Evolutionary Ecology of Fig Wasps Associated with the Port Jackson Fig

Home , Ficus hispida, Ficus rubiginosa, Megastigmus, Pegoscapus, Pleistodontes imperialis, Sycoscapter

Evolutionary ecology of fig wasps associated with the Port Jackson fig

Timothy Leonard Sutton BSc (Honours)

Thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Hawkesbury Institute for the Environment Western Sydney University

2016

Acknowledgements

I would first like to acknowledge my supervisor, James Cook, for his support and guidance throughout the last four years. Thanks for having the faith to take me on as a student at very short notice (and from the other side of the planet), for your willingness to chat science, sport and the world, and for teaching me not to take things too seriously.

Huge thanks to my co-supervisors, Markus Riegler and Jane DeGabriel, for all of their hard work and support throughout my candidature. Markus’ enthusiasm and work ethic are second-to-none and I am grateful for the encouragement he gave me to pursue a PhD. Jane not only taught me what real fieldwork is like, but was also a calming influence during stressful periods and an invaluable resource on how to forge a career in science. The amount of work she (and, by association, Clementine) put in during the final weeks of my candidature is unbelievable, and I’m not sure how I would have got over the line without it.

I am grateful for having received an Australian Postgraduate Award and top-up scholarships from Western Sydney University and Hawkesbury Institute for the Environment. I would like to acknowledge the Hawkesbury Foundation for granting me the F.G. Swain Award in 2013, which funded a large portion of the molecular work for this thesis. Thanks also to the Australian Biological Resources Study and the Australian Entomological Society for funding my attendance at numerous conferences in Australia and Europe.

I would like to thank my office mates of the past five years; Aidan Hall and Andrew Gherlenda. It’s been a pleasure to share this journey with you and see the world together, and I am sure many more adventures await us. Thanks also to my close friend and confidant, Adam Frew, for sharing his thoughts and support through the good times and the bad. This rollercoaster of a PhD was made a lot easier with friends like you.

I would like to acknowledge past and present members of the Cook and Riegler lab groups – Caroline Fromont, Ashley Montagu, Will Balmont, Lluvia Flores-Renteria,

Peter Kern, Jennifer Morrow, Duong Nguyen, Shannon Smith and Robert Müller – for sharing ideas, support and encouragement, and assisting with field and lab work throughout my candidature. Jennifer, in particular, was an extremely helpful during my early years in the molecular lab and was always willing to help me out or answer any questions I had (and there were a lot!).

Much of the fieldwork for this thesis was undertaken in collaboration with Jane DeGabriel, who led the fieldwork component in northern Queensland. Thanks also to Caroline Fromont, Will Balmont, Eliza Langford, Aidan Hall, Adam Frew, Nina White, Jennifer Hechinger and Ashley Montagu for their assistance in the field and/or setting up experiments. Aidan Hall, Andrew Gherlenda and Ashley Montagu gave helpful advice on manuscripts submitted during my PhD candidature. I also owe a big thanks to Paul Rymer for guiding me through the statistical component of population genetics in the early days of my candidature.

The research conducted at the Hawkesbury Institute for the Environment would not be possible without the wonderful work of the admin and technical staff. I would like to acknowledge David Harland, Jenny Harvey, Patricia Hellier, Lisa Davison, Gillian Wilkins, Gavin McKenzie, Kerri Sherriff, Nga Le, Markus Klein, Carolina Janitz and Jocelyn King for the hard work they put in everyday to ensure everything at the HIE runs smoothly.

I owe thanks to Leanne and David – my ‘weekend parents’ – for feeding and housing me for at least two-sevenths of the past three-and-a-bit years. Thanks for your kind words of support and encouragement, and for letting me wander in at ungodly hours of the morning after running all-night experiments.

To my family, particularly my parents – Diane and Glenn – thank you for your amazing patience and support (both moral and financial) throughout all of my educational endeavours. You taught me the importance and value of education and hard work, and this is the culmination of those lessons. I appreciate everything you have done for me (though it doesn’t always show), and I am forever indebted to you.

Finally, my love and thanks to Eliza. I cannot begin to thank you enough for the amazing support, encouragement and love you have blessed me with. You believed in me when I didn’t believe in myself, and I could not have done this without you.

Statement of authentication

The work presented in this thesis is, to the best of my knowledge and belief, original except as acknowledged in the text. I hereby declare that I have not submitted this material, either in full or in part, for a degree at this or any other institution.

… ……….

Timothy Sutton March 2016

III

Preface

The experimental chapters presented in this thesis were written as a series of independent papers that have been published, or are intended for publication, in peer- reviewed journals. Consequently, each of the five experimental chapters was structured in the style of the chosen journal. As each chapter represents a stand-alone paper, and is therefore self-contained, each has its own more detailed introduction and discussion and there is minor repetition between them. Each chapter was co- authored; nevertheless, I am the principal author and contributor to this work. I conceptualised the research with guidance from my supervisory panel. Furthermore, I collected and analysed the data, and wrote the manuscripts with editorial advice from my supervisory panel. Because each chapter is written as it appears, or will appear, in publication, I have retained the use of the collective pronoun. In addition to the experimental chapters, I conclude this thesis by highlighting the significance of the research, and discussing its limitations and potential directions of future research (Chapter 7). Below are details of published papers and papers in review that have arisen from the research presented in this thesis. Details are accurate at the time of submission.

Chapter 2: Sutton, TL, Riegler, M and Cook, JM (2015) Characterisation of microsatellite markers for fig-pollinating wasps in the Pleistodontes imperialis species complex. Australian Journal of Zoology, 63(2), 122-126.

Chapter 3: Sutton, TL, Riegler, M and Cook, JM (2016) One step ahead: a parasitoid disperses farther and forms a wider geographic population than its fig wasp host. Molecular Ecology, 25(4), 882-894.

Chapter 4: Sutton, TL, DeGabriel, JL, Riegler, M and Cook, JM (in revision) Barriers from barcodes: how reliable are DNA barcodes in fig wasps? Heredity.

Abstract

Geography plays an important role in the study of evolutionary ecology. It has implications for the coevolution of species interactions, our ability to accurately and confidently delimit species boundaries, and can influence species’ adaptations that may provide resilience, or make them vulnerable, to climatic change. A geographic consideration of these issues can help to improve our understanding of the processes generating biodiversity and its fate under predicted climate change.

In this thesis, I investigated the genetic structure of a pollinator-parasitoid interaction over a wide geographic range, assessed the reliability of DNA barcoding under different geographic scenarios, and examined the resilience of a temperate fig- pollinating wasp species to climate change. Figs are keystone species in many regions and the fig – pollinator system is a classic example of an obligate mutualism. In this thesis, I explored the evolutionary ecology of pollinator and non-pollinator fig wasps associated with the Port Jackson fig (Ficus rubiginosa) in eastern Australia. This system is particularly interesting due to the sheer number of species that interact within F. rubiginosa fruits. There are five recognised pollinating wasps of F. rubiginosa: Pleistodontes imperialis sp. 1, 2, 3, 4 and 5 (sensu Haine et al. 2006). All of these are parasitised by Sycoscapter non-pollinating fig wasps, of which there are three species: Sycoscapter sp. A, B and C.

I begin this thesis by describing the development of microsatellite markers that I characterised for population genetic studies in Pleistodontes imperialis sp. 1 and demonstrating their utility in other species of the P. imperialis complex (Chapter 2). Next, I applied these markers to a comparative population genetic study of a fig- pollinating wasp (P. imperialis sp. 1) and its parasitoid (Sycoscapter sp. A) (Chapter 3). I found that the pollinator comprises two disjunct populations while the parasitoid exists as a single population throughout eastern Australia. These results suggest that the two pollinator populations have the potential to diverge ecologically in their responses to a common antagonist, which may facilitate speciation. Furthermore, I provide the first evidence that a fig wasp parasitoid likely disperses farther than its associated fig-pollinating wasp, challenging common perceptions about parasitoid fig

wasp dispersal that were based solely on measurements of flight heights in tropical rainforests.

I then shifted my focus to test the reliability of DNA barcoding for delimiting fig wasp species in sympatry and allopatry (Chapters 4 and 5), focussing on the P. imperialis complex of closely related wasp species associated with F. rubiginosa. I found that Bayesian clustering analyses help to identify potential F1 hybrids, but reveal minimal interspecific gene flow between sympatric fig wasp species. Thus, my results validate the use of DNA barcoding for species delimitation of fig wasps pollinating F. rubiginosa.

I then scaled genetic differentiation of allopatric P. imperialis sp. 1 fig wasp populations with reference to a closely-related species. The levels of divergence in nuclear and mitochondrial DNA sequences, as well as in microsatellite markers, suggest that these populations are actually incipient species. While my findings support the use of DNA barcoding for species delimitation in sympatric fig wasps, they also highlight some of the issues associated with DNA barcoding of allopatric taxa that are recently diverged. Based on my results, I recommend that exploratory DNA barcoding should be supported by multilocus estimates of genetic divergence and/or gene flow.

In Chapter 6, I departed from a molecular framework to examine the potential impact of one aspect of climate change (heatwaves) on a temperate fig wasp species. Laboratory experiments revealed severely reduced emergence and longevity at temperatures often experienced on hot days during Australian summers (39-42°C), but these were unaffected by temperatures up to 30°C and 36°C, respectively. Interestingly, longevity was significantly shorter under low humidity conditions across all temperatures. The core temperatures of figs growing on trees was higher (+0.94°C) in figs exposed to direct sunlight compared to figs that were protected by shade, indicating that figs in shaded microhabitats could provide protection for wasps in hot conditions. In this context, studies of tropical fig wasps have revealed extreme sensitivity to even small increases in temperature. In contrast, I showed that temperate fig wasps are less sensitive to small increases in temperature, but are at risk from extreme temperatures that are expected to increase in frequency and intensity with global warming. The risks that reduced humidity pose to fig wasps,

however, appear to be applicable to both tropical and temperate species. By reducing pollinator emergence and longevity, extreme temperatures and reduced humidity could limit the pollination potential of figs, which may have serious implications for the persistence of the mutualism.

The results presented in this thesis highlight the importance of population genetic approaches to coevolutionary studies and biodiversity assessments. Importantly, in estimating biodiversity, we must consider contemporary processes that are likely to be shaping the future of biodiversity, rather than focussing purely on past processes that have shaped biodiversity as we see it today. Only then can realistic predictions be made about the future of species under predicted climate change.

VII

Table of contents

Acknowledgements ...... I

Statement of authentication ...... III

Preface ...... IV

Abstract ...... V

List of tables ...... XII

List of figures ...... XIV

List of abbreviations ...... XVI

Chapter 1: Introduction ...... 1

1.1 The fig – fig wasp mutualism ...... 3

1.2 Coevolution of plant-insect interactions ...... 6

1.2.1 Population genetics and the ‘geographic mosaic theory of coevolution’ 8

1.3 The species delimitation problem ...... 9

1.3.1 The DNA barcoding ‘solution’ ...... 11

1.4 Geographic distribution and species’ responses to climate change ...... 13

1.5 Thesis aims and outline ...... 16

1.5.1 Study system: Port Jackson fig (Ficus rubiginosa) and its associated fig wasps ...... 17

1.5.2 Host-parasitoid coevolution: a biogeographic investigation ...... 19

1.5.3 Assessment of DNA barcoding in sympatric and allopatric lineages .... 22

1.5.4 The effect of climate change on temperate fig wasps ...... 23

1.6 Conclusion ...... 25

Chapter 2: Characterisation of microsatellite markers for fig-pollinating wasps in the Pleistodontes imperialis species complex ...... 26

2.1 Abstract ...... 27

2.2 Introduction ...... 28

2.3 Methods, results and discussion ...... 28 VIII

2.4 Conclusion ...... 35

Chapter 3: Comparative population genetic structure in a fig-pollinating wasp and its main parasitoid ...... 37

3.1 Abstract ...... 38

3.2 Introduction ...... 39

3.3 Materials and methods ...... 41

3.3.1 Fig wasp biology ...... 41

3.3.2 Sample collection ...... 42

3.3.3 Molecular protocols ...... 44

3.3.4 Genetic variability ...... 45

3.3.5 Genetic differentiation ...... 45

3.3.6 Population genetic structure ...... 46

3.3.7 Isolation by distance ...... 47

3.3.8 Power analysis ...... 48

3.4 Results ...... 48

3.4.1 Genetic variability ...... 48

3.4.2 Genetic differentiation ...... 49

3.4.3 Population genetic structure ...... 50

3.4.4 Isolation by distance ...... 54

3.4.5 Power analysis ...... 54

3.5 Discussion ...... 55

3.5.1 Overall patterns of population structure ...... 55

3.5.2 Incongruent structure of host and parasitoid populations ...... 58

Chapter 4: Quantifying gene flow and hybridisation among sympatric cryptic fig wasp species ...... 59

4.1 Abstract ...... 60

4.2 Introduction ...... 61

4.3 Materials and methods ...... 63

4.3.1 Study system ...... 63

4.3.2 Sample collection ...... 64

4.3.3 Assessment of the potential for hybridisation ...... 65

4.3.4 DNA extraction and mtDNA analyses ...... 65

4.3.5 Microsatellite analyses ...... 66

4.3.6 Nuclear DNA barcoding ...... 67

4.4 Results ...... 67

4.5 Discussion ...... 72

Chapter 5: Scaling genetic differentiation of allopatric fig wasp populations ...... 75

5.1 Abstract ...... 76

5.2 Introduction ...... 77

5.3 Methods ...... 79

5.3.1 Sample collection ...... 79

5.3.2 Molecular methods ...... 81

5.3.3 DNA sequence analyses ...... 81

5.3.4 Microsatellite analyses ...... 82

5.4 Results ...... 83

5.4.1 DNA sequence analyses ...... 83

5.4.2 Microsatellite analyses ...... 83

5.5 Discussion ...... 89

5.5.1 Conclusions ...... 93

Chapter 6: Vulnerability of temperate fig-pollinating wasps to heat and desiccation stress ...... 94

6.1 Abstract ...... 95

6.2 Introduction ...... 96

6.3 Materials and methods ...... 99

6.3.1 Study system ...... 99

6.3.2 Experiment 1: Effects of increasing ambient temperature on wasp emergence ...... 100

6.3.3 Experiment 2: Effects of temperature and relative humidity on wasp lifespan ...... 103

6.3.4 Experiment 3: Core temperature of figs in sun versus shade ...... 104

6.4 Results ...... 105

6.4.1 Experiment 1: Effects of increasing ambient temperature on wasp emergence ...... 105

6.4.2 Experiment 2: Effects of temperature and relative humidity on wasp lifespan ...... 106

6.4.3 Experiment 3: Core temperature of figs in sun versus shade ...... 109

6.5 Discussion ...... 109

Chapter 7: Discussion ...... 114

7.1 Key findings, implications and limitations ...... 115

7.1.1 Host-parasitoid coevolution ...... 115

7.1.2 Assessment of DNA barcoding in fig wasps...... 116

7.1.3 The resilience of temperate fig-pollinating wasp species to climate change ...... 118

7.2 Future research ...... 120

7.3 Concluding remarks ...... 122

References ...... 123

Appendices ...... 144

Appendix A ...... 145

Appendix B ...... 146

Appendix C ...... 147

List of tables

Table 1.1: Cryptic fig wasp species nomenclature ………………………………...18

Table 2.1: Characteristics of nine microsatellite loci for 30 Pleistodontes imperialis sp. 1 from a population in Sydney (Australia)...... 31

Table 2.2: Genetic diversity of two Pleistodontes imperialis sp. 1 populations at nine microsatellite loci...... 32

Table 2.3: Characteristics of nine microsatellite loci for Pleistodontes imperialis sp. 2 (N = 11) and Pleistodontes imperialis sp. 4 (N = 11)...... 36

Table 3.1: Pollinator and parasitoid samples sizes used for molecular analyses, and number of trees from which samples were collected...... 44

Table 3.2: Genetic characteristics of microsatellite loci...... 51

Table 3.3: Within-site global genetic variability...... 52

Table 3.4: Genetic differentiation between sites...... 53

Table 4.1: mtDNA and microsatellite diversity of three sympatric P. imperialis species...... 70

Table 4.2: Tests of selective neutrality under a population expansion model for three sympatric P. imperialis species based on mtDNA cytb sequences...... 70

Table 4.3: Genetic differentiation among P. imperialis (cytb) species ...... 70

Table 4.4: Potential hybrid alleles and their frequency in each P. imperialis MOTU……………………………………………………………………………….75

Table 5.1: mtDNA (cytb) diversity of P. imperialis sp. 1 and P. imperialis sp. 2 populations…………………………………………………………………………..86

Table 5.2: Tests of selective neutrality under a population expansion model for P. imperialis sp. 1 populations based on cytb sequences...... 86

XII

Table 5.3: Mean within- and between-population Jukes-Cantor distances based on a 311-315 bp fragment of the ITS2 region...... 86

Table 5.4: Genetic characteristics of microsatellite loci...... 88

Table 5.5: Global (across pooled loci) genetic variability...... 89

Table 5.6: Pairwise population genetic differentiation based on microsatellite data.89

Table 6.1: Site locations and summary climate data...... 102

Table 6.2: Samples sizes (no. of wasps) used to test the effects of temperature and humidity on fig wasp lifespan...... 104

Table 6.3: The effect of increasing temperature and low humidity on mortality rates...... 107

Table 6.4: Comparison of mortality rate between sites for each humidity treatment...... 107

XIII

List of figures

Figure 1.1: Development of a monoecious fig...... 5

Figure 1.2: The DNA barcoding gap...... 11

Figure 1.3: The distribution and relative frequencies of the five fig-pollinating wasps of the Pleistodontes imperialis species complex that pollinate F. rubiginosa...... 18

Figure 1.4: Distributions of (A) Sycoscapter sp. A and (B) Sycoscapter sp. B and C...... 20

Figure 1.5: Pleistodontes imperialis sp. 1 (bottom) and Sycoscapter sp. A (top). ... 22

Figure 2.1: Deviance information criterion (DIC) as a function of Kmax...... 33

Figure 2.2: TESS bar plot for K = 2 (a), K = 3 (b) and K = 4 (c)...... 34

Figure 3.1: Geographic distribution of sampled sites...... 43

Figure 3.2: Genetic cluster assignments of P. imperialis sp. 1 (above) and Sycoscapter sp. A (below) individuals assigned by TESS...... 54

Figure 3.3: Isolation by distance...... 55

Figure 4.1: Location of P. imperialis sample collection sites...... 64

Figure 4.2: Bayesian haplotype tree of P. imperialis haplotypes based on a 372 bp fragment of the mitochondrial cytb gene...... 69

