Life history changes with the colonisation of land by fish

Edward Richard Murray Platt

Supervised by: Terry Ord

THESIS SUBMITTED FOR THE DEGREE OF MASTER OF PHILOSOPHY Evolution and Ecology Research Centre School of Biological, Earth and Environmental Sciences Faculty of Science University of New South Wales March 2014

THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Platt

First name: Edward Other name/s: Richard Murray

Abbreviation for degree as grven an the University calendar: MPhil

School: Biological, Earth and Environmental Sciences F acuity: Science

Title:Mr

My thesis addressed two questions: whether survival was inferred to have improved for fish that moved onto land, and what the relative role of predation and density were for detenmining life history variation among populations within one of these land species. For the first question I used life history theory to examine whether survival was inferred to have improved in two fish families which have independently made the transition onto land: Gobiidae and Blenniidae. I examined growth and various aspects of reproductive investment among terrestrial and aquatic species, finding that differences varied according to the level of independence from water. This was consistent with improved survival for certain age classes on land. Nevertheless, the details of life history change differed in each family, with the greatest increases in survival implied for early age classes in Blenniidae, but older age classes in Gobiidae. This suggests fundamental differences in the way the colonization of land occurred in each family. For the second question I investigated the consequences of predataon and density on life hastory variation among frve populations of the Pacific leaping blenny A/ficus amoldorum. This frsh lives out of the water on rocks in the splash zone of Guam, where it is vulnerable to predation and likely sensitive to changes in population density which impact resource availability. I found the extent to which populations invested in reproductaon was inversely related to predation rate, while growth rate appeared to vary largely in response to population density. The differences in fife history among populations were also likely to be plastic an the Pacific leaping blenny. My results revealed the interaction between predation and density can lead to complex outcomes in life history, and inferred that plastic life history traits could allow populations to persist in new environments. For example, plasticity in life history may have been an important facilitator for these fish to make the transition onto land in the first place. As a whole, my thesis demonstrates how life history theory, by identifying differences in survival between species and populations respectively, provides a window into the ecological conditions impacting an organism.

Declaration relating to disposition of project thesis/dissertation

1 hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or m part in the University libraries in all fonms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertataon.

1 also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

. 3??/;."f/~ .? ..1.?...... Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptaonal cirrumstances and reauire the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

iii ORIGINALITY STATEMENT

'I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.'

Signed . h:.". 'fl . ,. ~ ......

Date ....3~;ftolS ......

i COPYRIGHT STATEMENT

'I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

signed ..... i'~ .. j?J!? .. ?......

Date ...... 3;1~

AUTHENTICITY STATEMENT

'I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.'

Signed .. d.:,zt; ..~ ...... 01 Date ... 3flj?:: 5 ......

ii Abstract

My thesis addressed two questions: whether survival was inferred to have improved for fish that moved onto land, and what the relative role of predation and density were for determining life history variation among populations within one of these land species. For the first question I used life history theory to examine whether survival was inferred to have improved in two fish families which have independently made the transition onto land:

Gobiidae and Blenniidae. I examined growth and various aspects of reproductive investment among terrestrial and aquatic species, finding that differences varied according to the level of independence from water. This was consistent with improved survival for certain age classes on land. Nevertheless, the details of life history change differed in each family, with the greatest increases in survival implied for early age classes in Blenniidae, but older age classes in Gobiidae. This suggests fundamental differences in the way the colonization of land occurred in each family. For the second question I investigated the consequences of predation and density on life history variation among five populations of the Pacific leaping blenny

Alticus arnoldorum. This fish lives out of the water on rocks in the splash zone of Guam, where it is vulnerable to predation and likely sensitive to changes in population density which impact resource availability. I found the extent to which populations invested in reproduction was inversely related to predation rate, while growth rate appeared to vary largely in response to population density. The differences in life history among populations were also likely to be plastic in the Pacific leaping blenny. My results revealed the interaction between predation and density can lead to complex outcomes in life history, and inferred that plastic life history traits could allow populations to persist in new environments. For example, plasticity in life history may have been an important facilitator for these fish to make the transition onto land in the first place. As a whole, my thesis demonstrates how life history theory, by identifying

iv differences in survival between species and populations respectively, provides a window into the ecological conditions impacting an organism.

v Table of Contents Page

CHAPTER 1: Introduction 1

CHAPTER 2: Predictable life history changes reveal enhanced 19 survival for marine fish living on land

ABSTRACT 20

INTRODUCTION 21

MATERIALS AND METHODS 29

RESULTS 39

DISCUSSION 53

LITERATURE CITED 59

CHAPTER 3: Population variation in the life history of a land fish 68 reveals the interacting impact of predation and conspecific density

ABSTRACT 69

INTRODUCTION 70

MATERIALS AND METHODS 78

RESULTS 85

DISCUSSION 94

vi LITERATURE CITED 97

CHAPTER 4: Conclusion 105

APPENDIX 110

CHAPTER 2 110

CHAPTER 3 129

ACKNOWLEDGEMENTS 131

vii LIST OF TABLES MAIN TEXT Page

CHAPTER 2 Table 1: Life history predictions for different changes in survival. 25 Table 2: Parameter values of various goby life history traits in 43 response to moving onto land. Table 3: Parameter values of various blenny life history traits in 48 response to moving onto land. CHAPTER 3 Table 1: AIC table of predictors of population differences in reproductive 89 investment. Table 2: AIC table of predictors of population differences in growth rate. 91

APPENDIX

CHAPTER 2 Table A1: Stages of oocyte classification. 115 Table A2: Parameter values of various goby life history traits in response 116 to moving onto land (for tropical and subtropical species only). Table A3: Parameter values of various blenny life history traits in response 117 to moving onto land. Table A4: Species and sample sizes of blennies and life history 118 characteristics of gobies examined in Chapter 2. CHAPTER 3

Table A1: Sample sizes processed for each population. 130

viii LIST OF FIGURES MAIN TEXT Page

CHAPTER 2 Figure 1: Life history strategies of gobies as a function of life style 42 (aquatic or terrestrial). Figure 2: Life history trade-offs in gobies. 44 Figure 3: The change in growth rate as blennies moved onto land. 47 Figure 4: The change in ovarian weight and development rate as 49 blennies moved onto land. Figure 5: The change in egg number as blennies moved onto land. 50 Figure 6: The change in egg size as blennies moved onto land. 51 Figure 7: Life history trade-offs in blennies. 52 CHAPTER 3 Figure 1: Predicted changes in reproductive investment and growth 77 rate in response to different mortality regimes. Figure: 2: The relationship between predation and adult density and 88 juvenile density. Figure 3: The change in population reproductive investment in 90 response to predation. Figure 4: The change in population growth rate in response to 92 adult density. Figure 5: Trade-off between reproductive investment and growth rate. 93

APPENDIX

CHAPTER 2 Figure A1: Growth rate of aquatic and terrestrial gobies, including 111 outlier species. Figure A2: Oocyte size frequency histogram (example 1). 112 Figure A3: Oocyte size frequency histogram (example 2). 113 Figure A4: Egg number for gobies from tropical and subtropical climates. 114

ix CHAPTER 1

Introduction

1 Maximising fitness

Central to the existence of living organisms is survival and reproduction; two basic tasks for which an enormous variety of strategies exist. At one extreme there are species such as the bristlecone pine Pinus longaeva, a plant which can live for over 4000 years, that does not appear to senesce, and produces countless seeds in its lifetime (Currey 1965;

Lanner & Connor 2001). A similar vertebrate example is the orange roughy Hoplostethus atlanticus, a fish that can live for over 100 years, producing 30 – 50 000 eggs a season

(Koslow et al. 1995; Andrews et al. 2009). At the other extreme, there are plants which live for a single year and produce numerous seeds in a single suicidal effort (e.g. Primack

1979). Similarly, the fish Eviota sigillata (Pygmy goby) lives for a maximum of 60 days and produces around 400 eggs in its lifetime (Depczynski & Bellwood 2005). These species illustrate the great diversity of life history strategies, each of which have presumably evolved to maximize an organism’s lifetime fitness (Gadgil & Bossert 1970;

Roff 1992; Stearns 1992); i.e. the number of offspring surviving to reproduce (Smith &

Fretwell 1974). Life history theory has been developed to understand the balance between survival and reproduction and has proven to be highly successful in accounting for much of the diversity in life history strategies in nature (Stearns 1992).

Life history theory examines how the balance between traits linked directly to survival and reproduction (life history traits) should shift, following a change in the probability of survival, to maximise fitness (Roff 1992; Stearns 1992). Key to life history theory is the notion of trade-offs: the necessary splitting of investment between different life history traits as a consequence of organisms having a finite amount of energy at any one point in time (Gadgil & Bossert 1970; Stearns 1992; Zera & Harshman 2001). For

2 example, investment in current reproduction is predicted to come at the expense of future survival (and consequently future reproduction; Pianka & Parker 1975). The predictions of life history theory lead directly from such trade-offs and predict what balance of investment should be made between traits (e.g. those for survival and those for reproduction) to optimise fitness in a given environment (e.g. Gadgil & Bossert 1970;

Roff 1992; Stearns 1992).

Central to predicting the optimum balance of survival and reproduction has been the age-dependent probability of survival; i.e. the survival of early ages relative to older ages (e.g. Gadgil & Bossert 1970; Charlesworth 1980; Stearns 1992; Ernande et al.

2004). Age-dependent survival takes two general forms (Gadgil & Bossert 1970; Taylor et al. 1974; Law 1979; Michod 1979; Hutchings 1993): (i) survival declines with age

(e.g., the probability of survival is higher for younger individuals compared to older individuals in the population); or (ii) survival increases with age (e.g., mortality is higher in younger individuals compared to older individuals). Additionally, a third scenario is also possible in which survival is not age-dependent and all age classes have an equal probability of survival. For each of these scenarios, life history theory provides explicit predictions (e.g., see Table 1 in Chapter 2).

It is predicted that a change in the survival of early ages relative to older ages will change the optimum balance of investment in survival and reproduction (Gadgil &

Bossert 1970; Stearns 1976; Law 1979; Michod 1979; Roff 1992; Stearns 1992; Ernande et al. 2004). To maximise fitness, reproduction is expected to shift to those age classes that experience increased survivorship (Michod 1979). For example, if survival for younger age classes improves, then reproductive effort should be concentrated in these

3 earlier life stages and emphasise current reproduction over future reproduction (high reproductive effort) at the compromise of growth rate (Gadgil & Bossert 1970; Pianka &

Parker 1975; Law 1979; Michod 1979; Hutchings 1993). In contrast, if improvements in survival were largely confined to older age classes, then investment in reproduction should be concentrated in older individuals and growth rate should generally be higher

(Gadgil & Bossert 1970; Law 1979; Michod 1979; Hutchings 1993). This emphasis on growth occurs because an individual would do better to invest more in growth and to delay reproduction to later in life to increase the chances of surviving through the earlier

(high mortality) life stages (Gadgil & Bossert 1970; Pianka & Parker 1975; Law 1979).

Alternatively, if survival changes uniformly for all age classes, the optimum distribution of reproductive effort (emphasis on current or future reproduction) should not change, and as a result no change in life history strategy is predicted (Gadgil & Bossert 1970;

Taylor et al. 1974; Law 1979; Michod 1979).

Life history theory therefore provides explicit predictions to be investigated in nature. Where empirical work has investigated these predictions, the results have been largely supportive (e.g. Promislow & Harvey 1990; Stibor 1992; Wellborn 1994;

Conover & Munch 2002). Furthermore, given that theory provides empirically supported predictions for how life history should change in response to age-dependent mortality, it should also be possible to use life history differences among populations and species to understand how age-specific survival might differ among these taxa. This is particularly useful for situations where it has been specifically hypothesised that age-specific survival should differ among taxa for ecological reasons; e.g. the colonisation of new habitats.

4 The colonisation of new habitats

Why organisms move into new habitats, particularly those with vastly different biotic and abiotic properties, is unclear. An ecological “release” or ecological “opportunity” are two hypotheses for why organisms might make such transitions. Specifically, the new environment offers reduced predation or competition for resources (an ecological release) or the new environment offers new resources that are absent or rare in the ancestral environment (an ecological opportunity; Wilson 1961; Lister 1976; Jeffries & Lawton

1984; Losos & Mahler 2010). In both cases, the new environment offers an organism a better chance of survival.

One of the most extreme instances of habitat colonisation has been the move by fish onto land (Sayer & Davenport 1991; Martin 1995). This transition has occurred across many different fish families, including some of the most diverse fish families:

Gobiidae and Blenniidae (Sayer & Davenport 1991; Martin 1995). The diversity of these two families makes it possible to investigate whether broad-scale life history changes accompanied independent moves onto land, and thus whether the colonisation of a new habitat more generally may have led to improved survival. Perhaps the most well-known of these colonists are the gobies known as mudskippers (Oxudercinae), all of which are highly amphibious (Murdy 1989; Polgar 2012). These fishes occupy mudflats and mangroves, spending the majority of their time on land following the water’s edge with the changes in tide (Colombini 1995; Polgar 2010). They are active predators, feeding upon a wide variety of small intertidal inhabitants, including insects and small crustaceans, and acquiring additional small food items by filtering sediment with their gills (Kruitwagen et al. 2007; Baeck 2008; Bucholtz et al. 2009; Takeda et al. 2011).

5 Less well known but offering a similar comparative opportunity are the blennies from the tribe Salariini, generally known as rockskippers (Bhikajee et al. 2006). These fishes display a wide spectrum of amphibious behaviour, spanning from fully aquatic through to exclusively terrestrial in lifestyle (Sayer & Davenport 1991, Bhikajee et al.2006; Shimizu et al. 2006; Patzner et al.2009; Ord & Hsieh 2011; Polgar 2012). These fish occupy rocky intertidal areas, feeding upon detritus and, to a lesser extent, the epilithic algae upon which the detritus collects (Wilson 2001, 2004; Patzner et al. 2009;

Hundt et al. 2014).

An obvious question is why should fish colonise land in the first place? An ecological release from competition or predation, or some new ecological opportunity

(e.g., in resources) found only (or more commonly) in a new habitat are classic explanations for why animals might make the transition into a new environment (Wilson

1961; Lister 1976; Jeffries & Lawton 1984; Losos & Mahler 2010). In either case, colonisation of a novel habitat implies improved survival for the coloniser compared to the ancestral habitat. According to life history theory, such a change in survival should be apparent in noticeable changes in life history. More specifically, if fish have colonised land because of some form of ecological release or opportunity, then the accompanying increase in survival should be reflected in predictable changes in life history between terrestrial/amphibious fish and aquatic fish.

Interpreting life history variation

Life history theory provides clear predictions for how life history traits should change in response to shifts in age-specific mortality. However, testing these predictions is

6 complicated by understanding the primary causes of mortality in nature and how it varies with age. Factors such as disease, parasitism, resource competition and predation all contribute to mortality and can impact individuals of certain age classes in different ways

(e.g., Abrams & Rowe 1996; Reznick et al. 2002; Gutowsky & Fox 2012; Schroder et al.

2014). Many studies of life history have focussed exclusively on predation, as it directly contributes to mortality and can have a predictable impact on different age classes (e.g. see Dowdall et al. 2012). However, interpreting age-specific mortality solely through the direct effect of predation on population mortality overlooks the indirect effects of predation on population density (Abrams & Rowe 1996). Specifically predation, by reducing the density of individuals, can in turn reduce competition for resources which also contributes to age-specific mortality. Predation can therefore reduce the adverse density-dependent effects of other factors, and thus positively affect the survival of those individuals which remain (Abrams & Rowe 1996; Walsh & Reznick 2008; Reznick et al.

2012). Furthermore these density-dependent effects can influence separate age classes differently, changing the relative survival of early and older ages (Charlesworth 1980).

This can make it difficult to interpret changes in life history solely from the direct effect of predation on age-specific mortality.

The effect of predation on density should be greatest when a population is isolated from other populations, preventing individuals from immigrating and replacing those lost through predation (e.g., Walsh & Reznick 2008; Reznick et al. 2012). An important quality of these ‘closed’ populations is that their physical isolation also results in genetic isolation. Genetic isolation is expected to facilitate adaptation (e.g. in life history), because selection will lead to the accumulation of advantageous traits over time (Garcia-

7 Ramos & Kirkpatrick 1997; Hendry et al. 2002; Lenormand 2002). For example, a population may develop a higher growth rate and lower level of reproductive investment in response to predation that is more intensely focused on early ages relative to older ages. Conversely, if immigration does occur (i.e. ‘open’ populations), then the resulting gene flow is expected to essentially homogenise the populations involved (e.g.

Lenormand 2002), preventing each population from adapting to their separate habitats.

‘Open’ populations should therefore be unable to change in response to the environment, unless they exhibit a degree of plasticity.

Plastic changes in life history characteristics have been documented (e.g. Reznick

& Yang 1993; Belk 1995; Gale et al. 2013), but there is an implied assumption that these changes are of secondary importance to those with a genetic basis (Riesch et al. 2013).

Specifically, life history adaptation is primarily responsible for differences between populations, while plasticity explains variation within populations (Hutchings 1993).

However, if gene flow inhibits an adaptive life history response, then plastic life history changes may become of central importance for increasing fitness. Note, the immigration of individuals from other populations, which causes high gene flow, should also reduce the density-dependence of predation’s effects on life history variation. Thus, it is not normally possible to investigate how the interaction of predation and density may shape plastic life history responses.

The Pacific leaping blenny Alticus arnoldorum is a fish endemic to the island of

Guam that lives on land, and is likely highly sensitive to both predation and density- dependent effects (e.g., on resource availability). To survive on land, the fish has developed unique adaptations that allow agile movement—an axial-tail twisting

8 behaviour that enables the fish to walk, hop and even jump large distances (Hsieh

2010)—as well as breathing through cutaneous respiration (Martin & Lighton 1989;

Brown et al. 1991). Central to maintaining cutaneous respiration is moisture, which places spatial and temporal limitations on this species (Ord & Hsieh 2011). Spatially, the

Pacific leaping blenny is restricted to rocky outcrops in the splash zone, above which there is too little moisture to survive, and below which is water, which this species actively avoids (Ord & Hsieh 2011). Temporally, tide and temperature play a huge role in when this species can be active. At low tide, the risk of desiccation forces this fish to shelter in rocky holes and crevices. Rocky shelters are therefore central to the survival of this species. At high tide, wave activity becomes violent, impeding the ability of the fish to move safely (Ord & Hsieh 2011). Furthermore, the Pacific leaping blenny shows a clear temperature preference (27-30oC) outside of which this fish remain in their rocky shelters. The cumulative effect of these limitations is a brief window of activity at mid- tide and at moderate temperatures that changes on a daily basis depending on the tides and weather (Ord & Hsieh 2011). The Pacific leaping blenny must concentrate the majority of its feeding (as well as social behaviour and reproduction) to within this variable activity window of ~ 2-4 hours a day (Ord & Hsieh 2011). Importantly, feeding takes place out on highly exposed rocky platforms, where there is considerable predation risk (e.g. from birds; Morgans & Ord 2013). Consequently, it is likely that a delicate balance exists between feeding and predation risk. Additionally, because rock holes are finite in nature, the Pacific leaping blenny is likely to be particularly sensitive to changes in population density. For these reasons, the Pacific leaping blenny is ideal for investigating predation and density-dependent effects on life history variation.

9 The Pacific leaping blenny is also unique in that populations on land are ecologically isolated from one another, yet experience extensive inter-population gene flow. Separating each population on land are large stretches of unsuitable habitats, such as beaches. However, gene flow between these populations does occur because of active or passive dispersal of aquatic larvae (Cooke et al. In Review). Specifically, larvae hatch from eggs laid in rock holes on land and enter the water (either by being washed out or by swimming) where they are subsequently dispersed around the island (Cooke et al. In

Review). Once individuals have transitioned from their marine larvae stage to their terrestrial life stage (an average period of 28 days; ERM Platt unpublished data), migration among post-settlement populations is believed to be severely constrained

(Cooke et al. In Review). This has three important implications for a life history study.

Firstly, populations of Pacific leaping blenny offer an ideal opportunity to examine the interacting effects of predation and density on age-specific mortality in largely closed post-settlement populations. Secondly, the Pacific leaping blenny also allows an assessment of the extent to which life history variation can occur plastically among populations that experience extensive gene flow (e.g., see Morgans et al. 2014). Thirdly, because larvae are aquatic, factors affecting mortality in larvae are probably consistent for all populations. It can therefore be deduced that any variation in life history among post-settlement populations is directly attributable to variation in age-specific mortality after the transition onto land.

Thesis overview

My thesis aims to examine life history variation in fish, including those which live on

10 land. Specifically this thesis deals with two key questions:

1. Whether survival is inferred to have improved for fish that colonised land

compared to species in aquatic environments.

2. What the relative role of predation or density is in determining life history

variation among populations.

For the first question I compared the life history traits of fish from two families: Gobiidae and Blenniidae. For the gobies I took a broad scale approach, comparing species with either terrestrial or aquatic lifestyles. I then took a fine-scale approach for the blennies, comparing species from a continuum of lifestyles from aquatic to terrestrial. The results of this work form the body of Chapter 2. To answer the second question, I compared indicators of survival (growth) and reproduction (reproductive investment) between populations of the Pacific leaping blenny, and their correlation with predation and density. These results form the body of Chapter 3. In my fourth and final chapter, I summarise my key findings and discuss their broader implications.

Literature Cited

Abrams, P. A., and L. Rowe. 1996. The effects of predation on the age and size of

maturity of prey. Evolution 50:1052-1061.

Andrews, A. H., D. M. Tracey, and M. R. Dunn. 2009. Lead–radium dating of orange

roughy (Hoplostethus atlanticus): validation of a centenarian life span. Canadian

Journal of Fisheries and Aquatic Sciences 66:1130-1140.

