishes.

Acknowledgements We thank Lizard Island Research Station staff. We also like to thank E. C. McClure, M. De Brauwer for field assistance. S.P. Blomberg for statistical advice and R. Q-Y. Yong for reading drafts of this manuscript. Funding was provided by Sea World Research and Rescue Foundation, Australia (SWRRFI) to D.S, T.H.C and A.S.G and an Australian Research Council Discovery Grant to A.S.G and M.G.M.

Chapter 4: Cleaner wrasse influence habitat selection of young damselfish

Derek Sun, Kare L. Cheney, Johanna Werminghausen, Eva C. McClure, Mark G. Meekan, Mark I.

McCormick, Thomas H. Cribb & Alexandra S. Grutter

Submitted to a journal for review

Abstract The presence of bluestreak cleaner wrasse, Labroides dimidiatus, on coral reefs increases total abundance and biodiversity of reef fishes. The mechanism(s) that cause such shifts in population structure are unclear, but it is possible that young fish preferentially settle into microhabitats where cleaner wrasse are present. To investigate this possibility, we conducted aquarium experiments to examine whether settlement-stage and young juveniles of ambon damselfish, Pomacentrus amboinensis, selected a microhabitat near a cleaner wrasse (adult or juvenile). Both settlement-stage (0 d post-settlement) and juvenile (~ 5 weeks post-settlement) fish spent a greater proportion of time in a microhabitat adjacent to L. dimidiatus than in one next to a control fish (a non-cleaner wrasse, Halichoeres melanurus) or one where no fish was present. This suggests that cleaner wrasse may serve as a positive cue during microhabitat selection. We also conducted focal observations of cleaner wrasse and counts of nearby damselfishes (1 m radius) to examine whether newly-settled fish obtained direct benefits, in the form of cleaning services, from being near a cleaner wrasse in the field. Although abundant, newly-settled recruits (< 20 mm TL) were rarely (2 %) observed being cleaned per 20 min observations compared with larger damselfishes (58 %). Individual damselfish that were cleaned were significantly larger than the median size of the surrounding nearby non-cleaned conspecifics; this was consistent across four species. The selection by settlement-stage fish of a microhabitat adjacent to cleaner wrasse in the laboratory, despite only being rarely cleaned, suggests that even rare cleaning events and/or indirect benefits may drive their settlement choices. This behaviour may also explain the decreased abundance of young fishes on reefs from where cleaner wrasse had been experimentally removed. This study reinforces the potentially important role of mutualism during the processes of settlement and recruitment of young reef fishes.

Introduction On coral reefs there are many mutualistic cleaners that remove parasites from other organisms, including coral reef fishes. One of the most obvious of these is the bluestreak cleaner wrasse, Labroides dimidiatus (Labridae), a cleaner wrasse common throughout the Indo-Pacific (Randall et al. 1997). Labroides dimidiatus has been shown to reduce the ectoparasite load of individual client fishes (Gorlick et al. 1987; Grutter 1999a; Clague et al. 2011), aggression by predators (Cheney et al. 2008), stress (measured in cortisol levels) via physical contact (Bshary et al. 2007; Soares et al. 2011) and increase growth and size of client fishes (Clague et al. 2011; Waldie et al. 2011). In addition, these cleaner wrasse affect local populations of reef fishes, so that in their absence there are reductions in the diversity and abundance of site-attached adult residents, visitors (fishes that move between patch reefs) and juvenile visitor fishes (Bshary 2003; Grutter et al. 2003; Waldie et al. 2011). Although the positive effects of the presence of L. dimidiatus on individuals and populations have been demonstrated in experimental manipulations of patch reefs ranging in size from ~61 to 285 m2, the mechanisms involved, particularly at small spatial scales (1 m–10s of m), remain largely unstudied (Waldie et al. 2011). One of the most critical periods that effects the population dynamics of coral reef fishes occurs during the settlement of young fish from the plankton to the reef and the few days or weeks of life in benthic habitats that immediately follow this event (post-settlement). Processes at this time ultimately determine which individuals are recruited to the population of juveniles. Young fish can actively choose the habitat into which they will settle and the processes that underlie these choices are complex (Hoey and McCormick 2004; Heinlein et al. 2010). Additionally, the mortality rate at this time is high; indeed, it has been estimated that 57 % of individuals die within the first 1−2 days of settlement (Almany and Webster 2006), which could be due to microhabitat selection. Thus, the choice of a suitable microhabitat at settlement is critical for successful recruitment. Selecting the optimum microhabitat depends on numerous factors, which may include the structure of the microhabitat (Tolimieri 1995), and the presence of potential predators (Vail and McCormick 2011) and conspecifics (Sweatman 1985). It was recently shown that the experimental removal of L. dimidiatus was associated with a decrease in the abundance of new recruits on patch reefs, relative to control reefs where cleaner wrasse were present (Sun et al., in review, Chapter 3). However, it is unknown whether the presence of L. dimidiatus can serve as a positive cue for microhabitat choice at settlement. Our study used experimental and observational approaches to determine whether the presence of juvenile or adult cleaner wrasse influenced the choice of microhabitat by settling juveniles of the ambon damselfish, Pomacentrus amboinensis. In the laboratory, we examined whether the presence of cleaner wrasse influences habitat choice of P. amboinensis. To determine

the extent to which recently-settled damselfishes were cleaned by L. dimidiatus juveniles, and thus whether the benefits of settling near a cleaner wrasse could be related to the cleaning services they provide or to other indirect benefits, focal observations of the cleaning interactions between L. dimidiatus and damselfishes were conducted in the field. Lastly, to determine whether juvenile cleaner wrasse are size selective in their choice of clients, we compared the size of individuals of the four most abundant damselfishes that were cleaned by the wrasse with the median size of non- cleaned conspecifics.

Materials and methods Laboratory experiment: Does the presence of L. dimidiatus affect microhabitat choice in young damselfish? We conducted an aquarium experiment at Lizard Island Research Station, GBR, Australia, to investigate whether the presence of L. dimidiatus influenced microhabitat choice of P. amboinensis, a species chosen because of its high abundance at the study site (Sun et al. in review, Chapter 3). Pomacentrus amboinensis were collected between November 2011 and January 2012 using light- traps (see Meekan et al. 2001 for design) moored overnight approximately 100 m from the reef. These fish were juveniles caught just prior to settlement from the plankton and thus had no previous exposure to L. dimidiatus. Fish were held in groups of ~ 20 in clear plastic holding aquaria (29 × 17 × 12 cm) with a constant flow of seawater. Fish collected in November were held for around 5 weeks (at which point they were classified as juveniles, i.e. ~ 5 weeks post-settlement), whereas fish collected in January were tested within a day of capture (classified as settlement-stage, i.e. 0 days post-settlement). Fish were fed ad libitum twice a day with live nauplii of brine shrimp (Artemia sp.). Both adult and juvenile cleaner wrasse (L. dimidiatus) and a non-cleaner control fish (pinstripe wrasse, Halichoeres melanurus), were collected from the Lizard Island lagoon using hand-nets, barrier nets and anaesthetic clove oil mixed with alcohol and seawater (10 % clove oil; 40 % ethanol; 50 % seawater). Halichoeres melanurus was chosen as a control due to similarity in body shape to L. dimidiatus (Grutter 2004a; Cheney et al. 2008; Cheney et al. 2009). All wrasses were maintained in aquaria (43 × 32 × 30 cm) with running unfiltered seawater. Clear Perspex sheets were placed at either end of a glass experimental aquaria (35 × 36 × 65 cm), with flow-through seawater, to create two end compartments (10 cm away from the end of the aquarium). The middle section of the aquarium was divided into three 15 cm subsections, using vertical black lines drawn on the front and back of the aquarium. Three polyvinyl chloride (PVC) tube shelters (6 × 5 cm diameter) were placed in the centre of each subsection (Fig. 1). Aquaria

were covered with black plastic on three sides to isolate the fish from external activity, and a green plastic shade cloth hung in front of the tanks to create an observation blind. In each end compartment a L. dimidiatus, a control fish (H. melanurus) or no fish was placed, depending on the treatment: 1) L. dimidiatus and H. melanurus; 2) L. dimidiatus and no fish; or 3) H. melanurus and no fish. These combinations were presented to a randomly-selected P. amboinensis ‘client’ and the side in which a cleaner or control (or no fish) was placed was randomised, but balanced. The ‘no fish’ treatment at one end compartment of the aquarium was used to control for the ‘client’ potentially avoiding any heterospecifics, irrespective of identity. A P. amboinensis was placed in the middle section of the aquarium in a bottomless clear plastic container (8 × 8 × 8 cm) with numerous small (5 mm) ventilation holes for 20 min prior to commencing the trial, allowing the fish to acclimate to the new surroundings. The bottomless container was removed using an attached monofilament string by the observer, who was positioned behind the observation blind. Each trial was conducted for 20 min, and the position of P. amboinensis within the aquarium (left, middle, or right section) was recorded every 15 s. It is inferred that the degree of proximity to the stimulus represented its choice of microhabitat. Experiment 1 was divided into three parts (A, B, C) with different ontogenetic stages of cleaner, control and P. amboinensis per treatment (see Table 1 for a summary of the fish stage and species). For each part, we used 8 cleaner and 8 control fish that were randomly selected, and 75 P. amboinensis (n = 25 individuals per treatment). At the end of each trial, cleaner and control fish were replaced. Individual P. amboinensis were only used once, so that their behaviour was not affected by previous exposure. Part A used adult L. dimidiatus and H. melanurus, and Part B used juveniles of both species. These experiments tested whether the ontogenetic stage/size of L. dimidiatus affected habitat choice of juvenile P. amboinensis. Part C was conducted when settlement-stage fish became available in the light-traps; juvenile L. dimidiatus and H. melanurus were used for this component. The aim of this experiment, relative to Part B, was to determine whether P. amboinensis’ ontogenetic stage/size affected its habitat choice.

Figure 1 Front view of the laboratory experiment tank design. A: clear Perspex sheet placed on either side to create compartments for stimuli, B: vertical dashed lines on front and back of aquarium to create subsections, and C: PVC tube shelters placed in each subsection. Not drawn to scale.

Table 1 Ontogenetic stage and size of fish used in each part for laboratory experiment. Size is mean ± SE, TL.

Labroides Halichoeres Pomacentrus Part Date dimidiatus melanurus amboinensis (cleaner) (control) (client)

Adult Adult A December 2011 Juvenile (79.1 ± 4.0 mm) (86.1 ± 0.8 mm) (19.3± 0.3 mm)

Juvenile Juvenile Juvenile B December 2011 (24.2 ± 0.5 mm) (24.8 ± 0.4 mm) (18.0 ± 0.3 mm)

Juvenile Juvenile Settlement-stage C January 2012 (23.1 ± 0.7 mm) (24.3 ± 0.8 mm) (15.2 ± 0.1 mm)

Field observations: Do juvenile L. dimidiatus clean recently-settled damselfish recruits and are they selective in their choice of clients? To assess whether recently-settled damselfish recruits are cleaned by L. dimidiatus, and thus gain direct fitness benefits from settling near them, we conducted 20 min behavioural observations on juvenile L. dimidiatus (n = 79) on a section of continuous reef (~ 200 m) at Coconut Beach (1440S, 14528E), Lizard Island. Although cleaner wrasse of all sizes mainly consume ectoparasites (Grutter 2000), we observed juvenile L. dimidiatus, rather than adults, due to the higher frequency of juvenile L. dimidiatus interacting with young reef fish (Robertson 1974, D. Sun, unpublished data). Before observations commenced, observers learned to accurately estimate fish size (TL) underwater by using printed and laminated outlines of model fish (12−74 mm TL). The fish models were placed on the reef at varying distances from the observer, and sizes were estimated and compared with the actual length of the model until observers were more than 80 % accurate at estimating length. Rulers printed onto dive slates were also used as a guide for estimating fish length. For each observation, a juvenile L. dimidiatus was located and its estimated TL was recorded. A 1 m radius was measured around the area where the L. dimidiatus was first sighted using a 1 m string and marked using five lead sinkers (7 mm) attached to floats (25 mm) and all fish within this area were recorded. All fish present were recorded to account for the availability of other sizes of fish as potential clients that could influence the cleaners’ choice of client. Surrounding fish were allowed to acclimate (~ 5 min) to the presence of the markers and observer, before the commencement of observations. Cleaning interactions were recorded as any physical contact between cleaner and client, or significant inspection by the cleaner (> 1 s). For each interaction, we recorded the client species (where possible), cleaning duration, and estimated client size (TL, mm). To compare client size and the median size of nearby (non-cleaned) conspecifics, the size of all other fish present within the marked observation site (TL = > 10 to < 80 mm), but not involved in any cleaning interaction, was also estimated. All other fish were identified to species where possible. To ensure that each L. dimidiatus was observed only once, observers began at one end of the reef and continued without backtracking. In contrast to adults, juvenile L. dimidiatus stay within a highly restricted home range of around 4 m2 (Robertson 1974), further reducing the possibility of observing an individual twice. Although every individual within the observation area was identified to species, ultimately, in order to adequately perform statistical analyses, only the most abundant damselfish species (both cleaned and non-cleaned conspecifics) from the observations were used. These four common species were Pomacentrus amboinensis, P. moluccensis, P. nagasakiensis and P. wardi.

Statistical analyses Laboratory experiment A full linear mixed-effects (LMER) model with a restricted maximum likelihood (REML) procedure was used to examine whether P. amboinensis spent more time near a L. dimidiatus than near a control fish or no fish compartment. Two analyses were conducted; the first model used data from parts A and B (Table 1) to determine whether ontogenetic stage of the cleaner had an effect on juvenile P. amboinensis’ habitat choice, while the second model used data from part B and C, and included ontogenetic stage (settlement-stage and juvenile P. amboinensis) of the client as a factor. In the first model, cleaner ontogenetic stage (adult or juvenile), settlement stimulus type to which the microhabitat was adjacent (cleaner, control, no fish), aquarium side the cleaner was placed (left or right) and fish stimulus treatment combination (cleaner/control, cleaner/no fish or control/no fish) were used as fixed factors. Replicate trial identity was added as a random factor, to account for an order effect. In both models, proportion of time spent in each section of the aquarium (right, left and centre) was transformed by taking the arcsine of the square root of the proportion to meet the assumptions of normality.

Field observations A LMER model with a REML procedure was used to test whether there was a significant difference between client fish and the median size of the nearby conspecifics. The terms treatment (client or nearby conspecifics), nearby conspecific size and species (four damselfish species: P. amboinensis, P. moluccensis, P. nagasakiensis and P. wardi) were fixed factors, and the identity of the individual cleaner fish and the cleaning event identity (specifying which client and nearby conspecifics corresponded with one another) were random factors. To determine whether client size varied with cleaner size or the size of all other nearby species, separate (LMER) analyses were run with cleaner size and other nearby fish size as covariates and cleaner identity as a random factor. A general linear model (GLM) with a binomial distribution was used to determine whether there was a significant difference between recruits that were cleaned and the abundance of other fish species and recruits within the 1 m radius observation area. To determine the best model in all analyses, we compared the full model with models in which one of the explanatory terms was dropped using the ‘drop1’ function (Chambers 1992). The term was dropped if the analysis of variance found that a dropped term had no significant effect on the model using the Chi-square distribution. A Tukey-Kramer HSD test (TK-HSD) using the glht function in the ‘multcomp’ package (Horthorn et al. 2013) identified where differences occurred and a summary output table was reconstructed using the results from the ‘drop1’ function. Prior to all analyses, the assumptions of normality and homogeneity of variance were assessed using

histograms, residuals and quantile-quantile plots. All statistical analyses used R version 3.0.1 (R Development Core Team 2013).

Results Laboratory experiment: Does the presence of L. dimidiatus affect microhabitat choice in young damselfish? In all three parts (A-C) of the experiment, P. amboinensis spent significantly more time in the subsection or PVC tube that is adjacent to a L. dimidiatus than next to a control fish or no fish compartment (Fig. 2a-c). The first statistical analysis, which used data from parts A and B to examine whether cleaner ontogenetic stage had an effect, showed that the proportion of time that juvenile P. amboinensis spent next to the stimulus fish (cleaner, control or no fish) differed according to an interaction between fish stimulus and cleaner stage (LRT = 19.98, df = 2, P < 0.0001; Fig. 2a, b); a TK-HSD test showed that juvenile P. amboinensis spent significantly more time near a juvenile than an adult L. dimidiatus (P < 0.05). There was no significant effect of treatment combination (LRT = 0.474, df = 2, P = 0.788) or aquarium side (LRT = - 4.18, df = 1, P = 0.772). For the second statistical analysis, which used data from parts B and C to examine whether ontogenetic stage of the client affected habitat choice, the proportion of time that P. amboinensis spent next to a habitat did not differ with stage (LRT = - 4.40, df = 1, P = 0.749) but did differ according to fish stimulus (LRT = 144.22, df = 2, P < 0.0001; Fig. 2a, c); a TK-HSD test showed that P. amboinensis spent significantly more time near L. dimidiatus (P < 0.05). The proportion of time spent in a habitat differed according to treatment combination (LRT = 37.11, df = 2, P < 0.0001). A TK-HSD test showed that P. amboinensis spent more time in both chosen microhabitats (i.e. spent less time in the middle section) in the cleaner/no fish and control/no fish stimulus treatment combinations than in the cleaner/control treatment. The effect of aquarium side was not significant (LRT = -1.44, df = 1, P = 0.182). There was no significant effect of size between cleaner and control wrasse (> 0.05).