Figure 4.3: mtDNA mismatch distributions for P. imperialis sp. 2 (a), sp. 3 (b) and sp. 4 (c)...... 71

Figure 4.4: Genetic cluster assignments of 139 P. imperialis individuals based on nuclear microsatellite genotypes...... 71

Figure 4.5: Consensus Bayesian topology of P. imperialis inferred from a 338 bp fragment of the ITS2 region ...... 72

Figure 5.1: Geographic distribution of sampling efforts...... 80

XIV

Figure 5.2: Bayesian haplotype tree of three P. imperialis populations based on a 339 bp fragment of the mitochondrial cytb gene...... 85

Figure 5.3: Mismatch distributions of mtDNA lineages for the (a) southern and (b) northern P. imperialis sp. 1 populations...... 87

Figure 5.4: Schematic diagram of a 20 bp section of the ITS2 region showing the indel at position 102 of the sequence alignment (denoted by bracket) that is diagnostic for each of the three populations...... 89

Figure 5.5: PCA plot of P. imperialis individuals based on multilocus microsatellite genotypes...... 91

Figure 6.1: Mean (± S.E.) number of female P. imperialis sp. 1 wasps emerged from F. rubiginosa figs subject to various temperature/time treatments...... 106

Figure 6.2: Median lifespan duration (± S.E.) of adult P. imperialis sp. 1 females across temperature treatments for (a) Balmoral site and (b) Penrith site...... 108

Figure 6.3: Fig core temperature as a function of ambient temperature in shaded figs...... 109

List of abbreviations

AMOVA Analysis of molecular variance bp base pairs BSC Biological species concept CVH Climate Variability Hypothesis COI cytochrome c oxidase I cytb cytochrome b DIC Deviance information criterion ESU Evolutionary significant unit FDR False discovery rate HKY + G Hasegawa-Kishino-Yano model with gamma distributed rate variation HWE Hardy-Weinberg equilibrium IBD Isolation by distance ITS2 Internal transcribed spacer 2 LD Linkage disequilibrium MOTU Molecular operational taxonomic unit mtDNA mitochondrial DNA NPFW Non-pollinating fig wasp PCR Polymerase chain reaction PSC Phylogenetic species concept RH Relative humidity sp. Species

XVI

Chapter 1 Introduction

At the heart of evolutionary ecology is the motivation to understand the processes responsible for the generation and maintenance of biodiversity. The geographic structure and distribution of species is a fundamental theme in this field (Gaston 2009) that has implications for the coevolution of species interactions (Thompson 2005), how we estimate species diversity (Coyne & Orr 2004; Lukhtanov et al. 2015), and how it might be affected by predicted climate change (Samways et al. 1999; Gibson et al. 2010; Lynch et al. 2014). Understanding the interplay between these key issues is essential to creating a clearer picture of past, present and future patterns of biodiversity.

Thompson (2005) proposed that the geographic structure of species is integral to the evolution of their interactions. He argued that different selective pressures in different environments could result in divergent outcomes of an interaction across populations. Should gene flow between populations be restricted, ecological speciation may proceed. In many systems, however, genetically divergent populations do not exhibit obvious ecological (or even morphological) divergence, and may or may not exchange genes on secondary contact (Donnelly et al. 2013; Palmé et al. 2013; Giska et al. 2015; Martinsson et al. 2015). Consequently, our ability to confidently delimit species is highly-dependent on the geographic distribution of cryptic diversity.

Cryptic diversity that exists in allopatry poses significant problems to species delimitation. Specifically, genetic divergence in geographic isolation is not always indicative of reproductive isolation (e.g. Senn et al. 2010). Conversely, hybridisation and gene flow among sympatric cryptic taxa can be assessed using population genetic analyses (e.g. Giska et al. 2015). The geographic issues inherent to species delimitation are not trivial; inaccurate species delimitation affects our entire understanding of species interactions and how they evolve. Perhaps more importantly, though, the geographic distribution of cryptic species may influence their thermal tolerance adaptations (Stevens 1989), which has consequences for their resilience to predicted climate change (Addo-Bediako et al. 2000).

Plants and insects are extremely useful models in evolutionary ecology due to the extraordinarily diverse nature of their interactions. These interactions span the continuum from mutualistic (e.g. pollination) to antagonistic (e.g. herbivory), often

encompass multiple trophic levels (e.g. plant-herbivore-parasitoid) and range in their degree of specialisation. Furthermore, they exhibit high levels of geographic variation (e.g. Aigner 2005; Thompson & Fernandez 2006; Cosacov et al. 2008; Anderson & Johnson 2009; Gómez et al. 2009a), cryptic species (e.g. Haine et al. 2006; Griffiths et al. 2011; Williams et al. 2012), and are extremely sensitive to climatic change (e.g. Rader et al. 2013; Polce et al. 2014; Miller-Struttmann et al. 2015). In this thesis, I explore the geographic structure of interacting species, species delimitation under contrasting geographic scenarios, and the resilience of temperate insects to climate change using fig wasps as a model system.

1.1 The fig – fig wasp mutualism

One of the best studied examples of a mutualistic plant-insect relationship is the fig – fig wasp system. There are approximately 750 described fig (Ficus, Moraceae) species distributed across four subgenera and 19 sections, and Ficus is represented on all continents except Antarctica (Berg 1989). Figs exhibit marked variation in their reproduction and pollination mode, but all share an obligate relationship with species-specific fig-pollinating wasps (Hymenoptera: Agaonidae). Approximately half of all fig species are dioecious (Cook & Segar 2010), i.e. trees are either male or female, and the other half are monoecious. In dioecious species, pollinators cannot distinguish between male and female trees, and foundresses that enter syconia (enclosed inflorescences, usually referred to as ‘figs’, or ‘fig fruits’) of female trees pollinate ovules whose styles are too long for wasp oviposition and thus do not support the development of wasp offspring (Raja et al. 2008). Foundresses that enter figs of male trees can successfully reproduce and thus also facilitate tree reproduction, because female wasps are responsible for dispersing fig pollen. In monoecious species, each tree bears figs that contain male and female flowers that support the development of seeds and wasps.

Receptive figs emit unique volatile chemicals that attract pollen-bearing female fig wasps (Ware et al. 1993; Hossaert-McKey et al. 1994) which enter the receptive syconium through a narrow passageway called the ostiole. In monoecious figs (Figure 1.1), the female wasp (foundress) then pollinates fig ovules, depositing an

egg into a subset of ovules, and this initiates gall formation, which provides nourishment for her offspring. The foundress dies within the fig, and her offspring develop over the course of a few weeks, emerging at the time when male flowers produce pollen inside the fig. Wingless males exit their galls and create holes in the females’ galls in order to gain access and mate (Zammit & Schwarz 2000). As figs are normally founded by only one or a few females, mating commonly occurs between siblings. Consequently, fig wasp populations are typically highly inbred. Following insemination, males bore holes through the fig wall and subsequently die within the natal fig. Females emerge from their galls, collect pollen and exit through the holes created by the males in search of a receptive fig to enter and continue the reproductive cycle. The fig fruit ripens and changes colour soon after the wasp offspring emerge, which attracts vertebrate frugivores that disperse the fig seeds (Shanahan et al. 2001).

Pollination of fig trees by fig wasps can be active or passive. In actively-pollinated fig species, mated female fig wasps emerge from their galls and deliberately collect pollen by loading it into specialised structures called pollen pockets. Then, when a foundress enters a receptive fig, she deliberately unloads the pollen onto some of the fig ovules. In contrast, pollinators of passively-pollinated figs inadvertently collect pollen on their bodies. The pollen is then haphazardly and passively deposited onto fig ovules while the wasps are laying eggs inside the receptive fig (Galil & Neeman 1977; Jousselin & Kjellberg 2001).

Figure 1.1: Development of a monoecious fig. Receptive (B-stage) syconia emit volatile cues that attract pollen-bearing female fig-pollinating wasps. After oviposition and pollination, the foundress dies. Males chew exit holes during D-stage of development, through which pollen-bearing females exit. The syconium ripens at E-stage and is eaten by frugivores, which disperse its seeds. Figure reproduced from Harrison (2005).

The fig – fig wasp mutualism is exploited by non-pollinating fig wasps (NPFWs), which complete their life cycle within the fig but do not provide pollination services (Bronstein 1991). NPFWs are extremely diverse – species richness estimates exceed 7,000 (Cook & West 2005) – but Segar et al. (2013) characterised them into four functional guilds: small gallers, large gallers, small parasitoids and large parasitoids. Small gallers induce galls and feed within small flowers, and are preyed upon by small parasitoids. Large gallers can induce galls in fig flowers or in the fig wall, and are preyed upon by large parasitoids. Almost all NPFWs deposit eggs into the fig from the outside, and have thus evolved long external (or extrudable) ovipositors. Interestingly, the vast majority of NPFWs have not evolved the ability to create exit holes, and so rely on the male fig-pollinating wasps to release them from the natal fig.

1.2 Coevolution of plant-insect interactions

Plants and insects represent two of the most speciose eukaryotic groups on Earth (Mora et al. 2011), largely because of the diverse range of interactions between them. Indeed, insect pollination has been linked with the radiation of the angiosperms (Regal 1977; Mulcahy 1979; Labandeira et al. 1994), and coevolution of adaptive traits among plants and their insect pollinators is evidenced by pollination syndromes (i.e. matching suites of plant and pollinator traits that are produced by coadaptation; Fenster et al. 2004). Similarly, insect herbivores can induce the evolution of plant defence syndromes (Agrawal & Fishbein 2006), creating an antagonistic ‘coevolutionary arms race’ (Van Valen 1973) between plants and herbivores.

The fig/pollinator interaction has resulted in the correlated evolution of complex traits across more than 750 species within the genus Ficus and their associated insect pollinators. Ramírez (1974) first noted correlations between (i) differences in style length of female flowers in monoecious and dioecious figs and pollinator ovipositor length (later confirmed by Weiblen 2004), (ii) amount of pollen production and presence of corbiculae (pollen pockets) on pollinators, and (iii) ostiole shape and pollinator head morphology. Kjellberg et al. (2001) subsequently showed that passively-pollinated figs produce more pollen than actively-pollinated figs and that active pollinators have coxal combs (to pick up and brush off pollen) while passive pollinators do not. Furthermore, van Noort & Compton (1996) revealed convergent evolution of head morphology in a lineage of fig-pollinating and a lineage of internally-ovipositing NPFWs, such that the head shapes of wasps from different lineages on the same host plant were very similar, while closely related wasps from different host plants could have different head shapes.

Correlation between fig and fig wasp traits, along with observed extreme levels of host specificity in the fig – fig wasp system (Baker 1961; Wiebes 1961; Ramírez 1970) led to the early notion that each fig species is pollinated by a single wasp species, and vice versa, and that the two lineages diversified under strict-sense cospeciation (Herre et al. 1996). Furthermore, cophylogenetic studies supported strict-sense cospeciation of figs and their pollinators (Herre et al. 1996). However, advances in molecular tools and analyses over the past two decades changed the way 6

that coevolutionary studies within the fig – fig wasp system were undertaken. The use of genetic data to reconstruct phylogenetic relationships greatly improved the ability to test hypotheses regarding the evolution of coevolved morphological traits. Perhaps unexpectedly, phylogenetic studies revealed convergent evolution of several key fig (Jousselin et al. 2003) and wasp (Weiblen 2004) morphological traits, and acquisitions and reversions of pollination mode (Cook et al. 2004) and breeding systems (Machado et al. 2001). Hence, coevolved traits are not necessarily indicative of cospeciation and should not be used in isolation as evidence of cospeciation.

Subsequent studies have revealed that the taxonomic scale of sampling and the number of genetic markers employed affect the outcome of cophylogenetic analyses (Jackson 2004; Machado et al. 2005; Jackson et al. 2008). Specifically, sampling of deeply divergent taxa and the use of only one or two markers can lead to unrealistically high estimates of codivergence, while sampling of closely-related taxa and the application of multiple genetic markers often lead to lower estimates of cospeciation. The current, widely-accepted view of diversification in figs and their pollinators is one of broad-scale codivergence (at the level of fig section and wasp genus) with pollinator duplications and host switches among closely-related fig species (Weiblen & Bush 2002; Jackson 2004; Weiblen 2004; Machado et al. 2005; Jackson et al. 2008; Silvieus et al. 2008; Cruaud et al. 2012; McLeish & van Noort 2012). The existence of multiple pollinator species (co-pollinators) on a single host fig species is evidence of the duplication and host switching events that have contributed to pollinator diversification. Interestingly, Yang et al. (2015) revealed that all known co-pollinators of dioecious figs are sister taxa, while only a third of co-pollinators of monoecious figs are sister taxa. The tendency for duplicated co- pollinators to exhibit patterns of geographic replacement or allopatry suggests an allopatric origin of co-pollinators in dioecious figs (Cook & Segar 2010), while most co-pollinators of monoecious figs arise from host switches in sympatry.

These strong insights into the diversification of fig-pollinating wasps are in stark contrast to our current understanding of NPFW diversification. Far less effort has been applied to studying NPFWs, and current knowledge of their diversification is therefore poor. NPFWs are polyphyletic, having arisen from several chalcidoid wasp families (Rasplus et al. 1998), and occupy a range of different niches (Segar et al.

2013). Consequently, patterns of diversification are expected to vary among taxonomic groups (Cook & Segar 2010). Nevertheless, most cophylogenetic studies reject strict cospeciation of NPFWs with host (fig or wasp) species and reveal evidence for host switching (Lopez-Vaamonde et al. 2001; Weiblen & Bush 2002; Jackson 2004; Silvieus et al. 2008; McLeish et al. 2012; but see Marussich & Machado 2007) . McLeish et al. (2012), for example, showed that cryptic Arachonia parasitoids share host pollinators on multiple distantly related fig species from the sections Galoglychia and Sycomorus. The poor representation of NPFWs in the cophylogenetic literature is clear evidence that our understanding of the coevolutionary history of NPFWs is lacking. More detailed studies of NPFWs at finer taxonomic scales could help to elucidate patterns of diversification.

1.2.1 Population genetics and the ‘geographic mosaic theory of coevolution’

Thompson (1994, 2005) suggested that studying the genetics of populations is key to understanding the evolution of species interactions. Indeed, Hubby & Lewontin (1966) recognised population genetic variation as “the fundamental datum in evolutionary studies”, which can be used to explain the origins of genetic variation and predict its consequences. Thompson’s (2005) ‘geographic mosaic theory of coevolution’ hypothesises that differences in the genetic structure of interacting species can lead to spatial variation in the outcomes of such interactions (i.e. geographic mosaics), such as differences in parasitism resistance (e.g. Kraaijeveld & van Alphen 1995) or predator-prey dynamics (e.g. Brodie et al. 2002), which may result in ecological speciation. Thus, while cophylogenetic studies help to form hypotheses regarding long-term diversification of major lineages, population genetic studies of interacting species provide an overview of recent and contemporary gene flow which can help to formulate more accurate hypotheses regarding mechanisms of ongoing diversification and coevolution.

Thompson (2005) recognised the importance of geographic mosaics in the fig – fig wasp system. Fig wasps frequently mate among siblings, and fig wasp populations are thus typically highly inbred (e.g. Molbo et al. 2004). Consequently, reduced heterozygosity in fig wasp populations might be expected to accelerate genetic

divergence of sub-populations. Fig populations are likely to diverge at a much slower rate than pollinator populations, due to a longer generation time and relatively low levels of inbreeding, and thus exhibit potentially different patterns of genetic structure relative to pollinators (e.g. Liu et al. 2013). Yet despite the incredible insights that comparative population genetic research can provide, studies of this kind have been rare in this system. Liu et al. (2013) found that contemporary local habitat fragmentation resulted in highly structured populations in the pollinator wasp Wiebesia pumilae sp. 1, but not in its host, Ficus pumila. Such structure could result in the divergence of wasp populations inhabiting the same fig species. In contrast, Liu et al. (2015b) reported clinal geographic structure in F. pumila throughout southern China and Hainan Island, and no genetic structure in its pollinator, W. pumilae sp. 2, in the same region. Interestingly, whereas both F. hirta (Yu & Nason 2013) and its pollinator, Valisia javana (Tian et al. 2015), exhibit strong genetic structure between Hainan Island and mainland China, only F. hirta exhibits highly structured populations throughout mainland China. These studies highlight the potential importance of geographic mosaics for diversification in this system by demonstrating that patterns of geographic structure vary between interacting species. Yet, although NPFWs represent a significant evolutionary pressure within this system, they are noticeably absent from comparative population genetic studies.

1.3 The species delimitation problem

The accurate and reliable identification of species is integral to ecological and evolutionary studies, but the lack of a universally accepted ‘species concept’ presents major challenges to this process (Mallet 1995; de Queiroz 2005, 2007). Despite widespread calls for pluralist approaches to species delimitation that incorporate aspects of multiple species concepts (Mishler & Donoghue 1982; Ereshefsky 1992; Rosselló-Mora 2003), differences in the biological basis of many modern species concepts render them incompatible. The taxonomic grouping of organisms often depends on which concept is applied (Mallet & Willmott 2003), and this usually depends on the taxa being studied and/or the question being addressed.

Possibly the most familiar and widespread species concept is the Biological Species Concept (BSC), which was conceived in its earliest form by Dobzhansky (1935) and later refined by Mayr (1942). The BSC considers species as actually or potentially interbreeding organisms that produce fertile offspring. Although it proposes a clear and intuitive definition of species, there are major fundamental issues with the application of the BSC to actual ecological systems. For example, the BSC is not applicable for asexual organisms (as it relies on sex), and interbreeding is a poor indicator of conspecificity given that (i) some good species produce fertile hybrids in nature (e.g. Kahilainen et al. 2011) and (ii) reproductive barriers (e.g. assortative mating; Engeler & Reyer 2001) can exist between individuals within an interbreeding population.

However, perhaps the biggest issue with the BSC is how to treat geographically disjunct taxa. The BSC treats allopatric populations as distinct species (Mayr 1942), but widespread hybridisation of native and introduced species (e.g. Perry et al. 2001; Barbour et al. 2002; Riley et al. 2003; Pietri et al. 2011; Biedrzycka et al. 2012; Tang & Chen 2012; Saltonstall et al. 2014) poses a significant challenge to this definition of species. Coyne & Orr (2004) endorsed a revised definition of the BSC that describes species as “groups of interbreeding natural populations that are reproductively isolated from other such groups” (Mayr 1995). Consequently, naturally allopatric taxa are considered distinct species. While this is an improvement on the original BSC, it is unclear how this definition might treat taxa whose genetic identities have been eroded by ongoing long-term hybridisation.