Baeck, G., T. Takita, and Y. Yoon. 2008. Lifestyle of Korean mudskipper

Periophthalmus magnuspinnatus with reference to a congeneric species

11 Periophthalmus modestus. Ichthyological Research 55:43-52.

Belk, M. C. 1995. Variation in growth and age at maturity in bluegill sunfish: genetic or

environmental effects? Journal of Fish Biology 47:237-247.

Bhikajee, M., J. Green, and R. Dunbrack. 2006. Life history characteristics of Alticus

monochrus, a supratidal blenny of the southern Indian Ocean. African Zoology

41:1-7.

Brown, C. R., M. S. Gordon, and H. G. Chin. 1991. Field and laboratory observations on

microhabitat selection in the amphibious red sea rockskipper fish, Alticus kirki

(family Blenniidae). Marine Behaviour and Physiology 19:1-13.

Bucholtz, R. H., A. S. Meilvang, T. Cedhagen, J. T. Christensen, and D. J. Macintosh.

2009. Biological observations on the Mudskipper Pseudapocryptes elongatus in

the Mekong Delta, Vietnam. Journal of the World Aquaculture Society 40:711-

723.

Charlesworth, B. 1980. Evolution in age-structured populations. Cambridge University

Press, Cambridge, UK.

Colombini, I., R. Berti, A. Ercolini, A. Nocita, and L. Chelazzi. 1995. Environmental

factors influencing the zonation and activity patterns of a population of

Periophthalmus sobrinus Eggert in a Kenyan mangrove. Journal of Experimental

Marine Biology and Ecology 190:135-149.

Conover, D. O., and S. B. Munch. 2002. Sustaining fisheries yields over evolutionary

time scales. Science 297:94-96.

Cooke, G. M., Schlub, T. E., Sherwin, W. B., Ord, T. J. In Review. Understanding the

spatial scale of genetic connectivity at sea: unique insights from a land fish and a

12 meta-analysis.

Currey, D. R. 1965. An ancient bristlecone pine stand in eastern Nevada. Ecology

46:564-566.

Depczynski, M., and D. R. Bellwood. 2005. Shortest recorded vertebrate lifespan found

in a coral reef fish. Current Biology 15:288-289.

Ernande, B., U. Dieckmann, and M. Heino. 2004. Adaptive changes in harvested

populations: plasticity and evolution of age and size at maturation. Proceedings of

the Royal Society of London, Series B 271:415-423.

Gadgil, M., and W. H. Bossert. 1970. Life historical consequences of natural selection.

The American Naturalist 104:1-24.

Gale, B. H., J. B. Johnson, G. B. Schaalje, and M. C. Belk. 2013. Effects of predation

environment and food availability on somatic growth in the livebearing fish

Brachyrhaphis rhabdophora (Pisces: Poeciliidae). Ecology and Evolution 3:326-

333.

García-Ramos, G. and M. Kirkpatrick. 1997. Genetic models of adaptation and gene flow

in peripheral populations. Evolution 51:21-28.

Gutowsky, L. F. G., and M. G. Fox. 2012. Intra-population variability of life-history traits

and growth during range expansion of the invasive round goby, Neogobius

melanostomus. Fisheries Management and Ecology 19:78-88.

Hendry, A. P., E. B. Taylor, and J. D. McPhail. 2002. Adaptive divergence and the

balance between selection and gene flow: lake and stream stickleback in the misty

system. Evolution 56:1199-1216.

Hsieh, S. T. T. 2010. A locomotor innovation enables water-land transition in a marine

13 fish. PLoS ONE 5:e11197.

Hutchings, J. A. 1993. Adaptive life histories effected by age-specific survival and

growth rate. Ecology 74:673-684.

Hundt, P., Y. Nakamura, and K. Yamaoka. 2014. Diet of combtooth blennies

(Blenniidae) in Kochi and Okinawa, Japan. Ichthyological Research 61:76-82.

Jeffries, M. J., and J. H. Lawton. 1984. Enemy free space and the structure of ecological

communities. Biological Journal of the Linnean Society 23:269-286.

Koslow, J. A., J. Bell, P. Virtue, and D. C. Smith. 1995. Fecundity and its variability in

orange roughy: effects of population density, condition, egg size, and senescence.

Journal of Fish Biology 47:1063-1080.

Kruitwagen, G., I. Nagelkerken, B. R. Lugendo, H. B. Pratap, and S. E. Wendelaar

Bonga. 2007. Influence of morphology and amphibious life-style on the feeding

ecology of the mudskipper Periophthalmus argentilineatus. Journal of Fish

Biology 71:39-52.

Lanner, R. M., and K. F. Connor. 2001. Does bristlecone pine senesce? Experimental

Gerontology 36:675-685.

Law, R. 1979. Optimal life histories under age-specific predation. The American

Naturalist 114:399-417.

Lenormand, T. 2002. Gene flow and the limits to natural selection. Trends in Ecology &

Evolution 17:183-189.

Lister, B. C. 1976. The nature of niche expansion in West Indian Anolis lizards I:

ecological consequences of reduced competition. Evolution 30:659-676.

Losos, J. B., and D. L. Mahler. 2010. 'Adaptive radiation: the interaction of ecological

14 opportunity, adaptation, and speciation' in Evolution Since Darwin: The First 150

Years. Sunderland, MA, Sinauer Associates.

Martin, K. L. M. 1995. Time and tide wait for no fish: intertidal fishes out of water.

Environmental Biology of Fishes 44:165-181.

Martin, K. L. M., and J. R. B. Lighton. 1989. Aerial CO2 and O2 exchange during

terrestrial activity in an amphibious fish, Alticus kirki (Blenniidae). Copeia

1989:723-727.

Michod, R. E. 1979. Evolution of life histories in response to age-specific mortality

factors. The American Naturalist 113:531-550.

Morgans, C. L. and T. J. Ord. 2013. Natural selection in novel environments: predation

selects for background matching in the body colour of a land fish. Animal

Behaviour 86:1241-1249.

Morgans, C. L., Cooke, G. M., Ord, T. J. 2014. How populations differentiate despite

gene flow: sexual and natural selection drive phenotypic divergence within a land

fish, the Pacific leaping blenny. BMC Evolutionary Biology 14:97.

Murdy, E. O. 1989. A taxonomic revision and cladistic analysis of the oxudercine gobies

(Gobiidae: Oxudercinae), Australian Museum.

Ord, T. J., and S. Tonia Hsieh. 2011. A highly social, land-dwelling fish defends

territories in a constantly fluctuating environment. Ethology 117:918-927.

Patzner, R. A., E. J. Gonçalves, P. Hastings, and B. G. Kapoor. 2009. The biology of

blennies. United States of America, Science Publishers.

Pianka, E. R., and W. S. Parker. 1975. Age-specific reproductive tactics. The American

Naturalist 109:453-464.

15 Polgar, G. 2012. Ecology and evolution of mudskippers and oxudercine gobies

(Gobiidae: Oxudercinae): perspectives and possible research directions. Pp 117 –

137. In Sasekumar A., Chong, V. C. Mangrove and coastal environment of

Selangor, Malaysia. University of Malaya, Kuala Lumpur.

Polgar, G., A. Sacchetti, and P. Galli. 2010. Differentiation and adaptive radiation of

amphibious gobies (Gobiidae: Oxudercinae) in semi-terrestrial habitats. Journal of

Fish Biology 77:1645-1664.

Primack, R. B. 1979. Reproductive effort in annual and perennial species of Plantago

(Plantaginaceae). The American Naturalist 114:51-62.

Promislow, D. E. L., and P. H. Harvey. 1990. Living fast and dying young: a comparative

analysis of life-history variation among mammals. Journal of Zoology 220:417-

437.

Reznick, D., and A. P. Yang. 1993. The influence of fluctuating resources on life history:

patterns of allocation and plasticity in female guppies. Ecology 74:2011-2019.

Reznick, D., M. J. Bryant, and F. Bashey. 2002. R- and K-selection revisited: the role of

population regulation in life-history evolution. Ecology 83:1509-1520.

Reznick, D. N., R. D. Bassar, J. Travis, and F. H. Rodd. 2012. Life-history evolution in

guppies VIII: the demographics of density regulation in guppies (Poecilia

reticuata). Evolution 66:2903-2915.

Riesch, R., R. A. Martin, and R. B. Langerhans. 2013. Predation's role in life-history

evolution of a livebearing fish and a test of the Trexler-DeAngelis model of

maternal provisioning. The American Naturalist 181:78-93.

Roff, D. A. 1992. The evolution of life histories: theory and analysis. Chapman and Hall

16 Incorporated, New York, NY.

Sayer, M. D. J., and J. Davenport. 1991. Amphibious fish: why do they leave water?

Reviews in Fish Biology and Fisheries 1:159-181.

Schroder, A., A. van Leeuwen, and T. C. Cameron. 2014. When less is more: positive

population-level effects of mortality. Trends in Ecology and Evolution 29:614-

624.

Shimizu, N., Y. Sakai, H. Hashimoto, and K. Gushima. 2006. Terrestrial reproduction by

the air-breathing fish Andamia tetradactyla (Pisces; Blenniidae) on supralittoral

reefs. Journal of Zoology 269:357-364.

Smith, C. C., and S. D. Fretwell. 1974. The optimal balance between size and number of

offspring. The American Naturalist 108:499-506.

Stearns, S. C. 1976. Life-history tactics: a review of the ideas. The Quarterly Review of

Biology 51:3-47.

Stearns, S. C. 1992. The evolution of life histories. Oxford University Press, Oxford.

Stibor, H. 1992. Predator induced life-history shifts in a freshwater cladoceran. Oecologia

92:162-165.

Takeda, T., M. Hayashi, A. Toba, K. Soyano, and A. Ishimatsu. 2011. Ecology of the

Australian mudskipper Periophthalmus minutus, an amphibious fish inhabiting a

mudflat in the highest intertidal zone. Australian Journal of Zoology 59:312-320.

Taylor, H. M., R. S. Gourley, C. E. Lawrence, and R. S. Kaplan. 1974. Natural selection

of life history attributes: an analytical approach. Theoretical Population Biology

5:104-122.

Torres Dowdall, J., C. A. Handelsman, E. W. Ruell, S. K. Auer, D. N. Reznick, and C. K.

17 Ghalambor. 2012. Fine-scale local adaptation in life histories along a continuous

environmental gradient in Trinidadian guppies. Functional Ecology 26:616-627.

Walsh, M. R., and D. N. Reznick. 2008. Interactions between the direct and indirect

effects of predators determine life history evolution in a killifish. Proceedings of

the National Academy of Sciences 105:594-599.

Wellborn, G. A. 1994. Size-biased predation and prey life histories: a comparative study

of freshwater amphipod populations. Ecology 75:2104-2117.

Wilson, S. 2001. Multiscale habitat associations of detrivorous blennies (Blenniidae:

Salariini). Coral Reefs 20:245-251.

———. 2004. Growth, mortality and turnover rates of a small detritivorous fish. Marine

Ecology Progress Series 284:253-259.

Wilson, E. O. 1961. The nature of the taxon cycle in the melanesian ant fauna. The

American Naturalist 95:169-193.

Zera, A. J., and L. G. Harshman. 2001. The physiology of life history trade-offs in

animals. Annual Review of Ecology and Systematics 32:95-126.

18

CHAPTER 2

Predictable life history changes reveal enhanced survival for

marine fish living on land

Edward R. M. Platt, Ashley M. Fowler & Terry J. Ord

ERMP and TJO designed the study, ERMP collected and processed specimens with the assistance of AMF, ERMP analysed the data and wrote the manuscript under the supervision

and input of TJO and AMF.

A version of this paper has been submitted to The American Naturalist

19

Abstract

An ecological “release” or new ecological “opportunity” are two hypotheses for why organisms might colonise a novel habitat. In both cases, the chances of survival presumably increase following the transition into the new environment. Life history theory makes explicit predictions about how traits linked to survival and reproduction should change with shifts in age-specific survival, for example following the colonisation of a new environment. One of the most dramatic ecological transitions in the history of life was the colonisation of land by fish. Why fish made such a drastic transition in environment is unclear, but an ecological release or opportunity are likely explanations. Using predictions from life history theory, I tested whether survival was inferred to have increased following the colonisation of land in two contemporary fish families that have independently made the transition onto land:

Gobiidae and Blenniidae. I examined variation in growth and various aspects of reproductive investment among terrestrial and aquatic species. In both families I found that differences varied according to the level of independence from water, consistent with survival improving for certain age classes on land. Nevertheless, the details of life history change differed in each family, with the greatest increases in survival implied for early age classes in Blenniidae, but older age classes in Gobiidae. This suggests fundamental differences in the way the colonisation of land occurred in each family. Generally, though, my findings are consistent with the notion that an ecological release or ecological opportunity prompted fish to make the transition onto land. My study also shows how theoretical predictions and empirical investigations of life history variation can reveal important insights into the ecology of species that might otherwise remain hidden.

20

Introduction

It is now well established that the colonisation of novel habitats can be a vital ingredient for adaptation and potential speciation. This is because new environments frequently result in new selection pressures that can stimulate adaptive differentiation among populations (e.g.,

Munzing 1963; Rainey & Travisano 1998; Kornfield & Smith 2000; Streelman & Danley

2003; Rundle & Schluter 2004; Arbogast et al. 2006). However, why organisms should make the transition into a new environment in the first place—especially one that is drastically different in its biotic or abiotic conditions—remains an open question. At a general level, ecological “release” or ecological “opportunity” are two hypotheses for why an organism might colonise a novel habitat: the new environment offers reduced predation or competition for resources (an ecological release) or the new environment offers new resources that are absent or rare in the ancestral environment (an ecological opportunity; Wilson 1961; Lister

1976; Jeffries & Lawton 1984; Losos & Mahler 2010). In both cases, the new environment offers an organism a better chance of survival.

One of the most dramatic ecological transitions that has occurred in the history of life was the colonisation of land from an aquatic environment. In the case of vertebrates, the transition by fish from water onto land in the late Devonian was a critical evolutionary event that subsequently led to the evolution of all land animals in existence today (Laurin 2010).

Yet we have little idea why these early colonisers first ventured onto land. Ecological release and ecological opportunity seem probable explanations (Wilson 1961; Lister 1976; Jeffries &

Lawton 1984; Losos & Mahler 2010), but testing these hypotheses on extinct species is virtually impossible. However, there are a number of living examples of fish that have colonised—or are in the process of—colonising land (Sayer & Davenport 1991). Although direct parallels between living fish and the first vertebrate land invaders from the late

Devonian are difficult, studies on contemporary amphibious fishes have nevertheless helped

21

reveal the general challenges that follow the transition from water to land (Martin 1995;

Graham & Lee 2004; Sayer 2005). However, why these present-day fish that are clearly so well adapted to one environment (water) should make the shift into such a vastly different environment (land) remains unclear (Sayer & Davenport 1991; Vermeij & Dudley 2000;

Graham & Lee 2004). The emergence of amphibious behaviour in fish from closed water bodies (e.g., ponds, small lakes) might be explained by abiotic factors such as adverse fluctuations in water temperature or chemistry that push fish to temporarily strand themselves on land (Sayer & Davenport 1991). However, such abiotic factors are unlikely for species from open water environments such as marine fish. Instead, the best explanations centre on factors relating to a release from predation or competition or some new opportunity found only on land (Sayer & Davenport 1991).

If ecological release or ecological opportunity prompted present-day fish to make the transition onto land, then this should have resulted in increased survival in land fish compared to their marine relatives. Quantifying changes in survival directly is difficult without long-term study, but its signature should be apparent in differences in life history among taxa that can be measured over the short-term. In particular, there is an extensive body of life history theory that predicts how the probability of survival should impact the allocation of resources by an organism to characteristics that promote improved survival (e.g., growth) over current reproduction (e.g., offspring production; Gadgil & Bossert 1970; Stearns 1976;

Law 1979; Michod 1979; Stearns 1992). By comparing how investment in these life history characteristics has been made by different species, it should be possible to infer whether the colonisation of land might have improved survival and how it may have done so.

Life history theory works on the assumption that organisms have a fixed amount of energy at any one point in time that must be split between characteristics that increase survival or reproduction (Gadgil & Bossert 1970). For example, allocating resources to

22

reproduction will necessarily divert resources away from growth (Stearns 1992). Natural selection is expected to generate a balance in this allocation of resources that maximises the cumulative fitness of an organism over its lifetime (i.e., the total number of offspring produced that survive to reproduce; Smith & Fretwell 1974). Theory shows that this optimum balance in energy allocation depends on the probability of survival at different stages of an organism’s lifetime (Gadgil & Bossert 1970; Pianka & Parker 1975; Law 1979). Investment in current reproduction will also come at the expense of future survival (and consequently future reproduction; Pianka & Parker 1975). This leads to several predictions for how investment in reproduction (current versus future) should change because the chance of survival at a given age shifts for individuals within a population or species (Gadgil & Bossert

1970; Law 1979; Michod 1979; Hutchings 1993).

Age-dependent survival takes two general forms (Gadgil & Bossert 1970; Taylor et al. 1974; Law 1979; Michod 1979; Hutchings 1993): (i) survival declines with age (e.g., the probability of survival is higher for younger individuals compared to older individuals in the population); or (ii) survival increases with age (e.g., mortality is higher in younger individuals compared to older individuals). A third scenario is also possible in which survival is not age-dependent and all age classes have an equal probability of survival. Table 1 summarises the expected change in life history under each of these three scenarios because the probability of survival changes across age classes. In the context of the colonisation of land by fish, Table 1 outlines how life history traits would be expected to change across age classes as species move from an aquatic to terrestrial environment. Note, these predictions are based on the explicit assumption that some type of ecological release or ecological opportunity followed the shift onto land and as a result survival improved for the colonising individuals. If survival did not improve subsequent to the shift onto land, then life history traits are expected to be relatively consistent among species, or else variable but unrelated to

23

differences in habitat.

In general, theory predicts that reproduction should be concentrated in age classes that experience increased survivorship (Michod 1979). For example, if improvements in survival following the colonisation of land were largely confined to younger age classes (Scenario 1;

Table 1), then land species should concentrate reproduction in these earlier life stages and emphasise current reproduction over future reproduction (high reproductive effort) at the compromise of growth rate (Gadgil & Bossert 1970; Pianka & Parker 1975; Law 1979;

Michod 1979; Hutchings 1993). In contrast, if improvements in survival were largely confined to older age classes in the population (Scenario 2; Table 1), then investment in reproduction should be concentrated in older individuals and growth rate should generally be higher (Gadgil & Bossert 1970; Law 1979; Michod 1979; Hutchings 1993). This emphasis on growth occurs because an individual would do better to invest more in growth and to delay reproduction to later in life to increase the chances of surviving through the earlier (high mortality) life stages (Gadgil & Bossert 1970; Pianka & Parker 1975; Law 1979).

It is also possible—although probably rare in nature—that survival following the colonisation of land increased uniformly for all age classes (Scenario 3; Table 1; Law 1979).

In this scenario, the optimum distribution of reproductive effort (emphasis on current or future reproduction) should not have changed among marine and land species, and life history strategies would be broadly consistent among species (Gadgil & Bossert 1970; Taylor et al.

1974; Law 1979; Michod 1979). Here, the colonisation of land still might have been prompted by some improvement to survival, but there would be no way to distinguish this scenario from a situation in which the colonisation of land had little to do with an ecological release or opportunity (e.g., colonisation occurred by chance and neither increased nor decreased survival). The only circumstance where life history would differ among species would be where all age classes had increased survival on land specifically because of

24

Table 1: Predicted changes to age-specific survival and life history following the colonisation of land by fish

Scenario Expected change in life history Reproductive effort Size/age of maturity Growth rate Egg number Egg Size 1 Early age survival Increases Earlier Decreases No change, or No change, or increases relative to increase if juvenile decrease if juvenile older ages survival increases survival increases relative to adults relative to adults 2 Older age survival Decreases Later Increases No change, or No change or

25 increases relative to decrease if juvenile increase if juvenile

early ages survival decreases survival decreases relative to adults relative to adults 3 Survival increases No change, or No change, or earlier No change Increases Decreases uniformly across age increase if survival if survival increases classes increases because of because of resource resource availability availability

greater resource availability. If so, this should have resulted in higher reproductive effort at all ages in land species, with little or no change in growth rate between marine and land

(Gadgil & Bossert 1970).

In addition to the broad trade-off between reproduction and survival, there is another classic life history trade-off between the size and number of offspring produced. In general, an increase in offspring size will improve its chance of survival (Smith & Fretwell 1974), but offspring size comes at the expense of the number of offspring that can be produced (Smith &

Fretwell 1974). How offspring size improves an individual offspring’s chance of survival is predicted to be a result of either improving competitive ability, or by reducing the duration of time spent in a vulnerable life history stage (Smith & Fretwell 1974; Brockelman 1975; Sibly

& Calow 1983; Hutchings 1991). If the ecological release or opportunity associated with the transition onto land improved early age survival (Scenario 1; Table 1), either by reducing competition, predation, or increasing resource availability, then those species on land should produce a higher number of offspring at the cost of size (Smith & Fretwell 1974; Sibly &

Calow 1983). A matching outcome is expected if the move onto land improved survival across all age classes (Scenario 3; Table 1), because the absolute value of early age survival should still increase. These expectations are based on the assumption that higher resources, or lower competition or predation, should reduce the selective pressure for competitive ability in offspring, increasing the benefits of having a higher number of offspring at the expense of offspring size (Smith & Fretwell 1974; Brockelman 1975; Sibly & Calow 1983; Hutchings

1991).