Figure 2 The percentage of time spent on microhabitat adjacent to fish treatment by damselfish client Pomacentrus amboinensis. a) Part A: Juvenile P. amboinensis with adult cleaner/control fish, b) Part B: Juvenile P. amboinensis with juvenile cleaner/control fish, and c) Part C: Settlement- stage P. amboinensis with juvenile cleaner/control treatment.

Field observation: Do juvenile L. dimidiatus clean recently-settled damselfish recruits and are they selective in their choice of clients? A total of 1107 cleaning interactions were recorded during 79 observations of individual L. dimidiatus juveniles over a period of 1580 min. Cleaner wrasse cleaned 104 species of fish from 19 families, including 32 species of damselfishes of which the following were cleaned most often (listed from most to fewest): Acanthochromis polyacanthus, Stegastes apicalis, Amblyglyphidodon curacao, P. moluccensis, Plectroglyphidodon lacrymatus, P. amboinensis, P. nagasakiensis, P. wardi and Chromis lepidolepis (see Appendix 2 for a full list of client species). These interaction frequencies were not adjusted for relative abundance of each client species. The mean number (± SE) of clients cleaned and the duration spent cleaning per 20 min observation was 14 ± 1 fish and 83.8 ± 7.3 s, respectively. Overall, only 14 % of cleaning interactions were between juvenile L. dimidiatus and the four common damselfishes (P. amboinensis, P. moluccensis, P. nagasakiensis and P. wardi) that had the greatest abundances within the 1 m radius of a cleaner wrasse. A total of 46 % of interactions occurred with other damselfishes (excluding the latter common species listed), while the remaining 40 % of interactions occurred between non-damselfishes and juvenile L. dimidiatus. Overall, 60 % of the clients of juvenile L. dimidiatus were damselfishes. The range and median size (mm TL) of the common damselfishes that were cleaned were: P. amboinensis (20–80 and 50 mm), P. moluccensis (25–80 and 50 mm), P. nagasakiensis (15–90 and 40 mm), and P. wardi (50–120 and 90 mm); the duration (seconds, mean ± SE) of time spent cleaning per 20 min observation for the common species was 14 ± 2.5 s. Client size differed among the four species (LRT = 68.01, df = 3, P < 0.0001; Fig. 3a); a TK-HSD test showed that P. wardi clients were significantly larger than the other common client damselfishes. Damselfishes that were cleaned were significantly larger than nearby non-cleaned conspecifics occurring within a 1 m radius; this effect was consistent for all four species (LRT = 173.72, df = 1, P < 0.0001; Fig. 3a). Separate analyses showed that the size of the damselfish client was not correlated with the size of the cleaner wrasse (LRT = 0.17, df = 1, P = 0.675) nor was it correlated with the median size of all other fish species combined (excluding common damselfishes) within the 1 m radius of the cleaner (LRT = 2.55, df = 1, P = 0.110). Despite the presence of numerous recently-settled damselfishes (size range 15–20 mm) within the 1 m radius observation areas (Fig. 3b), only 2 % (22 out of 1107 cleaning interactions) of fishes that were cleaned were < 20 mm TL. There was a significant negative relationship between the number of recruits that were cleaned and the abundance of both other fish species (> 20 mm, TL) (LRT = 7.97, df = 1, P = 0.004, Fig. 4a) and recruits (LRT = 16.95, df = 1, P = < 0.0001, Fig, 4b) within the 1 m radius observation area. The probability of a recruit (< 20 mm) being cleaned by a juvenile cleaner wrasse in a 20 min period is 0.002 (95 % CI = 3.22 × 10-6 ̶ 0.22) when there are

nine other fishes and a probability of 0.003 (95 % CI = 6.77 × 10-5 ̶ 0.07) when there are three recruits within the observation area (Fig. 4a-b).

100 a

Client

)

m 80 Nearby

(

e 60

a 40

M 20

0 i is is is d s s s r n n n a e e ie w in c k . o c a P b lu s o a m g a m a . . n P P . P Fish species

r 8 e b

s i 6

a 4

c 2

u 0

b 0 5 0 5 0 5 0 5 0 5 0 5 0 5 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 a ------< 1 6 1 6 1 6 1 6 1 6 1 6 1 6 > 1 1 2 2 3 3 4 4 5 5 6 6 7 7

a e Size categories (mm)

Figure 3 Cleaning interactions of juvenile cleaner wrasse, Labroides dimidiatus and damselfishes. a) Size (TL) of damselfish clients and nearby non-cleaned conspecifics. Bars represent median; error bars represent 25th and 75th quantiles, b) size distribution of nearby non-cleaned damselfish per 1 m radius; bars indicate mean, error bars represent SE.

P. amboinensis C. rollandi

f 100 f 30

e * e *

1a 1b

r 80 r

c 60 c

n 40 n

u u 10

n 20 n

M 0 M 0 l l l l o a o a tr v tr v n o n o o m o m C e C e R R

P. moluccensis P. nagasakiensis

f 100 f 100

e e n.s. *

r r 80 80

c c 60 60

n n 40 40

n 20 n 20

M M 0 0 l l l l o a o a tr v tr v n o n o o m o m C e C e R R

P. adelus Total damselfish abundance

f 100 f 20

e e n.s. *

r r 80

e e 15

c 60

a a 10

n 40

5 n 20

M M 0 0 l l l l o a ro a tr v t v n o n o o m o m e C e C R R

Figure 4 The probability of an individual recruit (< 20 mm) being cleaned by a juvenile cleaner wrasse, Labroides dimidiatus, in relation to the abundance of a) other fish species and b) abundance of recruits per 20 min observations.

Discussion Our study provides experimental evidence that the presence of cleaner wrasse, L. dimidiatus, may directly influence the choice of microhabitats of young coral reef damselfish. In the laboratory, both settlement-stage and juvenile P. amboinensis preferentially selected microhabitats that were in close proximity to L. dimidiatus. This may explain the observed patterns of increased abundance of fishes in the presence of cleaner wrasse on the reef. However, contrary to expectations, our observations revealed that these young damselfish were rarely cleaned by L. dimidiatus. Thus, it is possible that

even these rare cleaning events and/or indirect benefits resulting from the presence cleaner wrasse may drive the selective behaviour of settling damselfish. We assessed microhabitat choice by settlement-stage (0 days post-settlement) and slightly older juvenile (~5 weeks post-settlement) P. amboinensis and found that individuals preferentially selected a microhabitat adjacent to L. dimidiatus. This indicates that the ‘attraction’ towards cleaner wrasse occurs in both stages. Habitat selectivity at settlement appears common in young damselfishes, with some species attracted to particular microhabitats such as coral heads that harbour conspecifics (Lecchini et al. 2007a; Coker et al. 2012a), while others are attracted to coral heads associated with heterospecifics (Green 1998; Almany 2003). Our study adds cleaner wrasse to the list of organisms that settlers are attracted to, so that the presence of cleaner wrasse could act as an indicator of microhabitat quality to settlement-stage fish. Furthermore, these results suggest that if juveniles move between microhabitats, that their site selection may also be influenced by the presence of L. dimidiatus. The apparent attraction of L. dimidiatus may explain the greater number of recently-settled recruits (Sun et al. in review, Chapter 3), juvenile visiting fishes and even adult damselfishes, found on patch reefs where L. dimidiatus are present relative to reefs without cleaner wrasse (Bshary 2003; Waldie et al. 2011). There are two ways in which these patterns may arise, using the presence of L. dimidiatus as a cue. Settlement-stage fish may either choose to settle on certain patch reefs, or engage in post-settlement movement between patch reefs as juveniles (Simpson et al. 2008). Both of these processes may additively result in damselfishes being more abundant on reefs with cleaner wrasse than on those without. The fish used in the laboratory experiment were naïve with respect to the shape or odour of L. dimidiatus, since they were collected by light-traps while still in their planktonic phase. Despite this naïveté, young fish were able to differentiate between L. dimidiatus and another species that was similar in size and shape but differed in colour, pattern and behaviour. The attraction to cleaner wrasse thus appears to be an innate behaviour. Slightly older recruits displayed the same abilities. However, we did not test whether P. amboinensis recognised L. dimidiatus as a cleaner per se. On some occasions, the fish did behave like a client by remaining in the same position and adopting a pose with spread fins when next to the compartment with a cleaner wrasse. This behaviour did not occur when they were adjacent to the control fish or an empty compartment. Prospective clients of cleaner fish often assume this posture when they attempt to attract a cleaner to inspect them (Côté et al. 1998). Because the cleaner wrasse was separated by Perspex, both the cleaner and client were unable to physically interact; such physical contact likely acts as a positive feedback (Bshary and Würth 2001). Losey et al. (1995) previously showed that newly-metamorphosed laboratory-reared, and hence ‘cleaner-naïve’ Hawaiian humbug damselfish, Dascyllus albisella, recognised the

Hawaiian cleaner wrasse, Labroides phthirophagus by adopting the cleaning ‘pose’. Our study thus confirms the hypothesis that client fish have an innate ability to recognise cleaners. It seems likely that P. amboinensis used visual cues to differentiate between the two stimulus fish that were presented. A combination of black and blue and/or yellow body patterns displayed by cleaner fish, including L. dimidiatus, allows clients to visually recognise cleaners (Cheney et al. 2009). Indeed, the particular colouration and patterns displayed by cleaner fish attract the higher numbers of client fish than other colour and pattern combinations. Client fish are also known to recognise cleaner fish based on their small body size and the presence of lateral stripes (Stummer et al. 2004). Pomacentrus amboinensis both chose adult and juvenile L. dimidiatus over the other stimuli. This occurred despite the two cleaner stages being differently coloured, with both having a black stripe contrasted by blue, but the adults also having the colours yellow and white. However, juvenile P. amboinensis spent more time near juvenile L. dimidiatus than adults implying the possibility of ontogenetic variation in choice of cleaner fish. Alternatively, this could simply be due to the difference in relative size between the client and cleaner at different ontogenetic stages. The same choice test was not performed using settlement-stage fish due to their limited numbers in the light traps. Previous studies have shown that fish larvae have good visual senses, which enables them to detect and recognise conspecifics, heterospecifics and predator fishes (Lecchini et al. 2005; Lecchini et al. 2014). Focal observations of juvenile L. dimidiatus on reefs revealed that only 2 % of their clients were < 20 mm TL. This suggests that small individuals are rarely cleaned by L. dimidiatus and so the interaction may contribute little to the cleaner wrasse’s diet. Small-bodied individuals, such as recently-settled damselfishes, are infected with very few ectoparasites (Grutter et al. 2010; Sun et al. 2012) and the presence of gnathiid isopods (the main food source of L. dimidiatus) is relatively low in juvenile fishes (only 3.5 % of recruit P. amboinensis are infected) (Grutter et al. 2011). From the cleaner wrasse’s perspective, these factors likely make settlers an unsuitable source of ectoparasites. It was therefore unsurprising that we found that the damselfish that were cleaned by juvenile L. dimidiatus were larger than the median size of non-cleaned conspecifics within a 1 m radius. This result was consistent with previous studies that show that L. dimidiatus selectively clean larger fish (Grutter et al. 2005; Clague et al. 2011), probably due to client size and parasite load being closely correlated (Grutter 1994,1995a; Grutter and Poulin 1998); L. dimidiatus may thus use size as an indicator of food availability. From the client’s perspective, the odds of an individual new recruit (< 20 mm, TL) being cleaned per 20 min, was low, with only two or three out of 1000 fish predicted to be cleaned when there were no other fishes or recruits (< 20 mm) within a 1 m radius, and this rate rapidly decreased

towards zero with increasing abundance of nearby other fish species and recruits. Over a day, however, the cumulative odds would be much higher. Such rare events could have disproportionately significant benefits as some parasites can have especially harmful effects on small individuals. While gnathiid isopods are generally considered micropredators, taking small amounts of blood from their significantly larger host (Lafferty and Kuris 2002), they are capable of consuming up to 85 % of the total blood volume of a settlement-stage P. amboinensis (Grutter et al. 2011). Indeed, infection by a single gnathiid isopod decreases the oxygen consumption and swimming performance of a recruit damselfish and reduces the likelihood of successful settlement on the reef (Grutter et al. 2011). If an infected fish has a gnathiid removed before it feeds on its blood, this could provide a major benefit for survival. Since observed cleaning interactions were brief (~ 1 s), it is unlikely that they obtained other direct benefit of cleaning, such as tactile stimulation which leads to a reduction in client stress (cortisol) levels (Soares et al. 2011). As direct cleaning interactions between cleaners and recently-settled recruits were relatively uncommon, it is possible that the benefits of preferentially selecting a microhabitat near cleaners may be at least partly indirect. For example, cleaner wrasse may reduce infection rate of parasites in the vicinity of a cleaning station. As each L. dimidiatus eats around 1200 gnathiids per day (Grutter 1996a), this could significantly reduce the number of ectoparasites in the immediate vicinity of a L. dimidiatus and hence lower the infection rate onto nearby settlers (Gorlick et al. 1987; Grutter 1999a; Clague et al. 2011). Other indirect benefits of the presence of cleaner wrasses may also exist for settling fish. ‘Safe havens’ from predators may be created by L. dimidiatus (Cheney et al. 2008) in that, in the presence of cleaners, some predators behave with notably reduced aggression. Settling reef fish may therefore prefer to settle in a habitat with L. dimidiatus present, due to reduced risk of predator aggression. Alternatively, or in addition, selection by young damselfish of microhabitats near cleaners may simply be an investment where the return occurs only when recruits have grown significantly. On reefs lacking cleaner wrasse, the size at age of the lemon damselfish, P. moluccensis, is smaller in older individuals (> 1 year old) and older fish have more parasitic copepods than fish on reefs where cleaners are present (Clague et al. 2011). Additionally, size frequency distributions of P. amboinensis and P. moluccensis are skewed towards smaller individuals on reefs without cleaner wrasse (Waldie et al. 2011). Ultimately, this might affect fecundity, and hence the supply of future settlers, since larger body sizes are correlated with egg production (Green 2008). In summary, our study provides evidence that settlement-stage reef fish may use the presence of cleaner wrasse as a cue when deciding what microhabitat to settle on. As the presence of this mutualistic interaction creates favourable environments, this may ultimately enhance the

survival and abundance of these species. This study demonstrates that mutualism contributes to the already-complex settlement processes of reef fishes.

Acknowledgements We thank F. Cortesi, M. De Brauwer, F. R. A. Jaine, J. E. Lange, A. Marshell, G. A. C. Phillips, and M. D. Mitchell for their help in the field; J. O. Hanson for statistical advice and R. Q-Y. Yong and S. C. Cutmore for comments on earlier drafts of the manuscript. We are particularly grateful for the support of the staff of the Lizard Island Research Station. This work was funded by a Sea World Research and Rescue Foundation, Australia (SWRRFI) awarded to D. Sun, A. S. Grutter and T. H. Cribb and funds from the Australian Research Council awarded to A. S. Grutter and M. G. Meekan and from The University of Queensland awarded to A. S. Grutter.