A group of species concepts that bypass issues with geography are the phylogenetic species concepts (PSCs), which use genetic data to identify species as clusters of organisms that are phylogenetically distinct from other clusters and exhibit patterns of ancestry and descent (Cracraft 1989). de Queiroz & Donoghue (1988) presented a refined PSC that identifies species as the smallest exclusively monophyletic groups, and Baum & Donoghue (1995) drew upon coalescent theory to define a species as a group of organisms “whose genes coalesce more recently with each other than with those of any organisms outside the group” (Coyne & Orr 2004).

PSCs do not consider the potential to interbreed as a defining feature of species, relying instead on the history of interbreeding recorded in DNA sequences, thus

avoiding the geographic issues pertaining to the BSC. However, this disregard for interbreeding and low intraspecific variation in asexual organisms may lead to overestimations of species diversity (oversplitting; Avise 2000). Furthermore, the phylogenetic history of genes is not always consistent with the phylogenetic history of species (Maddison 1997), and so accurate species delimitation with PSCs may require multilocus approaches.

1.3.1 The DNA barcoding ‘solution’

Hebert et al. (2003a) proposed, controversially, that a single fragment of mitochondrial DNA (mtDNA; namely, cytochrome c oxidase I, or COI) could be used to catalogue all animal life since each species possessed a unique COI sequence, or ‘barcode’. DNA barcoding offers a quick and easy means of identifying species by matching unknown sequences against a database of sequences from described species (Hebert et al. 2003a; Hebert et al. 2003b; Blaxter 2004). Furthermore, delimitation of undescribed or cryptic species is achieved by identifying the ‘barcoding gap’ – a non-overlapping bimodal distribution of intraspecific and interspecific genetic variation (Figure 1.2).

Figure 1.2: The DNA barcoding gap. Plotted here is the theoretical frequency distribution of intraspecific (black) and interspecific (grey) genetic distances. The barcoding gap is a region of the distribution that falls between the greatest intraspecific distance and the lowest interspecific distance.

DNA barcoding has undoubtedly increased the rate of species discovery by facilitating the identification of previously undetected cryptic diversity in a wide range of taxa (e.g. Hebert et al. 2004; Lara et al. 2010; Clare 2011; Saitoh et al. 2015) with identification rates of individuals often exceeding 90% within some taxa (Meyer & Paulay 2005; Kerr et al. 2007). However, the application of DNA barcoding to species delimitation and identification has received heavy criticism. Although it has been successfully applied to many groups of animals, there is no theoretical basis for the barcoding gap (Hickerson et al. 2006; Stoeckle & Thaler 2014), making it a contentious criterion for identifying species.

In addition, DNA barcoding fails empirically in some cases. For example, Wiemers & Fiedler (2007) and Kvist (2014) showed that the barcoding gap does not exist in some butterfly and annelid lineages, respectively. Rubinoff et al. (2006) provide comprehensive arguments against the use of mitochondrial genes in isolation for DNA barcoding and species discovery, citing small effective population size, introgression, lack of recombination, heteroplasmy, and selection on mitochondria driven by endosymbiont transmission as confounding factors in the DNA barcoding process.

Similar to the BSC, some of the issues with DNA barcoding relate to geographic patterns of genetic diversity. Lack of recombination can lead to the persistence of population structure long after the resumption of gene flow between previously isolated taxa. For example, Giska et al. (2015) revealed rampant gene flow between two deeply divergent mitochondrial lineages of the earthworm Lumbricus rubellus. Results such as these undermine the significance of mtDNA divergence and question the reliability of DNA barcoding in uncovering and identifying cryptic ‘species’.

Validating the use of DNA barcoding for species discovery in sympatric taxa can be relatively simple: multilocus population genetic studies can easily identify genetic exchange between divergent mitochondrial lineages (Bull et al. 2013; Bull & Sunnucks 2014; Giska et al. 2015). The task is more difficult for allopatric taxa as measures of gene flow are inconsequential due to geographic isolation. In some taxonomic groups, laboratory mating trials could address this issue. Gómez et al. (2007) validated the species status of allopatric cryptic species of the marine bryozoan Celleporella hyaline by showing that they are reproductively isolated in the

laboratory. This result somewhat invokes the BSC definition of species as reproductively isolated groups.

For many taxa, however, laboratory mating trials are not feasible. Fig wasps, for example, mate only within the confines of the syconium (Zammit & Schwarz 2000), and so mate choice cannot be controlled. Furthermore, the short lifespan of adult fig wasps (typically 1-2 days) would provide severe logistical challenges for experimental crosses between geographically disjunct populations. Consequently, while a number of molecular studies have reported cryptic fig wasp species (both in sympatry and allopatry), these studies typically lack multilocus estimates of gene flow among cryptic species (Molbo et al. 2003; Haine et al. 2006; Moe & Weiblen 2010; Sun et al. 2011; Chen et al. 2012; Darwell et al. 2014). Consequently, it is not known whether cryptic fig wasp species identified using only one or a few barcoding genes are actually reproductively isolated. At the very least, tests of gene flow among sympatric cryptic fig wasp lineages could provide a yardstick for the initial assessment of species status among allopatric lineages, and would go a long way towards improving our understanding of diversification within the fig – fig wasp system.

1.4 Geographic distribution and species’ responses to climate change

The discovery of cryptic species affects our understanding of not only evolutionary diversification, but also species’ ecology (Bickford et al. 2007; Joly et al. 2014). Indeed, the uncovering of cryptic diversity has important biodiversity consequences and may necessitate a reassessment of conservation status and a shift in how we perceive certain aspects of species interactions; for example, patterns of host association, niche partitioning and the structure of food webs (Bickford et al. 2007; Joly et al. 2014). Furthermore, the geographic distribution of cryptic species can have important implications for thermal adaptation (Stevens 1989; Addo-Bediako et al. 2000), which affect species’ resilience to predicted climate change (e.g. Samways et al. 1999; Gibson et al. 2010; Lynch et al. 2014).

Stevens (1989) showed that, compared to species at lower latitudes, animal species occupying higher latitudes generally: (i) have broader latitudinal ranges (Rapoport’s Rule) and (ii) experience greater variability in temperatures. The Climate Variability Hypothesis (CVH; Stevens 1989) posits that such broad latitudinal ranges are achieved by adaptation to a wide range of climatic conditions. In a meta-analysis, Addo-Bediako et al. (2000) confirmed that insects at higher latitudes exhibited a greater tolerance to a wide range of temperatures, although upper thermal limits remained fairly consistent across latitudes. Consequently, small increases in mean temperature are likely to have a greater impact on species at low latitudes than on those at high latitudes, which currently experience mean temperatures well below optimum. Despite the fact that temperate insects are predicted to be less negatively affected by increasing temperatures than those in the tropics, climate warming has had considerable consequences for pollination in temperate environments. In a number of pollination systems, increases in mean temperature have changed the timing of plant and insect development, thus reducing temporal overlap of crucial life history stages and resulting in decreased frequency of pollination events (e.g. Gordo & Sanz 2005; Doi et al. 2008; Kudo & Ida 2013; Kudo 2014).

Two important aspects of climate change that have received relatively little attention are extreme heat and reduced relative humidity. Alarmingly, Simmons et al. (2010) reported long-term decreases in relative humidity at low and mid-latitudes, while warming trends have been accompanied by increases in extreme heat events (Hughes 2003; Collins et al. 2013; Steffen 2015). Aseasonal pollination systems may be more susceptible to extreme climatic events because they are not reliant on seasonal cues for development and pollination. Similarly, generalist pollination systems are expected to be less sensitive to increases in mean temperature due to diverse pollinator assemblages that act as buffers to environmental change (Brittain et al. 2013). Despite the adverse effects of decreased relative humidity (Broufas et al. 2009; Lu & Wu 2011; Tochen et al. 2016) and extreme heat (Mironidis & Savopoulou-Soultani 2010; Nyamukondiwa et al. 2013; Jeffs & Leather 2014; Chiu et al. 2015; Ma et al. 2015) on insect fitness, very few laboratory studies have investigated their combined effects on insect pollinators. Consequently, we do not fully understand how temperate insect pollinators may fare in a changing climate.

Unlike most plants, figs typically exhibit flowering patterns that are aseasonal or only weakly seasonal (Patel 1996; McPherson 2005; Yeo & Tan 2009), and are thus less likely than other pollination systems to be affected by phenological mismatches. Nevertheless, Harrison (2001) reported local extinction of fig wasps due to drought- induced breaks in fig flower production in Borneo. Interestingly, pollinators of monoecious figs were less affected than pollinators of dioecious figs, possibly because the former could recolonise more effectively by long-distance dispersal (Harrison 2000). Furthermore, consistent with the CVH, Warren et al. (2010) demonstrated that the widespread African fig wasp Ceratosolen arabicus had a greater thermal tolerance and desiccation resistance than C. galili, which had a smaller geographic range and experiences cooler, more humid conditions. Perhaps unsurprisingly, then, laboratory studies have shown that small increases in temperature and decreases in humidity can significantly shorten the lifespan of tropical fig wasp species (Dunn et al. 2008b; Jevanandam et al. 2013; Wang et al. 2013).

While these laboratory studies have highlighted the potential risks that climate changes poses to some fig wasp species, many unanswered questions remain. It is not yet clear how the predicted increase in frequency of extreme heat events and a reduction in humidity might affect fig wasps with strictly temperate distributions. This is particularly relevant in Australia, which is predicted to become hotter and drier on average, with a greater incidence of extreme weather events (Hughes 2003; Steffen 2015). Intuitively, it could be suggested that, relative to tropical species, decreases in lifespan would require larger changes in temperature and humidity. However, there are no empirical data to support this. Moreover, adult longevity is an important factor for the persistence of fig-pollinator populations as both fig and wasp reproduction is dependent on female wasp dispersal and location of receptive figs; yet no study has explored the effect of increased temperature or reduced humidity on any other aspect of fig wasp biology.

In particular, while it is hypothesised that increasing temperatures may reduce pollination potential by reducing the lifespan of pollinators, perhaps more importantly, we do not know how high temperatures impact on fig wasp survival within the syconium and on rates of emergence. The ability of pollen-bearing female

fig wasps to effectively disperse fig pollen is contingent on their successful release from the natal fig, which is dependent on the ability of male fig wasps to create exit holes through which females can escape. If male fig wasps suffer sub-lethal effects due to high temperatures within the natal fig, they may fail to release the females (Zammit & Schwarz 2000). Of course, if similar effects are suffered by female fig wasps, they may fail to escape the natal fig regardless of whether or not males have created exit holes. Regardless, in both cases, ‘ecological death’ (i.e. failure to create exit holes) may precede ‘physiological death’ in the sense that female wasps would remain trapped within the natal fig and are unable to transport pollen to and reproduce within a receptive fig, even though they are no actually dead. While Patiño et al. (1994) demonstrated that temperatures inside a fig can exceed ambient temperature by up to 8°C, they failed to show what this meant for the pollinators themselves. Coupling experiments on fig wasp emergence and field measurements of internal fig temperature would provide a clearer picture of the potential impact of climate change on fig wasp biology.

1.5 Thesis aims and outline

The primary aim of this thesis is to explore the evolutionary ecology of fig wasps, specifically in relation to the biogeography of host-parasitoid interactions, including testing of the general applicability of DNA barcoding in closely-related species under different geographic scenarios. As figs represent keystone species in many ecosystems, I also investigated the resilience of a temperate fig-pollinating wasp to predicted climate change. The four key specific questions that I addressed were:

1. Do patterns of population genetic structure vary between a fig wasp and its associated parasitoid?

2. Is there substantial gene flow among sympatric cryptic species delimited by DNA barcoding?

3. Is genetic differentiation sufficient to suggest full species status for allopatric populations of a fig wasp that are not delimited as separate species by DNA barcoding?

4. What are the implications of the increasing frequency of extreme weather events for the emergence and longevity of fig wasps in temperate regions under predicted climate change?

This thesis explores the themes of coevolution, DNA barcoding and resilience to climate change. Below, I provide a detailed outline of my study system with reference to these themes. Each section contains an overview of the individual chapters of this thesis, with specific aims and hypotheses interspersed throughout.

1.5.1 Study system: Port Jackson fig (Ficus rubiginosa) and its associated fig wasps

Ficus rubiginosa (Desf. ex Vent.) is a monoecious actively-pollinated fig species that is naturally distributed along most of the east coast of Australia, and up to 200 km inland (Dixon et al. 2001). DNA barcoding efforts (Haine et al. 2006; Darwell et al. 2014) have revealed that F. rubiginosa is pollinated by five species of the Pleistodontes imperialis Saunders (Hymenoptera: Agaonidae) complex; referred to as P. imperialis sp. 1-5 (Table 1.1; sensu Haine et al. 2006; Darwell et al. 2014). These species have partially overlapping geographic ranges and, in many regions, multiple species exist in sympatry (Figure 1.3). For example, in Townsville, north Queensland, three P. imperialis species coexist at frequencies in excess of 10% (Darwell et al. 2014). Interestingly, P. imperialis sp. 1 (Figure 1.5) comprises two allopatric populations: a large population in the southern temperate range of F. rubiginosa and a small population in the Atherton Tablelands in north Queensland.

Ficus rubiginosa is also host to over 20 NPFW species (Darwell 2012), representing the four functional guilds defined by Segar et al. (2013). Two striking patterns have evolved independently in more than one NPFW lineage in this particular system: sympatric morphospecies differentiated by ovipositor length (Segar 2011) and latitudinal geographic replacement of ecologically similar species (Darwell 2012). For example, P. imperialis wasps are attacked by three (small) Sycoscapter parasitoids; referred to as Sycoscapter sp. A, B and C (Table 1.1; sensu Moore et al. 2008). Sycoscapter sp. A has a long ovipositor and is found throughout the wide geographic range of F. rubiginosa, while Sycoscapter sp. B and C have short ovipositors and occupy largely allopatric (northern and southern) geographic ranges

with a small area of overlap around Brisbane (Figure 1.4; Darwell 2012). Table 1.1 clearly sets out the nomenclature used for cryptic species investigated in this thesis.

Figure 1.3: The distribution and relative frequencies of the five fig-pollinating wasps of the Pleistodontes imperialis species complex that pollinate F. rubiginosa. The black dotted line denotes the western boundary of the geographic range of F. rubiginosa (Dixon et al. 2001). The locations of sampling sites used for the research presented in this thesis are shown. Figure modified from Darwell et al. (2014).

Table 1.1: Cryptic fig wasp species nomenclature.

Species group Trophic role No. recognised species Species designations Pleistodontes imperialis Pollinators 5 1, 2, 3, 4, 5 Sycoscapter spp. Parasitoids 3 A, B, C

The geographic distributions of the fig wasps hosted by F. rubiginosa make this a convenient system to address the specific aims of this thesis. Firstly, the existence of a widespread parasitoid species (Sycoscapter sp. A), attacking multiple conspecific

pollinator populations (i.e. P. imperialis sp. 1), warrants characterisation of population genetic structure to ascertain whether or not this situation may be a precursor to ecological speciation (Thompson 2005). Secondly, the local coexistence of multiple cryptic lineages in Townsville provides an excellent opportunity to assess gene flow among cryptic species delimited previously by DNA barcoding. Furthermore, allopatric P. imperialis sp. 1 populations facilitate tests of the DNA barcoding method in geographically-disjunct populations. Finally, the southern P. imperialis sp. 1 population is an ideal candidate for testing the effects of key factors associated with climate change (i.e. high temperatures and aridity) on the emergence and longevity of a temperate fig wasp, which has important ecosystem implications.

1.5.2 Host-parasitoid coevolution: a biogeographic investigation

I identified the interaction between P. imperialis sp. 1 and Sycoscapter sp. A as one of particular interest with regard to Thompson’s (2005) ‘geographic mosaic theory’ of coevolution (see section 1.2.1), due to the disjunct nature of P. imperialis sp. 1 populations and the broad range of Sycoscapter sp. A. Furthermore, my initial field sampling revealed that Sycoscapter sp. A was the most abundant parasitoid in this system and one for which microsatellite markers had previously been developed (Bouteiller-Reuter et al. 2009), reinforcing the suitability of studying this particular interaction.

Aim 1: Development of microsatellite markers for P. imperialis sp. 1

Microsatellites have arguably been the most popular tool in population genetics in the past 20 years, due to their ability to detect processes at short temporal and spatial scales (Sunnucks 2000). They are extremely useful for identifying not only patterns of gene flow among populations, but also the resumption of gene flow between historically isolated lineages (Randi 2008). The first aim of my thesis was therefore to develop a panel of microsatellite markers that were suitable for P. imperialis sp. 1 in particular, but that were also applicable to other members of the pollinator species complex.

In Chapter 2, I describe the development and characterisation of nine polymorphic microsatellite markers for P. imperialis sp. 1. I demonstrate the utility of these markers for population genetic studies in an analysis of samples collected from two

localities in southeastern Australia: Sydney and Newcastle (Figure 1.3). Although these markers were originally developed using individuals from the southern P. imperialis sp. 1 population, I also demonstrate their utility in two other P. imperialis species: P. imperialis sp. 2 and 4. Of the nine microsatellite loci, seven and six were polymorphic in P. imperialis sp. 2 and 4, respectively.

Figure 1.4: Distributions of (A) Sycoscapter sp. A and (B) Sycoscapter sp. B and C. Inset: Sycoscapter Long (A) and short (B) ovipositor morphs. The long ovipositor morph is a single species (sp. A), while the short ovipositor morph comprises two species (sp. B and C) Figure modified from Darwell (2012).

Aim 2: Comparison of population genetic structure of a fig-pollinating wasp and its parasitoid

Having developed microsatellite markers for P. imperialis sp. 1, the second major aim of my thesis was to compare the population genetic structure of a fig-pollinating wasp (P. imperialis sp. 1) and its parasitoid (Sycoscapter sp. A) to highlight any differences that may lead to the formation of geographic mosaics. In this study, detailed in Chapter 3, I co-sampled pollinator and parasitoid populations from field sites in four regions of New South Wales (Sydney, Newcastle, Port Macquarie and Byron Bay) and one region in north Queensland (Atherton Tablelands) (Figure 1.3). This sampling regime ensured that matched pollinator and parasitoid samples were collected from throughout the entire geographic range of the pollinator.