Finally, it is possible that mortality changed after the initial shift onto land, of which the most likely change is an increase in mortality (e.g., as predators adapted to exploit a new prey resource). An increase in mortality could have the effect of increasing or decreasing the relative mortality of younger versus older age classes, or else leave the relative mortality of

26

age classes unchanged while increasing mortality overall. While it seems unlikely that changes in mortality after the shift onto land were greater than those which resulted from colonising a new habitat, it is possible they have contributed to the life history variation in terrestrial species. If their contribution was great enough, these changes may have reduced the degree to which the life history traits of terrestrial species reflect reductions in mortality that occurred as a result of moving onto land. They may also explain any findings in which the life history traits of terrestrial species do not match those predicted from variation across species with aquatic and amphibious behaviours (e.g., see Blennies below). Consequently, the possibility that mortality significantly increased following the shift onto land, while unlikely, is an important caveat to be considered when interpreting the results.

Using the predictions from life history theory presented in Table 1, I evaluated the evidence for improved survival in land fish relative to their marine relatives from two families: Gobiidae and Blenniidae. These two families have experienced extensive ecological and species diversification (Patzner et al. 2009; Thacker & Roje 2011; Near et al. 2012). Of special relevance to my study, there are select species in both families that exhibit independent transitions from marine to terrestrial environments. In the family Gobiidae, this transition has occurred in the iconic group of “mudskippers” (Oxudercinae; Murdy 1989).

Although mudskippers are highly amphibious relative to their marine relatives, their terrestrial behaviour is surpassed by several species in the family Blenniidae known as

“rockskippers” (tribe Salariini; Bhikajee et al. 2006; Patzner et al. 2009). These land blennies are exceptional in their degree of territoriality: land blennies are highly agile on land (Hsieh

2010), rarely return to water voluntarily and spend the vast majority of their adult life out on the rocks in the splash zone above the waterline (Ord & Hsieh 2011). Furthermore, the blenny family as a whole offer the full spectrum of lifestyles, from species occupying the ancestral marine environment, to transitional amphibious species that spend varying degrees

27

of their daily life in the water and out on land, to fully terrestrial species that rarely return to water (Patzner et al. 2009; Hsieh 2010; Ord & Hsieh 2011; Morgans & Ord 2013). Taken together, these two families offer a replicated opportunity to examine how life history traits change following the colonisation of land by fish, and the extent that this remarkable ecological transition might have been promoted by an ecological release or new ecological opportunities on land.

To this end, I collected information from the literature (Gobiidae) or directly from fresh specimens caught in the field (Blennidae) on reproductive investment (the relative weight of reproductive versus somatic tissue), growth rate, egg number and egg size. These data were then used to evaluate changes in life history strategy among species following the progressive colonisation of land (Table 1). Life history changes between marine and land species that matched (or contradicted) those outlined in Table 1 would provide strong support for (or question) the expected improvements in survival following the colonisation of land by fish. Such changes in life history would be broadly consistent with the notion that fish on land experienced a release from predation or competition or were provided with new ecological opportunities on land (although it would not distinguish between these causal factors). If life history characteristics were broadly consistent among species or variable among species but unrelated to shifts in habitat, then it would be unlikely that fish made the transition onto land because of the improved survival conveyed by an ecological release or ecological opportunity.

28

Materials and Methods

Gobiidae

Literature Searches

A broad literature search was conducted using the Web of Sciences and Google Scholar with variations of the search term goby (“gobiidae”, “gobies”, “goby”) combined with one of the following terms “growth”, “age”, “reproduction”, “GSI”, “gonads”, “eggs”, or “spawning”

(searches were completed on 27th of September 2013). Searches were restricted to marine species because life history characteristics were likely to differ in freshwater species, as were the causes of any terrestrial behavior developed therein (Sayer & Davenport 1991; see introduction). Additional articles were also found by going through the reference lists of journal articles found by electronic searches.

Data Collection

Of the 83 research articles found, 68 had measured one or more of the desired traits through a shared, and thus comparable, method. These traits were body length (measured as the standard or total length of the largest individual reported by a study for a given species), growth rate (using the curvature (K) and asymptote (L∞) parameters of von bertalanffy growth models), reproductive investment (gonadosomatic index which is the weight of reproductive tissue relative to the body), and egg/oocyte size or egg/oocyte number of mature females. Details on each of these variables are given below. I also classified whether species were tropical, subtropical, or temperate depending on the location of where species were studied because life history may vary by latitude (especially tropical or subtropical versus temperate). The lifestyle of species was classified as aquatic (coded as 0) or terrestrial (1) depending on whether the species exhibited a significant degree of amphibious behavior, such as commonly spending time out of water. In instances where more than one study

29

reported information on the same life history characteristic, I took a weighted average across studies. All life history data were natural log-transformed prior to analyses unless stated otherwise.

Growth rate: From Von Bertalanffy growth models, I converted the reported K and L∞ parameters into a growth performance index (GPI), which overcomes complications with comparing growth among taxa that might differ in the shape of their growth curves or size range (Pauly 1979; Brey 1999). GPI was calculated using formula presented in Moreau et al.

(1986; see also Pauly 1991). L∞ was first changed into millimeters if not already presented in that format in the original study. GPI was not natural log transformed for analyses. Note all species from the study Hernaman & Munday 2005 were removed prior to analysis as they were clear outliers (Fig. A1).

Reproductive effort: I used the highest mean gonadosomatic index (GSI) value reported for a species as an estimate of reproductive effort. GSI values were only taken from those studies that reported GSI over several months, with the highest GSI value representing the spike in reproductive investment that occurs during the spawning period. There are a number of different methods for computing GSI and I used values calculated by the two most common methods (gonad weight/total body weight or gonad weight/somatic body weight) and included a factor code (0 or 1, respectively) for each method in linear models.

Egg number and size: Egg number analyses compared batch fecundity; i.e., the number of eggs released in a single spawn (Duponchelle et al. 2008). This number can be attained by counting either the number of oocytes in a gonad prior to spawning or number of eggs that were released at spawning. For changes relating to lifestyle, I took the maximum estimate

30

from either pre- or post-spawning depending on whichever value was larger. For correlations between egg number and size (see below), pre- and post-spawning data were treated separately.

Estimates of egg size were based on the two dimensional area reported by a study, and was either the mean or maximum egg area for a species. In analyses, a factor code for mean

(0) or maximum (1) egg size was included in the model.

Statistical Analyses

Growth rate, GSI, egg size and egg number were each tested separately in a linear model with factors for lifestyle, species total length and their interaction using the program R version 3.0.1 (R Development Core Team). These analyses were then repeated on the subset of tropical and subtropical species to confirm relationships were not confounded by climate.

Note, small sample sizes prevented the separate analyses of tropical and subtropical species to further investigate any confounding effects of climate. Species body length was included in these analyses as a covariate to control for differences in mean body size of the aquatic and terrestrial groups. All length measurements were natural log transformed, and where analyses combined total and standard lengths mean-centering was performed to standardize both measures (all analyses except GSI; NB: although total lengths can be expected to be larger than standard lengths—which is removed by mean-centering—I had no reason to expect the variance across species to differ between the two types of measurements; hence why data was mean-centered rather than a conversed to z-scores).

Finally, the trade-off between growth and reproduction was examined by comparing natural log transformed GSI values against the growth performance index in a Pearson correlation analysis in R. The trade-off predicted between offspring size and number was also tested by correlating egg number and egg size for both pre- and post-spawning estimates. To

31

control for species body size effects, I first regressed egg data on body length across species and computed the residuals. These “size-free” residuals of egg number and egg size were then used in subsequent Pearson correlation analyses.

Blenniidae

Specimen Collection

Sampling was conducted at various locations around the island of Guam in July 2012: Pago

Bay (13o25’39”N, 144o47’56”E), Taga’chang (13o24’16”N, 144o46’53”E), Talofofo

(13o20’34”N, 144o46’21”E), Umatac (13o17’40”N, 144o39’29”E), Inarajan (13o16’18”N,

144o44’53”E), Tanguisson (13o32’43”N, 144o48’32”E), and the Adelup Point (13o28’52”N,

144o43’44”E). This time of year is known to overlap with the breeding period for blennies

(Carlson 1992; Sunobe et al. 1995; Shimizu et al. 2006).

Collections were made along transects that traversed the water-land interface from below the waterline in the intertidal zone to above the waterline in rock pools and exposed rocks in the splash zone. Specimens representative of a broad range of body sizes for five blenny species were collected, each of which was classified as marine (Istiblennius lineatus), mildly amphibious (Praealticus labrovittatus and Praealticus poptae), highly amphibious

(Entomacrodus striatus), or terrestrial (Alticus arnoldorum) depending on the proportion of individuals caught in or out of the water (e.g., marine species were always found below the waterline, while the terrestrial species was always found out of water on the rocks above the waterline). Additionally, a sixth species, the marine Istiblennius edentulus, was collected opportunistically, but was not a main focus of this study.

Fish were euthanized and then preserved in 10% neutral buffered formalin. Prior to preservation the length and weight of all A. arnoldorum specimens was recorded for the purpose of evaluating subsequent shrinkage from formalin. These measurements indicated a

32

small amount of shrinkage did occur (~2 %) but importantly this was consistent across body sizes and was therefore unlikely to impact my life history estimates compared among differ sized species. Sample sizes were based on a preliminary work conducted on a large sample of

A. arnoldorum. This work indicated variance in the life history data was low, and it was possible to obtain very similar results with lower sample sizes provided included the full spectrum of body sizes. Subsequently, I used this preliminary work to establish minimum sample sizes for the current study. Note, sample sizes for each life history characteristic also varied depending on the availability of specimens, and the time required to prepare specimens for certain types of measurement.

Body Measurements and Dissections

Standard length, body weight, and gonad weight were recorded for female specimens only.

Standard length was measured as the length (mm) from the front of the top jaw to the caudal perduncle using digital calipers (Traceable Digital Caliper, Control Company, >0.03 mm accuracy, 0.01 mm resolution). Body weight was measured using a semi-micro analytical balance (A&D Company, Ltd, model GR-200, e=1 mg, d=0.1 mg) after all surface moisture had been removed from the fish. Specimens were then dissected to confirm sex and to estimate approximate level of maturity based on the general appearance of the gonads.

Gonads were then removed, dried of all surface moisture, and weighed to the nearest mg using the semi-micro analytical balance. In the case of very small gonads (< 0.002 g), a Cahn c-33 microbalance was used to obtain a weight to nearest 10 µg. Gonads were then stored in

10 % buffered formalin and later used for oocyte analysis. To age fish, saggital otoliths were removed and stored dry in microtubes.

33

Determining fish age

Age (in days) was determined by counting the number of microincrements visible in longitudinal sections of the right sagittal otolith. Microincrements appear as alternating light and dark bands, and are known to directly reflect daily cycles of otolith deposition for nearly all fish species examined (Brothers et al. 1976; Campana and Neilson 1985; Morales-Nin

2000), including blennies (Hernandez-Miranda et al. 2009; Mansur et al. 2013; Carvalho et al. 2014). While the species in the present study have not yet been confirmed to deposit microincrements on a daily basis, their close relatedness should insure a relatively consistent rate of deposition amongst them. Note too, there is some evidence that microincrement deposition still occurs on a daily basis in fish species which spend time out of water

(Bhikajee et al. 2006; Kruitwagen et al. 2006; Mazlan & Rohaya 2008). Sagittal sections were prepared by mounting the otolith on a microscope slide using a thermoplastic glue.

Otoliths were then sanded using 3, 9, and sometimes 12 micron grade lapping film. During this process the thermoplastic glue was normally re-heated to free the otolith so it could be flipped and sanded on alternate sides. The sanded otolith was then treated with cedar oil and left for a minimum of five minutes to increase its transparency prior to counting microincrements. Counts were made at 600x magnification using the oil iris of an Olympus

BX50 microscope mounted with a Spot Flex microscope camera (Diagnostic Instruments,

Inc. model #15.2 64 Mp Shifting Pixel) connected to a computer running the program

ImagePro Plus version 5.1 (Media Cybernetics Inc.). Because of the curvature of the otolith, microincrement counts could not be made from a single image. Instead, a series of images were taken from the core to the post-rostrum or in some cases the rostrum. These images were then landmarked based on physical features on the otolith such that successive microincrement counts could be made from one image to the next to obtain a total microincrement count across the entire otolith. Total counts were taken repeatedly until a

34

roughly consistent number was attained (counts differing by no more than ~3 microincrements), which were then averaged to provide a single estimate of age for a given fish. In rare cases, a consistent count could not be obtained and these individuals were excluded from subsequently analyses of age.

Quantifying oocyte size and number

Oocyte size and number was measured for at least 14 females per species, and only one of the two ovaries used where possible. Using the dissecting microscope, oocytes were separated from the ovarian tissue in a petri dish using a probe and combined with dilute soapy water to keep oocytes separated. Oocyte development stage was then estimated based on a subsample of oocytes representing every size and phenotype found in the ovum. These oocytes were photographed through the dissecting microscope at 63x magnification and the images used to classify oocytes into one of five stages (see criteria outline in Table A1; Selman et al. 1993;

Tyler & Sumpter 1996; Clelland et al. 2007; Nunez & Duponchelle 2009): Stage I (primary growth), Stage II (cortical alveolus stage), Stage III (vitellogenesis), and Stage IV (a combination of maturation/mature), and Stage V (atresia; almost no oocytes were found in this stage). To determine the size and number of all the oocytes, ImagePro Plus software was used to first filter out non-oocyte objects, such as ovarian tissue. This filtering was determined by visually adjusting settings in ImagePro Plus to ensure exclusion of non-oocyte material was maximized without inadvertent loss of actual oocytes from subsequent estimates. After filtering, the number and mean diameter of each detected object was inferred to represent the number and size of oocytes found in the ovum.

To determine the number and size of oocytes for each developmental stage, data were plotted as a frequency histogram and evaluated visually for distinct peaks and troughs in the data to identify oocyte ‘cohorts’ of different size classes and that presumably reflected

35

oocytes of different developmental stages (see Fig. A2 and A3 for examples). To confirm what the specific developmental stage was, the corresponding size range for a given oocyte cohort was cross-referenced with the size range of oocytes for that female that had been classified by development stage (see above). The total number and mean diameter of oocytes in each cohort was subsequently used to estimate the number and size of oocyte for each developmental stage. Where oocytes were only removed from one ovum, total oocyte number for a female was estimated by calculating the number of oocytes per unit mass of the ovum examined, multiplied by the total ovarian weight (i.e., the combined weight of both ovaries).

Finally, all females for a species were categorized by the probable latest oocyte stage present in their ovaries. This was done either by direct evaluation of oocytes as described above or, for those females for which ovaries were not examined in detail, based on detailed descriptions of ovaries noted during the initial dissection of the gonads. This latter classification was validated by cross-referencing ovary descriptions to females that had been subject to detailed oocyte analysis (see above) to confirm development stage.

Statistical Analyses

Statistical analyses were performed by running multivariate regressions using the lme4 package, version 0.99999911-8 (Bates 2008) in R.

Growth rate: Growth for each species was initially assessed separately to determine the best model for describing growth rate for each species. This was done by regressing standard length on total microincrement count (age) using several different regression functions

(linear, ln-linear, quadratic and power). The power function was consistently found to be the best fitting model for all species (based on computed r2 values). Data on juveniles, immature females, and females for all species was subsequently entered into a multivariate power

36

function with a random effect for species and fixed effects for age, lifestyle (an index ranging from 0 to 3; aquatic to fully terrestrial) and their interaction.

Reproductive investment: A linear regression of natural log-transformed ovarian weight on natural log-transformed standard length was used to quantify the relative investment of tissue to reproduction. This analysis was performed using data on juveniles, immature females and females and provided an estimate of the proportion of body weight dedicated to reproductive tissue (represented by the intercept value) and the rate at which reproductive tissue developed

(comparable to the onset of sexual maturity; represented by the slope value). All species were entered into this regression model with a random effect for species and fixed effects for standard length, lifestyle and their interaction.

To assess the impact of any difference in the maturity of individuals that happened to be surveyed for a given species, a second more complex model was applied that included an additional random effect for the estimated maturity of each female (based on the probable latest developmental stage of oocytes in the reproductive tract).

Oocyte size and number: Total number of eggs likely to be laid during the breeding season was measured as the largest number of oocytes at any development stage found in the ovaries for an individual. This removed the potential confound on egg number that might occur if females had already begun to lay eggs prior to being caught, and also controlled for differences in the reproductive stage of females that might occur among species (due to seasonal effects). Oocyte number was natural log-transformed and data for all species were entered into a regression model with a random effect for species and fixed effects for standard length (natural log-transformed), lifestyle (see Table 3) and their interaction. In this model, intercept values corresponded to the relative number of eggs produced on average

37

among species, while slope values measured the relative rate of egg production among species.

Egg size on laying was determined by modeling the trajectory of oocyte development within females using a regression of natural log-transformed oocyte size on egg development stage. Here, all species were entered into a single regression with a random effect for species and fixed effects for egg development stage, standard length and lifestyle, within separate interactions between lifestyle and egg stage and length. In this model, intercept values measured the relative size of oocytes among species, while the slope value of egg stage measured the rate of development of eggs. Differences in oocyte size and development by body size were controlled for by the covariate of standard length entered into the model.

Given that standard length was not found to be a significant predictor of oocyte size or development, a second analysis was performed with standard length removed to provide more accurate parameter estimates of the relationship between lifestyle and egg size and development.

Life history trade-offs: Finally, the trade-off predicted by life history theory between growth and reproduction (Stearns 1992) and between offspring size and number (Smith & Fretwell

1974), was examined by correlating the intercept values computed from regression models described above using a Pearson correlation in R (NB: these intercept values controlled for body size and differences in maturity among species).

38

Results

Gobiidae

Life History Variation as a Function of Lifestyle

Growth: There was a significant interaction between species mean total length and lifestyle implying growth rates may have increased substantially as a function of body size in terrestrial species compared to aquatic species (Fig. 1A; Table 2A). For example, the largest terrestrial species exhibited significantly faster growth rates compared to similar sized aquatic species. This relationship was qualitatively unchanged in analyses that excluded temperate species (Table A2; NB: this second analysis had a greatly reduced sample size, so interpretation of statistical significance is confounded by low power, but parameter estimates in the model were remarkably similar to those in Table 2A as were the pattern of relationships in non-temperature species highlighted in Fig. 1A). This suggests that growth rate may have increased in goby fishes following the transition from an aquatic to terrestrial environment.

Reproduction: Reproductive investment did not differ between fully aquatic and terrestrial gobies (Table 2B). Clearly the power of this analysis was very low given data were only available for three terrestrial species (Fig. 1B, which also prevented analysis of reproductive investment that excluded temperate species). However, it was interesting to note that the terrestrial species sat below or on the regression line, which becomes relevant for assessing the relationship between growth and reproduction (see below).

Egg Number and Size: Analyses of egg number did not reveal a significant effect of lifestyle

(Table 2C), which was exclusively predicted by the body size of species (Fig. 1C). When the analysis included only tropical and subtropical species, the positive relationship between egg

39

number and body size remained, but lifestyle subsequently became a significant predictor of egg number, such that terrestrial species appeared to produce more eggs for their given body size than aquatic species (Table A2B). Inspection of the plot comparing only tropical and subtropical species provides some visual support for this pattern as well, but the trend seems weak dependent on sample size (Fig. A4). No mean egg size values were available for temperate species, and as a result all analyses were performed using only tropical and subtropical species. Interestingly the results for egg number found lifestyle to be a significant predictor of egg size (Table 2D). When visually inspected, a difference can be observed in the distribution of aquatic and terrestrial species, possibly indicating that egg size changed with body length for aquatic species but not terrestrial species (Fig. 1D). It is also suggestive that egg size may be larger for terrestrial species. Interestingly the number of eggs produced by terrestrial species was not any lower than that for aquatics as predicted by the theory (Fig.

1C, and terrestrial species might actually produce more eggs compared to aquatics; Table

A2B and Fig. A4).

Given the low sample size for terrestrial species in analyses of both egg number and size (and GSI for that matter), interpretation of trends should be made cautiously.

Trade-offs Between Life history characteristics

Growth versus reproduction: A significant negative correlation between reproduction and growth was found for goby fishes (Pearson correlation: N = 11, r = -.56, p = .035 one-tailed;

Fig. 2A). However, this relationship disappeared when comparing size-free residuals (N = 11, r = -.003, p = .495 one-tailed; Fig. 2B), but interpretation of this result is complicated by the differences in the effect of body size on GSI and growth (Table 2A and B). In the case of raw estimates of reproduction and growth (Fig. 2A), the distribution of aquatic and terrestrial species along the trend line implies contrasting strategies: terrestrial species seem to invest

40

more in growth rate at the expensive of reproduction, whereas the opposite seems to occur in aquatic species. While a significant positive correlation was found between growth rate and lifestyle (Table 2A), results for reproductive effort were equivocal given the small number of terrestrial species with GSI data. However, at least two of the three terrestrial species had lower than expected GSI for their given body size (Fig. 1B). If the negative relationship between growth and reproduction is believed (Fig. 2A) and is considered together with results on growth rate as it relates to lifestyle, it is possible that terrestrial gobies may actually invest less in reproduction compared to aquatic species.

Egg number versus egg size: There was a prominent negative correlation between the number of oocytes and size of oocytes found in ovaries prior to laying (Pearson correlation: N = 12, r

= -.76, p < 0.005 one-tailed), which was also reflected between the number and size of eggs post-laying despite the acute reduction in sample size (Pearson correlation: N = 5, r = -.82, p

= 0.045 one-tailed; Fig. 2D). There was not enough data on terrestrial species to make any interpret of differences in egg production between terrestrial and aquatic species.

41

Figure 1: Life history strategies of gobies as a function of life style (aquatic or terrestrial).

Shown are data on growth rate (A), reproductive investment (B), and egg production (C and

D).

Note: all lengths displayed in Figure 1 are mean centered.