Chapter 5: Diel patterns of infection by gnathiid isopods on the ambon damsel, Pomacentrus amboinensis, and the influence of the presence of cleaner wrasse

Derek Sun, Karen L. Cheney, Johanna Werminghausen, Thomas H. Cribb & Alexandra S. Grutter

Abstract That the presence of bluestreak cleaner wrasse, Labroides dimidiatus, affects fish abundance and fish traits is well known, yet little is known of the causal mechanism (s) involved. Since juvenile gnathiid isopods are the main ectoparasite eaten by cleaner wrasse, a difference in their infection number as a consequence of cleaner wrasse presence could be one such causal mechanism. To test this hypothesis we used ambon damselfish, Pomacentrus amboinensis as a model host to estimate the cumulative number of gnathiid isopods on patch reefs where the presence and absence of cleaner wrasse had been experimentally manipulated for 12 years. Gnathiids were collected using sentinel traps containing three live juvenile P. amboinensis as hosts; traps were sampled after 12 h at dawn and dusk. There was no significant difference in the estimated cumulative number of gnathiids per trap between reefs with and without L. dimidiatus. The probability of capturing one or more gnathiid per trap on reefs without L. dimidiatus was 0.167 (95% CI = 0.112–0.240) compared with 0.148 (95% CI = 0.096–0.221) on control reefs. However, the probability of capturing gnathiids in traps was higher during the night [0.222 (95% CI = 0.144–0.327)], than during the day [0.088 (95% CI = 0.061 – 0.126)]; or a 2.934 (95% CI = 1.725–4.985) times higher probability during the night than during the day. We conclude that the rate of gnathiid isopod attraction/detection measured here does not explain the patterns observed involving the effect of the presence of L. dimidiatus on this fish reported in other studies.

Introduction Parasites are often difficult to observe, yet many studies have shown that they may play a vital role in influencing ecosystem functions (Hudson et al. 2006), regulate host populations (Scott and Dobson 1989), and influence overall biodiversity (Hudson et al. 2006). On coral reefs, fish parasites can have negative effects on their hosts (Ferguson et al. 2011; Binning et al. 2013; Krkošek et al.

2013). Understanding how fish manage parasite infections is thus critical to understanding their ecology. Cleaning behaviour, which involves organisms that remove ectoparasites from cooperating clients, may be one way that reduces parasitic loads. Indeed, the frequent nature of cleaning undergone by individual client fish, as often as every ~ 5 min for some species (Grutter 1995a), suggests that the removal of ectoparasites is important to them. On Indo-Pacific coral reefs, L. dimidiatus is the most common of the five Labroides species (Randall et al. 1997), all of which are obligate cleaner fish that remove ectoparasites and sometimes tissue from the surfaces of other fishes, known as clients. The presence of the cleaner wrasse L. dimidiatus greatly influences the richness and abundance of adult (Bshary 2003; Grutter et al. 2003), juvenile (Waldie et al. 2011) and recruiting (Sun et al. in review, Chapter 3) reef fish. After 4 ̶ 20 months, in the Red Sea, the diversity of reef fish on patch reefs was reduced when cleaner wrasse disappeared or were removed (Bshary 2003). At Lizard Island, Great Barrier Reef, the species richness and abundance of visitors (fish species that move between patch reefs) were reduced 18 months after cleaner wrasse had been removed (Grutter et al. 2003); after 8.5 years, there was a reduction in the abundance and diversity of both resident and visitor as well as adult herbivorous (Acanthuridae) fishes (Waldie et al. 2011). Only recently, it was found that the abundance of damselfish recruits (species that are site-attached), specifically Chrysiptera rollandi and Pomacentrus amboinensis, were reduced on these same patch reefs after 12 years of regular removal of cleaner wrasse (Sun et al. in review, Chapter 3). These studies highlight the importance of understanding the mechanism (s) via which a relatively uncommon and small fish maintains the diversity and abundance of so many coral reef fishes. An obvious way by which fish may benefit from the presence of cleaner fish is by obtaining health benefits from a reduction in ectoparasites, which ultimately leads to changes in individuals or communities. How ectoparasites can be influenced by cleaner fish will be determined by the ecology and behaviour of the parasite and particularly the level of association between the parasite and the host. There are two main types of ectoparasites, those that remain on the host for almost their entire lives, i.e. ‘permanent’ parasites such as copepods (Finley and Forrester 2003) and cymothoid isopods (Adlard and Lester 1994; Roche et al. 2013a), and those that remain on the host only briefly while feeding, i.e. ‘temporary’ parasites such as gnathiid (Smit et al. 2003; Ota et al. 2012) and corallanid isopods (Grutter and Lester 2002), which are often classified as mobile micropredators (Lafferty and Kuris 2002). Surprisingly, only a few studies have tested the effect of L. dimidiatus on parasite loads. The removal of L. dimidiatus did not have an effect on number of copepods per fish after 6 months on a damselfish P. moluccensis at Lizard Island (Grutter 1996b). After 2 y the number of copepods

on P. vaiuli in the Marshall Islands was also not affected but the number of larger copepods (and hence overall biomass) was higher (Gorlick et al. 1987). At Lizard Island, Clague et al. (2011) showed that copepod number, summed across 10 fish sampled per reef, was higher on P. moluccensis on reefs without a cleaner wrasse for 8.5 y, but only on larger P. moluccensis. While copepods are not often consumed by cleaner wrasse (Grutter 1997) and individual damselfish are cleaned relatively infrequently (Grutter 1995a), that larger P. moluccensis individuals are preferentially cleaned (Clague et al. 2011; Sun et al. in review, Chapter 4) could explain why an effect on copepod number was only apparent in larger individuals. The presence of cleaners also has an effect on parasitic isopods. For caged thicklip wrasse, Hemigymnus melapterus, gnathiid number per fish was higher on reefs without L. dimidiatus after 24 h and 12 d (Grutter 1999a), while on the same removal reefs the prevalence and number of corallanid isopods was higher after 24 h, while a noticeable shift towards smaller isopods was detected after 12 d (Grutter and Lester 2002). In a very different system involving cleaner gobies (Elacatinus spp.) in the Caribbean, gnathiid number for Stegastes diencaeus damselfish captured using handnets was lower on individuals that had cleaning stations within their territory than on individuals that lived further away from a cleaning station (Cheney and Côté 2001). Gnathiids can inhibit the growth (Jones and Grutter 2008), swimming performance and oxygen consumption (Grutter et al. 2011) and decrease the survival rate (Grutter et al. 2008; Grutter et al. 2011) of damselfish juveniles. High numbers of gnathiids can cause anaemia (Paperna 1977; Grutter 2008) and mortality in adult fish (Mugridge and Stallybrass 1983; Marino et al. 2004). Thus, a reduction in gnathiids could ultimately lead to the higher abundance of damselfishes observed on reefs where cleaner wrasse are present (Bshary 2003; Waldie et al. 2011; Sun et al. in review, Chapter 3). The size frequency distribution of both P. amboinensis and P. moluccensis was shifted towards smaller individuals on reefs where cleaner wrasse have been removed for 8.5 years (Waldie et al. 2011). The shift in size distribution could be due to decreased rate of growth in individuals on reefs without a cleaner wrasse. Although P. moluccensis had a lower growth rate and higher copepod numbers on reefs without cleaners, this was only apparent in larger individuals (Clague et al. 2011). Thus, it is possible that other ectoparasites such as gnathiid isopods may explain the overall shift in size distribution seen between reefs with and without cleaner wrasse. As L. dimidiatus can consume an average of 1200 ectoparasites a day with the majority of their diet being made up of gnathiid isopods (99.7 %) (Grutter 1996a), it is possible that their effect on parasites is likely to be the greatest on this group. Gnathiid isopods are one of the main ectoparasites that infect coral reef fishes (Grutter et al. 2000). Unlike copepods, gnathiids are mobile temporary parasites, emerging either during the day or at night to locate a host to feed

(Sikkel et al. 2006). Gnathiids usually detach from a host once they are fully fed (Grutter 2003) or are disturbed (Grutter 1995b). Due to this unique life cycle and behaviour, several measures have been used to quantify the ‘abundance’ of gnathiids. These terms are defined as density of cumulative number of emerging gnathiids (i.e. number per reef area/per time), number (number per fish) and infestation (number present per unit time). Therefore, the sampling method used to quantify gnathiid isopod ‘abundance’ on fish could also affect the estimate; for example due to differences in their emergence from the benthos (Grutter 1999b; Sikkel et al. 2006), the time of day that a fish are sampled will affect how many parasites are counted. Damselfishes also tend to have few gnathiids (Grutter 1994; Sun et al. 2012) and so the odds of sampling them on fish are low. A method that increases the time they are sampled on fish would increase these odds. The use of caged fish in previous studies to quantify the number or infestation of gnathiids (Grutter 1999a; Sikkel et al. 2006; Coile and Sikkel 2013) does not exclude the potential direct interaction between cleaner and the caged fish and hence the removal of gnathiids. The use of cages also does not prevent the potential loss of gnathiids once they are fed and have dropped off. Therefore, it is possible that the initial number of gnathiids in previous studies was similar on caged fish placed on reefs either with or without cleaner wrasse and that a lower infection was recorded simply due to the direct cleaning interaction between cleaners and the caged fish. As gnathiids can easily detach themselves from a host, conventional methods of capturing fish (handnets) to examine the number of gnathiids are not ideal (Grutter 1995b). Such detachment could have contributed to the low prevalence rate and number of gnathiids that has been reported on damselfishes (Grutter and Poulin 1998; Cheney and Côté 2001; Grutter et al. 2011; Sun et al. 2012). The high emergence rates of gnathiids (41.7 ± 6.9 m-2 day-1) (Grutter 2008) means it is likely that hosts are put under constant physical stress, as well-fed gnathiids are constantly replaced by unfed individuals. Although previous studies have shown an effect of cleaner fish presence on the number of gnathiids on a fish (Grutter 1999a; Cheney and Côté 2001), the results are limited in that they only provide information of the number of gnathiids on a host at a given time, when the host was captured. As a result, these studies may not provide a true representation of overall number in relation to cleaner fish presence. Therefore, an alternative sampling technique is required: one which prevents cleaner wrasse from interacting with the host fish, retains gnathiids once they have dropped off, and provides as estimate of the cumulative number of gnathiids on fish over time. One such method involves the use of ‘sentinel traps’ which hold ‘bait’ fish and allow gnathiids, but not cleaner fish, to enter and feed off hosts, but which do not allow gnathiids to escape (Sikkel et al. 2011). Despite damselfishes being relatively sedentary on coral reefs and unable to move between reefs to seek a cleaner wrasse, their abundance is higher in the presence of a cleaner wrasse (Sun et

al. in review, Chapter 3). Yet, the mechanism(s) remains unclear. Currently, it is not known if the cumulative number of gnathiids on damselfishes differ according to cleaner wrasse presence. As cleaner wrasse can consume large quantities of gnathiid isopods, we therefore predict that the cumulative number of gnathiids will be lower on reefs where cleaner wrasse are present. In this study, we examined whether the presence of cleaner wrasse has an effect on the cumulative number of gnathiid isopods per trap using relatively large juvenile ambon damsel, P. amboinensis. Pomacentrus amboinensis were placed in sentinel traps and placed on experimental patch reefs with or without cleaner wrasse. Because cleaner fish are only active during the day (Grutter 1996a) and thus could potentially affect only ‘diurnal gnathiids’, sampling was divided between the day and the night, over a 24 h time period.

Materials and methods Study site This study was conducted at Lizard Island (1440S, 14528E) in the northern GBR, Australia during late summer (January and February) of 2013. We used 16 isolated patch reefs (3−7 m depth) located at two sites: the Lagoon (11 reefs) and at Casuarina Beach (5 reefs) (Grutter et al. 2003; Clague et al. 2011). Reefs were randomly allocated as either removal or control treatment reefs in September 2000. Subsequently, reefs were checked at several-month intervals and were inspected for cleaner wrasse L. dimidiatus; any new recruits on removal reefs were removed using barrier and handnets. Control reefs were surveyed periodically for L. dimidiatus by swimming around the reef several times and the abundances of cleaners recorded.

Study species The ambon damselfish, Pomacentrus amboinensis, is a common coral reef fish species on the GBR. As much is known about its ecology, including habitat preference, recruitment, and feeding habits (Kerrigan 1996; McCormick 2006; McCormick et al. 2010), this species has been extensively used as a model species in both field and laboratory experiments and it is abundant on all of the experimental reefs.

Sampling design A total of 72 large juvenile P. amboinensis (mean TL per fish ± SE = 50.01 ± 0.38 mm) were collected at Lizard Island in January 2013 using barrier and handnets. Fish were initially maintained in aerated 40 L tanks, then, due to their known tendency for aggressive behaviour towards each other, were transferred to individual clear plastic holding aquaria (29 × 17 × 12 cm) with a constant

flow of unfiltered seawater and maintained on a diet of commercial fish flakes. Fish were then randomly divided into four groups (A, B, C, and D), each group consisting of 18 fish. Within each group, the fish were then further divided into 6 labelled subgroups (numbered from 1–24) with 3 fish in each subgroup. These three fish were used as the sampling unit per fish trap. Gnathiids mainly locate their hosts using olfactory cues (Sikkel et al. 2011) so three fish per trap were used in order to maximise the amount of chemical cue emitted from the trap. A total of 14 individually labelled sentinel traps were constructed using a T-junction PVC pipe (370 mm long × 150 mm diameter) with a threaded lid attached to each of the three ends (Figure 1) (see Sikkel et al. 2011 for a base design). A 100 mm diameter hole was cut out of each lid and a clear plastic funnel (10 mm diameter opening) attached to the inside surface of the lid, using plastic cable ties and aquarium grade silicone glue. To facilitate filling and emptying the water in the tubes, three 32 mm diameter holes were drilled into the sides of the tube. The drainage holes were covered with 100 μm plankton mesh glued inside the tube; this allowed only water (no gnathiids) to circulate in and out of the trap. The six openings allowed sufficient water circulation to maintain adequate oxygen levels. Twelve traps in a paired design, with 6 traps placed each time on a removal and control reef, were used per day. This was repeated on all 16 experimental reefs; due to the odd number of reefs at each of the two sites sampled, this meant only one reef was sampled on one day at each of the sites, resulting in 9-day sampling periods. For each sampling day, two fish groups (A and B, or C and D) were used. The traps were placed on the reef at dusk, sampled (traps emptied and redeployed with same fish) at dawn the following day (~ 12 h after initial placement) and the second and final sample collected at dusk later that day (~ 12 h after first sampling). The sampling design allowed each group to be used a total of 4 different times throughout the experiment, with each group being re-used every 48 h. This provided each group with a recovery period of at least 24 h between deployments. To determine whether gnathiids were attracted to traps and not the fish in them, we deployed empty traps as a control. Empty trap sampling was conducted with 6 traps placed on a control and 6 on a removal reef, one day before the commencement (day 0) and the day after the conclusion (day 10) of the 9-day experiment. As an additional control, every second day throughout the experiment, an extra empty trap was placed alongside the six traps containing fish. The entire experiment was replicated after a one-week recovery period, with the same groups of fish, in the same order, for the same duration (Table 1).

Table 1 Number of traps (with and with no fish) sampled per sampling day, site, reef pair cleaner presence treatment, and day and night. Total n per host = 382 traps.

Experiment Sampling Site Reef pair cleaner Number of traps with Number of traps with round day treatment fish no fish (Control/Removal) (Control/Removal) Day Day Day Night Casuarina One 0 Control/Removal - - 6/6 6/6 Beach Casuarina One 1 Control/Removal 6/6 6/6 0/1 0/1 Beach Casuarina One 2 Control/Removal 6/6 6/6 1/1 1/1 Beach Casuarina One 3 Control 6/6 6/6 0/1 0/1 Beach One 4 Lagoon Control/Removal 6/6 6/6 1/1 1/1 One 5 Lagoon Control/Removal 6/6 6/6 0/1 0/1 One 6 Lagoon Control/Removal 6/6 6/6 1/1 1/1 One 7 Lagoon Control/Removal 6/6 6/6 0/1 0/1 One 8 Lagoon Control/Removal 6/6 6/6 1/1 1/1 One 9 Lagoon Control 6/6 6/6 0/1 0/1 One 10 Lagoon Control/Removal - - 6/6 6/6 Casuarina Two 0 Control/Removal - - 6/6 6/6 Beach Casuarina Two 1 Control/Removal 6/6 6/6 0/1 0/1 Beach Casuarina Two 2 Control/Removal 6/6 6/6 1/1 1/1 Beach Casuarina Two 3 Control 6/6 6/6 0/1 0/1 Beach Two 4 Lagoon Control/Removal 6/6 6/6 1/1 1/1 Two 5 Lagoon Control/Removal 6/6 6/6 0/1 0/1 Two 6 Lagoon Control/removal 6/6 6/6 1/1 1/1 Two 7 Lagoon Control/Removal 6/6 6/6 0/1 0/1 Two 8 Lagoon Control/Removal 6/6 6/6 1/1 1/1 Two 9 Lagoon Control 6/6 6/6 0/1 0/1 Two 10 Lagoon Control/Removal - - 6/6 6/6

Figure 1 Photograph of the sentinel trap used in the experiment. Traps were made from plastic PVC material. A) Each lid is fitted with a transparent plastic funnel, allowing gnathiid isopods to enter the trap while providing fish with adequate water circulation; B) Each trap is fitted with a 1.4 kg dive weight, which is secured with a dive belt at the front opening; C) Three drainage holes (3.2 cm in diameter) are drilled at the back of the traps. Photo: Maarten De Brauwer.