This study had two main goals: (i) to contrast patterns of gene flow among pollinator and parasitoid populations and, (ii) to assess the relative dispersal ability of the parasitoid species. The first goal has implications for the coevolutionary dynamics of this interaction; specifically, discordant gene flow between interacting species may lead to variation in the outcome of host-parasitoid interactions and, potentially, ecological divergence (Thompson 1994, 2005). The second goal is of broader interest to improving the understanding of NPFW biology (Compton et al. 2000; Harrison 2003) and filling a species biology knowledge gap. Fig-pollinating wasps are renowned for very long-distance passive dispersal on wind currents (Harrison & Rasplus 2006; Ahmed et al. 2009) and broad geographic populations when host plant distribution is sufficiently continuous (Kobmoo et al. 2010), but very little is known about NPFW dispersal. Given that pollinators and NPFWs colonising the same fig species have been shown to have similar flight heights (Compton et al. 2000; Harrison 2003), it might be expected that they too are capable of long-distance dispersal. I therefore predicted that the pollinating wasp P. imperialis sp. 1 would exhibit weak genetic structure in the large southern population, but would be highly structured between the disjunct southern and northern populations. In contrast, I hypothesised that the parasitoid Sycoscapter sp. A would exhibit weak genetic structure throughout its wide, more or less continuous, geographic range (Figure 1.4).

Figure 1.5: Pleistodontes imperialis sp. 1 (bottom) and Sycoscapter sp. A (top).

1.5.3 Assessment of DNA barcoding in sympatric and allopatric lineages

Chapter 3 confirmed my hypothesis of large genetic differentiation of allopatric pollinator populations. This raised questions about the reliability of DNA barcodes as indicators of species status, because Darwell et al. (2014) concluded that these populations were the same species. While tests of gene flow between these widely separated populations are practically impossible due to the intricate biology of the fig/pollinator interaction, the local coexistence of multiple P. imperialis lineages provides an ideal opportunity to quantify genetic exchange between barcoding lineages in this system.

Aim 3: Testing the reliability of DNA barcoding methods for fig wasps

In Chapters 4 and 5, I address the third aim of this thesis: to explore the reliability of DNA barcoding for sympatric and allopatric P. imperialis lineages. In Chapter 4, I describe a study to test the validity of DNA barcoding in sympatric P. imperialis lineages using the microsatellite markers developed in Chapter 2. I chose to conduct field sampling in Townsville, north Queensland, because three P. imperialis species (P. imperialis sp. 2, 3 and 4) coexist in this region at relatively high frequencies

(Figure 1.3). The purpose of this study was to estimate the proportion of figs that contained broods from multiple fig wasp lineages, quantify gene flow between mtDNA lineages of P. imperialis, and conduct tests for genetic signals of population expansion that are generally characteristic of hybrid zones. I hypothesised that gene flow and hybridisation between lineages would be rare due to a fairly low, but not trivial, proportion of figs containing more than one P. imperialis lineage.

In Chapter 5, I revisited the P. imperialis sp. 1 populations investigated in Chapter 3, which are considered conspecific based on DNA barcoding (Darwell et al. 2014), but exhibit very high levels of divergence at microsatellite loci. The difficulties in delimiting allopatric taxa are well-known (Coyne & Orr 2004; Lukhtanov et al. 2015), but rarely studied. In this chapter, I scaled the genetic divergence of allopatric P. imperialis sp. 1 populations with reference to P. imperialis sp. 2 using mitochondrial and nuclear DNA sequences and microsatellites. I hypothesised that P. imperialis sp. 1 populations would exhibit patterns of genetic differentiation consistent with an incipient speciation process: very low (if any) divergence based on nuclear DNA sequences, moderate mtDNA divergence and very high microsatellite divergence.

1.5.4 The effect of climate change on temperate fig wasps

During field collections in January 2013, I observed widespread heat-induced population knockdown of fig wasp populations throughout the Greater Sydney region. On two separate days temperatures exceeded 40°C in Sydney, including the hottest day on record, when temperatures exceeded 46°C (Bureau of Meteorology 2016). These extreme events were also likely contributors to the collapse of other insect populations (e.g. psyllids; Hall et al. 2015). Following these events, I was unable to recover any live pollinators from ripe F. rubiginosa figs for over two months, despite intensive sampling. This raised the question of how extreme weather events, such as high temperatures and aridity, affect the survival of fig wasps, both inside and outside the natal fig. This has important implications for their dispersal potential, which is crucial to the persistence of pollinator populations. This is particularly significant for pollinators of monoecious figs, as the sparse distribution

of trees in monoecious fig populations often necessitates the long-distance dispersal of pollinators (Chapter 3 of this thesis; Ahmed et al. 2009; Kobmoo et al. 2010).

Aim 4: Assessing the implications of extreme weather events on a temperate fig- pollinating wasp species

I subsequently expanded the aims of my thesis to address the pertinent and very pressing issue of climate change. In Chapter 6, I depart from the molecular methods employed in earlier chapters for an exploration of the resilience of temperate fig wasp species to climate change. Numerous studies have demonstrated the sensitivity of tropical fig wasp species to small climatic changes (Dunn et al. 2008b; Jevanandam et al. 2013; Wang et al. 2013), yet there have been no studies of this kind on temperate fig wasps. Furthermore, while previous studies measured adult longevity in response to changes in temperature and humidity, they failed to address another important aspect of fig wasp biology: emergence from the natal fig.

The final aim of my thesis was to assess the impact of extreme weather events on different aspects of the biology of a temperate fig-pollinating wasp species. To address this, I tested the effects of (i) increased temperature on the emergence of wasps from the natal fig, and (ii) different temperature × humidity regimes on adult pollinator longevity. Specifically, I wanted to identify critical temperatures that cause widespread mortality in temperate fig wasp populations. Moreover, I took field measurements of internal fig temperatures to estimate the potential danger for wasps inside sun-exposed figs relative to wasps in shade-protected figs. I used the southern P. imperialis sp. 1 population to address these aims as it is the only population in this complex that is found exclusively in temperate southeastern Australia.

I hypothesised that (i) wasp emergence would increase with temperature, before dropping sharply at a critical temperature, (ii) adult longevity would decrease at extreme temperatures (i.e. >5°C above summer mean maximum temperature), (iii) decreases in humidity would cause further significant reductions in adult longevity, and (iv) sun-exposed figs would experience significantly higher internal temperatures than shade-protected figs. Reduced longevity presumably leads to a reduced dispersal potential, and under extreme conditions, the failure of pollinators to locate and pollinate receptive figs may lead to the local collapse of this pollination system

(Harrison 2000), which potentially has massive implications for the ecosystems it supports.

1.6 Conclusion

Following the experimental chapters contained herein, the final chapter of my thesis provides a brief overview of the major findings of my research. Importantly, I discuss the implications of these findings for our understanding of fig wasp evolutionary ecology. Finally, I address the limitations of this study and outline areas of research that deserve attention in the future.

Chapter 2 Characterisation of microsatellite markers for fig-pollinating wasps in the Pleistodontes imperialis species complex

Sutton, TL, Riegler, M and Cook, JM (2015) Characterisation of microsatellite markers for fig-pollinating wasps in the Pleistodontes imperialis species complex. Australian Journal of Zoology, 63(2), 122-126.

2.1 Abstract

We characterised a set of nine polymorphic microsatellite loci for Pleistodontes imperialis sp. 1, the pollinator wasp of Port Jackson fig (Ficus rubiginosa) in southeastern Australia. Characterisation was performed on 30 female individuals collected from a population in Sydney, Australia. The average number of alleles per locus was 7.33, and eight loci were not in Hardy–Weinberg equilibrium. This was expected as fig wasps are known to be highly inbred. A test of genetic differentiation between two natural populations of P. imperialis sp. 1 (Sydney and Newcastle, Australia – some 120 km apart) yielded a very low FST value of 0.012, suggesting considerable gene flow. Bayesian clustering analysis using TESS 2.3.1, which does not assume Hardy–Weinberg equilibrium, however, indicated potential spatial substructuring between the Sydney and Newcastle populations, as well as within the Sydney population. The described loci were also characterised for two other species in the P. imperialis complex: P. imperialis sp. 2 (Townsville, Australia) and P. imperialis sp. 4 (Brisbane, Australia). Seven and six of the nine loci were polymorphic for P. imperialis sp. 2 and P. imperialis sp. 4, respectively.

2.2 Introduction

The Port Jackson fig (Ficus rubiginosa) is a monoecious species with a natural distribution spanning Australia’s east coast, from Cape York in the north to Eden in the south (Dixon et al. 2001). Pleistodontes imperialis Saunders is the formally described pollinating wasp of F. rubiginosa (Lopez-Vaamonde et al. 2002). Extensive DNA barcoding efforts, using mitochondrial and nuclear DNA loci, have recently found that P. imperialis actually comprises five cryptic species with partially overlapping geographic ranges (Haine et al. 2006; Darwell et al. 2014).This system therefore provides excellent opportunities to explore the ecological and evolutionary dynamics of the fig/pollinator symbiosis, including the impact of different pollinator species. Consequently, there is a strong motivation to develop wasp microsatellite markers to explore population level interactions and infer patterns of dispersal. Such tools have previously been utilised for studies of fig wasp species in other genera from Africa and Asia (Molbo et al. 2004; Lin et al. 2008; Greeff et al. 2009; Kobmoo et al. 2010).

2.3 Methods, results and discussion

Using DNA from 20 individuals of P. imperialis sp. 1 collected from a single fig in Melbourne, Australia, we created a DNA library enriched for dinucleotide microsatellites (GA and CA repeats) following the double-enrichment method described in Hale et al. (2002). Two microlitres of the double-enriched library were transformed into 100 µL QAComp-C01 competent cells (Qbiogene) and plated onto Luria-Bertani agar plates containing 50µg/mL ampicillin. Plasmids from individual colonies were prepared using a QIAprep® Spin Miniprep Kit (Qiagen). Plasmid inserts were then sequenced using BigDye Terminator Cycle Sequencing chemistry (Applied Biosystems) and sequences were detected on an ABI 310 Prism® genetic analyser. Forty clones were sequenced, of which eleven contained a microsatellite with ten or more repeats. PCR primers were designed from the flanking sequences of each microsatellite locus using the Primer3 program (Koressaar & Remm 2007; Untergasser et al. 2012) (Table 2.1). These eleven loci were screened for polymorphism in 30 Pleistodontes imperialis sp. 1 wasps (sensu Haine et al. 2006)

collected in Sydney (Australia). To maximise independence of samples, only a single female wasp was collected from any one fig fruit, and 1-4 fig fruits were sampled per tree. The Sydney population comprised samples from 15 trees that were between 10 m and 40 km apart.

We extracted DNA from whole female insects in 100 l of extraction solution (5% Chelex, 0.01% proteinase K), incubated at 56°C for 35 min then at 96°C for 15 min, vortexed briefly and centrifuged for 5 min at 3500 rpm. We used 15 µL PCR reactions containing 1x Taq buffer, 2.5 mM MgCl2, 0.2 mM each dNTP, 0.33 M each primer, 0.6 U Taq DNA polymerase (Promega), with 1 µL template DNA. Thermal cycling conditions were 95ºC for 12 min, 10 cycles of 94ºC for 15 s, annealing temperature (Ta) (C) for 15 s (Table 2.1), 72ºC for 15 s, followed by 30 cycles of 89°C for 15 s, Ta (°C) for 15 s, 72°C for 15 s, and a final extension of 72ºC for 10 min. Forward PCR primers were fluorescent-labelled with HEX, NED and 6- FAM dyes (Table 2.1) and detected on an ABI 3700 genetic analyser with LIZ600 internal size standard. Fragments were analysed and sized using GeneMapper Software v4.1 (Applied Biosystems). PCR amplification of two loci did not reliably yield products, but the remaining nine amplified well and were polymorphic (Table 2.1).

We used ARLEQUIN v3.5 (Excoffier & Lischer 2010) to test for deviation from Hardy-Weinberg equilibrium (HWE), and to test for linkage disequilibrium among loci followed by correction for multiple testing using the False Discovery Rate

(FDR) (Benjamini & Hochberg 1995). Observed (HO) and expected (HE) heterozygosity were calculated in ARLEQUIN v.3.5 (Excoffier & Lischer 2010), and we used MICRO-CHECKER v.2.2.3 (van Oosterhout et al. 2004) to test for the presence of null alleles. F-statistics were estimated in FSTAT (Goudet 1995).

The number of alleles ranged from 4 to 11, with an average of 7.33 per locus (Table 2.1). Eight loci were not in HWE, but this was expected because fig-pollinating wasps are well known for high levels of sib-mating (e.g. Molbo et al. 2004). Following correction for FDR, all loci were found to be inherited independently (P > 0.02). Observed heterozygosity of loci ranged from 0.167 to 0.567, with an average of 0.422. Seven of the nine loci were suggested to have null alleles due to homozygote excess (Table 2.1). However, a more likely explanation of homozygote

excess in this case is inbreeding as the median number of foundresses entering F. rubiginosa figs is only two (J.M. Cook, unpublished data). Unfortunately, in most fig-pollinating wasps, it is not possible to confirm this by genotyping both mother and offspring because once a mother enters a fig to lay eggs and pollinate she does not leave it. By the time the offspring have developed, several weeks later (Xiao et al. 2008), the body of the mother has decomposed and disintegrated (Cook & Power 1996).

We estimated diversity between two natural populations of P. imperialis sp. 1 separated by approximately 120 km; Sydney (N = 30) and Newcastle (N = 30) (Australia) (Table 2.2). The Newcastle population comprised samples from 14 trees separated by distances of 10 m to 10 km. A single wasp was collected from each fig fruit, with 1-5 fig fruits sampled per tree. Overall allelic diversity was similar between the Sydney (NA = 69) and the Newcastle (NA = 63) population, although number of alleles differed between loci. Eight and six loci were not in HWE for the Sydney and the Newcastle population, respectively, five of which were common between both populations (Table 2.2). The larger number of loci not in HWE for the Sydney population could be due to the pooling of individuals from sites separated by approximately 40 km, but this requires further study. There were private alleles for six and four loci for the Sydney and the Newcastle population, respectively. FST was very low (0.012) indicating little genetic differentiation between the two populations.

Such FST values are typical within breeding fig wasp populations. For example, FST values for Ceratosolen fusciceps, the pollinator of Ficus racemosa, ranged from

0.001 to 0.039 in south-east Asia (Kobmoo et al. 2010). High values of FIS (0.344) and FIT (0.352) were expected as few females lay eggs within each fig, and mating between siblings is therefore very common (Hamilton 1979; Cook & West 2005).

Table 2.1: Characteristics of nine microsatellite loci for 30 Pleistodontes imperialis sp. 1 from a population in Sydney (Australia). Repeat Locus Primer Sequence (5’-3’) Dye Product size range (bp) T (°C) N H H motif a A O E F: AACCAGCGCGTATGTTTCTC Pli2 (AG)11 6-FAM 260-266 52 4 0.367 0.55 NS R: TTGGAACTTCGTCGTCTCG F: CGCATCAAGCTGCAGAAAAG Pli3 (CT)13 HEX 98-110 52 7 0.4 0.666 * R: TATAGTGCGCTGCCAAACTG F: TTCTCCCTTGCAAACGATTC Pli4 (AG)13 HEX 255-263 58 5 0.433 0.555 NS R: CCGAATTCACCAGTTTTTGG F: AGCGCATTTTCCTCTCACTC Pli5 (CT)18 HEX 118-138 52 10 0.567 0.87 * R: GCTCGCTAATCCAGTTCACC F: CGTAATTAGCAGGCGATTCC Pli6 (CT)11 NED 219-241 52 8 0.4 0.811 * R: ATAAACGCAAACGACGTTCC F: CTGCGTGTCTCGGTTTCTTC Pli7 (CT)10 HEX 190-232 52 7 0.167 0.587 * R: AGCGGAGGGTTTTAATTTCG F: GCGCAACGCAAAAGTAAGTC Pli8 (CT)12 6-FAM 174-194 58 11 0.6 0.907 * R: GTGCAGCTGTTAAGCTTTCG F: CAACACAAACGATGATGACG Pli10 (CT)14 6-FAM 130-150 58 8 0.467 0.793 * R: GCCGCGAAGATAAAGATGG F: CGATACTACCTCGTGAAGCTG Pli11 (CT)10 6-FAM 215-227 50 6 0.4 0.618 NS R: CGTTTTAATCTCGTAGGCGTATC DNA sequences deposited in Genbank with accession numbers KM229708 - KM229716; Locus name (bold indicates potential presence of null alleles); F, forward primer (bold denotes fluorescent-labelled primer); R, reverse primer; bp, base pairs; Ta, annealing temperature; NA, number of alleles; HO, observed heterozygosity; HE, expected heterozygosity; asterisk denotes significant deviation from Hardy-Weinberg equilibrium (P < 0.05) (NS: non significant).

Table 2.2: Genetic diversity of two Pleistodontes imperialis sp. 1 populations at nine microsatellite loci. Sydney (N = 30) Newcastle (N = 30)

Locus NA NP HO/HE NA NP HO/HE Pli2 7 0 0.37/0.55 4 0 0.33/0.57* Pli3 7 1 0.40/0.67* 6 0 0.53/0.65

Pli4 5 0 0.43/0.56 5 0 0.40/0.63 Pli5 10 2 0.57/0.87* 12 4 0.70/0.89

Pli6 8 2 0.40/0.81* 7 1 0.47/0.68

Pli7 4 2 0.17/0.59* 2 0 0.10/0.35*

Pli8 11 0 0.60/0.91* 14 3 0.73/0.92*

Pli10 11 4 0.47/0.79* 8 1 0.60/0.64

Pli11 6 1 0.40/0.62 5 0 0.30/0.38

NA, number of alleles; NP, number of private alleles; HO, observed heterozygosity; HE, expected heterozygosity; asterisk denotes statistically-significant deviation from Hardy-Weinberg equilibrium (P < 0.05); F-statistics measured across both populations and all loci in FSTAT.

We also analysed these data within the Bayesian framework of TESS v.2.3.1 (Chen et al. (2007) to test for genetic clustering in our data. TESS does not require loci to be in HWE and utilises individual spatial coordinates without assuming predefined populations. It uses Voronoi tessellation to link individuals, creating a ‘neighbourhood system’ that can be modified by the user to reflect realistic barriers to gene flow. Due to suspected long distance dispersal in fig-pollinating wasps (Ahmed et al. 2009), we modified the neighbourhood system so that all individuals were linked. Like other Bayesian clustering methods, TESS is used to estimate the number of genetic clusters in a dataset and the genetic make-up of each individual in the dataset.