42

Table 2: Variation in life history among goby species as a function of lifestyle and other factors

Variable 95% CI β t (lower, upper) 2 A. Growth rate: F3,20 = 11.75, p < .001, R = .64 Intercept 3.77 3.61, 3.93 47.63 Length .28 -.02, .58 1.84 Lifestyle .04 -.38, .46 .21 Lifestyle*Length 1.05 .14, 1.96 2.25 2 B. Reproductive effort: F3,11 = 4.74, p < .02, R = .56 Intercept 3.54 2.34, 4.74 5.83 Length -.32 -.59, -.05 -2.37 Lifestyle -.27 -.58, .04 -1.69 GSI code -.08 -.31, .15 -.67 2 C. Egg number: F2,27 = 67.15, p < .001, R = .83 Intercept 7.59 7.31, 7.87 53.26 Length 1.75 1.38, 2.12 9.29 Lifestyle .56 -.31, 1.43 1.27 2 D. Egg size: F3,12 = 5.03, p = .02, R = .56 Intercept -1.56 -1.93, -1.19 -8.20 Length -1.04 -1.68, -.40 -3.18 Lifestyle 1.03 .23, 1.83 2.53 Length*Lifestyle 1.04 -.09, 2.17 1.80 Parameter values with 95 % confidence intervals that do not overlap zero are statistically significant at p < .05. The lifestyle of species was coded 0 or 1 depending on whether species were aquatic or terrestrial, respectively. GSI code was included in B to account for studies using a slightly different computation of GSI values (see Materials and Methods). An interaction term between lifestyle and length was initially included in B and C, but failed to account for significant variation in the dependent variable and was subsequently removed from the final model. Note only tropical and subtropical species were available to compute model D.

43

Figure 2: Life history trade-offs in aquatic and terrestrial gobies as it relates to growth and reproduction (A and B) and egg production (C and D).

Note: plots (C) and (D) consist of separate but overlapping combinations of species.

44

Blenniidae

Life History Variation as a Function of Lifestyle

Growth: There was a significant interaction between age and lifestyle indicating that growth rate progressively decreased in species with an increasingly terrestrial lifestyle (Fig. 3; Table

3A). For example, although Alticus arnoldorum (terrestrial species) probably have larger larvae on hatching than Istiblennius lineatus (a marine species), Alticus have significantly slower growth rates such that its largest adult body size is considerably smaller than the largest adult body size of the rapidly growing Istiblennius.

Reproduction: Not surprisingly, larger females had larger ovaries (Fig. 4; Table 3B).

However, lifestyle was also an important factor accounting for differences in ovarian weight among species. Although ovarian weight seemed to decrease significantly with an increasingly terrestrial lifestyle, this was offset by a nearly a two fold increase in the development of the ovaries in the terrestrial species (Fig. 4; Table 3B). This pattern was consistent regardless of whether the maturity of individuals within species was controlled for or not (Tables 3B and A3). This implies a significant shift towards increasingly earlier reproduction as fish have moved from an aquatic environment to one on land.

Egg Number and Size: Egg number increased significantly with body size such that differences in egg number among species could be entirely attributed to differences in overall body size (Fig. 5; Table 3C). Consequently, lifestyle had no impact on the number of eggs produced by species.

The only factor predicting egg size was the developmental stage of eggs, with eggs becoming increasingly larger at each developmental stage. Neither body size nor lifestyle accounted for any variance in egg size among species (Fig. 6; Table 3D).

45

Trade-Offs Between Life History Characteristics

Growth versus reproduction: I used the parameter values computed for each species from the models reported in Table 3A and 3B to provide an estimate of growth rate and ovarian development rate that controlled for body size and developmental stage (see also Fig. 3B and

4B). There was a prominent, negative trend between the rate of growth and the rate of ovarian development displayed by species (Pearson correlation: N = 6, r = -.92, p = .005 one-tailed;

Fig. 7A). Species either had fast ovarian development and slow growth (e.g. Alticus arnoldorum) or slow ovarian development and high growth (e.g. Istiblennius lineatus). These contrasting life history strategies were also consistent with the lifestyle exhibited by species: the slow reproductive development and high growth rate found in aquatic species progressively shifted to high reproductive development and slow growth rate as species become increasingly more terrestrial (Fig. 7A; see also Table 3A and 3B, Figs. 3-4).

Egg number versus egg size: I used the models reported in Table 3C and 3D recomputed without lifestyle (which did not account for significant variation in egg number or size) to compute parameter estimates of egg number and size controlling for body size and developmental stage (Fig. 7B). There was a prominent, negative trend between the number and size of eggs produced by species (Pearson correlation: N = 6, r = -0.83, p = 0.02 one- tailed; Fig. 7B). Species either produced many, small eggs (e.g., Entomacrodus striatus) or few, large eggs (e.g., Praealticus labrovittatus). Not surprisingly given the lack of relationship between lifestyle and egg production (Table 3C and 3D), there was no obvious relationship between which strategy species exhibited and lifestyle.

46

Figure 3: Life history strategies in blennies: show here is the progressive reduction of growth rate as fish have moved from an aquatic

environment onto land (A). Age was measured as the total number of microincrements in an individual’s otolith. Also shown (B) are species

coefficients computed from the fitted model reported in Table 3A. Dashed lines for around coefficients represent 95 % confidence intervals.

Symbols are shaded according to their lifestyle, from white (aquatic) to black (terrestrial). 47

Table 3: Variation in life history among blenny species as a function of lifestyle and other factors

Variable β 95% CI t (lower, upper) A. Growth rate Intercept -.47 -.95, .01 -1.93 Age .85 .75, .95 16.81 Lifestyle .60 .33, .87 4.28 Age*Lifestyle -.13 -.19, -.07 -4.62 N observations, species 119, 6 B. Reproductive effort Intercept -17.69 -23.19, -12.19 -6.31 Length 3.85 2.60, 5.10 6.03 Lifestyle -7.61 -11.71, -3.51 -3.64 Length*Lifestyle 1.95 0.96, 2.94 3.85 Nobservations, species 161, 6 C. Egg number Intercept -3.30 -6.77, .17 -1.87 Length 2.62 1.83, 3.41 6.54 Lifestyle -.19 -.43, .05 -1.59 Nobservations, species 77, 6 D. Egg size Intercept 4.43 3.90, 4.96 16.45 Length .07 -.06, .20 .99 Egg stage .53 .48, .57 22.32 Lifestyle .04 -.01, .08 1.69 Nobservations, species 170, 6 Parameter values with 95 % confidence intervals that do not overlap zero are considered statistically significant effects (p < 0.05). The lifestyle of species was coded 0 to 3 depending on the prepotency of terrestrial behavior exhibited by the species. Interaction terms between length and lifestyle, or between egg stage and lifestyle, were initially included in C and D, but failed to account for significant variation in egg number or size and were subsequently removed from final models.

48

Figure 4: Life history strategies in blennies: fish on land have accelerated ovarian development compared to aquatic species (A). Ovarian weight

was the weight of an individual’s ovaries. Also shown (B) are species coefficients computed from the fitted model reported in Table 3B. See

Figure 3 legend for other details. 49

Figure 5: Life history strategies in blennies: the number of eggs produced by blenny fishes was entirely dependent on body size and was

unrelated to lifestyle (A). Also shown (B) are species coefficients computed from the fitted model reported in Table 3C. See Figure 3 legend for

other details. 5 0

Figure 6: Life history strategies in blennies: egg size was dependent on developmental stage (A) but not species lifestyle (B). Also shown in B

are species coefficients computed from the fitted model reported in Table 3D. See Figure 3 legend for other details. 51

Figure 7: Life history trade-offs in blennies between (A) growth and reproduction, and (B) egg size and number. Plots show species coefficients

computed from the fitted model reported in Table 3 (NB: parameter estimates for oocyte size and number were taken after removing lifestyle

from the models as lifestyle did not account for significant variation in size or number). 52

Discussion

My study found evidence of consistent and predictable changes in life history among closely related species in two families of fish that have each made independent transitions from an aquatic to terrestrial environment. Species in both families exhibited significant changes in growth rate following the colonisation of land, and a corresponding change in reproductive effort (certainly in blennies, but possibly in gobies as well). This trade-off between growth and reproduction is common in the natural world (Robinson & Doyle 1985; Berglund &

Rosenqvist 1986; Ryser 1989; Geber 1990; Schwarzkopf 1993) and consistent with life history theory (Gadgil & Bossert 1970; Stearns 1992; see introduction). Also consistent with life history theory were the specific changes in growth and reproduction that were predicted if the colonisation of land resulted in increased survival for certain age classes relative to others

(Table 1). It is unlikely then that fish simply colonised land by happenstance and experienced no special adaptive advantages on doing so. Instead, my results broadly support the view that some form of ecological release (e.g., from predators or competitors in the aquatic environment) or new ecological opportunity (e.g., new resources on land) prompted fish in both families to independently colonise land.

The direction of change in growth and reproduction was found to be different in the two families. This suggests that either the specific causal factors promoting colonisation were different in each family or that other factors unique to each family led to divergent life history trajectories between the two groups. Theory predicts that reproductive investment will be shifted away from age classes experiencing low survival, which will in turn have flow-on effects for growth rate given the predicted (and in this case confirmed for blennies; Figs. 2A and 7A) trade-off between reproduction and growth (Michod 1979; see introduction). In blennies, the progressive colonisation of land has resulted in increased reproductive effort

(specifically, rapid gonad development in females) and reduced growth rates, whereas the

53

opposite appears to have occurred in gobies with land species exhibiting high growth rates and (possibly) low reproductive effort. These differences in life history would typically occur if there were differences in age-specific survival (Table 1). Specifically, colonising land has conveyed the greatest benefits in survival to early age classes in blennies, but in gobies it has increased survivorship in older age classes. These differences in age-specific survival might occur if, for example, one group colonised land to avoid resource competition in aquatic environments that tended to reduce the survival of older individuals (e.g., Osenberg et al.

1992), while the other group colonised land to avoid predation in aquatic environments that tended to target younger individuals (e.g., see Belk 1995).

There are other important differences between the two families that may have been influential as well. Firstly, the conditions experienced on land by gobies and blennies are quite different and have been met with different behaviours. Terrestrial gobies (mudskippers) occupy the waterline of mudflats and mangroves which are tide-dominated and lack wave action (Polgar et al. 2010, 2012; TJ Ord, personal observation). It is common for mudskippers to remain active day and night regardless of tide as mudskippers are able to move with the changing waterline (Colombini et al. 1995; Polgar et al. 2010). On the other hand, terrestrial blennies live out on exposed rocks in the splash zone, a few feet above the waterline, where wave action and extensive fluctuations in tide level and temperature can severely impact their activity (see Ord & Hsieh 2011). Note, terrestrial blennies do exhibit some nocturnal behaviour, but this should also be restricted by wave and tidal action (Ord & Hsieh 2011).

Secondly, gobies and blennies differ in diet. Mudskippers are active predators, feeding upon a wide variety of small intertidal inhabitants, including insects and small crustaceans, and acquiring additional small food items by filtering sediment with their gills

(Takeda et al. 2011; Bucholtz et al. 2009; Baeck 2008; Kruitwagen et al. 2007). Furthermore, by moving with the tideline (intertidal migration) mudskippers may maximise their exposure

54

to food in what is commonly a food-rich environment (Gibson 2003). In contrast, the diet of terrestrial blennies appears to consist of detritus and, to a lesser extent, the epilithic algae

(Bhikajee & Green 2002; Shimizu et al. 2006; Ord & Hsieh 2011). Differences in the quality and abundance of food can have obvious and important implications for growth (e.g. Van der

Veer & Witte 1993), which may have contributed to the observed life history differences exhibited between terrestrial gobies and terrestrial blennies.

Thirdly, blennies and gobies differ in lifespan. I rarely found blennies older than a year (Fig. 1), whereas many of the goby species reported in the literature—including mudskippers—live for several years (Nash 1982; Melo & Clus 2005; Kruitwagen et al. 2006;

Abdoli et al. 2010). This is particularly significant when considering the aforementioned differences in habitat and behaviour. For terrestrial blennies, which my results suggest only survive for one spawning season, the restriction that wave and tidal action impose on activity, including presumably courtship and reproduction, may make it more advantageous to increase reproductive effort at the cost of growth. In contrast for mudskippers, which are not generally restricted as to when they can be active and live long lives exposing them to multiple spawning seasons, it may be more advantageous to distribute reproductive effort more selectively over time (e.g. see “bet-hedging” Stearns 1976). Further study of the interaction between lifespan, habitat, and life history characteristics may provide additional insights on the adaptive benefits of fish living on land.

There was neither an intuitive, nor a consistent change in egg production between aquatic and terrestrial species, as predicted (Table 1). This implies that there was no change in egg or juvenile survival with species colonising land (e.g., see Brockelman 1975; Sibly &

Calow 1983; Hutchings 1991), whereas the changes in growth and reproduction suggest a significant increase in survival for certain age classes of sexually mature individuals. It seems likely that the lack of compelling change in egg production probably reflects broad

55

similarities in egg or larval ecology between aquatic and land species, although this will need to be investigated further in gobies. In the case of amphibious and terrestrial blennies, females lay egg clutches in rock holes above the waterline (Bhikajee & Green 2002; Shimizu et al. 2006), but these rock holes are frequently inundated with water by waves (ERM Platt &

TJ Ord, personal observation) such that the conditions experienced by eggs are probably similar to those of aquatic egg clutches. Furthermore, all blenny and goby species—aquatic and terrestrial—likely have a pelagic larval stage (Bhikajee & Green 2002; Patzner 2009,

2011; Polgar 2012). Although it is unknown how long hatched larvae of terrestrial species remain in burrows or rock holes, genetic data suggests extensive gene flow among populations for the land blenny, Alticus arnoldorum, that almost certainly reflects pelagic larval dispersal (Morgans et al. 2014; Cooke et al, In Review). That is, at least for this land species, larvae seem to spend part of their life stage floating freely in the water column in the same way as pelagic larvae of aquatic species, before subsequently emerging onto land as juveniles (probably within 36 days of hatching based on the estimated settlement stage for A. arnoldorum; ERM Platt & TJ Ord, unpublished data). That is, while the colonisation of land by blennies and gobies resulted in a substantial change in environment for juveniles and adults, there may have been very little change in the conditions experienced (and subsequently survivorship) of eggs and larvae. This is also supported by the trade-off between egg number and egg size found in both families (Figs. 2 C, D and 7B), which suggests that differences in egg or larval survival do exist among species but it was unrelated to the colonisation of land (species that produce large eggs improve the chance of survival of their offspring at the expense of the number of offspring that can be produced; Smith &

Fretwell 1974).

Some suggestive evidence was gathered hinting that the egg size of terrestrial gobies might differ from their aquatic counterparts. The nature of this difference appeared to be

56

either a larger egg size in terrestrials, or else the absence of a change in eggs size with body length as occurred in aquatic species. Because there was no difference in egg number (or perhaps even an increase in egg number) between terrestrial and aquatic species, and a clear negative trend between egg number and size among species gobies generally, either explanation for egg size differing in terrestrial species would violate the life history assumption of trade-offs. Although rare, this can occur under highly unusual circumstances

(e.g. Reznick et al. 2000), and life history theory also predicts an increase in resources, such as food, should lead to greater reproductive effort at all ages for resource limited organisms

(Gadgil & Bossert 1970; see introduction). Regardless, the absence of any compensatory changes in egg number (or a change in the opposite direction that results in the production of many, large eggs), as well as the implied reduction in GSI and the absence of similar patterns in blennies, all suggest that any inferences made on egg production in gobies are premature until sample sizes for gobies can be increased by future study.

Another difficulty highlighted by my study relates to the complex nature of measuring reproductive investment among species. The classical method of using GSI in fish to assess reproductive investment (Gadgil & Bossert 1970) would have failed to find the important change in the rate of ovarian development in terrestrial blennies. In fact, exclusive interpretation of the relative weight of the ovaries to body size would have implied a reduction in reproductive investment with an increasingly terrestrial lifestyle (Fig. 4). This would have also been at odds with my findings on growth rate. Yet the significant acceleration of ovarian development rate in amphibious and terrestrial blennies signifies earlier maturation and greater production of reproductive tissue per unit of body size over time. That is, amphibious and terrestrial blennies actually exhibit higher reproductive investment than their aquatic relatives. Given the difficulty in using relative ovarian weight to body size as a measure of reproductive investment, it is perhaps not surprising that results

57

using GSI in gobies were equivocal (although suggestive; this was also not helped by the lack of information on terrestrial gobies more generally).

While my study was not designed to identify the specific causal factors that prompted fish to move onto land (although my results do suggest these factors were probably different in the two families studied), it has shown there was likely an important increase in survival for certain age classes post colonisation on land. These changes in survival were evident in the particular life history strategies exhibited by land fish and presumably resulted in a strong adaptive advantage for fish that left their ancestral aquatic environment. The most likely hypothesis for organisms to make such dramatic transitions in environment is some sort of selection advantage conveyed by an ecological release or new ecological opportunities

(Wilson 1961; Lister 1976; Jeffries & Lawton 1984; Losos & Mahler 2010). More generally, my study provides strong empirical support for theoretic models of life history. I found consistent trade-offs among different life history characteristics that were explicitly predicted by theory. Indeed, these trade-offs underlie my interpretation that changes in life history reflect enhanced survival following the colonisation of land. If these trade-offs had not been broadly supported, then it would have been impossible to infer anything about the possible consequences of life on land. My study shows that the comparison of life history strategies among closely related taxa occupying different environments can provide new insights into the underlying causes and processes of colonisation.

AKNOWLEDGEMENTS

I would like to thank Alex Kerr and Brett Taylor for logistical support, Georgina Cooke for her field training and sampling contributions, Iain Suthers for the use of his lab equipment,

Ash Fowler for his laboratory training and project support, and my supervisor Terry Ord for his extensive project support.

58

Literature Cited

Abdoli, L., E. Kamrani, A. Abdoli, and B. Kiabi. 2010. Age and growth of the Mudskipper,

Scartelaos tenuis (Day, 1876) (Gobiidae: Oxudercinae), in the coastal areas of the

Persian Gulf at Bushehr province, Iran. Zoology in the Middle East 51:113-115.

Arbogast, B. S., S. V. Drovetski, R. L. Curry, P. T. Boag, G. Seutin, P. R. Grant, B. R. Grant

et al. 2006. The origin and diversification of Galapagos mockingbirds. Evolution

60:370-382.

Baeck, G., T. Takita, and Y. Yoon. 2008. Lifestyle of Korean mudskipper Periophthalmus

magnuspinnatus with reference to a congeneric species Periophthalmus modestus.

Ichthyological Research 55:43-52.

Bates, D. 2008. lme4: linear mixed-effects models using S4 classes: R package, 0.999375–28,

http://lme4.r-forge.r-project.org/.

Belk, M. C. 1995. Variation in growth and age at maturity in bluegill sunfish: genetic or

environmental effects? Journal of Fish Biology 47:237-247.

Berglund, A., and G. Rosenqvist. 1986. Reproductive costs in the prawn Palaemon

adspersus: effects on growth and predator vulnerability. Oikos 46:349-354.

Bhikajee, M., and J. M. Green. 2002. Behaviour and habitat of the Indian Ocean amphibious

blenny, Alticus monochrus. African Zoology 37:221-230.

Bhikajee, M., J. Green, and R. Dunbrack. 2006. Life history characteristics of Alticus

monochrus, a supratidal blenny of the southern Indian Ocean. African Zoology 41:1-

7.

Brey, T. 1999. Growth performance and mortality in aquatic macrobenthic invertebrates,

Pages 153-223 in P. A. T. A.J. Southward, and C. M. Young, eds. Advances in

Marine Biology, Academic Press.

Brockelman, W. Y. 1975. Competition, the fitness of offspring, and optimal clutch size. The

59

American Naturalist 109:677-699.

Brothers, E. B., C. P Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from

larval and adult fishes. Fisheries Bulletin 74:1-8.

Bucholtz, R. H., A. S. Meilvang, T. Cedhagen, J. T. Christensen, and D. J. Macintosh. 2009.

Biological observations on the mudskipper Pseudapocryptes elongatus in the Mekong

Delta, Vietnam. Journal of the World Aquaculture Society 40:711-723.

Campana, S. E., and J. D. Neilson. 1985. Microstructure of fish otoliths. Canadian Journal of

Fisheries and Aquatic Sciences 42:1014-1032.

Carlson, B. A. 1992. The life history and reproductive success of the coral blenny, Exallias

brevis (Kner, 1868). PhD thesis: University of Hawaii.

Carvalho, M. G., A. S. Moreira, C. Moreira, H. Queiroga, P. T. Santos, and A. T. Correia.

2014. Validation of otolith daily increments in early juveniles of shanny Lipophrys

pholis. Journal of Fish Biology 84:1234-1239.

Clelland, E. S., Q. Tan, A. Balofsky, R. Lacivita, and C. Peng. 2007. Inhibition of premature

oocyte maturation: a role for bone morphogenetic protein 15 in zebrafish ovarian

follicles. Endocrinology 148:5451-5458.

Colombini, I., R. Berti, A. Ercolini, A. Nocita, and L. Chelazzi. 1995. Environmental factors

influencing the zonation and activity patterns of a population of Periophthalmus

sobrinus Eggert in a Kenyan mangrove. Journal of Experimental Marine Biology and

Ecology 190:135-149.

Duponchelle, F., E. Paradis, A. J. Ribbink, and G. F. Turner. 2008. Parallel life history

evolution in mouthbrooding cichlids from the African Great Lakes. Proceedings of the

National Academy of Sciences 105:15475-15480.

Gadgil, M., and W. H. Bossert. 1970. Life historical consequences of natural selection. The

American Naturalist 104:1-24.

60

Geber, M. A. 1990. The cost of meristem limitation in Polygonum arenastrum: negative

genetic correlations between fecundity and growth. Evolution 44:799-819.