Sampling protocol The three fish were held individually in subcages within the trap. Subcages reduced any potential negative interaction amongst the fish within the trap and were constructed from flexible corrugated plastic pipe (100 diameter × 50 length mm) with mesh (15 mm) covering each circular end (Figure 2). Although mesh size was large enough to allow gnathiids to enter the subcage, ~10 larger holes (3 mm) were also drilled in the mesh to increase the opportunity for gnathiids to enter the subcage. The subcages had small holes bored along its length in order to increase the entry points for gnathiid isopods. Prior to transporting fish to the field site, each fish was placed individually into subcages labelled with the corresponding subgroup identification. Fish were transported to the field in a 60 L container, which was aerated with regular water changes. At the field site, each subgroup was placed in a trap. Traps were then passed to a snorkeler and were placed carefully on the reef at dusk. Traps were secured with a 1.4 kg dive weight attached to the outside of each trap using a dive belt and buckle. Traps were subsampled at dawn the following day. To ensure no trapped gnathiids were lost while moving traps off the reef to the boat, all funnel holes were blocked using a 1 cm diameter ball of household gum (Blu-tack, Thomastown, Australia). The contents of each trap were emptied and well-rinsed into correspondingly-numbered 10 L buckets. Subcages, still containing a fish, from each trap were given a 1-minute freshwater bath (3 L) in a separate bucket, and the contents then correspondingly 91

placed into the aforementioned 10 L buckets. Freshwater baths are a standard method used to remove gnathiids that may be attached on a fish (Soares et al. 2008). Subcages and fish were then returned to the traps and placed on their original locations on the reef. This procedure was repeated the next dusk but using the next pair of reefs and the next two fish groups. Twelve buckets holding the trap, rinse and bath water samples were brought back to the research station, along with (after the dusk sampling) the recently-used fish in their subcages. For the dusk samples, an airstone was placed into each bucket for 3 hours to keep the fish alive while allowing any remaining gnathiids to drop off naturally. Each fish subgroup was also given a final 1- minute freshwater bath, during which their body was gently rubbed before being placed back into their allocated aquaria. The combined water samples were filtered through a 100 μm plankton sieve into vials to remove any gnathiids, and then preserved with a 10% formaldehyde solution in seawater. During the experiment a total of six fish died (two traps, each on a control and removal reef) and were replaced with new ones.

Figure 2 Individual subcages used to hold Pomacentrus amboinensis in the sentinel traps

Statistical analysis Since 53 of the 62 samples with gnathiids present contained only one individual, we used the presence/absence of gnathiids as the response variable. A generalized linear mixed effect model (GLMER) with a binomial distribution error was used to examine whether gnathiid presence differed with cleaner wrasse presence. The terms treatment (cleaner or no cleaner), site (Lagoon or Casuarina Beach), time (day or night) and experiment replicate (Experiment one or two) were fixed factors, and reef ID, sampling day, and fish group as random factors. Fish subgroup identity was nested with fish group. The sum of the fish weight of the three fish per subgroup was included as a covariate, as gnathiids are known to use fish scent to locate a potential host (Sikkel et al. 2011) and we expected that the scent of fish would be correlated with the mass of fish.

In a separate analysis, the same model was used to estimate the rate at which gnathiids entered empty traps (no fish), but without fish weight, fish group, and fish subgroup identity as factors. To determine the best model for the data, we compared the full model with models in which one of the explanatory terms was dropped using the “drop1” function (Chambers 1992). If the maximum likelihood ratio test (LRT) found that a dropped term had no significant effect on the model by using the Chi-square distribution, then the term was dropped. All statistical analyses were performed using R version 3.0.1 (R Development Core Team 2013). As information regarding species-level taxonomy of gnathiid isopods is scarce, gnathiids encountered in this study were not identified further.

Results Diel patterns Of the 382 trap samples containing fish, 62 contained a total of 72 gnathiids, with a maximum of three gnathiids in a single trap (Table 2). The presence of gnathiid isopods differed according to time of day (LRT= 13.61, df = 1, P = 0.0002), with more gnathiids being caught at night than in the day. Traps at night had a higher probability of capturing a gnathiid 0.222 (95% CI = 0.144–0.327), than those set during the day 0.088 (95% CI = 0.061–0.126); or a 2.934 (95% CI = 1.725–4.985) times higher probability during the night than during the day.

Cleaner wrasse presence The simplified model showed that the presence/absence of gnathiid isopods did not differ according to cleaner wrasse presence (LRT = 0.216, df = 1, P = 0.642), site (LRT =0.050, df = 1, P = 0.822), fish weight (LRT = 1.185, df = 1, P = 0.276) or experiment replicate (LRT = 0.299, df = 1, P = 0.583). Traps had a 0.167 (95% CI = 0.112–0.240) probability of capturing a gnathiid on removal reefs, compared with 0.148 probability (95% CI = 0.096–0.221) on control reefs, or a 1.153 (95% CI = 0.731–1.818) times higher probability on removal than control reefs. Of the 128 trap samples containing no fish, 11 contained a sum of 12 gnathiids, with a maximum of two gnathiids per trap (Table 2). The presence/absence of gnathiids did not differ according to cleaner wrasse presence (LRT = 0.136, df = 1, P = 0.711), time of day (LRT = 0.239, df = 1, P = 0.624), site (LRT = 2.43, df = 1, P = 0.118) and experiment replicate (LRT = 1.22, df = 1, P = 0.268).

Table 2 Number of sentinel traps with gnathiids, day and night trap samples and the number of gnathiids per trap collected from the sentinel trap experiment on Lizard Island.

Number of traps with Number of fed/not # Gnathiid isopods per trap isopods (day/night) fed gnathiids

Total Traps with Treatment Day Night Fed Not fed 0 1 2 3 number of gnathiids traps

Fish traps

Control 33 11 22 30 12 181 25 7 1 382 Removal 29 7 22 22 8 139 28 1 0

No fish traps Control 6 3 3 2 4 58 6 0 0 128 Removal 5 3 2 2 4 59 4 1 0

Discussion Our study found that the cumulative number of gnathiid isopods per trap using the three ambon damsel did not differ between patch reefs that have cleaner wrasse present and those from which cleaners had been experimentally removed over a 12-year period. There was a clear difference in the number of gnathiids according to day and night. The results from this study suggest that factors related to cleaner wrasse presence other than gnathiids alone in this system may explain the benefits that occur to damselfishes.

Diel patterns Although the number of gnathiids did not differ with cleaner wrasse presence or absence, it did differ according to day or night. There was a 2.9 times higher probability that traps sampled at night would capture a gnathiid than traps that where sampled during the day. The diel variation seen in this study is consistent with findings using other sampling methods, which have also reported that the number of gnathiids on caged fish is higher at dawn than at dusk (Grutter and Hendrikz 1999; Sikkel et al. 2006). Although, gnathiids clearly can emerge and infect fish during the day, this is usually at a much lower rate than during the night, with peak activity being recorded at night and at pre-dawn periods (Grutter and Hendrikz 1999; Sikkel et al. 2006). The higher abundance of gnathiid isopods active at night may have arisen as a result of selective pressures from diurnal parasite predators, principally cleaner fish, that feed on them during the day (Grutter 2002). As different ontogenetic stages (Grutter et al. 2000; Sikkel et al. 2006) and species (Jones and Grutter 2007) of gnathiids are known to emerge differentially during either the day or at night, it is possible that some species have evolved to emerge at night when there are fewer diurnal planktivores, cleaners and microinvertebrate feeders (Helfman 1986; Grutter 1996a), therefore, reducing the risk of predation. Although predation of gnathiid isopods can still occur from nocturnal cleaning activities as observed in the cleaner shrimp, Urocaridella sp. on coral reefs (Bonaldo et al. 2014) and on monogeneans by Lysmata amboinensis in tanks (Militz and Hutson 2015), gut contents analyses reveal that cleaner shrimps consume few gnathiids (Becker and Grutter 2004). Therefore, it is unlikely that cleaning activities by shrimps will have a major impact on the abundance of gnathiids, compared with the cleaning activities of cleaner fish.

Cleaner wrasse presence We found that there was no difference in the probability that a gnathiid was present in a trap on reefs with (0.148) and without cleaner wrasse (0.167). As gnathiid presence did not differ according to cleaner wrasse presence, it may therefore not explain some of the positive effects observed on fish from reefs with cleaner wrasse present. For example, the smaller size distribution

(Waldie et al. 2011) of P. amboinensis and P. moluccensis and reduced growth rate (Clague et al. 2011) reported in earlier studies which used the same experimental reefs may not be due to differing levels of gnathiids on these fishes. Indeed Clague et al. (2011) showed that the number of copepods was higher on larger P. moluccensis from reefs that have had cleaners removed than on reefs with cleaners and suggested this could explain their reduced growth rate. There is evidence that infection by copepods can affect the growth, gonad mass and survival of fish (Finley and Forrester 2003; Palacios-Fuentes et al. 2012). Other ectoparasites such as cymothoid isopods have also been shown to affect growth and survival in reef fishes (Adlard and Lester 1994; Fogelman and Grutter 2008). It is thus possible that other ectoparasites such as copepods and cymothoid isopods, in isolation or even in synergy with gnathiids, may cause this downward shift in size. However, we did not find any cymothoids in the traps. Similarly, the recent study that showed that the abundance of recruiting damselfishes was higher on reefs where cleaner wrasse are present (Sun et al. in review, Chapter 3) is thus unlikely due to higher gnathiid numbers. Both the prevalence and number of gnathiid isopods was low, with only 16 % of traps over a 24 h time period having at least one gnathiid, with the number of gnathiids ranging only from one to three gnathiids per trap. This low prevalence and number of gnathiids is consistent with previous studies that demonstrated that small juvenile P. amboinensis (mean TL ± SE = 14.28 ± 0.26) are rarely infected with gnathiids, with instantaneous prevalence being 3 and 3.5 % and never more than one gnathiid per infected fish (Grutter et al. 2011; Sun et al. 2012). The feeding success of a micropredator should increase as host mass increases, as the likelihood of being detected and thus potentially being preyed upon decreases (Kuris and Lafferty 2000). This could explain the lower number of gnathiids we recorded infecting the relatively small-sized P. amboinensis. Due to the low prevalence of gnathiids infecting these fish, it is unlikely that gnathiids are the likely mechanisms that have caused the higher abundance of damselfish recruits on reefs with cleaner wrasse than on reefs without. Therefore, it is likely that other factors related to cleaner wrasse presence could explain those results. For example, decreased aggression from predators when in the presence of cleaner wrasse (Cheney et al. 2008) or reduced stress via tactile stimulation provided by cleaner wrasse (Soares et al. 2011). It is also possible that we could not detect a difference due to cleaner wrasse presence with the sentinel traps due to the sampling method. The amount of sampling time (24 h) may have been too limited to detect slight differences in gnathiid abundance. Convincingly, further sampling would detect a small yet significant difference in gnathiid abundance between cleaned and non-cleaned reefs, such a small difference might still be important. We emphasize that although small fish may only be attacked by a few gnathiids over its life time this might be enough to reduce their growth and potentially affect their survival.

A speculative scenario that agrees with our results as well as the observed beneficial effects on fishes, is that the advantage to fish for being on reefs with L. dimidiatus could be due to the length of infection (feeding period) being reduced, rather than the number of gnathiids attacking a fish. This could occur if cleaners disturb gnathiids and reduced the time they spend (and the blood they remove) on fish. Previous studies have shown that the feeding rate (time spent ingesting blood) of gnathiid isopods can vary among individuals, with gnathiids remaining on a host for as little as 30 min or up to 10 h in order to feed (Grutter 2003; Smit et al. 2003; Grutter et al. 2011). The efficiency of L. dimidiatus in seeking and consuming gnathiids, coupled with the frequency at which clients are cleaned (e.g. barred rabbit fish, Siganus doliatus, are cleaned on average every 5 min in a day), means that the risk of predation can be exceedingly high for some gnathiids (Grutter 1995a). Grutter (2003) showed that for third-stage larval gnathiids, the likelihood of red blood being present in the gut increased over time, with 20 % of gnathiids being engorged at 15 minutes and 46 % engorged at 30 min. Given that clients such as S. doliatus are cleaned frequently, the amount of red blood that they may lose from gnathiids should be significantly reduced on reefs with cleaner wrasse. Gnathiids may not be able to become fully engorged on red blood within the shortened time frame, due to the client being cleaned (gnathiid predation). This scenario fits the paradigm of cleaner wrasse driving changes in gnathiid feeding behaviour, ultimately limiting their ability to adversely affect their host and thus conferring benefits upon fish clients. Although studies have shown a clear size preference by L. dimidiatus for larger host individuals (Grutter 1995a; Grutter et al. 2005; Clague et al. 2011), it should be emphasised that a reduction in parasitic numbers would be beneficial to any host, regardless of how frequently it may be cleaned. A single gnathiid can consume up to 85 % of the total blood volume of a settlement- stage damselfish (Grutter et al. 2011) and might transmit blood parasites and other infectious agents (Davies et al. 2009; Curtis et al. 2013). Therefore, fish species that are cleaned less regularly (e.g. P. moluccensis, which are only cleaned on average once per 30 min) (Grutter 1995a) would still benefit from a reduction in gnathiid feeding time due to a reduced blood loss and exposure to parasite transmission. Previous studies have clearly demonstrated that there is an advantage for fish being on reefs with cleaners present (e.g. Grutter 1999a; Soares et al. 2011; Waldie et al. 2011). In this light the results of this study showing no differences in relation to cleaner wrasse presence are surprising. In consideration, these results suggest that the benefits gained from being on reefs with cleaner wrasse (such as increased recruitment), could relate to the minimisation of the infection that do occur. In summary, this study demonstrates that the relationship between cleaner wrasse and their clients is

complex and that factors other than the abundance of gnathiid isopods in the environment may explain some of the advantages to fish of being on reefs where cleaner wrasse are present.

Acknowledgements We thank E. C. McClure, M. De Brauwer, F. Cortesi, G. A. C. Phillips and A. E. Winters for their help in the field; S. P. Blomberg for statistical advice, P. C. Sikkel for advice on sentinel trap and experimental design and R. Q-Y. Yong for comments on earlier drafts of the manuscript. We are particularly grateful for the support of the staff of the Lizard Island Research Station. This work was funded by a Sea World Research and Rescue Foundation, Australia (SWRRFI) awarded to D. Sun, A. S. Grutter and T. H. Cribb and funds from the Australian Research Council awarded to A. S. Grutter.