We implemented an admixture model (Durand et al. 2009) and ran 50 iterations for

Kmax (maximum number of genetic clusters, K) values ranging from two to seven. Each run consisted of 60,000 sweeps with a burn-in period of 10,000. The most likely value of K was determined as the point at which the deviance information criterion (DIC) curve inflected upward when plotted against Kmax. The admixture coefficients of the 20 runs with the lowest DIC for K values 2 to 4 were averaged using the Greedy algorithm in CLUMPP v.1.1.2 (Jakobsson & Rosenberg 2007) and visualised with DISTRUCT v.1.1 (Rosenberg 2004).

The most likely value of K was determined to be four (Figure 2.1). Despite this, the bar plot for K = 2 (Figure 2.2a) was more informative than the plots for both K = 4 (Figure 2.2c) and K = 3 (Figure 2.2b). There is some evidence of differentiation between the Sydney and Newcastle populations (Figure 2.2a), as a higher proportion of individuals within the Sydney sites (particularly Penrith) were primarily assigned to the darker of the two genetic clusters. The incidences of private alleles between populations further suggest the possibility of differentiation between populations. Previous studies have shown that pollinating wasps from other monoecious fig species form large populations and are capable of long-distance dispersal (Ahmed et al. 2009; Kobmoo et al. 2010). Results here suggest that more complex spatial structuring could potentially occur at small scales (i.e. within the Sydney region). However, a more exhaustive sampling approach is required to confirm these patterns and uncover geographic structure across the wider range of P. imperialis sp. 1.

Figure 2.1: Deviance information criterion (DIC) as a function of Kmax. The DIC curve should stabilise, or vary slightly, close to the correct K value (Chen et al. 2007). Optimal K in this figure is 3.

To test for cross-species utilisation of the nine described loci, we screened 11 female wasps from each of two other species in the P. imperialis species complex; P. imperialis sp. 2 wasps from Townsville (Australia) and P. imperialis sp. 4 wasps from Brisbane (Australia). PCR amplification was performed under the same conditions as for P. imperialis sp. 1 (Table 2.1). PCR tests for all loci reliably produced products, although two loci were monomorphic in P. imperialis sp. 2 and three loci were monomorphic in P. imperialis sp. 4 (Table 2.3). Overall allelic diversity was higher in P. imperialis sp. 2 (NA = 48) than species 4 (NA = 31). There were private alleles for all loci, and some loci were not in HWE (Table 2.3). Potential null alleles were detected in four and three loci for P. imperialis sp. 2 and sp. 4, respectively.

Figure 2.2: TESS bar plot for K = 2 (a), K = 3 (b) and K = 4 (c). Each vertical line represents an individual and its assignment to each of the genetic clusters (denoted by different shades).

Deviation from HWE and the suggestion of null alleles in microsatellite loci is common and expected in fig wasps (Zavodna et al. 2002; Lin et al. 2007; Hao et al. 2008; Katabuchi et al. 2008; Liu et al. 2009; Kobmoo et al. 2010; Tian et al. 2011; Kusumi & Su 2014; Wang et al. 2014) due to the unique ecology of fig – pollinator interactions. Fig flowers are encased in an enclosed inflorescence (syconium), which is receptive for only a few days in the early stages of development. Pollen-bearing females enter the syconium through a narrow passageway (ostiole) which closes at the end of the receptive phase. Due to time (and often, space) constraints, it is common for only one or two foundresses to enter a syconium (Cook & West 2005). As mating takes place within the natal syconium, sib-mating is very common, resulting in homozygote excess at many loci.

2.4 Conclusion

These are the first microsatellite markers developed for fig wasp species within the Australasian Pleistodontes genus, and will facilitate future studies of population genetics in members of the P. imperialis species complex. Such tools have great scope to improve the use of fig wasps as a model system for studies of both insect/plant coevolution and mating systems (Herre et al. 1997).

Table 2.3: Characteristics of nine microsatellite loci for Pleistodontes imperialis sp. 2 (N = 11) and Pleistodontes imperialis sp. 4 (N = 11).

Locus Product size range (bp) NA HO HE P. imperialis sp. 2 Pli2 276-282 4 0.091 0.767 * (N = 11) Pli3 106-118 6 0.727 0.823 NS Pli4 251-291 9 0.545 0.857 * Pli5 116-120 3 0.273 0.515 NS Pli6 209-249 9 0.364 0.905 * Pli7 196-200 3 0.364 0.537 NS Pli8 174-208 12 0.727 0.931 * Pli10 126 1 ND ND Pli11 215 1 ND ND P. imperialis sp. 4 Pli2 264-266 2 0.091 0.091 NS (N = 11) Pli3 116-118 2 0.091 0.236 NS Pli4 261-271 5 0.364 0.64 * Pli5 118-128 5 0.636 0.64 NS Pli6 189 1 ND ND Pli7 196-202 4 0.364 0.706 NS Pli8 174-202 10 0.182 0.868 * Pli10 120 1 ND ND Pli11 217 1 ND ND

Locus name (bold indicates potential presence of null alleles) bp, base pairs; NA, number of alleles, HO, observed heterozygosity; HE, expected heterozygosity; asterisk denotes significant deviation from Hardy-Weinberg equilibrium (P < 0.05) (NS: non significant); ND: test not done due to monomorphism.

Chapter 3

Comparative population genetic structure in a fig-pollinating wasp and its main parasitoid

Sutton, TL, Riegler, M and Cook, JM (2016) One step ahead: a parasitoid disperses farther and forms a wider geographic population than its fig wasp host. Molecular Ecology, 25(4), 882-894.

3.1 Abstract

The structure of populations across landscapes influences the dynamics of their interactions with other species. Understanding the geographic structure of populations can thus shed light on the potential for interacting species to coevolve. Host-parasitoid interactions are widespread in nature, and also represent a significant force in the evolution of plant-insect interactions. However, there have been few comparisons of population structure between an insect host and its parasitoid. We used microsatellite markers to analyse the population genetic structure of Pleistodontes imperialis sp. 1, a fig-pollinating wasp of Port Jackson fig (Ficus rubiginosa), and its main parasitoid, Sycoscapter sp. A, in eastern Australia. Besides exploring this host-parasitoid system, our study also constitutes, to our knowledge, the first study of population structure in a non-pollinating fig wasp species. We collected matched samples of pollinators and parasitoids at several sites in two regions separated by up to 2,000 km. We found that pollinators occupying the two regions represent distinct populations, but, in contrast, parasitoids formed a single population across the wide geographic range sampled. We observed genetic isolation by distance for each species, but found consistently lower FST and RST values between sites for parasitoids compared with pollinators. Previous studies have indicated that pollinators of monoecious figs can disperse over very long distances, and we provide the first genetic evidence that their parasitoids may disperse as far, if not farther. The contrasting geographic population structures of host and parasitoid highlight the potential for geographic mosaics in this important symbiotic system.

3.2 Introduction

Thompson (1994) argued that coevolution due to interactions between species is a major driving force of evolution, and emphasised that the spatial distributions of interacting species affect the direction and strength of coevolution. Interacting species (e.g. hosts and parasites) may have different patterns of gene flow, and selection may favour different traits in different geographic locations (Thompson 2005). Such patterns may allow, for example, disconnected host populations to evolve independently in response to a common parasite population, thus producing mixed outcomes for the interaction across landscapes (geographic mosaics) (e.g. Kraaijeveld & van Alphen 1995; Thompson & Cunningham 2002; Gómez et al. 2009b). Knowledge of the genetic structure of interacting species can therefore help us to understand how gene flow influences, or could influence, trait coevolution across landscapes.

Interactions between plants and insects have played an enormous role in shaping extant global terrestrial biodiversity. For example, coevolution with insects is thought to have played a major role in the radiation of the angiosperms (Regal 1977; Mulcahy 1979), which, in turn, form the base of food webs in many terrestrial ecosystems. At a higher trophic level, parasitoid wasps (Hymenoptera) regulate many insect herbivore populations, and apply important top-down selection pressures on plant-insect interactions (Godfray 1994). Contrasting gene flow patterns for species involved in antagonistic interactions may affect the coevolutionary process by allowing adaptive traits to evolve in genetically isolated populations (e.g. Chaves- Campos et al. 2011).

The influence of host specificity on the genetic structure of insect herbivore populations appears inconsistent (Nason et al. 2002; Ruiz-Montoya et al. 2003; Downey & Nice 2011; Kohnen et al. 2011). Similarly, host insect specificity is generally not a good indicator of the genetic structure of parasitoid populations, and the influence of herbivore host-races on parasitoid population differentiation varies between systems (Vaughn & Antolin 1998; Jourdie et al. 2010). Furthermore, different parasitoid species feeding on the same host may exhibit markedly different patterns of genetic structure (Kankare et al. 2005).

The few direct population genetic comparisons of host and parasitoid species to date have revealed complex patterns of geographic structure (Althoff & Thompson 1999), vastly discordant responses to habitat fragmentation (Anton et al. 2007), but also, in some other cases, evidence of host-tracking (Johannesen & Seitz 2003; Nyabuga et al. 2012; Stone et al. 2012; Gebiola et al. 2014). Some parasitoid populations exhibit patterns of genetic structure consistent with multiple processes of ecological sorting, including host-tracking, introduction and host-shifts (Hayward & Stone 2006; Nicholls et al. 2010a). Extreme variation in genetic structure within and between host-parasitoid systems likely results from differences in aspects of parasitoid biology and the extent of anthropogenic disturbances, and evolutionary outcomes will likely vary geographically between different host-parasitoid interactions.

Figs (Ficus, Moraceae) share an obligate mutualism with their pollinating wasps (Hymenoptera: Agaonidae). This system is well-known for extreme levels of reciprocal partner specificity; however, numerous cases of pollinator sharing, co- pollination (multiple pollinators on a single host) and more or less cryptic fig wasp species have been discovered (Ramírez 1970; Molbo et al. 2003; Haine et al. 2006; Moe & Weiblen 2010; Lin et al. 2011; Sun et al. 2011; Cornille et al. 2012; Darwell et al. 2014; Liu et al. 2015a), suggesting a pervasive role for geographic mosaics in the evolution of fig – pollinator interactions. The fig – pollinator mutualism is exploited by diverse non-pollinating fig wasps (NPFWs), which do not pollinate fig flowers and have direct (parasitism) or indirect (competition) negative effects on the survival of the pollinator wasps (Bronstein 1991; West & Herre 1994; West et al. 1996; Yu 2001; Segar & Cook 2012). However, despite several studies of the genetic structure of pollinator populations (e.g. Molbo et al. 2004; Kobmoo et al. 2010; Liu et al. 2013), no study to date has investigated the genetic structure of NPFW populations.

Port Jackson fig (Ficus rubiginosa Desf. ex Vent.) is a monoecious species with a continuous distribution along most of the east coast of Australia (Dixon et al. 2001). Molecular barcoding of both nuclear and mitochondrial genes has revealed that F. rubiginosa is pollinated by five more or less cryptic species that form the Pleistodontes imperialis Saunders species complex (Haine et al. 2006; Darwell et al. 2014). Curiously, P. imperialis sp. 1 (sensu Haine et al. 2006) is common in two

disjunct regions (New South Wales and northern Queensland), separated by a stretch of approximately 1,500 km where the four other species of the P. imperialis complex are found (Figure 3.1) (Darwell et al. 2014). Wasps of the P. imperialis species complex are parasitised by wasps of the genus Sycoscapter Saunders (Hymenoptera: Pteromalidae) (Segar et al. 2014), but only one of these (Sycoscapter sp. A) is common throughout the wide geographic range of F. rubiginosa (Darwell 2012).

We investigated the population genetic structure of the pollinator, P. imperialis sp. 1, and its parasitoid, Sycoscapter sp. A, by co-sampling the two insect species across most of the geographic range of their shared host plant. Some pollinating wasps of monoecious figs are capable of very long-distance dispersal (Ahmed et al. 2009), and some form geographically broad populations at the landscape level, in the absence of geographic barriers (Kobmoo et al. 2010). Evidence suggests that the flight heights of NPFWs are similar to those of pollinators colonising the same species of fig (Compton et al. 2000; Harrison 2003), so there is reason to believe that parasitoids attacking pollinators of monoecious figs are also capable of long-distance dispersal, but this has not been tested. Furthermore, monoecious fig trees typically occur at low densities, which may necessitate long-distance dispersal for pollinators and parasitoids alike. Consequently, we predicted that the pollinators would exhibit weak genetic structure between sites within the large southern region, but high levels of differentiation between the spatially disjunct northern and southern regions, with lower genetic diversity in the smaller northern region. For the parasitoids, we predicted weak genetic structure across the entire sampled range due to their suspected high dispersal ability and apparently continuous geographic distribution along Australia’s east coast.

3.3 Materials and methods

3.3.1 Fig wasp biology

Figs and their pollinating wasps are completely dependent on each other for reproduction. Pollen-bearing female wasps locate receptive figs (syconia) via detection of species-specific volatile emissions. The female wasp (foundress) then enters the receptive fig through a narrow opening called the ostiole, which closes

after the receptive phase. The receptive phase lasts about 3-6 days in F. rubiginosa (Xiao et al. 2008). Inside the fig, the foundress pollinates fig ovules, a service that is performed actively or passively, depending on species. A single egg is also deposited into some fig ovules. This initiates the formation of a gall in which the developing larva feeds on the fig seed.

Foundresses die within the fig soon after pollination. Some six-to-eight weeks later (Harrison 2005; Xiao et al. 2008), males exit their galls and chew holes into galls containing females through which to mate. After mating, males chew exit holes through the fig wall and females collect pollen on their bodies (passively) or by inserting it into special pollen pockets (actively, as in our study species). Males die within their natal figs and the pollen-bearing females exit in search of receptive figs in which to lay eggs. The number of foundresses that enter a receptive fig is generally very low (modal foundress number in F. rubiginosa is two; J.M. Cook, unpublished data), and sib-mating is therefore very common in fig wasps.

The biology of NPFWs is poorly-understood (Cook & Segar 2010). In most species, females do not enter the fig to lay eggs. Instead, females have long ovipositors, which are used to drill through the fig wall to lay their eggs inside. The length of the ovipositor varies between species and has been linked to the stage of fig development at which different NPFW species lay eggs (Ghara et al. 2011). Little is known about foundress number in NPFWs, but in some species (including Sycoscapter spp.; J.M. Cook, personal observation) several females have been observed probing a single fig simultaneously, so inbreeding is probably less common than in pollinators.

3.3.2 Sample collection

Pleistodontes imperialis sp. 1 (henceforth ‘pollinator’) and Sycoscapter sp. A (henceforth ‘parasitoid’) were co-sampled during field trips between March 2012 and August 2014 from both the southern and northern regions where the pollinator is found. To avoid confusion over the term “population” (which we reserve for reference to inferred biological populations), we refer to the two levels in our sampling hierarchy as ‘regions’ and ‘sites’. The southern region comprised four sites in New South Wales (Sydney, Newcastle, Port Macquarie and Byron Bay) and the

northern region comprised a single site in Queensland (Atherton Tablelands) (Figure 3.1). For the Atherton site, where other pollinator species also occur on the same host plant, thus reducing sampling efficiency, we also used a small number of insects collected during field trips in 2005 and 2007. Samples collected across different years were pooled for analysis.

Figure 3.1: Geographic distribution of sampled sites. The dotted line denotes the geographic range of F. rubiginosa and Sycoscapter sp. A (adapted from Dixon et al. 2001), and white hatching denotes the range of P. imperialis sp. 1. Letters in the inset represent sampled sites: (A) Sydney, (B) Newcastle, (C) Port Macquarie, (D) Byron Bay, and (E) Atherton. Circles represent the location of sampled trees. N (trees) and N (wasps) are detailed in Table 1.

Within sites, distance between trees typically ranged from 10 m to 30 km, but some trees in the Port Macquarie site were separated by up to 120 km. Near-ripe figs, from which wasps were about to emerge, were collected from naturally-occurring and planted trees, and placed into individual plastic specimen pots. Emergent wasps were preserved in 99% ethanol prior to DNA extraction. Since most figs are entered by only one-to-three pollinator foundresses, the likelihood that two wasps emerging from the same fig are related is high. To maximise the independence of samples, no more than one wasp per species from any given fig was used for molecular analyses, and samples were collected from multiple trees within each site (Table 3.1). In some instances, we collected foundress pollinators from receptive figs when ripe figs were not available. Receptive figs were sliced open and live female pollinators were transferred to ethanol for preservation.

Table 3.1: Pollinator and parasitoid samples sizes used for molecular analyses, and number of trees from which samples were collected.

Pollinator Parasitoid Population No. No. syconia per No. No. syconia per N trees tree (range) N trees tree (range) Sydney 30 15 1 – 4 30 15 1 – 3 Newcastle 30 14 1 – 5 30 17 1 – 3 Port Macquarie 30 13 1 – 5 30 13 1 – 4 Byron Bay 11 4 2 – 5 17 5 1 – 5 Atherton 35 8 1 – 9 17 6 1 – 4

3.3.3 Molecular protocols

DNA was extracted from whole female wasps using a Chelex method (West et al. 1998) in a volume of 100 µL. All samples were sequenced for the cytb gene using CB1/CB2 primers, following the PCR protocol outlined by Jermiin & Crozier (1994) with an annealing temperature of 45°C. Pollinator samples positively identified as P. imperialis sp. 1 (N = 136) based on cytb were used for subsequent analyses. Positive identification of parasitoids was based on cytb sequences and/or measurement of ovipositor sheath length relative to the length of the hind tibia (Segar et al. 2014).

We genotyped each pollinator at nine microsatellite loci shown to be polymorphic in P. imperialis sp. 1 (Sutton et al. 2015). Multiplex PCR amplification was performed using the Platinum® Multiplex PCR Master Mix (Life Technologies). Loci were pooled into three reactions in the following combinations: (1) Pli3, Pli6 and Pli10; (2) Pli4, Pli5 and Pli8; (3) Pli2, Pli7 and Pli11. Parasitoid samples were initially genotyped at nine microsatellite loci (Bouteiller-Reuter et al. 2009). However, we had difficulty obtaining consistent amplification for three of these, and they were subsequently omitted from further analyses. Each of the six successfully amplified loci were amplified independently and subsequently combined into two pooled samples for genotyping; (1) JMC5, JMC71 and JMC85; (2) JMC48, JMC89 and JMC95. Genotyping was performed on a 3500 Genetic Analyzer (Applied Biosystems) and electropherograms were visualised and scored using GeneMapper v4.1 (Applied Biosystems).