Gibson, R. N. 2003. Go with the flow: tidal migration in marine animals. Hydrobiologia

503:153-161.

Graham, Jeffrey B., and Heather J. Lee. 2004. Breathing air in air: in what ways might extant

amphibious fish biology relate to prevailing concepts about early tetrapods, the

evolution of vertebrate air breathing, and the vertebrate land transition? Physiological

and Biochemical Zoology 77:720-731.

Hernaman, V., and P. Munday. 2005. Life-history characteristics of coral reef gobies. I.

growth and life-span. Marine ecology progress series 290:207-221.

Hernández-Miranda, E., R. Veas, C. Valeria Espinoza, S. R. Thorrold, and F. P. Ojeda. 2009.

The use of otoliths and larval abundance for studying the spatial ecology of the blenny

Scartichthys viridis (Valenciennes, 1836) in coastal central Chile. Revista de Biologia

Marina y Oceanografia 44:619-633.

Hsieh, S. T. T. 2010. A locomotor innovation enables water-land transition in a marine fish.

PloS ONE 5:e11197.

Hutchings, J. A. 1991. Fitness consequences of variation in egg size and food abundance in

brook trout Salvelinus fontinalis. Evolution 45:1162-1168.

———. 1993. Adaptive life histories effected by age-specific survival and growth rate.

Ecology 74:673-684.

Jeffries, M. J., and J. H. Lawton. 1984. Enemy free space and the structure of ecological

communities. Biological Journal of the Linnean Society 23:269-286.

Kornfield, I., and P. F. Smith. 2000. African cichlid fishes: model systems for evolutionary

biology. Annual Review of Ecology and Systematics 31:163-196.

Kruitwagen, G., T. Hecht, H. B. Pratap, and S. E. Wendelaar Bonga. 2006. Changes in

61

morphology and growth of the mudskipper (Periophthalmus argentilineatus)

associated with coastal pollution. Marine Biology 149:201-211.

Kruitwagen, G., I. Nagelkerken, B. R. Lugendo, H. B. Pratap, and S. E. Wendelaar Bonga.

2007. Influence of morphology and amphibious life-style on the feeding ecology of

the mudskipper Periophthalmus argentilineatus. Journal of Fish Biology 71:39-52.

Laurin, M. 2010. How vertebrates left the water. University of California Press, United States

of America.

Law, R. 1979. Optimal life histories under age-specific predation. The American Naturalist

114:399-417.

Lister, B. C. 1976. The nature of niche expansion in West Indian Anolis lizards I: ecological

consequences of reduced competition. Evolution 30:659-676.

Losos, J. B., and D. L. Mahler. 2010. 'Adaptive radiation: the interaction of ecological

opportunity, adaptation, and speciation' in Evolution Since Darwin: The First 150

Years. Sunderland, MA, Sinauer Associates.

Mansur, L., D. Catalan, G. Plaza, M. F. Landaeta, and F. P. Ojeda. 2013. Validations of the

daily periodicity of increment deposition in rocky intertidal fish otoliths of the south-

eastern Pacific Ocean. Revista de Biologia Marina y Oceanografia 48:629-633.

Martin, K. L. M. 1995. Time and tide wait for no fish: intertidal fishes out of water.

Environmental Biology of Fishes 44:165-181.

Mazlan, A. G., and M. Rohaya. 2008. Size, growth and reproductive biology of the giant

mudskipper, Periophthalmodon schlosseri (Pallas, 1770), in Malaysian Waters.

Journal of Applied Ichthyology 24:290-296.

Melo, Y. C., and F. Le Clus. 2005. Growth and reproduction of the pelagic goby Sufflogobius

bibarbatus off the Orange River, Southern Africa. African Journal of Marine Science

27:265-273.

62

Michod, R. E. 1979. Evolution of life histories in response to age-specific mortality factors.

The American Naturalist 113:531-550.

Morales-Nin, B. 2000. Review of the growth regulation processes of otolith daily increment

formation. Fisheries Research. 46:53-67.

Moreau, J., Bambino, C., Pauly, D. 1986. A comparison of four indices of overall fish growth

performance, based on 100 tilapia populations (Fa. Cichlidae), pp 201 – 206. In

Maclean, J. L., Dizon, L. B., Hosillo, L. V. The First Asian Fisheries Forum. Asian

Fisheries Society, Manila, Phillipines.

Morgans, C. L., and T. J. Ord. 2013. Natural selection in novel environments: predation

selects for background matching in the body colour of a land fish. Animal Behaviour

86:1241-1249.

Morgans, C. L., Cooke, G. M., Ord, T. J. 2014. How populations differentiate despite gene

flow: sexual and natural selection drive phenotypic divergence within a land fish, the

Pacific leaping blenny. BMC Evolutionary Biology 14:97.

Münzing, J. 1963. The evolution of variation and distributional patterns in european

populations of the three-spined stickleback, Gasterosteus aculeatus. Evolution

17:320-332.

Murdy, E. O. 1989. A taxonomic revision and cladistic analysis of the oxudercine gobies

(Gobiidae: Oxudercinae), Australian Museum.

Nash, R. D. M. 1982. The biology of Fries’ goby, Lesueurigobius friesii (Malm), in the Firth

of Clyde, Scotland, and a comparison with other stocks. Journal of Fish Biology

21:69-85.

Near, T. J., R. I. Eytan, A. Dornburg, K. L. Kuhn, J. A. Moore, M. P. Davis, P. C.

Wainwright et al. 2012. Resolution of ray-finned fish phylogeny and timing of

diversification. Proceedings of the National Academy of Sciences 109:13698-13703.

63

Núñez, J., and F. Duponchelle. 2009. Towards a universal scale to assess sexual maturation

and related life history traits in oviparous teleost fishes. Fish Physiology and

Biochemistry 35:167-180.

Ord, T. J., and S. Tonia Hsieh. 2011. A highly social, land-dwelling fish defends territories in

a constantly fluctuating environment. Ethology 117:918-927.

Osenberg, C. W., G. G. Mittelbach, and P. C. Wainwright. 1992. Two-stage life histories in

fish: the interaction between juvenile competition and adult performance. Ecology

73:255-267.

Patzner, R. A., E. J. Gonçalves, P. Hastings, and B. G. Kapoor. 2009. The biology of

blennies. United States of America, Science Publishers.

Patzner, R. A., Van Tassel, J. L., Kovacic, M., Kapoor, B. G. 2011. The biology of gobies.

Science Publishers, Enfield, NH, USA.

Pauly, D. 1979. Gill size and temperature as governing factors in fish growth: a

generalization of von bertalanffy’s growth formula. Berichte am dem Znstitut fur

Meereskunde Kiel 63: 1 – 156.

———. 1991. Growth performance in fishes: rigorous description of patterns as a basis for

understanding causal mechanisms. ICLARM 4:3–6.

Pianka, E. R., and W. S. Parker. 1975. Age-specific reproductive tactics. The American

Naturalist 109:453-464.

Polgar, G. 2012. Ecology and evolution of mudskippers and oxudercine gobies (Gobiidae:

Oxudercinae): perspectives and possible research directions. Pp 117 – 137. In

Sasekumar A., Chong, V. C. Mangrove and coastal environment of Selangor,

Malaysia. University of Malaya, Kuala Lumpur.

Polgar, G., A. Sacchetti, and P. Galli. 2010. Differentiation and adaptive radiation of

amphibious gobies (Gobiidae: Oxudercinae) in semi-terrestrial habitats. Journal of

64

Fish Biology 77:1645-1664.

Rainey, P. B., and M. Travisano. 1998. Adaptive radiation in a heterogeneous environment.

Nature 394:69-72.

Reznick, D., L. Nunney, and A. Tessier. 2000. Big houses, big cars, superfleas and the costs

of reproduction. Trends in Ecology & Evolution 15:421-425.

Robinson, B. W., and R. W. Doyle. 1985. Trade-off between male reproduction (amplexus)

and growth in the amphipod Gammarus lawrencianus. The Biological Bulletin

168:482-488.

Rundle, H. D., and D. Schluter. 2004. Natural selection and ecological speciation in

sticklebacks. Adaptive speciation 19:192-209.

Ryser, J. 1989. Weight loss, reproductive output, and the cost of reproduction in the common

frog, Rana temporaria. Oecologia 78:264-268.

Sayer, M. D. J. 2005. Adaptations of amphibious fish for surviving life out of water. Fish and

Fisheries 6:186-211.

Sayer, M. D. J., and J. Davenport. 1991. Amphibious fish: why do they leave water? Reviews

in Fish Biology and Fisheries 1:159-181.

Schwarzkopf, L. 1993. Costs of reproduction in water skinks. Ecology 74:1970-1981.

Selman, K., R. A. Wallace, A. Sarka, and X. Qi. 1993. Stages of oocyte development in the

zebrafish, Brachydanio rerio. Journal of Morphology 218:203-224.

Shimizu, N., Y. Sakai, H. Hashimoto, and K. Gushima. 2006. Terrestrial reproduction by the

air-breathing fish Andamia tetradactyla (Pisces; Blenniidae) on supralittoral reefs.

Journal of Zoology 269:357-364.

Sibly, R., and P. Calow. 1983. An integrated approach to life-cycle evolution using selective

landscapes. Journal of Theoretical Biology 102:527-547.

Smith, C. C., and S. D. Fretwell. 1974. The optimal balance between size and number of

65

offspring. The American Naturalist 108:499-506.

Stearns, S. C. 1976. Life-history tactics: a review of the ideas. The Quarterly Review of

Biology 51:3-47.

———. 1992. The evolution of life histories. Oxford University Press, Oxford.

Sunobe, T., T. Ohta, and A. Nakazono. 1995. Mating system and spawning cycle in the

blenny, Istiblennius enosimae, at Kagoshima, Japan. Environmental Biology of Fishes

43:195-199.

Takeda, T., M. Hayashi, A. Toba, K. Soyano, and A. Ishimatsu. 2011. Ecology of the

Australian mudskipper Periophthalmus minutus, an amphibious fish inhabiting a

mudflat in the highest intertidal zone. Australian Journal of Zoology 59:312-320.

Taylor, H. M., R. S. Gourley, C. E. Lawrence, and R. S. Kaplan. 1974. Natural selection of

life history attributes: an analytical approach. Theoretical Population Biology 5:104-

122.

Thacker, C. E., and D. M. Roje. 2011. Phylogeny of Gobiidae and identification of gobiid

lineages. Systematics and Biodiversity 9:329-347.

Todd Streelman, J., and P. D. Danley. 2003. The stages of vertebrate evolutionary radiation.

Trends in Ecology & Evolution 18:126-131.

Tyler, C. R., and J. P. Sumpter. 1996. Oocyte growth and development in teleosts. Reviews in

Fish Biology and Fisheries 6:287-318.

Van der Veer, H. W., and J. I. Witte. 1993. The ‘maximum growth/optimal food condition’

hypothesis: a test for 0-group plaice Pleuronectes platessa in the Dutch Wadden Sea.

Marine Ecology Progress Series 101:81-90.

Vermeij, G. J., and R. Dudley. 2000. Why are there so few evolutionary transitions between

aquatic and terrestrial ecosystems? Biological Journal of the Linnean Society 70:541-

554.

66

Wilson, E. O. 1961. The nature of the taxon cycle in the melanesian ant fauna. The American

Naturalist 95:169-193.

67 CHAPTER 3

Population variation in the life history of a land fish reveals the

interacting impact of predation and conspecific density

Edward R. M. Platt & Terry J. Ord

ERMP and TJO designed the study, ERMP collected and processed specimens with the

assistance of TJO, ERMP analysed the data and wrote the manuscript with input from TJO.

A version of this paper has been submitted to Ecology

68

Abstract

Life history variation is predicted to reflect differences in age-specific mortality within populations, with reproductive effort shifted away from ages experiencing increased mortality. It is often difficult to determine what the key determinants of age-specific mortality might be for natural populations. Many investigators of life history in vertebrates have focused exclusively on the direct impact of predation on mortality. However, predation may have unexpected impacts on population density that in turn prompt unexpected changes in life history. I investigated the consequences of predation and density on life history variation among five populations of the Pacific leaping blenny Alticus arnoldorum. This fish from the island of Guam spends most of its life out of the water on rocks in the splash zone, where it is vulnerable to predation and likely sensitive to changes in population density that impact resource availability. I found that the extent that populations invested in reproduction was inversely related to predation rate, while growth rate appeared to vary largely in response to population density. These differences in life history among populations were also likely to be plastic, given past study has revealed extensive contemporary gene flow among these populations. The influence of predation and density on life history was unlikely to have operated independently of each other, with predation rate tending to be associated with reduced population densities. Taken together, my results show that the interaction between predation and density can lead to complex outcomes in life history, and infer that plastic life history traits could allow populations to persist in new or rapidly changing environments.

69

Introduction

Life history theory predicts that the characteristics of reproduction and survival that typify an organism will reflect a strategy selected by the environment to maximise fitness (Gadgil &

Bossert 1970; Roff 1992; Stearns 1992); i.e. the total number of offspring produced that survive to reproduce. Underpinning these strategies are trade-offs. Organisms have a fixed amount of energy at any one point in time that can be allocated to characteristics that either increase survival or reproduction, but rarely both (Gadgil & Bossert 1970; Zera & Harshman

2001). For example, resources invested in reproduction will necessarily divert resources away from growth (Stearns 1992). What determines the optimal strategy (balance between life history traits) is the age-specific probability of mortality. Life history theory more specifically predicts that fitness will be maximised by shifting reproductive effort away from ages with increased mortality (Michod 1979; Roff 1992; Stearns 1992; Ernande et al. 2004).

For example, if early age survival is low relative to older ages, the greatest gains in fitness will come from surviving to reproduce later in life. This should result in lower overall reproductive effort and higher growth rate (which should also improve the chances of survival for younger age classes to reach those older more fecund ages; Pianka & Parker

1975; Hutchings 1993). Conversely if early age survival increases relative to older ages, the reverse outcome is predicted because fitness will be maximised if resources are committed to current acts of reproduction despite the almost inevitable negative impact on growth (Gadgil

& Bossert 1970; Hutchings 1993). In the case where survival changes uniformly for all age classes (i.e., there is no change in age-specific mortality) the optimal distribution of reproductive effort by age should remain the same and no life history changes are expected

(Gadgil & Bossert 1970; Law 1979; Michod 1979).

The difficulty in testing these predictions from life history theory is understanding the primary causes of mortality in nature and how they might vary with age. There are many

70 factors that contribute to mortality and can impact individuals of certain age classes in different ways. For example, predation might target younger individuals in a population, whereas mortality driven by increased competition for resources might be more relevant for older individuals of the same population. This makes it difficult to predict what the overall age-specific mortality of the population might be and how it might subsequently vary with changes in either predation or competition in the population (e.g., see Osenberg et al. 1992;

Charlesworth 1980). Many studies of life history have focussed exclusively on predation given its direct contribution to mortality and given that its impact on different age classes can be reasonably predicted for certain species (e.g., in the case of gape-limited predators that target smaller, younger individuals in a population; see Johansson et al. 2001; Dowdall et al.

2012; Furness et al. 2012; Ghani et al. 2013). However, interpreting age-specific mortality solely through the direct contribution of predation on population mortality also overlooks the indirect effects of predation on population density (Walsh & Reznick 2008; Gale et al. 2013).

Density has increasingly been shown to be important for driving changes in life history as well (Ernande et al. 2004; Gutowsky & Fox 2012; Reznick et al. 2012; Cameron et al. 2013), and sometimes in the opposite direction to what might have been expected based solely on the effects of predation (e.g., Reznick et al. 1996).

This is because predation might increase mortality in a given age class because predators preferentially target individuals of that age (e.g., Wellborn 1994), and by doing so the density of that age class is sometimes reduced. This can in turn reduce the adverse density-dependent effects of other factors operating on that age class (Charlesworth 1980;

Abrams & Rowe 1996; Schroder et al. 2014). In particular, reductions in density may reduce competition for resources that subsequently increase the chances of survival for the remaining members of that age class. These indirect density-dependent effects of predation (and other factors, such as disease) should be most pronounced in “closed” systems, in which

71 populations are largely isolated and experience limited immigration from other populations

(e.g., Reznick et al. 1996). For these populations, there is more likely to be a negative relationship between predation pressure and the density of targeted age classes. This makes closed (or partially closed) systems especially useful for understanding the relative effects of predation and density on age-specific mortality and life history. Isolated populations are also those experiencing limited gene flow, implying that evolutionary change is also potentially more likely to occur in life history characteristics if age-specific mortality changes in some way (Garcia-Ramos & Kirkpatrick 1997; Hendry et al. 2002; Lenormand 2002).

In contrast, the negative relationship between predation and density is less likely to occur among connected populations because of the diluting effect of continued recruitment

(immigration) from outside the population. Predation and density-dependent effects on age- specific mortality through variables such as resource competition may still be influential in these populations, but the relative effects of predation and density on life history variation in open systems may be quite different to that occurring in less porous systems. Furthermore, the diluting effect of subsequent gene flow from immigrating individuals can potentially limit evolutionary change, which leads to an alternative question of whether life history characteristics can change via plasticity as a function of changes in age-specific mortality within an animal’s lifetime (Stearns & Koella 1986; Ernande et al. 2004). Plastic changes in some life history characteristics have been documented (e.g., offspring number in iteroparous species (Reznick & Yang 1993) or growth rate if resources become limiting early in life

(Belk 1995) or vary over time (Gale et al. 2013)). Nevertheless, there is often an implied assumption in many studies of life history variation that such plastic changes are of secondary importance to those with a genetic basis (Riesch et al. 2013). That is, life history adaptation is primarily responsible for differences between populations, while plasticity explains life history variation within populations (Hutchings 1993). The extent this is true is difficult to

72 judge, given that there has been a general focus on documenting life history variation among isolated populations, for which it might be reasonably assumed (if not directly confirmed) that gene flow is minimal or absent (e.g. Natsumeda et al. 2013; Riesch et al. 2013).

I investigated the role of predation and density in driving variation in reproduction and growth among populations of a unique fish, the Pacific leaping blenny (Alticus arnoldorum). This fish is endemic to the island of Guam and is unusual because it spends most if its life living on land. To be able to do so, it has evolved several adaptations that allow agile movement on land—an axial-tail twisting behaviour that enables the fish to walk, hop and even jump large distances (Hsieh 2010)—as well as breathing through cutaneous respiration (Martin & Lighton 1989; Brown et al. 1991).

Remaining moist is of special importance for cutaneous respiration, as is avoiding desiccation more generally, for the survival of the fish out of the water. This restricts the

Pacific leaping blenny to rocky outcrops within the splash zone around the island, but also limits the temporal activity of the fish to specific times of day (Ord & Hsieh 2011). At low tide and during high midday temperatures, desiccation is a serious problem and the fish subsequently retreats into rock holes and crevices for shelter. The availability of these rocky shelters are therefore a crucial resource for these fish. High tide also poses a problem as violent wave action against the rocks impedes the ability of the fish to move safely out on the rocks. The result is a brief temporal window of activity at mid-tide and at moderate temperatures that changes on a daily basis depending on the tides and weather (Ord & Hsieh

2011). The Pacific leaping blenny must concentrate the majority of its feeding, social behaviour and reproduction within this variable activity window (~ 2-4 hours a day; see Ord

& Hsieh 2011). Moreover, out on the rocks, the fish is exposed to predation from birds, lizards and land crabs (Morgans & Ord 2013). Presumably, then, the balance between being active on the rocks and avoiding predators is acute for the Pacific leaping blenny.

73

Additionally, given that the availability of rock shelters is finite (rock holes are also used in reproduction; see below), the fish is likely to be particularly sensitive to changes in population density through its impacts on competition for these shelters. For these reasons, the Pacific leaping blenny is especially well suited for a study of predation and density- dependent effects on resource availability and subsequent life history variation.

The Pacific leaping blenny is also unique in the respect that the terrestrial populations are ecologically isolated from one another around the island. Populations are limited to isolated sections of rock outcrops that are interspersed around the island by uninhabitable beaches and other inhospitable environments (Morgans et al. 2014). Yet, there is extensive gene flow among these populations as a result of active or passive dispersal by pelagic larvae

(Cooke et al. In Review). Females typically lay eggs in rock holes that are above the waterline, but these holes are frequently inundated by waves and king tides (TJ Ord, personal observation). On hatching, larvae are either frequently washed out or actively swim out of these nest holes into open water, where they subsequently disperse around the island to later emerge onto land as early-age juveniles. Once individuals have transitioned from their marine larvae stage to their terrestrial life stage (an average period of 28 days; ERP Platt unpublished data), migration among post-settlement populations is believed to be severely constrained

(Cooke et al. In Review).

This pattern of dispersal is important for my study for three reasons. Firstly, populations of Pacific leaping blenny offer an ideal opportunity to examine the interacting effects of predation and density on age-specific mortality in largely closed post-settlement populations. Secondly, the Pacific leaping blenny also allows an assessment of the extent to which life history variation might occur via plasticity among populations that experience extensive gene flow (e.g., see Morgans et al. 2014). Thirdly, because larvae are aquatic, factors affecting mortality in larvae are probably consistent for all populations. It can

74 therefore be deduced that any variation in life history among post-settlement populations is directly attributable to variation in age-specific mortality after the transition onto land.

My study focussed on five populations of Pacific leaping blenny on Guam for which the relative levels of predation and density had been recently quantified (Morgans et al.

2014). I began by evaluating the relationship between predation rate and population density, and confirmed that populations experiencing high predation were also those with lower population densities (see Results). I then tested the extent to which predation specifically, or its impacts on density more generally, has led to predictable changes in growth and reproductive investment. To do so, I formulated several predictions from life history theory on how changes in predation rate or population density (through its influence on resource competition) would impact age-specific mortality. These predictions are illustrated graphically in Figure 1 and were modelled on the assumption that, in general, mortality decreases with age in the Pacific leaping blenny. This general pattern of age-specific mortality is typical for small fishes (Peterson & Wroblewski 1984; Sogard 1997; Hampton

2000; Meekan et al. 2006) and there was no reason to expect the Pacific leaping blenny to be any different.