Chapter 6: General Discussion

The Preamble to this thesis referred to the complexity of interactions between three groups of organisms – hosts, parasites and cleaners. Although a lot of work has been done on the various elements of this system prior to my study, and although substantial progress has been made here, the questions arising from the study appear to leave at least as much to be done as when I started. From the study of the diversity of digenean trematodes of adult pomacentrid fishes at Lizard Island it was established that digenean richness is low relative to that of other coral reef fish groups. Host-specificity ranges dramatically from a few taxa that are strongly oioxenic through to several that are clearly stenoxenic and others that are seemingly euryxenic. In comparison with the known fauna of Heron Island, the fauna at Lizard Island was generally comparable but with several key distinctions. Questions arising from these findings start with the problems of identification. It was a surprise to find evidence of so much cryptic richness among specimens of just one of the 19 morphospecies in the study. Hysterolecitha nahaensis clearly is deserving of and requires dedicated independent study. There is no a posteriori explanation for why this one species has radiated so dramatically. In terms of the discovery of the total richness of the system, the distribution of oioxenic species presents a challenge. Not only do we not understand why only one faustulid is present in one species of Chromis, one bivesiculid in Acanthochromis polyacanthus and Lepotrema adlardi in just one species of Abudefduf – we are left uncertain as to what richness may remain undetected in the pomacentrid species that have never been examined. In terms of beta diversity, it might have been predicted that, as found, there would be some relatively low-level differences between the fauna found at Heron and Lizard Islands. What remains completely unexplained is why the species that have limited distributions are so restricted thus. Further, the striking differences in mean richness and mean prevalence between Heron and Lizard Islands remain unexplained. For convincing explanations to emerge on such topics we will need at least to understand the life-cycles of these taxa. Presently not a single life-cycle of a pomacentrid-infecting digenean is known. This study has also highlighted the importance of the presence of cleaner wrasse in the ecology of young reef fishes. I demonstrated that the positive effects of cleaner wrasse presence on adult fish population size reported in other studies (Bshary 2003; Grutter et al. 2003; Waldie et al. 2011), begins during the recruitment period by enhancing the abundance of damselfish recruits. Thus, we now know that cleaner wrasse can influence the population dynamics from the start of the recruitment period. Although cleaner wrasse species are important key stone mutualist on coral reefs, they are collected in large numbers for the aquarium trade; 20,000 individuals have been caught in Sri Lanka in one year alone. The removal of such a key organism in large numbers should

100

have a detrimental effect on the biodiversity of reef fishes. Therefore, studies on reef fish recovery on coral reefs should consider the beneficial role of cleaner wrasse alongside that of other key organisms. However, it remains unclear why the abundance of only two species of damselfishes (Chrysiptera rollandi and Pomacentrus amboinensis) which differ from each other in regards to ecology (e.g. habitat and social structure), were affected by the presence of a cleaner wrasse, when similar species showed no differences. It is possible that factors other than the presence of cleaner wrasse dominate when it comes to microhabitat selection during their recruitment. Despite the overall higher abundance of recruiting damselfishes on reefs where cleaner wrasse are present, it is not understood how these fish are selecting these particular reefs. As cleaner wrasse are inactive and shelter within the reef matrix at night when the majority of settlement-stage fish settle, it is possible that they are selecting reefs based on olfactory in addition to visual cues. This hypothesis is yet to be tested, however, doing so will provide greater insight into the dynamics that occur during the recruitment period. The results from chapter 4 showed that when given the option, a young damselfish would preferentially select a microhabitat that is adjacent to a cleaner wrasse. This finding indicates that cleaner wrasse may influence small-scale post-settlement migration. Field observations indicated that small recruits are rarely cleaned by juvenile cleaner wrasse. It was therefore unsurprising that the damselfish that were preferentially cleaned by juvenile L. dimidiatus were larger. It is thus likely that indirect rather than direct benefits are the driver for such small recruits seeking proximity to a cleaner. As cleaner wrasse are known to consume large quantities of parasites per day such as gnathiids, a reduction in gnathiid abundance within cleaning stations of juvenile cleaner wrasses could be one possible benefit. However, it is not yet known if the abundance of gnathiids are actually reduced within cleaning stations. Thus, further studies are required to explore this scenario. It is also not clear what other indirect benefits a recruit may obtain from being in a microhabitat that is near a cleaner wrasse. If gnathiid infestation is lower in the vicinity of a cleaner wrasse, we should expect both growth and survival to increase, thus possibly explaining the results found in this study. However, alternatively, or in addition, selection by young damselfish of microhabitats near cleaners may simply be an investment where the return only occurs when recruits have grown significantly. Given that past studies have shown reduction in gnathiid infestation on hosts and that cleaner wrasse are known to consume enormous numbers of gnathiids when cleaner wrasse are present, it was somewhat surprising to find no effect of cleaner wrasse presence in relation to the abundance of gnathiids caught in sentinel traps. As this study only examined the infection of one species of damselfishes, future studies should examine whether the infection of gnathiids is consistently the same or different for other species of damselfishes. It is possible that different

101

species of damselfishes may be infected with more gnathiids. However, the results shown in this study may suggest that the benefits of fish being on reefs with cleaner wrasse are not significantly from a reduction in gnathiid isopod load but due to exposure of gnathiid infection being minimised by cleaner wrasse presence. Further examination is required to test this idea. If supported, the implications could be an intriguing “minimal effect/win/win” outcome whereby the population of gnathiids is not materially affected by the presence of cleaners, the cleaners benefit from their presence by the provision of most of their food, and client fish benefit from the presence of the cleaners by the reduction of the impact of gnathiid parasitism. Could there be further complexities such as cleaners preferentially eating engorged gnathiids? Only further studies will tell.

102

References

Adam TC (2010) High-quality habitat and facilitation ameliorate competitive effects of prior residents on new settlers. Oecologia 166:121-130 Adamson ML, Caira JN (1994) Evolutionary factors influencing the nature of parasite specificity. Parasitology 109:S85-S95 Adlard RD, Lester RJG (1994) Dynamics of the interaction between the parasitic isopod, Anilocra pomacentri, and the coral reef fish, Chromis nitida. Parasitology 109:311-324 Aiken HM, Bott NJ, Mladineo I, Montero FE, Nowak BF, Hayward CJ (2007) Molecular evidence for cosmopolitan distribution of platyhelminth parasites of tunas (Thunnus spp.). Fish Fish 8:167-180 Allen GR (2008) Conservation hotspots of biodiversity and endemism for Indo-Pacific coral reef fishes. Aquat Conserv 18:541-556 Almany GR (2003) Priority effects in coral reef fish communities. Ecology 84:1920-1935 Almany GR, Webster MS (2006) The predation gauntlet: early post-settlement mortality in reef fishes. Coral Reefs 25:19-22 Anderson GR, Barker SC (1998) Inference of phylogeny and taxonomy within the Didymozoidae (Digenea) from the second internal transcribed spacer (ITS2) of ribosomal DNA. Syst Parasitol 41:87-94 Anderson MJ, Gorley RN, Clarke KR (2008) PERMANOVA+ for Primer: Guide to software and statistical methods, Plymouth Arnal C, Côté IM (2000) Diet of broadstrip cleaning gobies on a Barbadian reef. J Fish Biol 57:1075-1082 Arnal C, Morand S (2001) Importance of ectoparasites and mucus in cleaning interactions in the Mediterranean cleaner wrasse Symphodus melanocercus. Mar Biol 138:777-784 Barber I, Hoare D, Krause J (2000) Effects of parasites on fish behaviour: a review and evolutionary perspective. Rev Fish Biol Fish 10:131-165 Barker SC, Bray RA, Cribb TH (1993) Preptetos cannoni n. sp. (Digenea: Lepocreadiidae) from Siganus lineauts (Teleostei: Siganidae) from the southern Great Barrier Reef, Australia. Syst Parasitol 26:151-155 Barker SC, Cribb TH, Bray RA, Adlard RD (1994) Host-parasite associations on a coral reef: Pomacentrid fishes and digenean trematodes. Int J Parasitol 24:643-647 Barnett LJ, Miller TL, Cribb TH (2014) A review of the currently recognised zoogonid cercariae, including the identification and emergence ecology of Cercaria capricornia XI (Digenea: Zoogonidae) from Nassarius olivaceus (Gastropoda: Nassariidae) in Central Queensland, Australia. Folia Parasitol (Praha) 61:322-330 Bartoli P, Gibson DI (2007) Synopsis of the life cycles of Digenea (Platyhelminthes) from lagoons of the northern coast of the western Mediterranean. J Nat Hist 41:1553-1570 Bartoli P, Gibson DI, Bray RA (2005) Digenean species diversity in teleost fish from a nature reserve off Corsica, France (Western Medterranean), and a comparison with other Mediterranean regions. J Nat Hist 39:47-70 Bartoli P, Bray RA, Gibson DI (2006) Four closely related but forgotten species of Rhipidocotyle Diesing, 1858 (Digenea : Bucephalidae) in fishes from European seas. Syst Parasitol 65:129-149 Bascompte J, Jordano P (2007) Plant-animal mutualistic networks: The architecture of biodiversity. Annu Rev Ecol Evol Syst 38:567-593 Bay LK, Jones GP, McCormick MI (2001) Habitat selection and aggression as determinants of spatial segregation among damselfish on a coral reef. Coral Reefs 20:289-298 Beasley VR, Faeh SA, Wikoff B, Eisold J, Nichols D, Cole R, Schotthoefer AM, Staehle C, Greenwell M, Brown LE (2005) Risk factors and the decline of the northern cricket frog, Acris crepitans: evidence for involvement of herbicides, parasitism, and habitat modifcations. In: Lanoo MJ (ed) Amphibian declines: the conservation status of United States species. University of California Press, Berkeley, CA, Becker JHA, Grutter AS (2004) Cleaner shrimp do clean. Coral Reefs 23:515-520 Becker JHA, Curtis LM, Grutter AS (2005) Cleaner shrimp use a rocking dance to advertise cleaning service to clients. Curr Biol 15:760-764 Biani NB, Mueller UG, Wcisio WT (2009) Cleaner mites: Sanitary mutualism in the miniature ecosystem of neotropical bee nests. The American Naturalist 173:841-847 Binning SA, Roche DG, Layton C (2013) Ectoparasites increase swimming costs in a coral reef fish. Biol Lett 9:20120927 Blair D, Bray RA, Barker SC (1998) Molecules and morphology in phylogenetic studies of the Hemiuroidea (Digenea: Trematoda: Platyhelminthes). Mol Phylogenet Evol 9:15-25 Bonaldo RM, Grutter AS, Sazima I, Krajewski JP (2014) 24/7 service: nocturnal cleaning in a tropical Indo-Pacific reef. Mar Biodivers:1-2 Bond WJ (1994) Keystone species. Springer Berlin Heidelberg Bonin MC, Almany GR, Jones GP (2011) Contrasting effects of habitat loss and fragmentation on coral-associated reef fishes. Ecology 92:1503-1512

103

Bosiger YJ, McCormick MI (2014) Temporal links in daily activity patterns between coral reef predators and their prey. PLoS ONE 9:e111723 Bott NJ, Cribb TH (2009) Prosorhynchine trematodes (Digenea: Bucephalidae) from epinephelines (Perciformes: Serranidae) on the Great Barrier Reef, Australia. Syst Parasitol 72:57-69 Bott NJ, Miller TL, Cribb TH (2013) Bucephalidae (Platyhelminthes: Digenea) of Plectropomus (Serranidae: Epinephelinae) in the tropical Pacific. Parasitol Res 112:2561-2584 Bowles J, Hope M, Tiu WU, Liu X, McManus DP (1993) Nuclear and mitochondrial genetic markers highly conserved between Chinese and Philippines Schistosoma japonicum. Acta Trop 55:217-229 Bray R, Cribb T (2002) Lepocreadiinae lepocreadiids (Digenea) from tetraodontiform fishes in the waters off Tasmania, Queensland and Moorea, French Polynesia. Acta Parasitol 47:6-13 Bray R, Cribb T (2004) Species of Lecithocladium Luhe, 1901 (Digenea, Hemiuridae) from Australian marine fishes, with a description of a new species from various hosts off eastern Australia. Acta Parasitol 49:3-11 Bray RA (1982) Two new species of Bacciger Nicoll, 1914 (Digenea: Fellodistomidae) from mullet in Australia. J Nat Hist 16:23-29 Bray RA (2001) Family Tandanicolidae Johnston, 1927. In: Gibson DI, Jones A, Bray RA (eds) Keys to the Trematoda, Volume 1. CABI Publishing and the Natural History Museum, Wallingford, Bray RA (2002) Superfamily Gymnophalloidea Odhner, 1905. In: Gibson DI, Jones A, Bray RA (eds) Keys to the Trematoda Vol 1. CAB International, Wallingford, pp243-244 Bray RA (2008) Family Faustulidae Poche, 1926. In: Bray RA, Gibson DI, Jones A (eds) Keys to the Trematoda Vol 3. CABI Publishing and the Natural History Museum, Wallingford, pp509-522 Bray RA, MacKenzie K (1990) Aponurus laguncula Looss, 1907 (Digenea: Lecithasteridae): a report from herring, Clupea harengus L., in the easter English Channel and a review of its biology. Syst Parasitol 17:115-124 Bray RA, Cribb TH (1996) Two Lepotrema Ozaki, 1932 species (Digenea: Lepocreadiidae) from marine fishes from the southern Great Barrier Reef, Queensland, Australia. Syst Parasitol 35:111-117 Bray RA, Cribb TH (1998) Lepocreadiidae (Digenea) of Australian coastal fishes: new species of Opechona Loos, 1907, Lepotrema Ozaki, 1932 and Bianium Stunkard, 1930 and comments on other species reported for the first time or poorly known in Australian waters. Syst Parasitol 41:123-148 Bray RA, Cribb TH, Barker SC (1993a) The Hemiuroidea (Digenea) of pomacentrid fishes (Perciformes) from Heron Island, Queensland, Australia. Syst Parasitol 24:159-184 Bray RA, Cribb TH, Barker SC (1993b) The Lepocreadiidae (Digenea) of pomacentrid fishes (Perciformes) from Heron Island, Queensland, Australia. Systematic Parasitology 26:189-200 Bray RA, Cribb TH, Barker SC (1994) Fellodistomidae and Lepocreadiidae (Platyhelminthes: Digenea) from chaetodontid fishes (Perciformes) from Heron Island, southern Great Barrier Reef, Queensland, Australia. Invertebr Taxon 8:545-581 Bray RA, Littlewood DTJ, Herniou EA, Williams B, Henderson RE (1999) Digenean parasites of deep-sea teleosts: a review and case studies of intrageneric phylogenies. Parasitology 119:S125-S144 Bray RA, Waeschenbach A, Cribb TH, Weedall GD, Dyal P, Littlewood DTJ (2009) The phylogeny of the Lepocreadioidea (Platyhelminthes, Digenea) inferred from nuclear and mitochondrial genes: Implications for their systematics and evolution. Acta Parasitol 54:310-329 Bronstein JL (1994) Conditional outcomes in mutualistic interactions. Trends Ecol Evol 9:214-217 Bronstein JL, Dieckmann U, Ferrière R (2004) Evolutionary Conservation Biology. Cambridge University Press, Cambridge Brown BL, Creed RP, Skelton J, Rollins MA, Farrell KJ (2012) The fine line between mutualism and parasitism: complex effects in a cleaning symbiosis demonstrated by mutliple field experiments. Oecologia 170:199-207 Brown SP, Renaud F, Guegan JF, Thomas F (2001) Evolution of trophic transmission in parasites: the need to reach a mating place? J Evol Biol 14:815-820 Bshary R (2003) The cleaner wrasse, Labroides dimidiatus, is a key organism for reef fish diversity at Ras Mohammed National Park, Egypt. J Anim Ecol 72:169-176 Bshary R, Würth M (2001) Cleaner fish Labroides dimidiatus manipulate client reef fish by providing tactile stimulation. Proc R Soc Biol Sci Ser B 268:1495-1501 Bshary R, Oliveira RF, Oliveira TS, Canário AV (2007) Do cleaning organisms reduce the stress response of client reef fish? Front Zool 4:21 Bykhovskaya-Pavlovskaya I, Gusev A, Dubinia M, Izyumova N, Smirnova T, Sokolovskaya I, Shtein G, Shul'man S, Epshtein V (1962) Key to parasites of freshwater fish of the USSR Akad. Sci USSR Zool Inst Keys to fauna of the USSR 80:180-223 Caley MJ, Carr MH, Hixon MA, Hughes TP, Jones GP, Menge BA (1996) Recruitment and the local dynamics of open marine populations. Annu Rev Ecol Syst 27:477-500 Candolin U, Voigt HR (2001) No effect of a parasite on reproduction in stickleback males: a laboratory artefact? Parasitology 122:457-464 Carrassón M, Cribb TH (2014) Benign effect of the fish parasitic isopod Ceratothoa cf. imbricata on Selenotoca multifasciata (Scatophagidae) from Australia. Dis Aquat Organ 110:173-180