3.3.4 Genetic variability

We used GENEPOP v4.2 (Raymond & Rousset 1995; Rousset 2008) to test for deviation from Hardy-Weinberg equilibrium (HWE) and test for linkage disequilibrium (LD) for each site to ensure all loci were inherited independently. Following LD tests, we used the False Discovery Rate (FDR) method (Benjamini &

Hochberg 1995) to correct for multiple testing. We calculated observed (HO) and expected (HE) heterozygosity at each locus for each site and across all loci for each site. We further calculated allelic richness (NA) and private allelic richness (NP) with rarefaction in ADZE v1.0 (Szpiech et al. 2008) to account for differences in sample size between sites (Table 3.1).

3.3.5 Genetic differentiation

Genetic differentiation between sites was estimated by calculating pairwise FST

(Weir & Cockerham 1984) and RST (Slatkin 1995) values in GENALEX v6.501 (Peakall & Smouse 2006, 2012). Under the stepwise mutation model, microsatellite alleles evolve in a stepwise fashion by the addition or subtraction of repeat units. RST

takes into account the number of repeat units in estimating the extent of genetic differentiation between populations, and is more efficient than FST at identifying distinct gene pools

We further tested whether RST explained our data better than FST using allele size permutation tests (Hardy et al. 2003) in SPAGEDI v1.4 (Hardy & Vekemans 2002)

(20,000 permutations). Hardy et al. (2003) noted that significant tests imply that RST outperforms FST, and result when the mutation rate for a stepwise mutation process is ≥ 1/t (t = number of generations of drift). Pairwise tests of genotypic differentiation (G-tests; Goudet et al. 1996) between sites were performed in GENEPOP v4.2 (Raymond & Rousset 1995; Rousset 2008) and corrected for multiple testing using the FDR method.

3.3.6 Population genetic structure

We used the Bayesian clustering algorithm implemented in TESS v2.3.1 (Chen et al. 2007) to identify barriers to gene flow between sites. Although STRUCTURE (Pritchard et al. 2000) has been the dominant software program for implementing Bayesian models for population genetics (François & Durand 2010), a primary assumption of these models is that populations are in HWE (Kalinowski 2011). Fig wasps reproduce under local mate competition (Hamilton 1979; Cook & West 2005) with frequent inbreeding and populations are therefore expected to deviate from HWE. Furthermore, TESS, which does not assume HWE, has been shown to be more sensitive to geographic clines when genetic differentiation is low (François & Durand 2010).

TESS links individuals by generating a neighbourhood system by Voronoi tessellation, which can be modified by the user to reflect suspected barriers to gene flow. Our study sites follow the east coast of Australia and represent a roughly linear series from north to south (Figure 3.1). To represent biologically likely patterns of dispersal and gene flow, the neighbourhood system for both pollinators and parasitoids was modified so that each individual was linked to all other individuals from within the same site, as well as to every individual in the site(s) immediately adjacent. Links connecting pollinator samples from the Atherton site to individuals in

any other site were removed to represent the disjunct nature (separated by 1,500 km) of the two regions.

We conducted 100 iterations consisting of 60,000 sweeps of MCMC with a burn-in period of 10,000 sweeps. Convergence of the Bayesian algorithms within the burn-in period was confirmed by viewing graphical outputs of log-likelihood versus iteration number for twenty runs selected at random. We used an admixture model (CAR model; Durand et al. 2009) and compared a maximal number of clusters (Kmax) ranging from 2 to 8. Following this, for each value of Kmax we plotted the deviance information criterion (DIC) for the 20% runs with the lowest DIC value against Kmax, and estimated the optimal Kmax value as the point at which the DIC curve stabilised or inflected upwards. We then averaged the estimated admixture coefficients for the optimal Kmax using the Greedy algorithm in CLUMPP v1.1.2 (Jakobsson & Rosenberg 2007) and visualised the data using DISTRUCT v1.1 (Rosenberg 2004). If the DIC curve did not stabilise or inflect, we averaged admixture coefficients for the 20% runs with the lowest DIC for each Kmax and plotted the results independently for each Kmax value. Following the TESS manual, the optimal Kmax was determined by viewing barplots for increasing values of Kmax, beginning at Kmax = 2, until an increase in Kmax did not result in a clear additional cluster. The highest value of Kmax to yield a clear additional cluster was considered optimal.

3.3.7 Isolation by distance

Population-based measures of isolation by distance (IBD) require dense sampling from geographically well-defined populations. Monoecious fig populations tend to have very low population densities and are thus distributed over large geographic areas. Asynchronous fruiting and low fruiting densities within trees often necessitate long-distance dispersal by pollinating wasps. Consequently, few wasps were collected from individual trees, and trees sampled within each site were often sparsely distributed. In this case, as in the pollinator study of Kobmoo et al. (2010), measures of distance between individuals (not sites) are more pertinent. This approach is most appropriate when samples have patchy geographic distributions and there is no a priori knowledge of the scale of genetic structuring (Heywood 1991).

To investigate the strength of IBD processes, we conducted regressions on pairwise kinship coefficients (F) (Hardy & Vekemans 1999), as calculated in Loiselle et al. (1995), against both logarithmic and linear geographic distance. Analyses were performed for each species within each site and across their wider ranges. As the pollinator is found in two disjunct regions, range-wide IBD regressions were performed only on samples from sites in the southern region. For parasitoid samples, an initial range-wide IBD regression was performed, followed by a regression for sites in the southern region only for a matched comparison with pollinators. Mantel tests with 10,000 permutations were conducted to test the significance of all correlations. All analyses were performed in SPAGEDI v1.4.

3.3.8 Power analysis

The statistical power of the microsatellite loci was assessed using POWSIM v4.1 (Ryman & Palm 2006). This software allows the user to test the efficiency of loci at detecting genetic differentiation at a range of predefined FST values. We ran simulations for FST values ranging from 0.001 to 0.025 (2000 replicates each) with an effective population size of 4000, and the empirical allele frequencies and sample sizes obtained from our real data. The number of generations of drift was varied according to the target FST, and the null hypothesis of genetic homogeneity was tested using Fisher’s exact test and a chi-squared test.

3.4 Results

3.4.1 Genetic variability

Tests for LD showed that three locus pairs (Pli3/Pli8, Pli4/Pli10 and Pli5/Pli8) were not inherited independently for pollinators from the Byron Bay site, which was the site with the lowest number of both wasps and trees sampled (Table 3.1). All other locus pairs across sites, for both pollinators and parasitoids, were found to be inherited independently (P > 0.02) following correction using the FDR method. As expected, there were many cases where loci deviated from HWE for pollinators; in fact, each locus deviated from HWE in at least two sites (Table 3.2). Pollinators from 48

the Byron Bay site were monomorphic at one locus, and Atherton pollinators were monomorphic at three loci. There were far fewer instances of deviation from HWE for parasitoid loci (Table 3.2).

When tested across all (pooled) loci, pollinators and parasitoids deviated from HWE at all sites, although the degree of statistical significance was consistently higher for pollinators (Table 3.3). Allelic richness (NA) in pollinators was substantially lower for the Atherton site compared to other sites, but was consistent across all sites for parasitoids. Within each species, the number of private alleles per locus (NP) was consistent across sites. For pollinators, the inbreeding coefficient (FIS) ranged from 0.27 to 0.53 and was largest in the Byron Bay and Atherton sites (Table 3.3). For parasitoids, FIS ranged from 0.08 to 0.25, and was significantly lower than the corresponding values for pollinators (t = 3.153, DF = 4.549, P = 0.028).

3.4.2 Genetic differentiation

For pollinators, pairwise tests of genotypic differentiation were significant for all site comparisons except for Sydney/Newcastle (Table 3.4). Pairwise FST was low between sites in the southern region and high between sites in different regions. RST was generally lower than FST, with three exceptions; Sydney/Newcastle,

Sydney/Atherton and Newcastle/Atherton. RST was also larger for comparisons involving Atherton, except for an unusually low RST of 0.113 for the Port Macquarie/Atherton comparison.

For parasitoids, only three site comparisons were significant: Newcastle/Byron Bay,

Port Macquarie/Byron Bay, and Port Macquarie/Atherton (Table 3.4). Pairwise FST was lower in parasitoids than in pollinators for all comparisons, and there was no correlation between pollinator and parasitoid FST among the southern sites (Mantel test: P = 0.17). RST was weak for all pairwise comparisons, and was not correlated with pollinator RST (Mantel test: P = 0.42).

2 FST and RST were significantly correlated for both pollinators (R = 0.66, F1,8 = 18.33, 2 P = 0.003) and parasitoids (R = 0.33, F1,8 = 5.42, P = 0.048). Allele size permutation

tests were non-significant (pollinators: P = 0.38; parasitoids: P = 0.95), indicating that F-statistics perform better on our data than R-statistics.

3.4.3 Population genetic structure

The Bayesian clustering analysis indicated strong regional structuring in the pollinator dataset (Figure 3.2). The optimal number of clusters was unclear from the DIC curve (data not shown), and was therefore determined by visualising barplots for

Kmax values ranging from 2 to 8. There was unambiguous evidence for three genetic clusters and there were no clear additional clusters for Kmax > 3 (Appendix A), so the optimal number of clusters was determined to be three. Pollinator samples from Atherton comprised a single genetic cluster, and individuals from the southern region were assigned to one of two genetic clusters; one which dominates the most southerly sites and another (much smaller) cluster that appears clinally in the two northernmost sites of the southern region (Port Macquarie and Byron Bay). The optimal Kmax of two for the parasitoid dataset was determined by visualising barplots for Kmax values from 2 to 8. There was no genetic structure evident from the parasitoid data (Figure 3.2). It is important to note that the methods implemented in

TESS do not allow for the determination of Kmax = 1.

Table 3.2: Genetic characteristics of microsatellite loci.

Sydney Newcastle Port Macquarie Byron Bay Atherton

HO/HE FIS HO/HE FIS HO/HE FIS HO/HE FIS HO/HE FIS Pli2 0.37/0.55*** 0.34 0.33/0.57** 0.42 0.23/0.35 0.34 monomorphic - monomorphic - Pli3 0.40/0.67** 0.40 0.53/0.65 0.18 0.57/0.70 0.20 0.09/0.78*** 0.89 0.03/0.03 0.00

Pli4 0.43/0.56 0.22 0.40/0.63* 0.37 0.53/0.60 0.11 0.18/0.64*** 0.73 0.40/0.67** 0.41

Pli5 0.57/0.87*** 0.35 0.70/0.89* 0.21 0.60/0.88** 0.32 0.36/0.84** 0.58 0.31/0.67*** 0.53 Pli6 0.40/0.81*** 0.51 0.47/0.68** 0.32 0.43/0.61 0.20 0.45/0.82* 0.89 monomorphic -

Pollinator Pli7 0.17/0.59*** 0.72 0.10/0.35*** 0.71 0.33/0.58** 0.38 0.36/0.33 -0.11 0.26/0.47* 0.45 Pli8 0.60/0.91*** 0.34 0.73/0.92** 0.21 0.70/0.92* 0.24 0.55/0.87* 0.39 0.43/0.88*** 0.52 Pli10 0.47/0.79*** 0.42 0.60/0.64 0.06 0.60/0.83* 0.28 0.36/0.82** 0.57 monomorphic - Pli11 0.40/0.62* 0.36 0.30/0.38 0.21 0.27/0.37 0.28 0.45/0.65** 0.32 0.11/0.52*** 0.78 JMC5 0.73/0.84 0.13 0.80/0.79 -0.02 0.90/0.85 -0.06 0.71/0.87* 0.19 0.65/0.84** 0.23

JMC71 0.17/0.16 -0.04 0.23/0.26 0.12 0.07/0.19** 0.65 0.24/0.21 -0.10 0.18/0.17 -0.07 JMC85 0.57/0.72** 0.21 0.67/0.77 0.13 0.70/0.73 0.04 0.71/0.78 0.09 0.59/0.73 0.20

JMC48 0.27/0.24 -0.12 0.13/0.13 -0.04 0.17/0.16 -0.06 0.06/0.06 0.00 0.12/0.12 -0.02 Parasitoid JMC89 0.43/0.47 0.07 0.40/0.44 0.09 0.37/0.45** 0.19 0.29/0.36 0.18 0.29/0.62** 0.53 JMC95 0.13/0.24** 0.46 0.13/0.25** 0.46 monomorphic - 0.18/0.17 -0.04 0.12/0.11 -0.03

Observed (HO) and expected (HE) heterozygosity with significance for deviation from HWE (*P < 0.05,**P < 0.01, ***P < 0.001) after correction for multiple testing, and inbreeding coefficient (FIS).

Table 3.3: Within-site global genetic variability.

Pollinator Parasitoid

HO/HE FIS NA NP HO/HE FIS NA NP Sydney 0.42/0.70*** 0.41 6.55 ± 0.72 2.12 ± 0.63 0.38/0.44*** 0.14 6.03 ± 1.68 1.66 ± 0.69 Newcastle 0.46/0.62*** 0.27 6.05 ± 0.85 1.97 ± 0.73 0.39/0.43** 0.10 6.19 ± 1.77 1.40 ± 0.75 Port Macquarie 0.48/0.63*** 0.27 6.48 ± 0.87 2.51 ± 0.68 0.37/0.39** 0.08 5.32 ± 1.80 1.10 ± 0.57 Byron Bay 0.31/0.61*** 0.52 6.33 ± 1.05 2.55 ± 0.55 0.36/0.40* 0.11 5.83 ± 1.96 1.20 ± 0.87 Atherton 0.17/0.35*** 0.53 3.54 ± 0.94 2.05 ± 0.62 0.32/0.42** 0.25 6.17 ± 1.96 1.10 ± 0.67

Observed (HO) and expected (HE) heterozygosity with significance for deviation from HWE (*P < 0.05,**P < 0.01, ***P < 0.001) after correction for multiple testing, inbreeding coefficient (FIS), allelic richness with rarefaction (NA) (± standard error), and private allelic richness (NP) (± standard error) across all loci

Table 3.4: Genetic differentiation between sites.

Pollinator Sydney Newcastle Port Macquarie Byron Bay Atherton Sydney - NS * ** ** Newcastle 0.012/0.065 - * ** ** Port Macquarie 0.023/-0.004 0.023/-0.002 - ** ** Byron Bay 0.069/0.000 0.122/0.065 0.048/-0.015 - ** Atherton 0.416/0.564 0.458/0.500 0.465/0.113 0.501/0.434 -

Parasitoid Sydney Newcastle Port Macquarie Byron Bay Atherton Sydney - NS NS NS NS Newcastle 0.016/0.020 - NS * NS Port Macquarie 0.006/0.019 0.004/-0.013 - ** ** Byron Bay 0.009/0.007 0.026/0.037 0.025/0.022 - NS Atherton 0.013/0.001 0.038/0.022 0.026/0.010 0.006/-0.034 -

Population pairwise FST/RST are given in the lower diagonal, and significance for G-tests of genotypic differentiation is given in the upper diagonal (*P < 0.05, **P < 0.001, NS: non-significant).

3.4.4 Isolation by distance

Regression analyses did not yield any significant negative correlation between the kinship coefficient (F) and geographic distance within any site for pollinators or parasitoids (Appendix B). For pollinators, broad-scale analysis across all sites in the southern region produced significant slopes for regressions of the kinship coefficient against logarithmic (P < 0.001) (Figure 3.3a) and linear geographic distance (P < 0.001) (Figure 3.3c). Regression analysis also produced a significant slope for parasitoids across all sites for logarithmic (P < 0.01) (Figure 3.3b) and linear geographic distance (P = 0.026) (Figure 3.3d). IBD regressions for parasitoids from sites in the southern region alone did not produce significant slopes.

Figure 3.2: Genetic cluster assignments of P. imperialis sp. 1 (above) and Sycoscapter sp. A (below) individuals assigned by TESS. Sites are arranged from south (left) to north (right). Each vertical bar represents an individual and the colours represent the probability of assignment to each genetic cluster.

3.4.5 Power analysis

The power analysis indicated that the pollinator microsatellite loci could detect a true

FST value of 0.01 with a probability of 99.80% (Appendix C). In addition, parasitoid microsatellite loci detected a true FST value of 0.013 with a probability of 95.70%.

Global FST in our datasets were 0.036 for pollinators (southern region only) and 0.013 for parasitoids.

3.5 Discussion

3.5.1 Overall patterns of population structure

Bayesian clustering analyses support our initial predictions about population structure, revealing limited regional gene flow for pollinators, but high regional connectivity in parasitoids. The assignment of some pollinators in the southern region to a third genetic cluster could indicate clinal variation or intraspecific hybridisation; however, there is no clear genetic evidence for hybridisation among sympatric P. imperialis species, despite frequent crossing opportunities provided by multiple species often occurring within the same fig (T.L. Sutton, unpublished data). A more likely explanation is the small sample size (and low heterozygosity) and the geographic location of samples from Byron Bay on the edge of the southern pollinator population. In further exploratory analyses (data not shown), we found that the probability of assignment of these genotypes to the third genetic cluster decreases when they are provided with geographic coordinates for a more centrally-located site.

Figure 3.3: Isolation by distance. Regression of kinship coefficient (F) against the log geographic distance for (a) pollinator (southern sites only) and (b) parasitoid, and against linear geographic distance for (c) pollinator and (d) parasitoid.

Kobmoo et al. (2010) revealed low differentiation between sites for Ceratosolen fusciceps, the pollinator of Ficus racemosa, in southeast Asia, but high differentiation between disjunct populations in southeast Asia and Australia. Our results emulate this over a shorter geographic distance (~2,000 km versus ~6,000 km), but over a stronger climatic gradient, since the range is essentially a latitudinal transect. Kobmoo et al. (2010) further found that the smaller, more isolated Australian C. fusciceps population was less genetically diverse than larger Asian populations. Consistent with this, we found that genetic diversity was lower in the smaller northern P. imperialis sp. 1 population, as evidenced by reduced HE and NA. Allopatric fig wasp populations that are genetically differentiated could represent incipient species (Kobmoo et al. 2010; Moe & Weiblen 2010) and this warrants further morphological and ecological investigation in P. imperialis sp. 1.

In contrast to the pollinator case, our data suggest the presence of a single Sycoscapter sp. A population throughout eastern Australia. This is supported by mtDNA data from throughout its entire range, including localities where P. imperialis sp. 1 is not found (Darwell 2012), suggesting that it likely also parasitises other species of the P. imperialis complex. Fewer parasitoid than pollinator loci deviated from HWE, which may reflect less sib-mating in parasitoids. Female parasitoids do not need to enter figs to lay eggs, and so more mothers may lay eggs in a given fig. Consequently, while mating still occurs within the natal fig, we expect it to occur less often between siblings. The generally lower inbreeding coefficients for parasitoids relative to pollinators (Table 3.3) are also consistent with this explanation. Indeed, data suggest that the number of Sycoscapter foundresses typically ranges from one to five (Cook et al. in preparation), while the number of P. imperialis foundresses ranges from one to three (James Cook, unpublished data).