Given this general pattern of mortality, there were three possible scenarios for how changes in predation or density among populations might impact relative mortality among age classes within populations. Firstly, if predation or density effects increased older age mortality relative to early ages, theory predicts that reproductive investment will increase at the expense of decreased growth rate (Scenario A, Fig. 1). Secondly, if predation or density instead increased mortality in earlier ages relative to older ages, reproductive investment should decrease and allow growth rate to increase (Scenario B, Fig. 1). Thirdly, if predation or density results in uniform changes in mortality across all age classes, there should be no changes in reproductive investment or growth rate (Scenario C, Fig. 1; NB: this last scenario

75 is very unlikely in most natural systems). Following these tests of predation and density on population life history variation, I examined the relationship between reproductive investment and growth rate to assess the extent to which these life history characteristics were negatively correlated with one another, a central assumption of life history theory (Stearns

1992; Zera & Harshman 2001).

76

Figure 1: Changes in reproduction and growth predicted for (A) older age biased mortality,

(B) early age biased mortality, or (C) unbiased mortality.

77

Materials and Methods

Data Collection

Sampling was conducted at various locations around the island of Guam in July 2012. These locations were: Pago Bay (13o25’39”N, 144o47’56”E), Taga’chang (13o24’16”N,

144o46’53”E), Talofofo (13o20’34”N, 144o46’21”E), Umatac (13o17’40”N, 144o39’29”E), and the Adelup Point (13o28’52”N, 144o43’44”E). This time of year is expected to correspond with the peak breeding period for the Pacific leaping blenny (e.g., see Carlson

1992; Sunobe et al. 1995; Shimizu et al. 2006). Sampling at each site consisted of catching individuals observed on rocks in the splash zone using a small aquarium hand net.

Considerable effort was made to obtain specimens across the full range of body sizes seen in the population to ensure individuals from all age classes were sampled. Following capture, specimens were euthanized by first anaesthetizing them using clove oil and then storing them under ice. All specimens were later preserved in 95 % ethanol before transporting back to

UNSW for processing.

Predictors of Life History Variation

Information on post-settlement predation rate and the density of individuals (adult and juvenile) of each site were taken from Morgans et al. (2014), which sampled these same populations 10 months prior to my study. It was assumed that the relative levels of predation and density between sites would have remained consistent over this arguably short time because these factors are dependent on a sites physical characteristics (e.g., the availability of rock holes; see introduction). It was therefore deemed appropriate to use these measurements in the current study. Note, in the event that either predation or density had changed following sampling, it was expected to add noise to the data. A brief outline of how this data was collected is given below.

78

Morgans et al. (2014) estimated predation rate by deploying plasticine model replicates of small to large blennies made from casts of fish caught from Taga’chang. The models were highly realistic in appearance and colour (data presented in Morgans & Ord

2013), and previous experiments have shown that they provide biologically meaningful estimates of probable predation rates on the Pacific leaping blenny (Morgans & Ord 2013).

Models were anchored to the rocks at each site and left for three days, after which evidence of predation was recorded for each model (see Morgans & Ord 2013 for details). Predation level was then calculated as the overall percentage of models exhibiting signs of predation, and this was found to vary significantly among populations (Morgans et al. 2014).

Population density was estimated by Morgans et al. (2014) by setting out eight 50 cm by 50 cm quadrats and periodically recording the number of individuals present over one week and across different tide levels. These quadrats were placed above the high tide mark in the splash zone where the Pacific leaping blennies were predominantly active (Ord & Hsieh

2011). Individuals were classed as adult or juvenile depending on whether they were above or below 4 cm in length, as this corresponds to the size of sexual maturity for this species (ERP

Platt unpublished data). Density was then calculated as the number of individuals per m2 for each quadrat, averaged across all quadrats to generate a population average.

Body Measurements

Processing of specimens focussed on females because investment in reproduction for females is higher than males (e.g., the production of eggs is more costly than the production of sperm). The balance between reproductive investment and growth rate should therefore be especially important for females compared to males (this is not to say that reproduction is not costly for males). For each of my five populations, I examined at least 16 specimens (adult and juvenile females, plus some early age juveniles that could not be sexed because gonad

79 development had yet to occur; overall range in sample size: 16-32; Table A1). For each specimen, I obtained measures of body size, reproductive investment, and age.

Body size: Pre-anal length was measured as the length (mm) from the front of the top jaw to the anus, recorded to two decimal places using digital calipers (Traceable Digital Caliper,

Control Company, > 0.03 mm accuracy, 0.01 mm resolution).

Reproductive investment: Specimens were dissected and the gonads carefully extracted, dried of all surface moisture, and weighed to the nearest mg using the semi-micro analytical balance unless very small (> 0.002 g). These smaller gonads were weighed using a Cahn c-33 microbalance and recorded to the nearest 10 µg. The appearance of the ovaries was then used to obtain an estimate of the likely reproductive state of the individual. This was done by cross-referencing the appearance of the ovaries with those of a separate study (Chapter 2) in which detail descriptions of the ovaries had been made, followed by subsequent confirmation of reproductive state based on a detail analysis of the development stage of oocytes found inside the ovaries. Using these descriptions and classification scheme, I categorized gonads as: stage 1 (primary growth), stage 2 (cortical alveolus stage), stage 3 (vitellogenesis), stage 4

(a combination of maturation/mature), and stage 5 (atresia) in which there was evidence of left-over oocytes from spawning that have normally started degenerating (see criteria outline in Table A1 in Chapter 2 and Selman et al. 1993; Tyler & Sumpter 1996; Clelland et al. 2007;

Nunez & Duponchelle 2009).

Age: The age of females (in days) was determined by counting the number of microincrements visible in longitudinal sections of the right sagittal otolith. Microincrements appear as alternating light and dark bands, and are known to directly reflect daily cycles of

80 otolith deposition for nearly all fish species examined (Brothers et al. 1976; Campana and

Neilson 1985; Morales-Nin 2000), including blennies (Hernandez-Miranda et al. 2009;

Mansur et al. 2013; Carvalho et al. 2014). Note too, that while not yet confirmed, there is some evidence that microincrement deposition still occurs on a daily basis in fish species that spend time out of water (Bhikajee et al. 2006; Kruitwagen et al. 2006; Mazlan & Rohaya

2008). Saggital otoliths were dissected from each fish and stored dry in microtubes until processing. Otolith were prepared by taking longitudinal sections by first mounting the otolith on a microscope slide using a thermoplastic glue, and sanding it using 3, 9, and, in a few instances, 12 micron grade lapping film. During this process the otolith was normally flipped so that opposing sides could be sanded. Once sanded, cedar oil was applied and the otolith left for a minimum of five minutes to increase the otolith’s transparency. Age was then determined by counting the total number of microincrements from the right otolith

(where possible). Microincrement counts were made at 600x magnification using the oil iris of an Olympus BX50 microscope, with a Spot Flex microscope camera (Diagnostic

Instruments, Inc. model #15.2 64 Mp Shifting Pixel), and the software program Image Pro

Plus, software version 5.1 (Media Cybernetics Inc.). A series of images were taken for each otolith from the core to the post-rostrum. In a few instances this was not possible and images were taken from the core to the rostrum, but this should have had no influence on the final estimate of microincrement counts. Landmarks evident in each image were then used to allow microincrement counts across the series of images to obtain an estimate of the total number of microincrements for that otolith. This number provides a reasonable proxy for the total number of days the individual had been alive up until the point of capture (Hernaman et al.

2000). Repeat counts were taken for the same otolith until a consistent number was attained

(counts differing by no more than ~3 microincrements). A minimum of two counts was made for each otolith, and the average subsequently taken to provide the final estimate of fish age.

81

Statistical Analyses

All statistical analyses were performed using the statistical program R, version 3.0.1 (R

Development Core Team). All morphological characteristics were natural log transformed prior to analyses.

Relationship between predation and density: The extent predation impacts population density will determine the extent predation has other (potentially conflicting) indirect effects on age- specific mortality. I tested this using a Pearson correlation analysis of Morgan et al. (2014)’s estimates of predation rate against both adult and juvenile density. I examined both adult and juvenile density given that predation might impact age classes in different ways (e.g., compared Scenarios A and B in Fig. 1).

Assessing the potential causes of life-history variation among populations: I examined whether predation, adult density, or juvenile density explained variation in reproductive investment and growth rate using a model selection approach. Juvenile density was included here because competition for resources might be more pronounced or different to that in adults (e.g., Scenario B in Fig. 1). I created a number of biologically plausible statistical models. Some of these models consisted of a single predictor (predation, adult density, juvenile density), while others included a combination of predictors (predation with adult or juvenile density). A null model was also evaluated that did not include any predictor variable and effectively tested the extent to which population variation in reproductive effort or growth rate was random or related to factors other than predation or density.

Akaike’s Information Criterion with a correction for sample size (AICc) was used to evaluate the level of support for each model. The model with the lowest AICc value was interpreted as the best supported model, although any model within two AICc units of this

82 value might be considered equally plausible (Burnham & Anderson 2002). Model weights,

AICw, were also calculated to indicate the relative support of different models to one another

(Johnson & Omland 2004). Values of AICw range from 1.0 (exclusive support for a given model) to 0.0 (virtually no support for a given model).

To examine population differences in reproductive investment and growth rate I ran multivariate regressions using the lme4 package, version 0.99999911-8 (Bates 2008) in R.

All models described above included population as a random effect.

For models estimating reproductive investment, I included pre-anal length and the predictors of predation and density as fixed effects to estimate whether the proportion of tissue dedicated to reproduction (reflected by the intercept value in models) varied among populations as a function of predation or density. Interaction terms between pre-anal length and the predictors were also included in some models to determine whether the development rate of the ovaries (i.e. the effort put into developing reproductive tissue) varied as a function of predation or density (see Chapter 2). To control for the potential impact of any differences in the maturity of individuals surveyed for each population, I created a second set of models in which an additional random effect for the estimated maturity of each female was included

(based on the probable latest egg stage found in the reproductive tract).

For models estimating growth rate, I included age and the predictors of predation and density as fixed effects and their interaction with age. The interaction term was of specific interest because it indicates whether growth rate changed as a function of predation or density. I fitted two model sets. In the first set, populations were allowed to vary in their intercept values, which meant the size of larvae on hatching were assumed to vary among populations. In the second set, intercepts were forced through the origin and populations were subsequently assumed not to vary in larval size.

83

Evaluating Life-history trade-offs: I used a Pearson correlation in R of population parameter estimates of reproductive investment and growth rate to investigate whether these two traits varied inversely with one another as predicted by life history theory (Stearns 1992). The specific model used to provide population estimates of reproductive investment and growth rate was dependent on the model that received the most support in the analyses described above. For reproductive investment, the specific parameter of interested was found to be the relative weight of the ovaries (intercept), while for growth rate the only relevant parameter was the estimated slope (i.e. the change in length with age).

84

Results

My analyses were conducted in three parts. First I examined whether predation had the expected negative effect on population density (adult or juvenile). Second, I determined whether predation or density explained most of the variation in reproductive investment and growth rate among populations. Third, I investigated whether population variation in reproductive investment and growth rate were inversely related to one another.

The relationship between predation and population density

Neither the correlation between Morgans et al. (2014)’s estimates of predation rate and adult density, nor that between predation rate and juvenile density, was statistically significant (N =

5; p = .29 and .78 two-tailed, respectively). However, it is worth nothing the relationship between predation rate and adult density was noticeably negative (r = -.59; Fig. 2A), a pattern that was also weakly repeated for predation rate and juvenile density (r = -.18; Fig. 2B). It is possible that the absence of significant results was simply due to low sample size and therefore reduced statistical power. Thus, when also considering the magnitude of the trend in

Figure 2A, it seems reasonable not to completely dismiss the possibility that an interaction exists between predation rate and adult density. If an interaction were to exist, my results imply populations experiencing high predation rate would also tend to be those with lower adult densities (and vice versa). The obvious exception was Taga’chang, which had much higher population densities (both adults and juvenile) compared to the other four populations.

Causes of life history variation among populations

Reproductive investment: Predation was consistently supported as the best predictor of reproductive investment, irrespective of whether potential differences in the maturity of females sampled for populations was controlled for or not (Table 1). As predation increased,

85 populations decreased in their relative ovarian weight (Fig. 3). This finding was consistent with a scenario in which increased predation has had a disproportionally higher impact on younger females compared to older females (Scenario B, Fig. 1).

Growth rate: Both the null model and adult density were consistently ranked as the highest models for growth rate (Table 2). In the case of adult density, growth rate tended to decrease with increased adult density (Fig. 4), but the effect of density was dependent on whether populations were assumed to vary in larvae size or not (i.e., the starting position of the growth trajectory). A reduction in growth rate as a function of increased adult density was consistent with a situation in which density dependent effects (e.g., competition) results in disproportionately higher mortality in older females relative to younger females (Scenario A,

Fig. 1).

Life history trade-offs

I used the parameter values computed from the top supported models to provide an estimate of growth and reproduction for each population. Specifically I used the population parameters computed for reproductive investment based on the predation model (specifically that reported in Table 1A) and for growth rate based on the adult density model (the model reported in Table 2A). There was a noticeable negative relationship between reproductive investment and growth rate, but the relationship was not statistically significant (Pearson correlation: N = 5, r = -.59, p = .15 one-tailed; Fig. 5). Taga’chang was an obvious outlier appearing to have a lower estimated growth rate for its estimated investment in reproduction.

However, the remaining four populations displayed either high reproductive investment and low growth (e.g. Pago Bay) or low reproductive investment and high growth (e.g. Adelup

Point; NB: because Taga’chang appeared to be a prominent outlier in Figures 2 and 5, a

86 follow up series of analyses with Taga’chang removed was conducted for all analyses, but results were qualitatively unchanged to those reported in Tables 1 and 2).

87

Figure 2: The relationship between predation and (A) adult density and (B) juvenile density. Symbols are shaded according to the level of

predation, from white (low predation) to black (high predation). 88

Table 1: Predictors of reproductive investment between populations of the Pacific leaping blenny. (A) Models that do not consider variation due

to maturity. (B) Models in which maturity has been controlled for.

Model t-value A*: ovarian weight adult pre-anal length* AICc ∆AIC AICw predation predation Nfemales,populations=106,5 density predation 277.4 .0 .52 -3.21 pre-anal length*predation 278.5 1.0 .31 .87 -1.06 adult density 281.4 4.0 .07 null (pre-anal length) 282.8 5.4 .03 pre-anal length*adult density 283.6 6.1 .02 juvenile density 283.9 6.5 .02 pre-anal length*juvenile density 286.0 8.6 .01 B**: ovarian weight

89 Nfemales,populations=106,5

predation 227.4 .0 .47 -2.63 pre-anal length*predation 229.5 2.1 .16 adult density 229.9 2.5 .13 null (Preanal length) 230.9 3.5 .08 juvenile density 230.9 3.5 .08 pre-anal length *adult density 232.0 4.7 .05 pre-anal length*juvenile density 232.9 5.5 .03 Model support was evaluated using Akaike’s Information Criterion (AICc). Effect sizes (t-values) indicating the magnitude and direction of

effects are reported for all models receiving the most support (∆AIC ≤ 2.00). Models that included a population random effect for pre-anal length

are highlighted with an asterisk. Models that included a population random effect for pre-anal length, egg stage and their interaction are

highlighted with a double asterisk.

Figure 3: Ovarian weight progressively decreased as a population experienced increased predation (A). Also shown (B) are population

coefficients computed using the fitted model for predation reported in Table 1. The dashed lines around coefficients represent 95 % confidence

intervals. Symbols are shaded according to the level of predation, from white (low predation) to black (high predation). 90

Table 2: Predictors of growth rate among five populations of the Pacific leaping blenny. (A) Models that assume populations differ in larval size.

(B) Models in which larval size was assumed to be consistent among populations.

Model t-value A.*growth rate adult age* age* AICc ∆AIC AICw predation Nfemales,populations= 107,5 density predation adult density age*adult density -219.8 .0 .34 1.91 -2.03 age -219.5 .3 .29 age*predation -218.5 1.3 .18 -1.37 1.50 age*juvenile density -217.3 2.6 .09 age*adult density + age*predation -216.4 3.5 .06 age*juvenile density + age*predation -215.3 4.5 .04 B.**growth rate

Nfemales,populations= 107,5 age -224.8 .0 .37 91 age*adult density -223.7 1.1 .22 1.19 -1.41 age*predation -222.6 2.1 .13 age*adult density + age*predation -222.4 2.3 .12 age*juvenile density -222.1 2.7 .10 age*juvenile density + age*predation -221.3 3.4 .07 Model support was evaluated using Akaike’s Information Criterion (AICc). Effect sizes (t-values) indicating the magnitude and direction of

effects are reported for all models receiving the most support (∆AIC ≤ 2.00). Models that included a population random effect for age are

highlighted with an asterisk. Models that included a population random slope for age, but assume populations share an intercept of zero are

highlighted with a double asterisk.

Figure 4: Population growth rate progressively decreased with increasing adult density (A). Also shown (B) are population coefficients

computed using the fitted model for adult density reported in Table 2. See Figure 3 legend for further details. 92

Figure 5: The relative ovarian weight of populations decreased as growth rate increased.

Plots show population coefficients computed from the fitted model reported in Tables 1 and

2. See Figure 3 and 4 legends for further details.

93

Discussion

I identified variation in life history among populations of the Pacific leaping blenny as a function of differences in predation rate and to some extent adult density. This would only be expected if predation and density have resulted in certain changes in age-specific mortality among these populations (Fig. 1). The effects of predation and density-dependent effects on mortality were also unlikely to have operated independently of each other: populations experiencing high predation were also populations that tended to have low densities (Fig. 2).

This appears to have led to a complex interaction between these two factors on age-specific mortality within populations, resulting in the relative impact of predation and density on reproduction and growth tending to vary: reproductive investment was exclusively predicted by predation rate (Table 1), whereas growth rates might have been affected by adult density, predation or neither of these factors directly (Table 2).

Reproductive investment seemed to be inversely related to growth rate (Fig. 5), which is a classic trade-off that underlies many of the predictions from life history theory (e.g., Fig

1; Stearns 1992; Zera & Harshman 2001). This trade-off also implies that increased predation has probably resulted in both reduced reproductive investment and high growth rate among populations. More specifically, predation on the Pacific leaping blenny has had a disproportionally higher impact on younger females in a population than older females

(Scenario B in Fig. 1; see also Gadgil & Bossert 1970; Law 1979; Michod 1979). Yet growth rate seemed to be more directly affected by changes in adult density rather than predation, and in a way that would imply a different pattern of age-specific mortality in populations: density-dependent effects (e.g., increased resource competition) have had a disproportionately higher impact on mortality in older females compared to younger females

(Scenario A in Fig. 1; see also Gadgil & Bossert 1970; Law 1979; Michod 1979). It is possible that predation and adult density do have these different impacts on age-specific

94

mortality within the Pacific leaping blenny. Alternatively, the effect of predation on the relative mortality of early versus older females might actually remain relatively constant as predation rate increases (e.g. Reznick et al. 1996). Instead, it was the reduction in adult density caused by predation more specifically (and not the density of younger individuals in populations; e.g., see Fig. 2) that increased the survival of older females and prompted the corresponding changes in life history. In either case, the interaction between predation and density on life history has clearly not been straightforward in the Pacific leaping blenny, which provides some insight into why variation in life history among taxa might not appear to conform to theoretical predictions when factors causing mortality are tested in isolation of each other (e.g. Blondel et al. 1993; Reznick et al. 1996; Ernande et al. 2004; see also Wilbur et al. 1974). The next obvious step for further work on the Pacific leaping blenny will be to determine the precise pattern of age-specific mortality exhibited by these populations in response to changes in predation and density.

Although the details remain unclear, my current research does confirm that predation has played some role in driving life history variation among populations of the Pacific leaping blenny: differences among populations in reproductive investment were consistently and predictably associated with predation, and that there was some evidence that growth rates of populations were dependent on predation rates as well. Moreover, the differences I found among populations in reproduction and growth in response to predation (or through its indirect effects on density) have occurred despite considerable gene flow among these populations (Cooke et al. In Review). This implies that the life history strategy expressed by populations in response to predation might be plastic in origin.

The ability to express different life history phenotypes depending on the environment has important implications for an organism’s ability to colonise new habitats (Gutowsky &

Fox 2012; Sol et al. 2012; Riesch et al. 2013). Plastic life history strategies would provide a

95

means for organisms to persist in novel habitats, especially those that are vastly different in the conditions experienced (Ernande et al. 2004). For example, the Pacific leaping blenny plainly has its evolutionary origins in an aquatic environment, where predation and density- dependent effects are likely to have been different than those experienced on land (e.g., see

Morgans & Ord 2013). In Chapter 2, I documented clear changes in life history following the colonisation of land by blennies. Specifically, as blenny species have progressively transitioned from a marine to terrestrial environment, species have increased their reproductive investment and reduced their growth rates. The findings reported in this chapter suggest that such adaptive changes in life history might have had their initial origins as plastic responses to changes in age-specific mortality while the transition was being made to life on land.

The finding that life history can vary among populations experiencing high contemporary gene flow is inconsistent with the notion that reproductive isolation is essential for life history variation to occur among populations (Blondel et al. 1992; Gale et al. 2013).

The Pacific leaping blenny is a short-lived fish (e.g., the oldest female included in my study was 289 days old) and likely reproduces only a handful of times in its lifetime (e.g., see

Bhikajee et al. 2006). The opportunity to change the relative investment in reproduction and growth via plasticity must therefore occur early in life. We know that post-settlement populations on land of the Pacific leaping blenny are probably ecologically isolated from one another, and most (if not all) dispersal among populations is limited to pelagic larvae.