104

Carreras-Aubets M, Repullés-Albelda A, Kostadinova A, Carrassón M (2011) A new cryptic species of Aponurus Looss, 1907 (Digenea: Lecithasteridae) from Mediterranean goatfish (Teleostei: Mullidae). Syst Parasitol 79:145-159 Ceccarelli D (2007) Modification of benthic communities by territorial damselfish: a multi-species comparison. Coral Reefs 26:853-866 Ceccarelli D, Jones G, McCook L (2005) Effects of territorial damselfish on an algal-dominated coastal coral reef. Coral Reefs 24:606-620 Chambers JM (1992) Linear models. In: Chambers JM, Hastie TJ (eds) Statistical models in S. Wadsworth & Brooks/Cole, Pacific Grove, California, pp95-144 Cheney KL, Côté IM (2001) Are Caribbean cleaning symbioses mutualistic? Costs and benefits of visiting cleaning stations to longfin damselfish. Anim Behav 62:927-933 Cheney KL, Côté IM (2003) The ultimate effect of being cleaned: does ectoparasite removal have reproductive consequences for damselfish clients? Behav Ecol 14:892-896 Cheney KL, Bshary R, Grutter AS (2008) Cleaner fish cause predators to reduce aggression toward bystanders at cleaning stations. Behav Ecol 19:1063-1067 Cheney KL, Grutter AS, Blomberg SP, Marshall NJ (2009) Blue and yellow signal cleaning behavior in coral reef fishes. Curr Biol 19:1283-1287 Chun SK, Chang DS, Park CK, Kim YG, Rho YG (1981) Basic studies for the production of the hard clam Meretrix lusoria (Roding) in Jeonnbug farming area. Bulletin of Fisheries Research and Development Agency 26:7-36 Clague GE, Cheney KL, Goldizen AW, McCormick MI, Waldie PA, Grutter AS (2011) Long-term cleaner fish presence affects growth of a coral reef fish. Biol Lett 6:863-865 Coile AM, Sikkel PC (2013) An experimental field test of susceptibility to ectoparasitic gnathiid isopods among Caribbean reef fishes. Parasitology 140:888-896 Coker DJ, Pratchett MS, Munday PL (2012a) Influence of coral bleaching, coral mortality and conspecific aggression on movement and distribution of coral-dwelling fish. J Exp Mar Biol Ecol 414-415:62-68 Coker DJ, Graham NAJ, Pratchett MS (2012b) Interactive effects of live coral and structural complexity on the recruitment of reef fishes. Coral Reefs 31:919-927 Cooper WJ, Smith JJ, Westneat MW (2009) Exploring the radiation of a diverse reef fish family: phylogenetics of the damselfishes (Pomacentridae), with new classifications based on molecular analyses of all genera. Mol Phylogenet Evol 52:1-16 Coppock AG, Gardiner NM, Jones GP (2013) Olfactory discrimination in juvenile coral reef fishes: Responses to conspecifics and corals. J Exp Mar Biol Ecol 443:21-26 Côté IM (2000) Evolution and ecology of cleaning symbioses in the sea. Oceanography and Marine Biology 38:311- 355 Côté IM, Arnal C, Reynolds JD (1998) Variation in posing behaviour among fish species visiting cleaning stations. J Fish Biol 53 (sA):256-266 Cribb T, Bray R (1994) A review of the Tandanicolinae (= Monodhelminthinae)(Digenea: Fellodistomidae), with a description of Prosogonarium angelae n. sp. Syst Parasitol 27:187-205 Cribb TH, Bray RA (2010) Gut wash, body soak, blender and heat-fixation: approaches to the effective collection, fixation and preservation of trematodes of fishes. Syst Parasitol 76:1-7 Cribb TH, Bray RA, Barker SC (1992) Zoogonidae (Digenea) from southern Great Barrier Reef fishes with a description of Steganoderma (Lecithostaphylus) gibsoni n. sp. Syst Parasitol 23:7-12 Cribb TH, Bray RA, Barker SC (1994a) Bivesiculidae and Haplosplanchnidae (Digenea) from fishes of the southern Great Barrier Reef, Australia. Syst Parasitol 28:81-97 Cribb TH, Anderson GR, Bray RA (1999) Faustulid trematodes (Digenea) from marine fishes of Australia. Syst Parasitol 44:119-138 Cribb TH, Anderson GR, Adlard RD, Bray RA (1998) A DNA-based demonstration of a three-host life cycle for the Bivesiculidae (Platyhelminthes: Digenea). Int J Parasitol 28:1791-1795 Cribb TH, Miller TL, Bray RA, Cutmore S, C. (2014a) The sexual adult of Cercaria praecox Walker, 1971 (Digenea: Fellodistomidae), with the proposal of Oceroma n. g. Syst Parasitol 88:1-10 Cribb TH, Miller TL, Bray RA, Cutmore SC (in press) The sexual adult of Cercaria praecox Walker, 1971 (Digenea: Fellodistomidae), with the proposal of Oceroma n. g. Syst Parasitol Cribb TH, Bray RA, Barker SC, Adlard RD, Anderson GR (1994b) Ecology and diversity of digenean trematodes of reef and inshore fishes of Queensland. Int J Parasitol 24:851- 860 Cribb TH, Bray RA, Littlewood DTJ, Pichelin SP, Herniou EA (2001) The Digenea. In: Littlewood DTJ, Bray RA (eds) Interrelationships of the Platyhelminthes. Taylor and Francis, London, pp168-185 Cribb TH, Bott NJ, Bray RA, McNamara MKA, Miller TL, Nolan MJ, Cutmore SC (2014b) Trematodes of the Great Barrier Reef, Australia: emerging patterns of diversity and richness in coral reef fishes. Int J Parasitol 44:929- 939 Curran SS, Tkach VV, Overstreet RM (2006) A review of Polylekithum Arnold, 1934 and its familial affinities using morphological and molecular data, with description of Polylekithum catahoulensis sp. nov. Acta Parasitol 51:238-248

105

Curtis LM, Grutter AS, Smit NJ, Davies AJ (2013) Gnathia aureamaculosa, a likely definitive host of Haemogregarina balistapi and potential vector for Haemogregarina bigemina between fishes of the Great Barrer Reef, Australia. Int J Parasitol 43:361-370 Danilowicz BS (1996) Choice of coral species by naive and field caught damselfishes. Copeia 3:735-739 Davies AJ, Smit NJ (2001) The life cycle of Haemogregarina bigemina (Adeleina: Haemogregarinidae) in South African hosts. Folia Parasitol (Praha) 48:169-177 Davies AJ, Curtis L, Grutter AS, Smit NJ (2009) Suspected viral erythrocytic necrosis (VEN) in a juvenile blackbar triggerfish, Rhinecanthus aculeatus, from Lizard Island, Great Barrier Reef, Australia. Marine Biodiversity Records 2:e149 Davies AJ, Smit NJ, Hayes PM, Seddon AM, Wertheim D (2004) Haemogregarina bigemina (Protozoa: Apicomplexa: Adeleorina) - Past, present and future. Folia Parasitol (Praha) 51:99-108 Derbel H, Chaari M, Neifar L (2011) Redescription of Rhipidocotyle galeata (Rudolphi, 1819) (Digenea, Bucephalidae), the type species of Rhipidocotyle Diesing, 1907. Zoosystema 33:133-139 Diaz EP, Bray RA, Cribb TH (2013) Paradiscogaster flindersi and P. oxleyi n. sp. (Digenea: Faustulidae): overlapping host and geographical distributions in corallivore chaetodontid fishes in the tropical Indo-west Pacific. Syst Parasitol 86:87-99 Diaz PE, Cribb TH (2013) Paradiscogaster leichhardti sp. nov. (Digenea: Faustulidae) in Chaetodontoplus meredithi (Perciformes: Pomacanthidae) from Heron Island, Great Barrier Reef. Memoirs of the Queensland Museum 58:49-53 Diaz PE, Bray RA, Cutmore SC, Ward S, Cribb TH (2015) A complex of species related to Paradiscogaster glebulae (Digenea: Faustulidae) in chaetodontid fishes (Teleostei: Perciformes) of the Great Barrier Reef. Parasitol Int Dimitrov GI, Bray RA, Gibson DI (1999) A redescription of Pseudobacciger harengulae (Yamaguti, 1938) (Digenea: Faustulidae) from Sprattus sprattus phalericus (Risso) and Engraulis encrasicholus ponticus Alexandrov off the Bulgarian Black Sea coast, with a review of the genus Pseudobacciger Nahhas & Cable, 1964. Syst Parasitol 43:133-146 Dyer WG, Williams EH, Williams LB (1985) Digenetic trematodes of marine fishes of the western and southern coasts of Puerto Rico. Proc Helminthol Soc Wash 52:85-94 Dyer WG, Williams EH, Bunkley-Williams L (1988) Digenetic trematodes of marine fishes of Okinawa, Japan. J Parasitol 74:638-645 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792-1797 Euzet L, Combes Cle (1980) Les problèmes de l'espèce chez les animaux parasites. Bull Soc Zool Fr 40:239-285 Feary DA, Almany GR, McCormick MI, Jones GP (2007) Habitat choice, recruitment and the response of coral reef fishes to coral degradation. Oecologia 153:727-737 Ferguson JA, Koketsu W, Ninomiya I, Rossignol PA, Jacobson KC, Kent ML (2011) Mortality of coho salmon (Oncorhynchus kisutch) associated with burdens of multiple parasite species. Int J Parasitol 41:1197-1205 Finley RJ, Forrester GE (2003) Impact of ectoparasites on the demography of a small reef fish. Mar Ecol Prog Ser 248:305-309 Fishelson L (1998) Behaviour, socio-ecology and sexuality in damselfishes (Pomacentridae). Ital J Zool 65:387-398 Fogelman RM, Grutter AS (2008) Mancae of the parasitic cymothoid isopod, Anilocra apogonae: early life history, host-specificty, and effect on growth and survival of preferred young cardinal fishes. Coral Reefs 27:685-693 Fogelman RM, Kuris AM, Grutter AS (2009) Parasitic castration of a vertebrate: effect of the cymothoid isopod, Anilocra apogonae, on the five-lined cardinalfish, Cheilodipterus quinquelineatus. Int J Parasitol 39:577-583 Froese R, Pauly D (2012) FishBase (version 02/2014), www.fishbase.org Gaevskaja AV (1996) [New records of trematodes from eastern Atlantic fishes] (In Russian). Parasitlogiya 30:504-509 Gaevskaja AV, Naidenova NN (1996) [Revision of Bacciger bacciger (Trematoda: Fellodistomidae) from fishes of the Black and Mediterranean Seas] (In Russian). Parazitologiya (St Petersb) 30:39-44 Gibson DI, Bray RA (1979) The Hemiuroidea: terminology, systematics and evolution. Bulletin of the British Museum Natural History (Zoology) 36:35-146 Gorlick DL, Atkins PD, Losey GS (1987) Effect of cleaning by Labroides dimidiatus (Labridae) on an ectoparasite population infecting Pomacentrus vaiuli (Pomacentridae) at Enewetak Atoll. Copeia 1:41-45 Green AL (1994) The early life history of labroid fishes at Lizard Island, Northern Great Barrier Reef. James Cook University, Green AL (1998) Spatio-temporal patterns of recruitment of labroid fishes (Pisces: Labridae and Scaridae) to damselfish territories. Environ Biol Fishes 51:235-244 Green BS (2008) Maternal effects in fish populations. Adv Mar Biol 54:1-105 Grorud-Colvert K, Sponaugle S (2011) Variability in water temperature affects trait-mediated survival of a newly settled coral reef fish. Oecologia 165:675-686 Grutter AS (1994) Spatial and temporal variations of the ectoparasites of seven reef fish species from Lizard Island and Heron Island, Australia. Mar Ecol Prog Ser 115:21-30 Grutter AS (1995a) Relationship between cleaning rates and ectoparasite loads in coral reef fishes. Mar Ecol Prog Ser 118:51-58

106

Grutter AS (1995b) A comparison of methods for sampling ectoparasites from coral reef fishes. Mar Freshw Res 46:897-903 Grutter AS (1996a) Parasite removal rates by the cleaner wrasse, Labroides dimidiatus. Mar Ecol Prog Ser 130:61-70 Grutter AS (1996b) Experimental demonstration of no effect by the cleaner Labroides dimidiatus (Cuvier and Valenciennes) on the host fish Pomacentrus moluccensis (Bleeker). J Exp Mar Biol Ecol 196:285-298 Grutter AS (1997) Spatiotemporal variation and feeding selectivity in the diet of the cleaner fish Labroides dimidiatus. Copeia 1997:346-355 Grutter AS (1999a) Cleaner fish really do clean. Nature 398:672-673 Grutter AS (1999b) Infestation dynamics of gnathiid isopod juveniles parasitic on the coral-reef fish Hemigymnus melapterus (Labridae). Mar Biol 135:545-552 Grutter AS (2000) Ontogenetic variation in the diet of the cleaner fish Labroides dimidiatus and its ecological consequences Mar Ecol Prog Ser 197:241-246 Grutter AS (2001) Parasite infection rather than tactile stimulation is the proximate cause of cleaning behaviour in reef fish. Proc R Soc Biol Sci Ser B 268:1361-1365 Grutter AS (2002) Cleaning symbioses from the parasites' perspective. Parasitology 124:65-81 Grutter AS (2003) Feeding ecology of the fish ectoparasite Gnathia sp. (Crustacea: Isopoda) from the Great Barrier Reef, and its implications for cleaning behaviour. Mar Ecol Prog Ser 2559:295-302 Grutter AS (2004a) Cleaner fish use tactile dancing behaviour as a preconflict management strategy. Curr Biol 14:1080- 1083 Grutter AS (2004b) Cleaner fish use tactile dancing behaviour as a preconflict management strategy. Current biology 14:1080-1083 Grutter AS (2008) Interactions between gnathiid isopods, cleaner fish and other fishes on Lizard Island, Great Barrier Reef. J Fish Biol 73:2094-2109 Grutter AS, Poulin R (1998) Intraspecific and interspecific relationships between host size and the abundance of parasitic larval gnathiid isopods on coral reef fishes. Mar Ecol Prog Ser 164:263-271 Grutter AS, Hendrikz J (1999) Diurnal variation in the abundance of juvenile parasitic gnathiid isopods on coral reef fish- implications for parasite-cleaner fish interactions. Coral Reefs 18:187-191 Grutter AS, Lester RJG (2002) Cleaner fish Labroides dimidiatus reduce 'temporary' parasitic corallanid isopods on the coral reef fish Hemigymnus melapterus. Mar Ecol Prog Ser 234:247-255 Grutter AS, Lester RJG, Greenwood JG (2000) Emergence rates from the benthos of the parasitic juveniles of gnathiid isopods. Mar Ecol Prog Ser 207:123-127 Grutter AS, Murphy J, Choat H (2003) Cleaner fish drives local fish diversity on coral reefs. Curr Biol 13:64-67 Grutter AS, Glover S, Bshary R (2005) Does client size affect cleaner fish choice of client? An empirical test using client fish models. J Fish Biol 66:1748-1752 Grutter AS, Pickering JL, McCallum H, McCormick MI (2008) Impact of micropredatory gnathiid isopods on young coral reef fishes. Coral Reefs 27:655-661 Grutter AS, Cribb TH, McCallum H, Pickering JL, McCormick MI (2010) Effects of parasites on larval and juvenile stages of the coral reef fish Pomacentrus moluccensis. Coral Reefs 29:31-40 Grutter AS, Crean AJ, Curtis LM, Kuris AM, Warner RR, McCormick MI (2011) Indirect effect of an ectoparasite reduce successful establishment of a damselfish at settlement. Funct Ecol 25:586-594 Grutter AS, Blomberg SP, Fargher B, Kuris AM, McCormick MI, Warner RR (in review) A reason for pelagic larval stages: Small size of coral reef fish at settlement and high mortality risk from gnathiid isopod micropredators Guégan JF, Lambert A, Lévêque C, Combes C, Euzet L (1992) Can host body size explain the parasite species richness in tropical freshwater fishes? Oecologia 90 Hall KA, Cribb TH, Barker SC (1999) V4 region of small subunit rDNA inidicates polyphyly of the Fellodistomidae (Digenea) which is supported by morphology and life-cycle data. Syst Parasitol 43:81-92 Harms KE, Wright SJ, Calderón O, Hernández A, Herre EA (2000) Pervasive density-dependent recruitment enhances seedling diveristy in a tropical forest. Nature 404:493-495 Hayes PM, Smit NJ, Grutter AS, Davies AJ (2011) Unexpected response of a captive blackeye thicklip, Hemigymnus melapterus (Bloch), from Lizard Island, Australia, exposed to juvenile isopods Gnathia aureamaculosa Ferreira & Smit. J Fish Dis 34:563-566 Heinlein J, M., Stier AC, Steele MA (2010) Predators reduce abundance and species richness of coral reef fish recruits via non-selective predation. Coral Reefs 29:527-532 Helfman GS (1986) Fish behaviour by day, night and twilight The behaviour of teleost fishes. Springer, pp366-387 Heupel MR, Bennett MB (1999) The occurrence, distribution and pathology associated with gnathiid isopod larvae infecting the epaulette shark, Hemiscyllium ocellatum. International Journal for Parasitology 29:321-330 Hixon MA, Brostoff WN (1983) Damselfish as keystone species in reverse: intermediate disturbance and diversity of reef algae. Science 220:511-513 Ho HW, Bray RA, Cutmore SC, Ward S, Cribb TH (2014) Two new species of Phyllodistomum Braun, 1899 (Trematoda: Gorgoderidae Looss, 1899) from Great Barrier Reef fishes. Zootaxa 3779:551-562 Hoey AS, McCormick MI (2004) Selective predation for low body condition at the larval-juvenile transition of a coral reef fish. Oecologia 139:23-29