The lack of correlation between pollinator and parasitoid pairwise FST, along with a non-significant IBD regression for parasitoids in the southern region, indicate that parasitoids may disperse farther than pollinators. This could reflect the fact that parasitoids do not enter figs and can thus disperse both before and after laying eggs in a given fig, while pollinators rarely exit the first fig they enter. Parasitoids also live longer, with most pollinators surviving 1 – 2 days (< 3 days; Jevanandam et al. 2013), while some NPFWs can live for a week or more without access to food (T.L.

Sutton, personal observation). Lifespan of NPFWs could also be extended if they feed as adults (Lysyk 1998, 2001). The behaviours of adult NPFWs are poorly- understood, and deserve further attention as adult feeding may reduce starvation and increase desiccation resistance, resulting in increased longevity and increasing the potential for long-distance dispersal.

While the current study is the first on the population genetic structure of a fig wasp parasitoid, there is genetic evidence for gene flow across substantial distances in other parasitoids. For example, Grillenberger et al. (2009) found non-significant genetic differentiation between Nasonia vitripennis (Pteromalidae) populations separated by 100 km, as well as a pattern of isolation by distance between populations up to 3,000 km apart. Although little is known about the mechanics of parasitoid dispersal, the patterns reported by Grillenberger et al. (2009) are broadly similar to those found in the current study, and suggest that some pteromalid parasitoids are capable of dispersal over great distances.

Overall, our data are consistent with previous studies suggesting that pollinators of monoecious figs disperse over long distances (Ahmed et al. 2009), thus forming large populations over expansive geographic areas (Kobmoo et al. 2010). In contrast, pollinators of dioecious figs generally disperse more locally (Harrison 2003; Harrison & Rasplus 2006) and may form relatively more structured populations (e.g. Liu et al. 2013; Liu et al. 2015b). A major finding of the present study is that parasitoids in this system disperse as far, if not farther, than pollinators. Harrison (2003) showed that (1) pollinators of monoecious figs dispersed at higher altitudes than pollinators of dioecious figs, and (2) NPFWs and pollinators of a dioecious fig species dispersed at the same height as each other, but much closer to the ground than pollinators of monoecious figs. Unfortunately, data on flight heights of parasitoids from monoecious figs are limiting. Many dioecious figs occur as understorey shrubs, and in forest gaps/edges, with high population densities and flowering frequencies (Harrison 2003). Long-distance dispersal may therefore not be required, and active pollinator dispersal beneath the tree canopy, where wind speed is slow, is suspected to be the rule (Harrison & Rasplus 2006; Alvarez et al. 2010). Monoecious figs, on the other hand, typically have relatively low flowering densities

(Janzen 1979), and so pollinators must usually disperse long distances to locate receptive figs.

3.5.2 Incongruent structure of host and parasitoid populations

Parasitoids regulate many insect herbivore species, and the relative structure of host and parasitoid populations can facilitate or mitigate adaptive responses of each species to the other (Thompson 1994). We report key differences in the population structure of a fig-pollinating wasp and its parasitoid. In particular, the pollinator comprises two genetically disparate allopatric populations, while the parasitoid appears to have a large continuous population over the 2000 km sampled range.

The pollinator situation is likely a result of processes undergone during the historical range expansion of F. rubiginosa. Darwell et al. (2014) hypothesised that diversification in this pollinator complex may be linked with adaptation to local climate. A similar line of reasoning was utilised to explain the abundance and distribution patterns in African fig wasps associated with F. sycomorus (Warren et al. 2010). In the latter case, physiological and behavioural adaptations that enhance resilience to heat shock and desiccation may explain the wider distribution of one of the two Ceratosolen fig wasp species studied.

Variation between interacting species in spatial gene flow can have significant evolutionary consequences. For example, restricted gene flow between P. imperialis sp. 1 populations could hypothetically facilitate a local adaptive response to parasitism in one (or both) of the two populations, which could have major implications for the mutualism in that region. Parasitoids are thought to contribute to stability of the fig – pollinator mutualism by reducing seed parasitism (Dunn et al. 2008a), so an increase in parasitism resistance might reduce fig reproductive output. However, the outcomes of genetic divergence between P. imperialis sp. 1 populations remain to be explored, and future studies should aim to understand ecological differentiation between these populations and how this affects the outcomes of this host-parasitoid interaction in order to envisage the potential evolutionary trajectory of the mutualism across its wider range.

Chapter 4

Quantifying gene flow and hybridisation among sympatric cryptic fig wasp species

Sutton, TL, DeGabriel, JL, Riegler, M and Cook, JM (in revision). Barriers from barcodes: how reliable are DNA barcodes in fig wasps? Heredity.

4.1 Abstract

DNA barcoding for species delimitation is a contentious endeavour. Mitochondrial DNA (mtDNA) barcoding has been utilised extensively for the discovery of cryptic ‘species’; but whether this diversity represents actual reproductive isolation or historical population structure often remains unexplored. We adopted a population genetics approach to examine gene flow between three sympatric cryptic ‘species’ of the fig-pollinating wasp Pleistodontes imperialis, which pollinate Port Jackson fig (Ficus rubiginosa) in northeastern Australia. Our data indicate strong reproductive isolation of the three lineages, despite ample opportunity for heterospecific mating, and the identification of seven potential F1 hybrids. We show that moderate mitochondrial differentiation is likely a reliable indicator of reproductive isolation in fig wasps. The discovery of cryptic diversity is becoming increasingly common, and prompts the question of whether sympatric species have diversified into subtly different niches, or pose a challenge to ecological theory due to their co-occupancy of the same niche.

4.2 Introduction

A common pattern in nature is the geographic replacement of closely related, but distinct, species (Wallace 1858; Darwin 1859). This helped inspire the development of modern evolutionary theory (Wallace 1858; Darwin 1859) and entrench widespread acceptance of allopatric speciation (Mayr 1963). Geographic isolation allows populations to diverge genetically as they respond to differential local ecological selection pressures (Schluter 2001), resulting in speciation. However, speciation can also arise in sympatry (Dieckmann & Doebeli 1999), and insects are important models due to their high host plant fidelity and tendency to form genetically differentiated ‘host races’ following shifts to a novel host plant (Berlocher & Feder 2002).

While geographic isolation can lead to speciation, it doesn’t always, and it may take a long time. Many species exhibit substantial genetic variation across their ranges and/or show signatures of past demographic events (Avise et al. 1987). In particular, many animals show patterns of deep mitochondrial DNA (mtDNA) variation that stem from ongoing dispersal barriers or past biogeographic events (e.g. ice age refugia) (e.g. (Schneider & Moritz 1999)). Secondary contact of such populations (e.g. with range expansion after climate change), can lead to sympatry of distinct mtDNA lineages with no barrier to interbreeding; i.e. they are still a single species (Hewitt 2004; Nicholls et al. 2010b).

Patterns of animal mtDNA variation have become very important with the rise of DNA barcoding (Hebert et al. 2003a), particularly for insects and other invertebrates, because many taxa combine high species diversity with limited alpha taxonomy. Explorations of insect diversity often use mtDNA to delimit molecular operational taxonomic units (MOTUs; (Blaxter 2004)), but rarely test if these MOTUs behave as reproductively isolated species. This issue becomes even more significant with the common discovery of cryptic ‘species’ (e.g. (Witt et al. 2006; Hubert et al. 2012; Saitoh et al. 2015)) that cannot be distinguished morphologically, but resolve into multiple MOTUs. Are these cryptic genetic lineages ‘good’ (reproductively isolated) species, or do they simply reflect current or historical population structure without speciation (Rubinoff et al. 2006)? Giska et al. (2015) provided an important case study, revealing rampant nuclear gene flow between two lineages of the earthworm

Lumbricus rubellus, despite deep mitochondrial divergence. In fact, the clustering of nuclear (microsatellite) genetic variation was predicted far better by geographic location than by mitochondrial lineage. How often are coexisting MOTUs that lack obvious morphological differentiation really different species? This question has many ramifications for our understanding of biodiversity.

If coexisting cryptic lineages are indeed distinct species, then this raises a fundamental ecological problem. Conventional ecological theory predicts that cryptic species cannot coexist in the same niche, and that all but one species will become extinct over time. Consequently, if MOTU-delimited cryptic species coexist at many sites over a wide geographic range, we should test if they (i) occupy subtly different ecological niches or (ii) actually readily exchange genes. The use of population genetic tools and Bayesian statistics to assess hybridisation between taxa (e.g. Bastos-Silveira et al. 2012) will often be easier that disentangling subtle differences in the ecology of cryptic species (Randi 2008), so the latter tests provide a logical starting point.

Figs (Ficus, Moraceae) and their pollinator wasps (Hymenoptera: Agaonidae) share an obligate reproductive mutualism renowned for high levels of partner specificity (Wiebes 1979). Cryptic pollinator ‘species’ are now known to exist on the same host fig both in sympatry and allopatry (Molbo et al. 2003; Haine et al. 2006; Sun et al. 2011; Chen et al. 2012; Darwell et al. 2014), although the loss of pollination behaviour in some pollinator lineages (Jandér & Herre 2010) suggests that selection can drive ecological divergence of sympatric species. Despite this, studies of cryptic fig wasp (and other insect) diversity rarely venture beyond DNA barcodes, and information on the strength and validity of inferred species boundaries is therefore lacking.

An extreme deviation from reciprocal partner specificity occurs with the Port Jackson fig (Ficus rubiginosa Desf. ex Vent.), which is pollinated by five cryptic wasp ‘species’ of the Pleistodontes imperialis Saunders complex along the Australian east coast (Haine et al. 2006; Darwell et al. 2014). Following expert taxonomic revision in 2002, these were considered a single species (Lopez- Vaamonde et al. 2002), but resolve into distinct MOTUs (Haine et al. 2006). For

three of these, no morphological differences are known, while the other two differ only slightly in colour or head length (J.-Y. Rasplus, personal communication).

Hybrid zones can represent areas of secondary contact following range expansion, and we identified and explored a potential hybrid zone around Townsville in northeastern Queensland, where three P. imperialis species ('species 2', 'species 3' and 'species 4' sensu Haine et al. 2006) coexist in sympatry. We used genetic markers (nuclear DNA sequences, mtDNA sequences and microsatellites) to explore patterns of gene flow between these three MOTUs and assess their evolutionary status. Specifically, we assessed (1) the opportunity for hybridisation; i.e. the frequency of figs (mating patches) containing offspring from different MOTUs, (2) nuclear microsatellite differentiation between these mtDNA lineages, (3) the prevalence of hybridisation, and (4) genetic signals of population expansion.

4.3 Materials and methods

4.3.1 Study system

Receptive figs attract pollen-bearing female fig wasps through the emission of species-specific volatiles. The female wasp (foundress) enters the receptive fig through a narrow opening (ostiole), which closes at the end of the receptive phase. Inside, the foundress pollinates the fig ovules, depositing a single egg into some of these, and dies within 24-48 hours. Egg deposition initiates gall formation, within which the developing larva feeds on the fig seed. Offspring develop over a period of about six-to-eight weeks (Harrison 2005; Xiao et al. 2008). Mature males exit their galls to locate galls containing females, through which they chew a hole in order to mate with them. After mating, the wingless males chew holes exit holes through the fig wall for the females, but do not disperse themselves. Females emerge from their galls, collect pollen and exit the fig to disperse in search of receptive figs in which to pollinate and lay eggs (Galil & Eisikowitch 1968).

Three P. imperialis MOTUs (‘species’ 2, 3 and 4 (Haine et al. 2006; Darwell et al. 2014)) are found in sympatry in parts of northeastern Australia. Females of one lineage (P. imperialis sp. 2) are easily distinguished based on colour (it is yellow,

while species 3 and 4 are black) (Haine et al. 2006; Darwell et al. 2014), but unique morphological identifiers for species 3 and 4 have not yet been found.

4.3.2 Sample collection

Pleistodontes imperialis wasps were collected from eight sites around Townsville in north Queensland (19°15’24” S, 146°49’3” E) between November 2013 and November 2014 (Figure 4.1). Near-ripe figs (N = 611) were collected from naturally- occurring trees and placed into individual plastic specimen pots. Emerged wasps were preserved in 99% ethanol prior to molecular analyses. No more than one wasp per fig was used for molecular analyses as only 1-4 foundresses enter a given fig, so there is a high probability that individuals emerging from the same fig are siblings. Samples from different sites and different time points were pooled for analyses.

Figure 4.1: Location of P. imperialis sample collection sites. The square on the inset denotes the location of Townsville in northeastern Australia. Letters on the main map denote sample collection sites (A = Magnetic Island, B = Cape Pallarenda, C = The Strand, D = Castle Hill, E = Hervey Range, F = Mt Stuart, G = Alligator Creek, H = Mt View Road).

4.3.3 Assessment of the potential for hybridisation

The potential for hybridisation was quantified as the proportion of figs which produced both yellow (‘species’ 2) and black (‘species’ 3 or 4) wasps. This method provides an estimate of the minimum frequency of co-occurrence of multiple species within a fig, and thus represents the minimum frequency of opportunities for heterospecific matings.

4.3.4 DNA extraction and mtDNA analyses

A subset of samples was used for DNA analyses, with sample sizes approximately proportional to the frequency of each MOTU (Table 4.1). DNA was extracted from whole female insects (N = 139) using a Chelex method (West et al. 1998) to a volume of 100 µL. The cytb gene was amplified and sequenced in all samples using CB1/CB2 primers and the protocol described in (Jermiin & Crozier 1994). Although COI is the standard barcoding gene in animals, most barcoding efforts for this species complex have utilised the cytb gene (Haine et al. 2006; Darwell et al. 2014) since COI amplifies inconsistently and is compromised as a marker in some fig wasp species by nuclear pseudogenes that amplify alongside the true COI sequence. DNA sequences were edited by eye in Sequencher v4.10.1 (Gene Codes) and aligned in Clustal X v2.0 (Larkin et al. 2007).

A cytb haplotype phylogeny was constructed in MrBayes v3.2 (Ronquist et al. 2012) to assign individuals to MOTUs according to Darwell et al. (Darwell et al. 2014). Reference sequences for each of the three MOTUs (Haine et al. 2006; Darwell et al. 2014) were included in the phylogeny, along with outgroup sequences for Pleistodontes nigriventris (Lopez-Vaamonde et al. 2001) and Ceratosolen galili (Kerdelhue et al. 1999). jModelTest v2 (Guindon & Gascuel 2003; Darriba et al. 2012) was used to determine that the HKY + G (Hasegawa et al. 1985) model was the most appropriate for our data. We considered the default model parameters sufficient as we were interested in broad scale groupings, rather than fine-scale and detailed phylogenetic relationships. Two independent runs of four Metropolis coupled Monte Carlo Markov chains for 10 million generations were performed with a relative burn-in of 25%, and sampling every 1,000 generations. The two runs were

assumed to have converged if the mean standard deviation of the split frequencies fell below 0.01.

Patterns of historical demography were inferred using Tajima’s D (Tajima 1989) and

Fu’s FS (Fu 1997) as implemented in ARLEQUIN v3.5.1.3 (Excoffier & Lischer 2010). Positive values for these statistics are interpreted as evidence for population bottlenecks or over-dominant selection, while negative values are considered evidence for population expansion or purifying selection (Simonsen et al. 1995). Values not significantly different from zero indicate no change in population size. These analyses were complemented with an analysis of mismatch distribution in performed in ARLEQUIN. Smooth unimodal distributions of mtDNA pairwise genetic differences are expected for populations having undergone recent population expansion (Slatkin & Hudson 1991; Ray et al. 2003). Multimodal distributions are expected for populations that have not experienced recent expansion or those that are genetically structured. Difficulties pertaining to population structure are common in population demographic analyses; for example, Bayesian skyline plots can yield false signals of population contraction when there is genetic structuring in the sampled data (Heller et al. 2013). While population structure can present issues for mismatch analyses, these are generally easily identified by observing the distribution itself.

4.3.5 Microsatellite analyses

Each individual was genotyped at nine microsatellite loci (Sutton et al. 2015). These were amplified in multiplex reactions, as detailed in (Sutton et al. 2016). All genotyping was performed on a 3500 Genetic Analyzer (Applied Biosystems), and electropherograms were visualised and scored using GeneMapper v4.1 (Applied Biosystems). One locus (Pli10) did not amplify in P. imperialis sp. 3 samples, and was therefore excluded from subsequent analyses. The number of alleles and number of private alleles with rarefaction were calculated in ADZE (Szpiech et al. 2008), deviation from Hardy-Weinberg equilibrium was tested in GENEPOP v4.2 (Raymond & Rousset 1995; Rousset 2008). Pairwise tests of genotypic differentiation (Goudet et al. 1996) were performed in GENEPOP and pairwise population FST was calculated in ARLEQUIN.

The Bayesian clustering algorithm implemented in TESS (Chen et al. 2007) was utilised to assess potential gene flow between the cytb-defined MOTUs. The CAR admixture model (Durand et al. 2009) was implemented and 100 iterations were conducted, consisting of 60,000 sweeps of MCMC with a burn-in period of 10,000 sweeps. The maximum number of clusters (Kmax) was restricted to three as our aim was to assess gene flow between the three cytb-defined MOTUs. The estimated admixture coefficients were averaged using the Greedy algorithm in CLUMPP v1.1.2 (Jakobsson & Rosenberg 2007) and the data were subsequently visualised in DISTRUCT v1.1 (Rosenberg 2004).

4.3.6 Nuclear DNA barcoding

For individuals whose highest cluster assignment probability from the TESS analysis was less than 96%, mtDNA species assignment was confirmed by sequencing the nuclear ITS2 region. ITS2 was amplified using ITS2-F (5’- ATTCCCGGACCACGCCTGGCTGA-3’) and ITS2-R (5’- TCCTCCGCTTATTGATATGC-3’) primers (White et al. 1990) in 20 µL reactions containing 1 mM MgCl2, 0.1 mM each dNTP, 0.36 µM each primer and 0.5 units Taq DNA polymerase (Promega). Thermal cycling conditions involved an initial denaturation step of 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, 52°C for 40 s and 72°C for 40s, and a final elongation step of 72°C for 10 min. The phylogenetic methods employed for cytb data were also used for ITS2. Two reference sequences for each of the three P. imperialis species present in Townsville (Darwell et al. 2014) were included in the phylogeny, and Pleistodontes froggatti (Lopez-Vaamonde et al. 2001) was included as an outgroup.