Adaptive plasticity in life history would seem to be ideal for this fish, as it would allow an individual to tailor its subsequent investment in reproduction or growth depending on the type of the environment that individual happens to finds itself in post-settlement on land. In this respect, it was perhaps not surprising that life history variation in the Pacific leaping blenny was probably plastic.

96

Taken together, I have shown that predation may have important and unanticipated effects on population density and life history variation among populations (e.g., see also

Walsh & Reznick 2008). This highlights the need to consider carefully what the potential influences on age-specific mortality might be within a population if the life history strategy expressed by that population is to be fully understood. Furthermore, my study showed that detectable and presumably adaptive, plastic changes in life history can occur among highly connected populations, although the extent to which this occurs in other organisms is unclear given the unique ecology of the Pacific leaping blenny. However, plastic life history strategies would improve population persistence for any species colonising new habitats, which may in turn facilitate other adaptive changes in phenotype and ultimately the formation of new species. Plasticity in life history might therefore play a greater role in the spread of organisms into new environments than is currently appreciated (e.g., see Sol et al. 2012).

AKNOWLEDGEMENTS

I would like to thank Jennifer McIlwain and Andrew Halford for logistical support, Matt

Taylor, Georgina Cooke, and Terry Ord for catching the specimens, Iain Suthers for the use of his laboratory equipment, Ash Fowler for his laboratory training and project support, and my supervisor Terry Ord for his extensive project support.

Literature Cited

Abrams, P. A. and L. Rowe. 1996. The effects of predation on the age and size of maturity of

prey. Evolution 50:1052-1061.

Bates, D. 2008. lme4: linear mixed-effects models using S4 classes: R package, 0.999375–28,

http://lme4.r-forge.r-project.org/.

Belk, M. C. 1995. Variation in growth and age at maturity in bluegill sunfish: genetic or

97

environmental effects? Journal of Fish Biology 47:237-247.

Bhikajee, M., J. Green, and R. Dunbrack. 2006. Life history characteristics of Alticus

monochrus, a supratidal blenny of the southern Indian Ocean. African Zoology 41:1-

7.

Blondel, J., P. C. Dias, M. Maistre, and P. Perret. 1993. Habitat heterogeneity and life-history

variation of mediterranean blue tits (Parus caeruleus). The Auk 110:511-520.

Blondel, J., P. Perret, M. Maistre, and P. Dias. 1992. Do harlequin Mediterranean

environments function as source sink for blue tits (Parus caeruleus L.)? Landscape

Ecology 6:213-219.

Brothers, E. B., C. P Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from

larval and adult fishes. Fisheries Bulletin 74:1-8.

Brown, C. R., M. S. Gordon, and H. G. Chin. 1991. Field and laboratory observations on

microhabitat selection in the amphibious red sea rockskipper fish, Alticus kirki (family

Blenniidae). Marine Behaviour and Physiology 19:1-13.

Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodal inference: a

practical information-theoretic approach. Springer, New York.

Cameron, T. C., D. O'Sullivan, A. Reynolds, S. B. Piertney, and T. G. Benton. 2013. Eco-

evolutionary dynamics in response to selection on life-history. Ecology Letters

16:754-763.

Campana, S. E., and J. D. Neilson. 1985. Microstructure of fish otoliths. Canadian Journal of

Fisheries and Aquatic Sciences 42:1014-1032.

Carlson, B. A. 1992. The life history and reproductive success of the coral blenny, Exallias

brevis (Kner, 1868). PhD thesis: University of Hawaii.

Carvalho, M. G., A. S. Moreira, C. Moreira, H. Queiroga, P. T. Santos, and A. T. Correia. 2014.

Validation of otolith daily increments in early juveniles of shanny Lipophrys pholis.

98

Journal of Fish Biology 84:1234-1239.

Charlesworth, B. 1980. Evolution in age-structured populations. Cambridge University Press,

Cambridge, UK.

Clelland, E. S., Q. Tan, A. Balofsky, R. Lacivita, and C. Peng. 2007. Inhibition of premature

oocyte maturation: a role for bone morphogenetic protein 15 in zebrafish ovarian

follicles. Endocrinology 148:5451-5458.

Cooke, G. M., Schlub, T. E., Sherwin, W. B., Ord, T. J. In Review. Understanding the spatial

scale of genetic connectivity at sea: unique insights from a land fish and a meta-

analysis.

Ernande, B., U. Dieckmann, and M. Heino. 2004. Adaptive changes in harvested populations:

plasticity and evolution of age and size at maturation. Proceedings of the Royal

Society of London, Series B 271:415-423.

Furness, A. I., M. R. Walsh, and D. N. Reznick. 2012. Convergence of life-history

phenotypes in a Trinidadian killifish (Rivulus hartii). Evolution 66:1240-1254.

Gadgil, M. and W. H. Bossert. 1970. Life historical consequences of natural selection. The

American Naturalist 104:1-24.

Gale, B. H., J. B. Johnson, G. B. Schaalje, and M. C. Belk. 2013. Effects of predation

environment and food availability on somatic growth in the livebearing fish

Brachyrhaphis rhabdophora (Pisces: Poeciliidae). Ecology and Evolution 3:326-333.

García-Ramos, G. and M. Kirkpatrick. 1997. Genetic models of adaptation and gene flow in

peripheral populations. Evolution 51:21-28.

Ghani, N. I. A., G. Herczeg, T. Leinonen, and J. Merila. 2013. Evidence for genetic

differentiation in timing of maturation among nine-spined stickleback populations.

Journal of Evolutionary Biology 26:775-782.

Gutowsky, L. F. G., and M. G. Fox. 2012. Intra-population variability of life-history traits

99

and growth during range expansion of the invasive round goby, Neogobius

melanostomus. Fisheries Management and Ecology 19:78-88.

Hampton, J. 2000. Natural mortality rates in tropical tunas: size really does matter. Canadian

Journal of Fisheries and Aquatic Sciences 57:1002-1010.

Hendry, A. P., E. B. Taylor, and J. D. McPhail. 2002. Adaptive divergence and the balance

between selection and gene flow: lake and stream stickleback in the misty system.

Evolution 56:1199-1216.

Hernaman, V., P. L. Munday, and M. L. Schläppy. 2000. Validation of otolith growth-

increment periodicity in tropical gobies. Marine Biology 137:715-726.

Hernández-Miranda, E., R. Veas, C. Valeria Espinoza, S. R. Thorrold, and F. P. Ojeda. 2009. The

use of otoliths and larval abundance for studying the spatial ecology of the blenny

Scartichthys viridis (Valenciennes, 1836) in coastal central Chile. Revista de Biologia

Marina y Oceanografia 44:619-633.

Hsieh, S. T. T. 2010. A locomotor innovation enables water-land transition in a marine fish.

PLoS ONE 5:e11197.

Hutchings, J. A. 1993. Adaptive life histories effected by age-specific survival and growth

rate. Ecology 74:673-684.

Johansson, F., R. Stoks, L. Rowe, and M. de Block. 2001. Life history plasticity in a

damselfly: effects of combined time and biotic constraints. Ecology 82:1857-1869.

Johnson, J. B. and K. S. Omland. 2004. Model selection in ecology and evolution. Trends in

Ecology & Evolution 19:101-108.

Kruitwagen, G., T. Hecht, H. B. Pratap, and S. E. Wendelaar Bonga. 2006. Changes in

morphology and growth of the mudskipper (Periophthalmus argentilineatus)

associated with coastal pollution. Marine Biology 149:201-211.

Law, R. 1979. Optimal life histories under age-specific predation. The American Naturalist

100

114:399-417.

Lenormand, T. 2002. Gene flow and the limits to natural selection. Trends in Ecology &

Evolution 17:183-189.

Mansur, L., D. Catalan, G. Plaza, M. F. Landaeta, and F. P. Ojeda. 2013. Validations of the daily

periodicity of increment deposition in rocky intertidal fish otoliths of the south-eastern

Pacific Ocean. Revista de Biologia Marina y Oceanografia 48:629-633.

Martin, K. L. and J. R. Lighton. 1989. Aerial CO2 and O2 exchange during terrestrial activity

in an amphibious fish, Alticus kirki (Blenniidae). Copeia 1989:723-727.

Mazlan, A. G., and M. Rohaya. 2008. Size, growth and reproductive biology of the giant

mudskipper, Periophthalmodon schlosseri (Pallas, 1770), in Malaysian Waters. Journal of

Applied Ichthyology 24:290-296.

Meekan, M., L. Vigliola, A. Hansen, P. Doherty, A. Halford, and J. Carleton. 2006. Bigger is

better: size-selective mortality throughout the life history of a fast-growing clupeid,

Spratelloides gracilis. Marine Ecology Progress Series 317:237-244.

Michod, R. E. 1979. Evolution of life histories in response to age-specific mortality factors.

American Naturalist 113:531-550.

Morales-Nin, B. 2000. Review of the growth regulation processes of otolith daily increment

formation. Fisheries Research. 46:53-67.

Morgans, C. L., Cooke, G. M., Ord, T. J. 2014. How populations differentiate despite gene

flow: sexual and natural selection drive phenotypic divergence within a land fish, the

Pacific leaping blenny. BMC Evolutionary Biology 14:97.

Morgans, C. L. and T. J. Ord. 2013. Natural selection in novel environments: predation

selects for background matching in the body colour of a land fish. Animal Behaviour

86:1241-1249.

Natsumeda, T., M. Yuma, and M. Hori. 2013. Variation in life-history traits of male Japanese

101

fluvial sculpin Cottus pollux in relation to nest abundance along a stream course.

Environmental Biology of Fishes 96:1123-1133.

Núñez, J. and F. Duponchelle. 2009. Towards a universal scale to assess sexual maturation

and related life history traits in oviparous teleost fishes. Fish Physiology and

Biochemistry 35:167-180.

Ord, T. J. and S. Tonia Hsieh. 2011. A highly social, land-dwelling fish defends territories in

a constantly fluctuating environment. Ethology 117:918-927.

Osenberg, C. W., G. G. Mittelbach, and P. C. Wainwright. 1992. Two-stage life histories in

fish: the interaction between juvenile competition and adult performance. Ecology

73:255-267.

Peterson, I. and J. S. Wroblewski. 1984. Mortality rate of fishes in the pelagic ecosystem.

Canadian Journal of Fisheries and Aquatic Sciences 41:1117-1120.

Pianka, E. R. and W. S. Parker. 1975. Age-specific reproductive tactics. The American

Naturalist 109:453-464.

Reznick, D. and A. P. Yang. 1993. The influence of fluctuating resources on life history:

patterns of allocation and plasticity in female guppies. Ecology 74:2011-2019.

Reznick, D. N., M. J. Butler IV, F. H. Rodd, and P. Ross. 1996. Life-history evolution in

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

selection. Evolution 50:1651-1660.

Reznick, D. N., R. D. Bassar, J. Travis, and F. H. Rodd. 2012. Life-history evolution in

guppies VIII: the demographics of density regulation in guppies (Poecilia reticuata).

Evolution 66:2903-2915.

Riesch, R., R. A. Martin, and R. B. Langerhans. 2013. Predation's role in life-history

evolution of a livebearing fish and a test of the Trexler-DeAngelis model of maternal

provisioning. The American Naturalist 181:78-93. 102

Roff, D. A. 1992. The evolution of life histories: theory and analysis. Chapman and Hall

Incorporated, New York, NY.

Schroder, A., A. van Leeuwen, and T. C. Cameron. 2014. When less is more: positive

population-level effects of mortality. Trends in Ecology and Evolution 29:614-624.

Selman, K., R. A. Wallace, A. Sarka, and X. Qi. 1993. Stages of oocyte development in the

zebrafish, Brachydanio rerio. Journal of Morphology 218:203-224.

Shimizu, N., Y. Sakai, H. Hashimoto, and K. Gushima. 2006. Terrestrial reproduction by the

air-breathing fish Andamia tetradactyla (Pisces; Blenniidae) on supralittoral reefs.

Journal of Zoology 269:357-364.

Sogard, S. M. 1997. Size-selective mortality in the juvenile stage of teleost fishes: a review.

Bulletin of Marine Science 60:1129-1157.

Sol, D., J. Maspons, M. Vall-llosera, I. Bartomeus, G. E. García-Peña, J. Piñol, and R. P.

Freckleton. 2012. Unraveling the life history of successful invaders. Science 337:580-

583.

Stearns, S. C. 1992. The evolution of life histories. Oxford University Press, Oxford.

Stearns, S. C., and J. C. Koella. 1986. The evolution of phenotypic plasticity in life-history

traits: predictions of reaction norms for age and size at maturity. Evolution 40:893-

913.

Sunobe, T., T. Ohta, and A. Nakazono. 1995. Mating system and spawning cycle in the

blenny, Istiblennius enosimae, at Kagoshima, Japan. Environmental Biology of Fishes

43:195-199.

Torres Dowdall, J., C. A. Handelsman, E. W. Ruell, S. K. Auer, D. N. Reznick, and C. K.

Ghalambor. 2012. Fine-scale local adaptation in life histories along a continuous

environmental gradient in Trinidadian guppies. Functional Ecology 26:616-627.

Tyler, C. R. and J. P. Sumpter. 1996. Oocyte growth and development in teleosts. Reviews in

103

Fish Biology and Fisheries 6:287-318.

Walsh, M. R., and D. N. Reznick. 2008. Interactions between the direct and indirect effects of

predators determine life history evolution in a killifish. Proceedings of the National

Academy of Sciences 105:594-599.

Wellborn, G. A. 1994. Size-biased predation and prey life histories: a comparative study of

freshwater amphipod populations. Ecology 75:2104-2117.

Wilbur, H. M., D. W. Tinkle, and J. P. Collins. 1974. Environmental certainty, trophic level,

and resource availability in life history evolution. The American Naturalist 108:805-

817.

Zera, A. J., and L. G. Harshman. 2001. The physiology of life history trade-offs in animals.

Annual Review of Ecology and Systematics 32:95-126.

104

CHAPTER 4

Conclusion

105 My thesis addressed two questions: whether survival may have improved for fish that moved onto land, and what the relative role of predation and density was for determining life history variation among populations within one of these land species.

For the first question (Chapter 2) I took advantage of the fish families Gobiidae and

Blenniidae, from which select species have adapted to life on land. In both cases I compared the life history traits of aquatic and terrestrial individuals, using life history theory to determine whether the difference in traits was consistent with the notion of improved survival on land. For Gobiidae, I conducted a broad-scale comparison of aquatic and terrestrial species via meta-analysis. In the case of Blenniidae, I conducted a fine-scale species comparison using a gradient of lifestyles from aquatic to terrestrial. In both cases I compared the life history traits of growth, reproductive investment, egg number, and egg size. It was expected that survival should have improved for fish following the transition onto land because it may have freed the colonists from environmental pressures such as predation (ecological release), or provided colonists with new or previously rare resources (ecological opportunity). This improvement in survival is predicted by theory to be reflected in changes in the investment in reproduction over growth relative to fish remaining in water. Furthermore, by contrasting life history changes between the two families, I could determine what type of change in survival might have accompanied the colonisation of land (e.g., increase survival for older age individuals compared to younger aged individuals).

I found distinct life history differences between aquatic and terrestrial species in both families that matched those predicted by life history theory under a scenario of improved survival for fish that have colonised land. Nevertheless, my results also indicated that the nature of improved survival probably differed among age classes in the two families. For gobies, reproductive investment decreased and growth rate increased with the move onto land, implying the transition benefited older age survival relative to early ages. In contrast,

106 the opposite life history changes were observed for blennies, implying in this family the move onto land benefited early age survival relative to older ages. These differences suggested that the way in which colonisation of land occurred in each family probably occurred quite differently, and I discussed several ideas on this topic.

For the second question (Chapter 3), I compared measurements of growth and reproductive effort between separate populations of the same fish living on land – the Pacific leaping blenny, Alticus arnoldorum. By focussing on this particular species, I was able to compare estimates of predation and density (Morgans et al. 2014) as a function of inter- population variation in growth and reproduction. Both theoretical work and past empirical studies inferred that predation would play a primary or singular role in determining life history variation among populations. By comparing model support for predation and density,

I was able to test whether predation had this singular strong effect over changes in density.

Furthermore, by determining how life history traits changed in response to predation and density, I was able to evaluate how these factors might have influenced age-specific mortality in the Pacific leaping blenny.

My results were complex, but generally indicated that predation has been an important selection pressure shaping life history variation among populations of the Pacific leaping blenny. My findings were therefore consistent with past empirical studies that found predation played a key role in life history variation (e.g. Reznick & Endler 1982), but also suggested density was likely important as well. Furthermore, the changes I observed in reproductive investment and growth rate occurred despite considerable gene flow between the populations (Cooke et al. In Review). This implies that the differences in life history among populations were likely plastic in origin, and such plastic changes in life history would presumably allow organisms to persist in new or rapidly changing environments.

107 In both Chapters 2 and 3 I demonstrated how life history theory provides a window into the ecological conditions surrounding an organism by identifying differences in survival between species and populations respectively. By using life history theory, I found evidence that an ecological release or new ecological opportunities might have allowed fish to move onto land. Notably, the key factors commonly associated with an ecological release and ecological opportunity are predation and density-dependent resource competition, which I found to be associated with predictable life history changes among populations of the land fish, the Pacific leaping blenny. Collectively, my results support the conclusion that predation and resource competition are key factors likely impacting the survival and reproduction of land fish. While my results do not provide definitive evidence that these factors specifically caused fish to move onto land, they do emphasise the plausibility of these explanations.

Additionally it is also possible, but as yet unclear, that plastic life history responses to predation and resource competition facilitated the move by fish onto land by allowing the first colonists to persist in their new habitat.

Finally, in this thesis I demonstrated how fish that live on land provide a unique opportunity to investigate life history and ecological theory (e.g., why animals colonise new environments). While my study focused on fish species with highly terrestrial lifestyles, there are a number of other fish that spend part of their lives out of the water (e.g. see Sayer &

Davenport 1991) and many invertebrates known to have also transitioned from aquatic to terrestrial environments (e.g. see Giomi et al. 2014). These systems will likely provide many more research opportunities into the future.

108 Literature Cited

Cooke, G. M., Schlub, T. E., Sherwin, W. B., Ord, T. J. In Review. Understanding the spatial

scale of genetic connectivity at sea: unique insights from a land fish and a meta-

analysis.

Giomi, F., M. Fusi, A. Barausse, B. Mostert, H.-O. Pörtner, and S. Cannicci. 2014. Improved

heat tolerance in air drives the recurrent evolution of air-breathing. Proceedings of the

Royal Society B: Biological Sciences 281: EarlyView.

Morgans, C. L., Cooke, G. M., Ord, T. J. 2014. How populations differentiate despite gene

flow: sexual and natural selection drive phenotypic divergence within a land fish, the

Pacific leaping blenny. BMC Evolutionary Biology 14:97.

Reznick, D., and J. A. Endler. 1982. The impact of predation on life history evolution in

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

Sayer, M. D. J., and J. Davenport. 1991. Amphibious fish: why do they leave water? Reviews

in Fish Biology and Fisheries 1:159-181.

109 APPENDIX

Chapter 2

110 Figure A1: The short-lived goby species from Hernaman & Munday 2005 were clear outliers in growth rate and were removed from analyses.

Note: aquatic species represented by black plus symbols are from temperate climates.

111 Figure A2: Oocyte size frequency histogram from a Praealticus labrovittatus individual with oocytes in predominantly stages 2 and 4. Note this histogram was generated using the oocytes from both ovaries.

112 Figure A3: Oocyte size frequency histogram from a Praealticus labrovittatus individual with oocytes in predominantly stages 2 and 3. Note this histogram was generated using the oocytes from both ovaries.

113 Figure A4: Egg number for gobies from tropical and subtropical climates.

114 Table A1: Stages of oocyte classification

Stage Oocyte Description 1 Primary Oocytes tiny and clear, although some may appear slightly opaque. If this is the Growth latest oocyte stage present then the gonads should be tiny, skinny, opaque, and (Immature take up very little space in the gut cavity. This stage was not included in the oocytes) oocyte size/number data. All oocytes <200 microns in diameter for all species. 2 Cortical Oocytes small and whitish-opaque to white in color. The smallest oocytes in this alveolus stage stage are not always as round as the later stages, with some appearing more (maturing rectangular or elliptical. The germinal vesicle is often visible, normally appearing oocytes) as a darker opaque zone in the middle of the oocyte. The space around the germinal vesicle extending to the oocyte wall is a thin zone compared to that of stage 3 oocytes, and may not be as well defined. The oocytes outer tissue layer is thin and faint. Ovaries in which there are no oocytes more developed than stage 2 are white in color and take up a small to moderate portion of the gut cavity. Oocyte size frequency plots taken from gonads containing oocytes no more developed that stage 2 have unimodal distribution with a broad range of sizes. 3 Vitellogenesis This stage was phenotypically diverse across species. Traits shared across species (advanced included being large in size, very rounded, having thick tissue walls, and maturation) occasionally having gained faint polarity. Yellowish coloration very common, although some were clear. Ovaries containing this oocyte stage were large and yellowish, taking up a considerable portion of the abdominal cavity. Oocyte size frequency plots for these ovaries were bimodal, with the stage 3 oocytes represented by a distinct peak of similar sized objects. 4 Maturation Oocytes large and ready or near ready for spawning. They are clear/translucent and Mature (although one species—Praealticus labrovittatus—appeared opaque yellow), and Oocytes normally have clear polarity (a top and bottom) as indicated by a well defined white cap at one end. The most advanced of these oocytes had a well-developed thick adhesive pad at one pole. The germinal vesicle had moved/attached to the oocyte wall. Large yellow bodies may be visible within the oocyte. Ovaries containing this stage of oocyte fill most of the abdominal cavity, with limited room left for the digestive tract. Very little effort is required to separate these oocytes as they are free within the ovaries. Oocyte size frequency plots of ovaries containing this stage, are bimodal with the stage 4 oocytes represented by a distinct peak that is very homogenous in size. 5 Left-over Oocytes not released at spawn. Few in number, these oocytes appear dull and oocytes from partially or wholly deflated. Ovaries containing this stage of oocyte are spawning recovering from spawning and beginning to develop the next batch of eggs. As a (spawned and result they do not take up as much of the gut cavity as would be initially expected recovering/ given the size of a female. Furthermore all oocytes other than those in stage 5 are resting) generally of an early developmental stage (e.g. stage 1).