107

Horthorn T, Bretz F, Westfall P, Hothorn MT (2013) Package 'multcomp' (ed. CRAN) See http://cran.r- project.org/web/packages/multcomp. Hudson PJ, Dobson AP, Lafferty KD (2006) Is a healthy ecosystem one that is rich in parasites? Trends Ecol Evol 21:381-385 Huijbers CM, Nagelkerken I, Lössbroek PA, Schulten IE, Siegenthaler A, Holderied MW, Simpson SD (2012) A test of the senses: Fish select novel habitats by responding to multiple cues. Ecology 93:46-55 Hunter JA, Cribb TH (2010) Transversotrematidae (Plathelminthes: Trematoda) are rich and abundant on Indo-Pacific fishes. Zootaxa 2442:25-38 Hunter JA, Cribb TH (2012) A cryptic complex of species related to Transversotrema licinum Manter, 1970 from fishes of the Indo-West Pacific, including descriptions of ten new species of Transversotrema Witenberg, 1944 (Digenea: Transversotrematidae). Zootaxa 3176:1-44 Ichihara A (1974) Hemiurid trematodes from marine fishes near the Tsushima Islands in the Sea of Japan. Third International Congress of Parasitology Ichihara A (1986) On the helminths of marine fishes in Sagami Bay (II) Parasitology - quo vadit? Sixth International Congress of Parasitology Jacoby CA, Greenwood JG (1988) Spatial, temporal, and behavioural patterns in emergence of zooplankton in the lagoon of Heron Reef, Great Barrier Reef, Australia. Mar Biol 97:309-328 Johnston SJ (1913) On some Queensland trematodes, with anatomical observations and descriptions of new species and genera. Q J Microsc Sci 59:361-400 Johnston TH (1927) New trematodes from an Australian siluroid. Trans R Soc S Aust 51:129-136 Jones CM, Grutter AS (2005) Parasitic isopods (Gnathia sp.) reduce haematocrit in captive blackeye thicklip (Labridae) on the Great Barrier Reef. Journal of Fish Biology 66:860-864 Jones CM, Grutter AS (2007) Variation in emergence of parasitic and predatory isopods among habitats at Lizard Island, Great Barrier Reef. Mar Biol 150:919-927 Jones CM, Grutter AS (2008) Reef-based micropredators reduce the growth of post-settlement damselfish in captivity. Coral Reefs 27:677-684 Jones GP (1988) Experimental evaluation of the effects of habitat structure and competitive interactions on the juveniles of two coral reef fishes. J Exp Mar Biol Ecol 123:115-126 Jordan LKB, Lindeman KC, Spieler RE (2012) Depth-variable settlement patterns and predation influence on newly settled reef fish (Haemulon spp., Haemulidae). PLoS ONE 7:e50897 Jordano P (2000) Fruits and frugivory. In: Fenner M (ed) Seeds: The ecology of regeneration in natural plant communities. CABI, Wallingford, UK, pp125-166 Jousson O, Bartoli P, Pawlowski J (2000) Cryptic speciation among intestinal parasites (Trematoda: Digenea) infecting sympatric host fishes (Sparidae). J Evol Biol 13:778-785 Justine JL (2010) Parasites of coral reef fish: how much do we know? With a bibliography of fish parasites in New Caledonia. Belg J Zool 140:155-190 Justine JL, Beveridge I, Boxshall GA, Bray RA, Moravec F, Trilles JP, Whittington ID (2010) An annotated list of parasites (Isopoda, Copepoda, Monogenea, Digenea, Cestoda and Nematoda) collected in groupers (Serranidae, Epinephelinae) in New Caledonia emphasizes parasite biodiversity in coral reef fish. Folia Parasitol (Praha) 57:237-262 Kamiya T, O'Dwyer K, Nakagawa S, Poulin R (2014) What determines species richness of parasitic organisms? A meta‐analysis across animal, plant and fungal hosts. Biological Reviews 89:123-134 Kerrigan BA (1996) Temporal patterns in size and condition at settlement in two tropical reef fishes (Pomacentridae: Pomacentrus amboinensis and P. nagasakiensis). Mar Ecol Prog Ser 135:27-41 Kim YG, Chun SK (1984) Studies on the life history of Bacciger harengulae. J Korean Fish Soc 17:449-470 King RE (1964) Three hemiurid trematodes from South Vietnam. Trans Am Microsc Soc 83:435-439 Knowlton N, Rohwer F (2003) Multispecies microbial mutualisms on coral reefs: The host as a habitat. The American Naturalist 162:51-62 Korotaeva VD (1969) [Helminths of some food-fishes of the order Clupeiformes in the sea near Australia and New Zealand] (In Russian). In: Markevich AP (ed) Problemy Parazitologii Trudy Nauchno Konferentsii Parazitologov USSR (6th) 237-238 Krebs CJ (1999) Ecological methodology. Benjamin/Cummings, Menlo Park, California Krkošek M, Revie CW, Gargan PG, Skilbrei OT, Finstad B, Todd CD (2013) Impact of parasites on salmon recruitment in the Northeast Atlantic Ocean. Proc R Soc Biol Sci Ser B 280:doi: 10.1098/rspb.2012.2359 Kuris A, Lafferty K (2000) Parasite-host modeling meets reality: adaptive peaks and their ecological attributes. Evolutionary biology of host–parasite relationships: theory meets reality:9-26 Kurochkin YV (1970) [New trematode species Discogasteroidinae (Trematoda, Fellodistomatidae) from marine fishes of Australia] (In Russian). Parazitologiya 4:111-115 Lafferty KD, Kuris AM (2002) Trophic strategies, animal diversity and body size. Trends Ecol Evol 17:507-513 Lecchini D, Planes S, Galzin R (2005) Experimental assessment of sensory modalities of coral-reef fish larvae in the recognition of their settlement habitat. Behav Ecol Sociobiol 58:18-26 Lecchini D, Planes S, Galzin R (2007a) The influence of habitat characteristics and conspecifics on attration and survival of coral reef fish juveniles. J Exp Mar Biol Ecol 341:85-90 108

Lecchini D, Peyrusse K, Lanyon RG, Lecellier G (2014) Importance of visual cues of conspecifics and predators during the habitat selection of coral reef fish larvae. Ethology 337:345-351 Lecchini D, Osenberg CW, Shima JS, St Mary CM, Galzin R (2007b) Ontogenetic changes in habitat selection during settlement in a coral reef fish: ecological determinants and sensory mechanisms. Coral Reefs 26:423-432 Lemly AD, Esch GW (1984) Effects of the trematode Uvulifer ambloplitis on juvenile Bluegill Sunfish, Lepomis macrochirus: ecological implications. J Parasitol 70:475-492 Lester RJG, Sewell KB (1990) Checklist of parasites from Heron Island, Great Barrier Reef. Aust J Zool 37:101-128 Lewis AR (1997) Recruitment and post-recruit immigration affect the local population size of coral reef fishes. Coral Reefs 16:139-149 Lienart GDH, Mitchell MD, Ferrari MC, McCormick MI (2014) Temperature and food availability affect risk assessment in an ectotherm. Anim Behav 89:199-204 Linton E (1910) Helminth fauna of the Dry Tortugas. II. Trematodes. Papers of the Tortugas Laboratory of the Carnegie Institute of Washington Publication 4:11-98 Littlewood DTJ, Curini-Galletti M, Herniou EA (2000) The interrelationships of Proseriata (Platyhelminthes: Seriata) tested with molecules and morphology. Mol Phylogenet Evol 16:449-466 Lo CM, Morand S, Galzin R (1998) Parasite diversity/host age and size relationship in three coral-reef fishes from French Polynesia. Int J Parasitol 28:1695-1708 Losey GS (1972) The ecological importance of cleaning symbiosis. Copeia 4:820-833 Losey GS (1979) Fish cleaning symbiosis: proximate cauases of host behaviour. Anim Behav 27:669-685 Losey GS, Mahon JL, Danilowicz BS (1995) Innate recognition by host fish of their cleaning symbiont. Ethology 100:277-283 Machida M (1984) Two new trematodes from gallbladder of tropical marine fishes, Myripristis and Abudefduf. Bulletin of the Natural Museum, Tokyo 10:1-5 Machida M (2005) Monorchiidae (Trematoda, Digenea) from fishes of Japanese and adjacent waters. Bulletin of the National Science Museum, Tokyo Series A Zoology 31:123-136 Machida M, Uchida K (2001) Digenean trematodes from pomacenthid fishes of Japan and Palau. Bulletin of the National Science Museum, Tokyo, Series A, Zoology 27:217-227 Maddison WP, Maddison DR (2009) Mesquite: a modular system for evolutionary analysis. Version 2.72 http://mesquiteproject.org Madhavi R (1975) Digenetic trematodes from marine fishes of Waltair coast, Bay of Bengal: Family Fellodistomidae. Revista di Parasitologia 36:115-128 Manter HW (1934) Some digenetic trematodes from deep-water fish of Tortugas, Florida. Papers from Tortugas Laboratory 28:257-345 Manter HW (1942) Monorchiidae (Trematoda) from fishes of Tortugas, Florida. Trans Am Microsc Soc 61:349-360 Manter HW (1947) The digenetic trematodes of marine fishes of Tortugas, Florida. The American Midland Naturalist 38:257-416 Manter HW (1969a) Three new genera hemiurid trematodes from Australian fishes. Journal of the Fisheries Board of Canada 26:787-792 Manter HW (1969b) Some digenetic trematodes of marine fishes of New Caledonia. Part IV. Hemiuridae and summary. Proc Helminthol Soc Wash 36:194-204 Marcogliese DJ (1995) The role of zooplankton in the transmission of helminth parasites to fish. Rev Fish Biol Fish 5:336-371 Marcogliese DJ (2002) Food webs and the transmission of parasites to marine fish. Parasitology 124:83-99 Marino F, Giannetto S, Paradiso ML, Bottari T, De Vico G, Macri B (2004) Tissue damage and haematophagia due to praniza larvae (Isopoda: Gnathiidae) in some aquarium seawater teleosts. Dis Aquat Organ 59:43-47 McCormick MI (2006) Mothers matter: crowding leads to stressed mothers and smaller offspring in marine fish. Ecology 87:1104-1109 McCormick MI, Hoey AS (2004) Larval growth history determines juvenile growth and survival in a tropical marine fish. Oikos 106:225-242 McCormick MI, Meekan MG (2007) Social facilitation of selective mortality Ecology 88:1562-1570 McCormick MI, Moore JAY, Munday PL (2010) Influence of habitat degradation on fish replenishment. Coral Reefs 29:537-546 McNamara M, Cribb T (2011) Taxonomy, host specificity and dietary implications of Hurleytrematoides (Digenea: Monorchiidae) from chaetodontid fishes on the Great Barrier Reef. Parasitol Int 60:255-269 McNamara M, Miller T, Cribb T (2014) Evidence for extensive cryptic speciation in trematodes of butterflyfishes (Chaetodontidae) of the tropical Indo-West Pacific. Int J Parasitol 44:37-48 McNamara M, Adlard R, Bray R, Sasal P, Cribb T (2012) Monorchiids (Platyhelminthes: Digenea) of chaetodontid fishes (Perciformes): biogeographical patterns in the tropical Indo-West Pacific. Parasitol Int 61:288-306 Medeiros PR, Souza AT, Ilarri MI (2010) Habitat use and behavioural ecology of the juveniles of two sympatric damselfishes (Actinopterygii: Pomacentridae) in the south-western Atlantic Ocean. J Fish Biol 77:1599-1615 Meekan M, Steven A, Fortin M (1995) Spatial patterns in the distribution of damselfishes on a fringing coral reef. Coral Reefs 14:151-161

109

Meekan MG, Wilson SG, Halford A, Retzel A (2001) A comparison of catches of fishes and invertebrates by two light trap designs, in tropical NW Australia. Mar Biol 139:373-381 Mensink PJ, Geange SW, Shima JS (2014) Reproductive success of parasitized males in a marine reef fish. Mar Biol 161:2689-2696 Milicich MJ, Doherty PJ (1994) Larval supply of coral reef fish populations - magnitude and synchrony of replenishment to Lizard Island, Great Barrier Reef. Mar Ecol Prog Ser 110:121-134 Militz TA, Hutson KS (2015) Beyond symbiosis: cleaner shrimp clean up in culture. PLoS ONE 10:e0117723 Miller MA, Pfeiler E, Schwartz T (2010) Creating the CIPRES Science Gateway for inference of large phylogenetic trees. In: Proceedings of the Gateway Computing Environments Workshop (GCE), 14 Nov. 2010, New Orleans, LA, pp. 1-8 Miller TL, Cribb TH (2007) Coevolution of Retrovarium n. gen. (Digenea: Cryptogonimidae) in Lutjanidae and Haemulidae (Perciformes) in the Indo-West Pacific. Int J Parasitol 37:1023-1045 Miller TL, Cribb TH (2008) Eight new species of Siphoderina Manter, 1934 (Digenea, Cryptogonimidae) infecting Lutjanidae and Haemulidae (Perciformes) off Australia. Acta Parasitol 53:344-364 Miller TL, Bray RA, Cribb TH (2011a) Taxonomic approaches to and interpretation of host-specificity of trematodes of fishes: lessons from the Great Barrier Reef. Parasitology 138:1710-1722 Miller TL, Bray RA, Cribb TH (2011b) Taxonomic approaches to and interpretation of host specificity of trematodes of fishes: lessons from the Great Barrier Reef. Parasitology Monod T (1926) Les Gnathiidae. Essai monographique (morphologie, biologie, systématique). Mémoires de la Société des Sciences Naturelles du Maroc 13:1-668 Morand S (2015) (macro-) Evolutionary ecology of parasite diversity: From determinants of parasite species richness to host diversification. International Journal for Parasitology: Parasites and Wildlife 4:80-87 Morand S, Cribb TH, Kulbicki M, Rigby MC, Chauvet C, Dufour V, Faliex E, Galzin R, Lo CM, Lo-Yat A, Pichelin S, Sasal P (2000) Endoparasite species richness of New Caledonian butterfly fishes: host density and diet matter. Parasitology 121:65-73 Mugridge RER, Stallybrass HG (1983) A mortality of eels, Anguilla anguilla L., attributed to Gnathiidae. J Fish Dis 6:81-82 Muñoz G, Cribb TH (2006) Parasite communities and diet of Coris batuensis (Pisces: Labridae) from Lizard Island, Great Barrier Reef. Mem Queensl Mus 52:191-198 Muñoz G, Grutter A, Cribb T (2006) Endoparasite communities of five fish species (Labridae: Cheilininae) from Lizard Island: how important is the ecology and phylogeny of the hosts? Parasitology 132:363-374 Muñoz G, Grutter AS, Cribb TH (2007) Structure of the parasite communities of a coral reef fish assemblage (Labridae): testing ecological and phylogenetic host factors. J Parasitol 93:17-30 Nahhas FM, Cable RM (1964) Digenetic and aspidogasterid trematodes from marine fishes of Curacao and Jamaica. Tulane Studies of Zoology 11:167-228 Nahhas FM, Wetzel JA (1995) Digenetic trematodes of marine fishes from Suva, Fiji: the family Gyliauchenidae Ozaki, 1933. Journal of Helminthological Society Washington 62:117-130 Nolan MJ, Cribb TH (2005) The use and implications of ribosomal DNA sequencing for the discrimination of digenean species. Adv Parasitol 60:101-163 Ohman MC, Munday PL, Jones GP, Caley MJ (1998) Settlement strategies and distribution patterns of coral-reef fishes. J Exp Mar Biol Ecol 225:219-238 Olson PD, Cribb TH, Tkach VV, Bray RA, Littlewood DTJ (2003) Phylogeny and classification of the Digenea (Platyhelminthes: Trematoda). Int J Parasitol 33:733-755 Ormond RFG, Roberts JM, Jan R-Q (1996) Behavioural differences in microhabitat use by damselfishes (Pomacentridae): implications for reef fish biodiversity. J Exp Mar Biol Ecol 202:85-95 Ostlund-Nilsson S, Curtis L, Nilsson G, Grutter AS (2005) Parasitic isopod Anilocra apogonae, a drag for the cardinal fish Cheilodipterus quinquelineatus. Mar Ecol Prog Ser 287:209-216 Ota Y, Hoshino O, Hirose M, Tanaka K, Hirose E (2012) Third-stage larva shifts host fish from teleost to elasmobranch in the temporary parasitic isopod, Gnathia trimaculata (Crustacea: Gnathiidae). Mar Biol 159:2333-2347 Overstreet RM (1969) Digenetic trematodes from Biscayne Bay, Florida. Tulane Stud Zool Bot 15:119-176 Ozaki Y (1932) A new trematode from File-fish: Lepotrema clavatum ngn sp. Proceedings of the Imperial Academy 8:44-47 Palacios-Fuentes P, Landaeta MF, Munoz G, Plaza G, Ojeda FP (2012) The effects of a parasitic copepod on the recent larval growth of a fish inhabiting rocky coasts. Parasitol Res 111:1661-1671 Pankov P, Webster BL, Blasco-Costa I, Gibson DI, Littlewood DTJ, Balbuena JA, Kostadinova A (2006) Robinia aurata n. g., n. sp. (Digenea: Hemiuridae) from the mugilid Liza aurata with a molecular confirmation of its position within the Hemiuroidea. Parasitology 133:217-227 Paperna I (1977) Preliminary data on the Gnathiidae (Isopoda) of the northern Red Sea, the Bitter Lakes and the Eastern Mediterranean and the biology of Gnathia piscivora n. sp. Rapports et Proces-Verbaux des Reunions- Commission Internationale pour l'Exploration Scientifique de la Mer Mediterranee (CIESM) Paris CB, Atema J, Irisson J-O, Kingsford M, Gerlach G, Guigand CM (2013) Reef odor: a wake up call for navigation in reef fish larvae. PLoS ONE 8:e72808 Parukhin AM (1976a) [Parasitic worms of benthic fishes of the southern seas] (In Russian). Naukova Dumka, Kiev 110