115 Table A2: Variation in life history among tropical and subtropical goby species as a function of lifestyle and other factors

Variable 95% CI β t (lower, upper) 2 A. Growth rate: F3,8 = 5.75, p = .02, R = .68 Intercept 3.77 3.39, 4.15 19.67 Length .34 -.36, 1.04 .96 Lifestyle .05 -.53, .63 .16 Lifestyle*Length .99 -.23, 2.21 1.59 2 B. Egg number: F2,15 = 112.5, p < .001, R = .94 Intercept 7.34 7.01, 7.67 43.58 Length 1.68 1.31, 2.05 8.86 Lifestyle .89 .08, 1.7 2.16 Parameter values with 95 % confidence intervals that do not overlap zero are considered statistically significant effects (p < 0.05). The lifestyle of species was coded 0 to 3 depending on the prepotency of terrestrial behaviour exhibited by the species. An interaction term between length and lifestyle was initially included in B but failed to account for significant variation in egg number and was subsequently removed from the final model.

116 Table A3: Variation in life history among blenny species as a function of lifestyle and other factors. Models account for (A) changes in length and ovarian weight with ovarian maturity, and (B) variation in egg size with egg stage.

Variable 95% CI β t (lower, upper) A. Reproductive effort Intercept -9.88 -11.77, -7.99 -10.27 Length 2.05 1.60, 2.50 8.93 Lifestyle -5.57 -7.92, -3.22 -4.64 Lifestyle*Length 1.35 .76, 1.94 4.48 N observations, species 161,6 B. Egg size Intercept 4.71 4.63, 479 57.83 Egg stage .52 .50, .54 22.65 Lifestyle .03 -.01, .07 1.39 N observations, species 170, 6 Note: length was removed from model B to provide more accurate parameter estimates for egg stage and lifestyle.

117 Table A4: Species and sample sizes of (A) blennies and life history characteristics of (B) gobies examined in this study.

Species: Lifestyle Growth Gonads Egg/ Egg/ Source /GSI oocyte oocyte number size A. Blenniidae Alticus arnoldorum Terrestrial (3) 26 18 14 14 field collected Entomacrodus Highly 17 23 15 15 field collected striatus amphibious (2) Praealticus Mildly 25 35 15 15 field collected labrovittatus amphibious (1) Praealticus poptae Mildly 23 37 15 15 field collected amphibious (1) Istiblennius Aquatic (0) 25 46 15 15 field collected lineatus Istiblennius Aquatic (0) 3 2 2 2 field collected edentulus B. Gobiidae Aphia minuta Aquatic (0) Y Y Iglesias et al. 1997; La Mesa 1999; Iglesias & Morales-Nin 2001 Asterropteryx Aquatic (0) Y Y Y Privitera 2001; semipunctata Privitera 2002 Boleophthalmus Terrestrial (1) Y Ravi & boddarti Rajagopal 2007 Boleophthalmus Terrestrial (1) Y Y Nanami & pectinirostris Takegaki 2005, Kim et al. 2011 Coryphopterus Aquatic (0) Y Kramer & dicrus Patzner 2008 Coryphopterus Aquatic (0) Y Kramer & thrix Patzner 2008 Coryphopterus Aquatic (0) Y Kramer & venezuelae Patzner 2008 Crystallogobius Aquatic (0) Y Caputo et al. linearis 2003 Deltentosteus Aquatic (0) Y Y Metin et al. quadrimaculatus 2011

118 Table A4: Continued. Species: Lifestyle Growth Gonads Egg/ Egg/ Source /GSI oocyte oocyte number size Eviota fasciola Aquatic (0) Y Privitera 2002 Eviota melasama Aquatic (0) Y Depczynski & Bellwood 2006 Eviota prasina Aquatic (0) Y Privitera 2002 Eviota prasites Aquatic (0) Y Privitera 2002 Eviota Aquatic (0) Y Depczynski & queenslandica Bellwood 2006 Eviota sigillata Aquatic (0) Y Depczynski & Bellwood 2006 Evorthodus lyricus Aquatic (0) Y Foster & Fuiman 1987 Favonigobius Aquatic (0) Y Taylor & lateralis Willis 1998 Gillichthys mirabilis Aquatic (0) Y Y Weisel 1947 Gobius bucchichii Aquatic (0) Y Sasal et al. 1996 Gobius cobitis Aquatic (0) Y Gibson 1970 Gobius niger Aquatic (0) Y Y Arruda et al. 1993; Iglesias et al. 1997; Filiz & Togulga 2009 Gobius paganellus Aquatic (0) Y Y Y Miller 1961; Azevedo & Simas 2000 Gobius roulei Aquatic (0) Y Y Y Y Kovacic 2001 Gobius vittatus Aquatic (0) Y Kovacic 2006 Lepidogobius Aquatic (0) Y Y Grossman lepidus 1979 Lesueurigobius Aquatic (0) Y Y Y Y Gibson & Ezzi friesii 1978; Nash 1982 Microgobius gulosus Aquatic (0) Y Y Y Provancha & Hall 1991 Neogobius Aquatic (0) Y Gumus & Kurt melanostomus 2009 Neogobius Aquatic (0) Y Y Y Engin & platyrostris Seyhan 2010

119 Table A4: Continued. Species: Lifestyle Growth Gonads Egg/ Egg/ Source /GSI oocyte oocyte number size Paragobiodon Aquatic (0) Y Lassig 1976 echinocephalus Paragobiodon Aquatic (0) Y Lassig 1976 lacunicola Paragobiodon Aquatic (0) Y Lassig 1976 xanthosoma Periophthalmodon Terrestrial (1) Y Y Y Tsuhako et al. schlosseri 2003; Mazlan & Rohaya 2008 Periophthalmus Terrestrial (1) Y Y Tsuhako et al. argentilineatus 2003; Kruitwagen et al. 2006 Periophthalmus Terrestrial (1) Y Y Y Etim et al. barbarus 2002; Udo 2002; Chukwu & Deekae 2010; Lawson 2010; Lawson 2011 Periophthalmus Terrestrial (1) Y Baek et al. magnuspinnatus 2008 Periophthalmus Terrestrial (1) Y Takeda et al. minutus 2011 Periophthalmus Terrestrial (1) Y Tsuhako et al. modestus 2003 Pomatoschistus Aquatic (0) Y Y Claridge et al. lozanoi 1985 Pomatoschistus Aquatic (0) Y Y Fouda et al. marmoratus 1993; Koutrakis & Tsikliras 2009 Pomatoschistus Aquatic (0) Y Arruda et al. microps 1993; Pampoulie 2001 Pomatoschistus Aquatic (0) Y Y Claridge et al. minutus 1985; Chukwu & Deekae 2010 Pomatoschistus Aquatic (0) Y Y Gibson & Ezzi norvegicus 1981

120 Table A4: Continued. Species: Lifestyle Growth Gonads Egg/ Egg/ Source /GSI oocyte oocyte number size Pseudapocryptes Terrestrial (1) Y Y Dinh 2008 elongatus Pseudogobius Aquatic (0) Y Y Gill et al.1996 olorum Scartelaos gigas Terrestrial (1) Y Y Y Park et al. 2008; Kim et al. 2011 Scartelaos tenuis Terrestrial (1) Y Abdoli et al. 2010 Sufflogobius Aquatic (0) Y Melo & Clus bibarbatus 2005 Thorogobius Aquatic (0) Y Dunne 1976 ephippiatus Tridentiger Aquatic (0) Y Kramer & trigonocephalus Patzner 2008 Trimma grammistes Aquatic (0) Y Y Sunobe 1995 Trimma okinawae Aquatic (0) Y Y Sunobe 1995 Trimmatom sp. Aquatic (0) Y Y Sunobe 1995 Zosterisessor Aquatic (0) Y Hajji et al. ophiocephalus 2013 Note: Life style was coded in statistical analyses from 0 to 3 for blennies and 0 to 1 for

gobies. Numbers under trait headings indicate sample sizes.

Literature Cited

Abdoli, L., E. Kamrani, A. Abdoli, and B. Kiabi. 2010. Age and growth of the

mudskipper, Scartelaos tenuis (Day, 1876) (Gobiidae: Oxudercinae), in the

coastal areas of the Persian Gulf at Bushehr province, Iran. Zoology in the Middle

East 51:113-115.

Arruda, L. M., J. Azevedo, and A. I. Neto. 1993. Abundance, age-structure and growth

and reproduction of gobies (Pisces, Gobiidae) in the Ria de Aveiro Lagoon

(Portugal). Estuarine, Coastal and Shelf Science 37:509-523.

121 Azevedo, J., and A. Simas. 2000. Age and growth, reproduction and diet of a sublittoral

population of the rock goby Gobius paganellus (Teleostei, Gobiidae).

Hydrobiologia 440:129-135.

Baeck, G., T. Takita, and Y. Yoon. 2008. Lifestyle of Korean mudskipper

Periophthalmus magnuspinnatus with reference to a congeneric species

Periophthalmus modestus. Ichthyological Research 55:43-52.

Caputo, V., M. L. Mesa, G. Candi, and P. N. Cerioni. 2003. The reproductive biology of

the crystal goby with a comparison to that of the transparent goby. Journal of Fish

Biology 62:375-385.

Chukwu, K., and S. Deekae. 2010. Growth, mortality and recruitment of Periopthalmus

barbarus (Linneaus 1766) in New Calabar River, Nigeria.

Claridge, P. N., M. W. Hardisty, I. C. Potter, and C. V. Williams. 1985. Abundance, life

history and ligulosis in the gobies (teleostei) of the inner Severn Estuary. Journal

of the Marine Biological Association of the United Kingdom 65:951-968.

Depczynski, M., and D. R. Bellwood. 2006. Extremes, plasticity, and invariance in

vertebrate life history traits: insights from coral reef fishes. Ecology 87:3119-

3127.

Dinh, T. D. 2008. Some aspects of the biology and population dynamics of the goby

Pseudapocryptes elongates (Cuvier 1816) in the Mekong delta. unpublished PhD

thesis, University of Malaysia Terengganu.

Dunne, J. 1976. Littoral and benthic investigations on the west coast of Ireland: V

(section A: faunistic and ecological studies). A contribution to the biology of the

leopardspotted goby, Thorogobius ephippiatus (Lowe) (Pisces: Teleostei:

122 Gobiidae). Proceedings of the Royal Irish Academy. Section B: Biological,

Geological, and Chemical Science 76:121-132.

Engin, S., and K. Seyhan. 2010. Age, growth, reproduction and diet of the flatsnout goby,

Neogobius platyrostris (Pallas, 1814), on the south-eastern Black Sea Coast of

Turkey. Zoology in the Middle East 50:59-66.

Etim, L., R. P. King, and M. T. Udo. 2002. Breeding, growth, mortality and yield of the

mudskipper Periophthalmus barbarus (Linneaus 1766) (Teleostei: Gobiidae) in

the Imo River estuary, Nigeria. Fisheries Research 56:227-238.

Filiz, H., and M. Toğulga. 2009. Age and growth, reproduction and diet of the black

goby,(Gobius niger) from Aegean Sea, Turkish Journal of Fisheries and Aquatic

Sciences 3:243-265.

Foster, N. R., and L. A. Fuiman. 1987. Notes on the behavior and early life history of

captive lyre gobles, Evorthodus Lyricus. Bulletin of Marine Science 41:27-35.

Fouda, M. M., M. Y. Hanna, and F. M. Fouda. 1993. Reproductive biology of a Red Sea

goby, Silhouettea aegyptia, and a Mediterranean goby, Pomatoschistus

marmoratus, in Lake Timsah, Suez Canal. Journal of Fish Biology 43:139-151.

Gibson, R. N. 1970. Observations on the biology of the giant goby Gobius Cobitis Pallas.

Journal of Fish Biology 2:281-288.

Gibson, R. N., and I. A. Ezzi. 1978. The biology of a Scottish population of fries' goby,

Lesueurigobius friesii. Journal of Fish Biology 12:371-389.

———. 1981. The biology of the Norway goby, Pomatoschistus norvegicus (Collett), on

the west coast of Scotland. Journal of Fish Biology 19:697-714.

123 Gill, H. S., B. S. Wise, I. C. Potter, and J. A. Chaplin. 1996. Biannual spawning periods

and resultant divergent patterns of growth in the estuarine goby Pseudogobius

olorum: temperature-induced? Marine Biology 125:453-466.

Grossman, G. 1979. Demographic characteristics of an intertidal bay goby (Lepidogobius

lepidus). Environmental Biology of Fishes 4:207-218.

Gümüş, A., and A. Kurt. 2009. Age structure and growth by otolith interpretation of

Neogobius melanostomus (Gobiidae) from Southern Black Sea. Cybium 33:29-37.

Haji, F., A. Ouannes-Ghorbel, M. Ghorbel, and O. Jarboui. 2013. Age and growth of the

grass goby Zosterisessor ophiocephalus Pallas, 1811 in the Gulf of Gabes

(Tunisia, Central Mediterranean). Acta Adriatica 54:27-40.

Iglesias, M., E. B. Brothers, and B. Morales-Nin. 1997. Validation of daily increment

deposition in otoliths. Age and growth determination of Aphia minuta (Pisces:

Gobiidae) from the northwest Mediterranean. Marine Biology 129:279-287.

Iglesias, M., and B. Morales-Nin. 2001. Life cycle of the pelagic goby Aphia minuta

(Pisces: Gobiidae). Scientia Marina 65:183-192.

Kim, J. K., H.-J. Baek, J.-W. Kim, D.-S. Chang, and J.-I. Kim. 2011. Sexual maturity and

early life history of the mudskipper Scartelaos gigas (Pisces, Gobiidae):

implications for conservation. Fisheries and Aquatic Sciences 14:403-410.

Koutrakis, E. T., and A. C. Tsikliras. 2009. Reproductive biology of the marbled goby,

Pomatoschistus marmoratus (Pisces, Gobiidae), in a northern Aegean estuarine

system (Greece). Folia Zoologica 58:447-456.

Kovacic, M. 2006. Age structure, growth and mortality of the striped goby, Gobius

vittatus (Gobiidae) in the northern Adriatic Sea. Scientia Marina 70:635-641.

124 KovačIć, M. 2001. The biology of roule's goby in the Kvarner area, northern Adriatic

Sea. Journal of Fish Biology 59:795-809.

Kramer, A., and R. A. Patzner. 2008. Morphological and developmental studies on the

eggs of four Coryphopterus species (Actinopterygii: Gobiidae) in the caribbean

sea. Acta Ichthyologica Et Piscatoria 38:97-102.

Kruitwagen, G., T. Hecht, H. B. Pratap, and S. E. Wendelaar Bonga. 2006. Changes in

morphology and growth of the mudskipper (Periophthalmus argentilineatus)

associated with coastal pollution. Marine Biology 149:201-211.

La Mesa, M. 1999. Age and growth of Aphia minuta (Pisces, Gobiidae) from the central

Adriatic Sea. Scientia Marina 63:147-155.

Lassig, B. 1976. Field observations on the reproductive behaviour of Paragobiodon spp.

(Osteichthyes: Gobiidae) at heron island great barrier reef. Marine Behaviour and

Physiology 3:283-293.

Lawson, E. 2011. Length-weight relationships and fecundity estimates in mudskipper,

Periophthalmus papilio (Bloch and Schneider 1801) caught from the mangrove

swamps of Lagos Lagoon, Nigeria. Journal of Fisheries and Aquatic Science

6:264-271.

Lawson, E. O. 2010. Maturation and histological characteristics of ovaries in mudskipper,

Periophthalmus papilio from Lagos Lagoon, Nigeria. The Journal of American

Science 6:965-976.

Mazlan, A. G., and M. Rohaya. 2008. Size, growth and reproductive biology of the giant

mudskipper, Periophthalmodon schlosseri (Pallas, 1770), in Malaysian waters.

Journal of Applied Ichthyology 24:290-296.

125 Melo, Y. C., and F. Le Clus. 2005. Growth and reproduction of the pelagic goby

Sufflogobius bibarbatus off the Orange River, Southern Africa. African Journal of

Marine Science 27:265-273.

Metin, G., A. T. İlkyaz, O. Soykan, and H. T. Kinacigil. 2011. Age, growth and

reproduction of four-spotted goby, Deltentosteus quadrimaculatus (Valenciennes,

1837), in İzmir Bay (central Aegean Sea). Turkish Journal of Zoology 35:711-

716.

Miller, P. 1961. Age, growth and reproduction of the rock goby, Gobius paganellus L., in

the Isle of Man. Journal of the Marine Biological Association of the United

Kingdom 41:737-769.

Nanami, A., and T. Takegaki. 2005. Age and growth of the mudskipper Boleophthalmus

pectinirostris in Ariake Bay, Kyushu, Japan. Fisheries Research 74:24-34.

Nash, R. D. M. 1982. The biology of Fries’ goby, Lesueurigobius friesii (Malm), in the

Firth of Clyde, Scotland, and a comparison with other stocks. Journal of Fish

Biology 21:69-85.

Pampoulie, C. 2001. Demographic structure and life history traits of the common goby

Pomatoschistus microps (Teleostei, Gobiidae) in a Mediterranean coastal lagoon

(Rhône River delta, France). Acta Oecologica 22:253-257.

Park, K. D., J. K. Kim, D. S. Chang, J. I. Kim, and C. W. Oh. 2008. Age and growth of

the mudskipper, Scartelaos gigas (perciformes, gobiidae) from Korea. Animal

Cells and Systems 12:305-311.

126 Privitera, L. 2002. Reproductive Biology of the coral-reef goby, Asterropteryx

Semipunctata, in Kaneohe Bay, Hawaii. Environmental Biology of Fishes 65:289-

310.

Privitera, L. A. 2001. Characteristics of egg and larval production in captive bluespotted

gobies. Journal of Fish Biology 58:1211-1220.

Provancha, M., and C. Hall. 1991. Ecology and life history of the clown goby inhabiting

the upper Banana River, Cape Canaveral, Florida. Environmental Biology of

Fishes 31:41-54.

Ravi, V. 2007. Age and growth of the mudskipper, Boleophthalmus boddarti (Pallas,

1770). Journal of Aquatic Biology 22:123-128.

Sasal, P., E. Faliex, and S. Morand. 1996. Population structure of Gobius bucchichii, in a

Mediterranean marine reserve and in an unprotected area. Journal of Fish Biology

49:352-356.

Sunobe, T. 1995. Embryonic development and larvae of three gobiid fish, Trimma

okinawae, Trimma grammistes and Trimmatom sp. Japanese Journal of

Ichthyology 42:11-16.

Takeda, T., M. Hayashi, A. Toba, K. Soyano, and A. Ishimatsu. 2011. Ecology of the

Australian mudskipper Periophthalmus minutus, an amphibious fish inhabiting a

mudflat in the highest intertidal zone. Australian Journal of Zoology 59:312-320.

Taylor, R. B., and T. J. Willis. 1998. Relationships amongst length, weight and growth of

north-eastern New Zealand reef fishes. Marine and Freshwater Research 49:255-

260.

127 Tsuhako, Y., A. Ishimatsu, T. Takeda, K. K. Huat, and K. Tachihara. 2003. The eggs and

larvae of the giant mudskipper, Periophthalmodon schlosseri, collected from a

mudflat in Penang, Malaysia. Ichthyological Research 50:178-181.

Udo, M. T. 2002. Morphometric relationships and reproductive maturation of the

mudskipper, Periophthalmus barbarus from subsistence catches in the mangrove

swamps of IMO estuary, Nigeria. Journal of Environmental Sciences 14:221-226.

Weisel, G. F., Jr. 1947. Breeding behavior and early development of the mudsucker, a

gobiid fish of California. Copeia 1947:77-85.

128 APPENDIX

Chapter 3

129 Table A1: Sample sizes processed for each population. Population Female Juvenile Juvenile Total Female Talofofo 15 4 1 20 Adelup 18 10 4 32 Pago 14 8 2 24 Umatac 9 7 0 16 Taga’chang 18 2 0 20

130 ACKNOWLEDGEMENTS

I would like to thank everyone who helped me over the course of my research.

Firstly thank you members of the Ord Lab for your support. In particular thank you

Courtney Morgans for all those tips that saved me a lot of time and stress. Furthermore thank you Georgina Cooke for all you taught me in the field, as well as for your constant encouragement and humour.

Thank you Iain Suthers and the FAMER Lab. Without access to your lab facilities I would have been unable to conduct my research.

Thank you Ash Fowler for your help, particularly in regards to laboratory work. If you had not shared with me your expertise in extracting and processing otoliths I would never have been able to age the fishes in my study.

Thank you Terry Ord for being a great supervisor. Your readiness to answer questions, help me understand concepts, and provide constant feedback was paramount to my completion of this thesis. This is further attested to by the fact that you lent me your laptop for nearly two years (which I am also very thankful for). Finally thank you for sending me to all those amazing field sites. Those were experiences I will never forget.

Thank you to my friends for always helping me put things in perspective.

Finally thank you to my family. I would not have reached this point in my studies if it was not for your loving support.

131