Parukhin AM (1976b) [Parasitic worms of commercially important fishes of the southern Seas] (In Russian), Kiev: Naukova Dumka Parukhin AM (1989) [Parasitic worms of bottom fishes of the southern seas] (In Russian). Naukova Dumka, Kiev 153 Pearson JC (1968) Observations on the morphology and life-cycle of Paucivitellosus fragilis Coil, Reid & Kuntz, 1965 (Trematoda: Bivesiculidae). Parasitology 58:769-788 Penfold R, Grutter AS, Kuris AM, McCormick MI, Jones CM (2008) Interactions between juvenile marine fish and gnathiid isopods: predation versus micropredation. Mar Ecol Prog Ser 357:111-119 Pleijel F, Jondelius U, Norlinder E, Nygren A, Oxelman B, Schander C, Sundberg P, Thollesson M (2008) Phylogenies without roots? A plea for the use of vouchers in molecular phylogenetic studies. Mol Phylogenet Evol 48:369- 371 Posada D (2008) jModelTest: Phylogenetic model averaging. Mol Biol Evol 25:1253-1256 Poulin R, Grutter AS (1996) Cleaning symbioses: Proximate and adaptive explanations. Bioscience 46:512-517 Poulin R, Morand S (2000) The diversity of parasites. Q Rev Biol 75:277-293 Poulin R, Cribb TH (2002) Trematode life cycles: short is sweet? Trends Parasitol 18:176-183 Pritchard MH (1963) Studies on digenetic trematodes of Hawaiian fishes, primarily families Lepocreadiidae and Zoogonidae. The Journal of Parasitology 49:578-587 Pritchard MH, Manter HW (1961) Studies of digenetic trematodes of Hawaiian fishes-family Haplosplanchnidae. Proc Helminthol Soc Wash 28:191 R Development Core Team (2013) R: A language and environment for statistical computing , R Foundation for Statistical Computing. http://www.R-project.org), Vienna Radford C, Stanley J, Simpson S, Jeffs A (2011) Juvenile coral reef fish use sound to locate habitats. Coral Reefs 30:295-305 Randall JE, Allen GR, Steene RC (1997) Fishes of the Great Barrier Reef and Coral Sea. Crawford House Publishing, Bathurst, N.S.W Rankin TL, Sponaugle S (2014) Characteristics of settling coral Rref fish are related to recruitment timing and success. PLoS ONE 9:e108871 Rizk MH, Taha SA, Ab El-Rahman A (1996) Helminth parasites of some fishes from Mediterranean Sea and their effect on the coefficient of condition. J Egypt Ger Soc Zool 19:129-144 Robertson DR (1974) A study of the biology and reproductive biology of the labrid fish, Labroides dimidiatus, at Heron Island, Great Barrier Reef. Ph.D, thesis. The University of Queensland, pp 174 Roche DG, Strong LE, Binning SA (2013a) Prevalence of the parasitic cymothoid isopod Anilocra nemipteri on its fish host at Lizard Island, Great Barrier Reef. Aust J Zool 60:330-333 Roche DG, Binning SA, Strong LE, Davies JN, Jennions MD (2013b) Increased behavioural lateralization in parasitized coral reef fish. Behav Ecol Sociobiol 67:1339-1344 Rohde K (2002) Ecology and biogeography of marine parasites. Adv Mar Biol 43:1-86 Rohner CA, Cribb TH (2013) Opecoelidae (Digenea) in northern Great Barrier Reef goatfishes (Perciformes: Mullidae). Syst Parasitol 84:237-253 Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Hohna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP (2012) MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol 61:539-542 Ros AF, Lusa J, Meyer M, Soares MC, Oliveira RF, Brossard M, Bshary R (2011) Does access to the bluestreak cleaner wrasse Labroides dimidatus affect indicators of stress and health in resident reef fishes in the Red Sea? Horm Behav 59:151-158 Russell BC (1983) Annotated checklist of the coral reef fishes in the Capricorn-Bunker group Great Barrier Reef, Australia. Great Barrier Reef Marine Park Authority, Mackay Sambrook J, Russell DW (2001) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring, New York Sasal P, Morand S, Guégan JF (1997) Determinants of parasite species richness in Meditteranean marine fishes. Mar Ecol Prog Ser 149:61-71 Scholz T (2002) Family Gymnophallidae Odhner, 1905. In: Gibson DI, Jones A, Bray RA (eds) Keys to the Trematoda, Volume 1. CABI Publishing and the Natural History Museum, Wallingford, Scott ME, Dobson A (1989) The role of parasites in regulating host abundance. Parasitol Today 5:176-183 Sikkel PC, Cheney KL, Côté IM (2004) In situ evidence for ectoparasites as a proximate cause of cleaning interactions in reef fish. Anim Behav 68:241-247 Sikkel PC, Schaumburg CS, Mathenia JK (2006) Diel infestation dynamics of gnathiid isopod larvae parasitic on Caribbean reef fish. Coral Reefs 25:683-689 Sikkel PC, Ziemba RE, Sears WT, Wheeler JC (2009) Diel ontogenetic shift in parasitic activity in a gnathiid isopod on Caribbean coral reefs. Coral Reefs 28:489-495 Sikkel PC, Sears WT, Weldon B, Tuttle BC (2011) An experimental field test of host-finding mechanisms in a Caribbean gnathiid isopod. Mar Biol 158:1075-1083 Simpson SD, Jeffs A, Montgomery JC, McCauley R, Meekan MG (2008) Nocturnal relocation of adult and juvenile coral reef fishes in response to reef noise. Coral Reefs 27:97-104 Smit NJ, Davies AJ (2004) The curious life-style of the parasitic stages of gnathiid isopods. Adv Parasitol 58:289-391 111

Smit NJ, Basson L, Van As JG (2003) Life cycle of the temporary fish parasite, Gnathia africana (Crustacea: Isopoda: Gnathiidae). Folia Parasitol (Praha) 50:135-142 Smith JW (2002) Family Sanguinicolidae Graff, 1907. In: Gibson DI, Jones A, Bray RA (eds) Keys to the digenean parasites of vertebrates. CAB International, Cambridge, Soares MC, Bshary R, Côté IM (2008) Does cleanerfish service quality depend on client value or choice options? Anim Behav 76:123-130 Soares MC, Oliveira R, Ros AFH, Grutter AS, Bshary R (2011) Tactile stimulation lowers stress in fish. Nat Commun 2:534 Srinivasan M (2003) Depth distibutions of coral reef fishes: the influence of microhabitat structure, settlement, and post-settlement processes. Oecologia 137:76-84 Stachowicz JJ (2001) Mutualism, facilitation, and the structure of ecological communities. BioScience 51:235-246 Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap algorithm for the RAxML web servers. Syst Biol 57:758-771 Stummer LE, Weller JA, Johnson ML, Côté IM (2004) Size and stripes: how fish clients recognize cleaners. Anim Behav 68:145-150 Sun D, Blomberg SP, Cribb TH, McCormick MI, Grutter AS (2012) The effects of parasites on the early life stages of a damselfish. Coral Reefs 31:1065-1075 Sun D, Bray RA, Yong R, Q-Y, Cutmore SC, Cribb TH (2014) Pseudobacciger cheneyae n. sp. (Digenea: Gymnophalloidea) from Weber’s chromis (Chromis weberi Fowler & Bean) (Perciformes: Pomacentridae) at Lizard Island, Great Barrier Reef, Australia. Syst Parasitol 88:141-152 Sweatman H (1985) The influence of adults of some coral reef fishes on larval recruitment. Ecol Monogr 55:469-485 Sweatman H (1988) Field evidence that settling coral reef fish larvae detect resident fishes using dissolved chemical cues. J Exp Mar Biol Ecol 124:163-174 Tolimieri N (1995) Effects of microhabitat characteristics on the settlement and recruitment of a coral reef fish at two spatial scales. Oecologia 102:52-63 Toman G, Kamegai S (1974) Some digenetic trematodes of marine fishes from Saipan, Micronesia Proceedings of the Third International Congress of Parasitology, Munich 1617-1618 Torres J, Miquel J, Casanova J-C, Ribas A, Feliu C, Morand S (2006) Endoparasite species richness of Iberian carnivores: influences of host density and range distribution. Biodivers Conserv 15:4619-4632 Vail AL, McCormick MI (2011) Metamorphosing reef fishes avoid predator scent when choosing a home. Biol Lett 7:921-924 Waldie PA, Blomberg SP, Cheney KL, Goldizen AW, Grutter AS (2011) Long-term effects of the cleaner fish Labroides dimidiatus on coral reef fish communities. PLoS ONE 6:e21201-21201-e21201-21207 Wang P, Wang Y (1993) Some digenetic trematodes from marine fishes from Fujian Province. Journal of Fujian Teachers University Natural Science Edition 9:66-73 Weeks P (2000) Red-billed oxpeckers: vampires or tickbirds? Behav Ecol 11:154-160 Wenger AS, Johansen J, Jones G (2011) Suspended sediment impairs habitat choice and chemosensory discrimination in two coral reef fishes. Coral Reefs 30:879-887 Whittington ID (1998) Diversity "down under": monogeneans in the Antipodes (Australia) with a prediction of monogenean biodiversity worldwide. Int J Parasitol 28:1481-1493 Wilson SK, Burgess SC, Cheal AJ, Emslie M, Fisher R, Miller I, Polunin NVC, Sweatman HPA (2008) Habitat utilization by coral reef fish: implications for specialists vs. generalists in a changing environment. J Anim Ecol 77:220-228 Yamaguti S (1934) Studies on the helminth fauna of Japan. Part 2. Trematodes of fishes. Japanese Journal of Zoology 5:249-541 Yamaguti S (1938) Studies on the Helminth Fauna of Japan. Part 21. Trematodes of Fishes, IV. Yamaguti, S., Kyöto Yamaguti S (1942) Studies on the helminth fauna of Japan. Part 39. Trematodes of fishes mainly from Naha. Transactions of the Biogeographical Society of Japan 3:329-398 Yamaguti S (1953) Parasitic worms mainly from Celebes. Part 3. Digenetic trematodes of fishes II. Acta Med Okayama 8:257-295 Yeo SE, Spieler RE (1980) Habitat effects on the occurence of parasites inhabiting the Sergeant Major, Abudefduf saxatilis (Linnealus), with a list of parasites of Caribbean damselfish. Bull Mar Sci 30:313-324

112

Appendices

Appendix I

All damselfish recruits observed on experimental patch reefs (Chapter 3).

Fish Genus Species

Acanthochromis polyacanthus Amblyglyphidodon curacao Amblyglyphidodon leucogaster Chromis ternatensis Chromis viridis Chrysiptera cyanea † Chrysiptera flavipinnis Chrysiptera rex * Chrysiptera rollandi Chrysiptera talboti Dascyllus aruanus Dascyllus reticulatus Dischistodus perspicillatus Dischistodus prospotaenia Dischistodus spp. Neoglyphidodon melas Neoglyphidodon nigroris Neopomacentrus bankieri Neopomacentrus cyanomos Pomacentrus adelus Pomacentrus amboinensis Pomacentrus bankanensis † Pomacentrus branchialis Pomacentrus chrysurus Pomacentrus coelestis Pomacentrus grammorhynchus Pomacentrus lepidogenys Pomacentrus moluccensis Pomacentrus nagasakiensis Pomacentrus pavo Pomacentrus simsiang

* Species only recorded at reefs with L. dimidiatus present.

† Species only recorded at reefs without L. dimidiatus present.

113

Appendix II

Species of fish cleaned by juvenile Labroides dimidiatus (Chapter 4).

Species Family

Ctenochaetus binotatus Acanthuridae Ctenochaetus striatus Acanthuridae Zebrasoma scopas Acanthuridae Zebrasoma veliferum Acanthuridae Apogon doederleini Apogonidae Apogon compressus Apogonidae Apogon cookii Apogonidae Apogon properupta Apogonidae Apogon trimaculatus Apogonidaae Cheilodipterus intermedius Apogonidae Cirripectes chelomatus Blennidae Cirripectes stigmaticus Blennidae Ecsenius aequalis Blennidae Ecsenius bicolor Blennidae Ecsenius mandibularis Blennidae Ecsenius stictus Blennidae Exallias brevis Blennidae Meiacanthus atrodorsalis Blennidae Plagiotremus rhinorhynchos Blennidae Plagiotremus tapeinosoma Blennidae Caesio caerulaurea Caesionidae Caesio teres Caesionidae Chaetodon aureofasciatus Chaetodontidae Chaetodon baronessa Chaetodontidae Chaetodon citrinellus Chaetodontidae Chaetodon klienii Chaetodontidae Chaetodon lunulatus Chaetodontidae Chaetodon plebius Chaetodontidae Chaetodon rainfordi Chaetodontidae Chaetodon ulietensis Chaetodontidae Amblygobius rainfordi Gobidae Valenciennea strigata Gobidae Neoniphon opercularis Holocentridae Neoniphon sammara Holocentridae Sarocentron cornutum Holocentridae Sargocentron diadema Holocentridae Sargocentron spiniferum Holocentridae Bodianus axillaris Labridae Cheilinus chlorourus Labridae Chlorurus sordidus Labridae Choerodon fasciatus Labridae Coris batuensis Labridae

114

Epibulis insidiator Labridae Gomphosus varius Labridae Halichoeres hortulanus Labridae Halichoeres malanurus Labridae Halichoeres melanurus Labridae Hemigymnus melapterus Labridae Labrichtys unilineatus Labridae Labroides dimidiatus Labridae Pseudocheilinus hexataenia Labridae Thalassoma lunare Labridae Scarus frenatus Labridae Scarus ghobban Labridae Scarus niger Labridae Stethojulis bandanensis Labridae Lutjanus carponotatus Lutjanidae Pervagor janthinosoma Monacanthidae Parupeneus multifasciatus Mullidae Scolopsis bilineatus Nemipteridae Centropyge bicolor Pomacanthidae Pomacanthus sexstriatus Pomacanthidae Abudefduf sexfaciatus Pomacentridae Abudefduf whitleyi Pomacentridae Acanthochromis polyacanthus Pomacentridae Amblyglyphididon curacao Pomacentridae Amblyglyphididon leucogaster Pomacentridae Chromis actripectoralis Pomacentridae Chromis lepidolepis * Pomacentridae Chromis viridis Pomacentridae Chrysiptera rex Pomacentridae Dischistodus melanotus Pomacentridae Dischistodus prosopotaenia Pomacentridae Dischistodus pseudochrysopoecilus Pomacentridae Hemiglyphidodon plagiometopon Pomacentridae Neoglyphidodon melas Pomacentridae Neopomacentrus azysron Pomacentridae Plectroglyphidodon lacrymatus Pomacentridae Pomacentrus adelus * Pomacentridae Pomacentrus amboinensis * Pomacentridae Pomacentrus bankanensis Pomacentridae Pomacentrus branchialis * Pomacentridae Pomacentrus coelestis Pomacentridae Pomacentrus chrysurus Pomacentridae Pomacentrus grammorhynchus Pomacentridae Pomacentrus imitator Pomacentridae Pomacentrus moluccensis Pomacentridae Pomacentrus nagasakiensis * Pomacentridae Pomacentrus simsiang Pomacentridae Pomacentrus wardi Pomacentridae Premnas biaculeatus Pomacentridae Stegastes apicalis Pomacentridae

115

Stegastes fasciolatus Pomacentridae Stegastes nigricans Pomacentridae Ogilbyina queenslandia Pseudochromidae Pseudochromis fuscus Pseudochromidae Cephalopholis boenak Epinephelidae Cephalopholis cyanostigma Epinephelidae Cephalopholis microprion Epinephelidae Siganus corallinus Siganidae Synodus variegatus Synodontidae Canthigaster papua Tetraodontidae Canthigaster valentini Tetraodontidae

* Recently-settled individuals (< 2 cm) of these species were cleaned by L. dimidiatus.

116