4.4 Results

At least 13.38% of F. rubiginosa fruits contained wasps from multiple MOTUs (N = 611), as evidenced by the presence of yellow and black P. imperialis wasps. One hundred and forty P. imperialis wasps were sequenced for a 372 bp cytb fragment, and each individual was unambiguously assigned to one of the three known P. imperialis MOTUs found in Townsville (Figure 4.2). We identified 52 unique

haplotypes across three P. imperialis MOTUs and nucleotide diversity within lineages ranged from 0.007 to 0.015 (Table 4.1). Intraspecific cytb genetic divergence ranged from 0.0 to 6.5% and interspecific divergence ranged from 8.9 to

16.1%. Tajima’s D was not significant for all three lineages, but Fu’s FS was significantly negative for P. imperialis sp. 2 and 3 (Table 4.2). Mismatch distributions were bimodal for each lineage and thus did not appear to support hypotheses of population expansion (Figure 4.3).

Microsatellite diversity (NA and HE) was greatest in P. imperialis sp. 2 and lowest in P. imperialis sp. 3, and over 50% of alleles were not shared between lineages (Table

4.1). All lineages were significantly differentiated with pairwise FST values ranging from 0.28 to 0.54 (Table 4.3). Bayesian clustering analyses revealed three distinct microsatellite genetic clusters corresponding to cytb MOTU assignments (Figure 4.4). Seven individuals had maximum cluster assignment probabilities lower than 96% (two of which were less than 70%), but sequencing of the ITS2 region (Figure 4.5) validated their initial cytb MOTU assignment. Examination of individual microsatellite genotypes revealed that these individuals were not F1 hybrids. In three individuals, we identified alleles that are rare within their MOTU, but common in at least one other MOTU, suggesting they could be later generation hybrids that have retained a conspecific allele from an ancestral hybrid mating (Table 4.4).

Figure 4.2: Bayesian haplotype tree of P. imperialis haplotypes based on a 372 bp fragment of the mitochondrial cytb gene. Posterior probabilities are indicated. The scale bar represents the number of substitutions per site.

Table 4.1: mtDNA and microsatellite diversity of three sympatric P. imperialis species.

mtDNA Microsatellites P. imperialis species N No. haplotypes Nucleotide diversity NA NP HO/HE 2 79 19 0.007 ± 0.004 10.71 ± 2.38 8.89 ± 2.36 0.47/0.76* 3 39 21 0.010 ± 0.006 4.02 ± 1.17 2.58 ± 1.11 0.18/0.40* 4 22 12 0.015 ± 0.008 6.25 ± 1.26 4.94 ±1.20 0.11/0.57*

N, sample size; NA, number of alleles (with rarefaction); NP, number of private alleles (with rarefaction); HO, observed heterozygosity; HE, expected heterozygosity. Asterisk denotes significant deviation from Hardy-Weinberg equilibrium (P < 0.05).

Table 4.2: Tests of selective neutrality under a population expansion model for three sympatric P. imperialis species based on mtDNA cytb sequences.

P. imperialis species Tajima's D Fu's FS 2 -0.66 -7.15* 3 -0.95 -11.84* 4 -0.90 -1.69 Asterisk denotes significant deviation from Hardy-Weinberg equilibrium (P < 0.05).

Table 4.3: Genetic differentiation among P. imperialis (cytb) species

sp. 2 sp. 3 sp. 4 sp. 2 - * * sp. 3 0.54 - * sp. 4 0.28 0.34 -

Pairwise FST is given in the lower diagonal and statistical significance for tests of genotypic differentiation are given in the upper diagonal (Asterisk denotes P < 0.001).

Figure 4.3: mtDNA mismatch distributions for P. imperialis sp. 2 (a), sp. 3 (b) and sp. 4 (c). Observed distributions are represented by open circles and solid lines. Simulated distributions are represented by black triangles and dotted lines. Parameters for tests of the sudden expansion model are given for each lineage.

Figure 4.4: Genetic cluster assignments of 139 P. imperialis individuals based on nuclear microsatellite genotypes. Each vertical column represents an individual, and the colours indicate its probability of assignment to each of the three genetic clusters. Samples are grouped by mtDNA assignment. Circles denote individuals whose highest cluster assignment probability was less than 0.96.

4.5 Discussion

We used a population genetic approach to assess the potential for hybridisation and gene flow among sympatric P. imperialis mtDNA lineages in northeastern Australia. Our results revealed that more than 13% of figs contained wasps from at least two mtDNA MOTUs. Despite ample opportunity for hybridisation, genetic data show that none of the samples analysed were hybrids. This study supports the reproductive isolation in local sympatry and species status of MOTUs delimited by Darwell et al. [24] using a few specimens from each of many sites over a wide geographic range.

Figure 4.5: Consensus Bayesian topology of P. imperialis inferred from a 338 bp fragment of the ITS2 region. Posterior probabilities are indicated. Pleistodontes froggatti was included as an outgroup. Reference sequences for ingroups are denoted by their accession numbers, followed by the species assignment made by Darwell et al. (2014). The scale bar represents the number of substitutions per site.

Table 4.4: Potential hybrid alleles and their frequency in each P. imperialis MOTU

Frequency in each MOTU Individual MOTU Locus Allele sp. 2 sp. 3 sp. 4 TSY12 sp. 2 Pli3 112 0.0833 0.2564 0.0000 Pli6 253 0.0192 0.2564 0.0000 TSY14 sp. 2 Pli4 271 0.0577 0.8718 0.4318 TNV3947 sp. 4 Pli4 277 0.1603 0.0000 0.0682 Pli5 118 0.6026 1.0000 0.1364 Pli7 198 0.3718 0.2180 0.0455

Bayesian clustering analyses indicated a clear concordance between mtDNA and multilocus microsatellite genotypes, despite some individuals exhibiting signs of genetic admixture. We found no evidence that any of the samples used here were F1 hybrids, and only very weak evidence to suggest that any hybridisation occurs between previously-identified MOTUs. Molbo et al. (2004) identified F1 hybrids of Pegoscapus fig wasps in Panama, but suggested that they were sterile due to the low frequency of hybrids, lack of introgression and the identification of diploid males, which indicates a breakdown in the haplodiploid sex determination system.

Tests of selective neutrality yielded conflicting results. Tajima’s D revealed no evidence for changes in population size in any of the three P. imperialis species, but

Fu’s FS provided evidence for population expansion in species 2 and 3. While both statistics can detect changes in historical demography, Tajima’s D is more efficient at detecting bottlenecks (Simonsen et al. 1995), while Fu’s FS is more sensitive to population expansion (Fu 1997). Bimodal mismatch distributions did not appear to support hypotheses of sudden population expansion. Population structure, such as the cytb sub-clade structure observed in some P. imperialis species, however, can produce bimodal mismatch distributions, making interpretation of these statistics difficult (Wei et al. 2014). Indeed, a number of factors, such as population structure, levels of DNA sequence polymorphism, and sample size, can all potentially confound the interpretation of mismatch analyses (Grant 2015). A deeper investigation of the population genetic structure and demographic history of all P. imperialis species would provide more intriguing insights into the evolution of this system.

There is a growing number of reports of cryptic ‘species’ in the fig – fig wasp mutualism, existing both in sympatry and allopatry (Molbo et al. 2003; Haine et al. 2006; Moe & Weiblen 2010; Lin et al. 2011; Sun et al. 2011). Moe & Weiblen (2010) identified cryptic Ceratosolen fig wasp ‘species’ in association with several widespread Ficus species, suggesting an extensive role for vicariance in the generation of cryptic biodiversity. This notion is supported by population genetic analyses of fig wasps in Asia and Australia (Kobmoo et al. 2010; Sutton et al. 2016). Despite the local coexistence of cryptic fig wasp ‘species’ in at least four genera (Molbo et al. 2003; Molbo et al. 2004; Haine et al. 2006; Lin et al. 2011; Sun et al. 2011; Darwell et al. 2014), true tests of hybridisation between such mtDNA lineages have only been undertaken in one of these. Molbo et al. (2004) used diagnostic microsatellite loci to identify hybrids of cryptic Pegoscapus hoffmeyeri lineages from less than 1% of figs sampled; suggesting that gene flow between lineages is rare.

Our data also highlight the need to address the question of how cryptic species that appear to be ecological equivalents can coexist on the same resources (Zhang et al. 2004). Zhang et al. (2004) proposed a model to address this fundamental ecological issue by building on the idea that stable coexistence is possible if species discriminatively direct appropriate behaviours towards conspecifics and heterospecifics (Chesson 1991). Alternatively, the onset of sympatry may have been too recent for competitive exclusion to cause local extinction of the least ‘fit’ species. Ecological factors, such as preferential oviposition in different subsets of flowers, might also explain the coexistence of seemingly cryptic fig wasp species. Another interesting possibility is that one species performs better during the dry season, while another performs better in the wet season.

We have shown that sympatric P. imperialis lineages do not exchange genes and thus represent ‘good’ species. Future work should aim to investigate potential differences in the intricate ecology and morphology of these species to elucidate potential points of divergence and to test the model of cryptic fig wasp coexistence proposed by Zhang et al. (2004).

Chapter 5

Scaling genetic differentiation of allopatric fig wasp populations

5.1 Abstract

Single locus DNA barcoding for species delimitation can be problematic, and there is overwhelming support for methods that integrate multiple sources of data. In the case of morphologically indistinguishable cryptic species, genetic approaches that utilise multiple markers are considered the most robust tool for species delimitation. However, these often fail to clearly resolve species from recent radiations. The task of species delimitation is made even more difficult when divergent lineages are allopatric, such that species boundaries cannot be reinforced by assortative mating or broken down by hybridisation.

In Australia, Port Jackson fig (Ficus rubiginosa) is pollinated by five fig wasp species of the Pleistodontes imperialis Saunders species complex. One of the pollinator species (P. imperialis sp. 1) comprises two disjunct populations separated by 1,500 km; one in the southern range of the host plant, and one in the northern range. Here, we use a comparative population genetics approach with multiple DNA markers to quantify genetic divergence of allopatric P. imperialis sp. 1 populations with reference to P. imperialis sp. 2, a closely related but taxonomically unambiguous species. Between P. imperialis sp. 1 populations, we observed small, but fixed, differences at a nuclear DNA locus (ITS2), reciprocal monophyly at a mitochondrial gene (cytochrome b) and extreme levels of genetic differentiation estimated from multilocus microsatellite genotypes. We also found strong support for rapid expansion of the geographically more widespread southern population, but not in the smaller, more isolated northern population.

We show that allopatric P. imperialis sp. 1 populations represent incipient species. Our study supports the utility of population genetic approaches in delimiting species boundaries, but highlights some limitations for the genetic delimitation of allopatric species. Most notably, estimates of genetic exchange between geographically isolated species are not possible. Where possible, genetic data should be combined with morphological, behavioural or ecological data when delimiting closely related or cryptic species.

5.2 Introduction

In the last decade, taxonomy has been revolutionised by DNA barcoding – the designation of an individual to species level based on a short mitochondrial DNA (mtDNA) sequence (Hebert et al. 2003a). The advancement of molecular-based species identification tools has uncovered a plethora of diversity that went unseen by traditional taxonomy. Despite early ‘successes’ in the discovery of cryptic diversity (Hebert et al. 2003a; Hebert et al. 2003b; Hebert et al. 2004), the use of DNA barcoding alone for species delimitation and discovery has also received heavy criticism (e.g. Ferguson 2002; Will & Rubinoff 2004a; Brower 2006; Rubinoff et al. 2006). Rubinoff et al. (2006) developed a compelling case against the use of mtDNA barcoding for species delimitation, citing (among others) reduced effective population size of mtDNA, introgression, selective sweeps (particularly those induced by intracellular bacteria, such as Wolbachia), lack of recombination, and the confounding effects of nuclear pseudogenes of mitochondrial origin as potential shortcomings of the barcoding philosophy.

Integrative approaches, which combine multiple types of data (e.g. molecular and morphological), are considered most effective and desirable for species delimitation (Will et al. 2005). However, in the case of morphologically indistinguishable cryptic species, genetic data may be the only viable option. In these cases, especially when taxa are closely related, the use of multiple genetic markers and dense taxon sampling is recommended (Roe & Sperling 2007; Dupuis et al. 2012). Multi-locus approaches are crucial for delimiting recently diverged species (Knowles & Carstens 2007; Dupuis et al. 2012), and tools commonly applied to population genetics studies are proving invaluable in this regard (e.g. Dépraz et al. 2009; Bernasconi et al. 2010; Hausdorf et al. 2011; Duminil et al. 2012). For example, Hausdorf et al. (2011) revealed vastly discordant patterns of population structure inferred from mtDNA and nuclear-encoded microsatellites in Astyanax fishes. Polymorphic microsatellite markers can further enhance our understanding of the structure of cryptic diversity by facilitating the quantification of gene flow between closely related ‘species’ inferred from mtDNA (e.g. Bernasconi et al. 2010).

Understanding how cryptic diversity is structured across landscapes provides valuable insights into the geographic processes involved in speciation (Coyne & Orr

2004). Indeed, Warren et al. (2014) argued that the primary focus of biogeographic studies should be to understand the geography, and not the ecology, of speciation. A geographic understanding of speciation requires geographically well-defined systems – something that is generally lacking from most studies, particularly in the tropics (McBride et al. 2009). A major difficulty for species delimitation is the treatment of genetically divergent allopatric demes (Lukhtanov et al. 2015). While molecular tools can be used to quantify genetic divergence between allopatric populations, whether or not they are reproductively isolated is a question that often cannot be easily addressed.

The monoecious Port Jackson fig, Ficus rubiginosa (Desf. ex Vent.) (Moraceae), is naturally distributed along most of Australia’s east coast, from the tropics of northern Queensland to the temperate regions of southern New South Wales (NSW) (Dixon et al. 2001). Extensive DNA barcoding surveys have revealed that F. rubiginosa is pollinated by five cryptic fig wasp species that form the Pleistodontes imperialis Saunders (Hymenoptera: Agaonidae) species complex, and which have partially overlapping geographic ranges (Figure 1.3; Haine et al. 2006; Darwell et al. 2014) and different Wolbachia infections (Haine et al. 2006). A peculiar feature of this complex is the existence of two disjunct populations of P. imperialis sp. 1 (sensu Haine et al. 2006); one which dominates the southern range of F. rubiginosa, and another found in the Atherton Tablelands in northern Queensland (Haine et al. 2006; Darwell et al. 2014). Although these populations are clearly discernible based on mitochondrial DNA sequences, two species delimitation methods – one distance- based method (jMOTU; Jones et al. 2011) and one phylogeny based method (GMYC; Pons et al. 2006) – fail to recognise the two populations as discrete species (Darwell et al. 2014). A recent study revealed high levels of genetic differentiation between these populations using multilocus microsatellite genotypes (Chapter 3; Sutton et al. 2016), but the extent of this differentiation has not yet been scaled among other species within the P. imperialis species complex.

In the present study, we used multiple genetic markers (nuclear and mitochondrial DNA sequences, and microsatellites) to quantify genetic variability between the two P. imperialis sp. 1 populations. This was done with reference to another closely related, but unambiguously different P. imperialis species (P. imperialis sp. 2 sensu

Haine et al. 2006) to determine if the extent of genetic differentiation is indicative of speciation. We further tested for genetic signals of population expansion using neutrality tests and mismatch distributions. We predicted that levels of genetic variation would be consistent with signals of incipient speciation: highest in the microsatellite markers (as they are the most sensitive to ruptured gene flow), and lowest in nuclear DNA (as it evolves much slower than mitochondrial DNA (Brown et al. 1979)). We expected patterns of expansion in the larger southern P. imperialis sp. 1 population due to its geographic size relative to the smaller northern population (Sutton et al. 2016) and that fig and fig wasp diversity (including species diversity in the P. imperialis complex; Darwell et al. 2014) is concentrated in the tropics, so poleward expansion might be expected.

5.3 Methods

5.3.1 Sample collection

Ficus rubiginosa has a continuous distribution along Australia’s east coast (Dixon et al. 2001). We sampled the southern P. imperialis sp. 1 population from four sites in NSW between March 2012 and August 2014 (Figure 5.1). Over the same time period, we sampled the northern P. imperialis sp. 1 population from throughout its small range in the Atherton Tablelands (Figure 5.1; see Figure 3.1 for more detail). Three cryptic P. imperialis species (1, 3 and 4 sensu Haine et al. 2006; Darwell et al. 2014) occur sympatrically in the Atherton Tablelands (Figure 1.3), which reduced sampling efficiency of P. imperialis sp. 1 in this region. We therefore included a small number of insects collected earlier in 2005 and 2007. The geographic range of Pleistodontes imperialis sp. 2 is restricted to Townsville, north Queensland (Figure 1.3; Darwell et al. 2014). We collected P. imperialis sp. 2 samples from field sites in Townsville between November 2013 and November 2014 (Figure 5.1; see Figure 4.1 for more detail) Although two other P. imperialis species (3 and 4 sensu Haine et al. 2006; Darwell et al. 2014) coexist with P. imperialis sp. 2 in Townsville, P. imperialis sp. 2 is the most common and also the easiest to identify as it lacks pigment and appears yellow, while the others are pigmented and appear dark brown or black.

We collected near-ripe figs and stored them in plastic specimen pots for 48 hours. We transferred emergent wasps to a tube containing absolute ethanol for preservation prior to molecular analyses. Since figs are entered by only a small number of foundresses (usually one or two), the probability of sampling siblings from the same fig is high. Therefore, we used only a single wasp from each fig for DNA analyses in order to maximise independence.

Figure 5.1: Geographic distribution of sampling efforts. The dotted line represents the western boundary of the range of Ficus rubiginosa. (Black circles = southern P. imperialis sp. 1, open circle = northern P. imperialis sp. 1, black square = P. imperialis sp. 2). Each point denoted on the map represents several trees from several sites. Note that none of the ranges of these populations/species overlap.

5.3.2 Molecular methods

We extracted DNA from whole female insects using a Chelex protocol to a volume of 100 µL (West et al. 1998). We amplified and sequenced mitochondrial cytb using the CB1/CB2 primers and protocol described in (Jermiin & Crozier 1994). We amplified and sequenced nuclear ITS2 in a subset of individuals using ITS2-F (5’- ATTCCCGGACCACGCCTGGCTGA-3’) and ITS2-R (5’- TCCTCCGCTTATTGATATGC-3’) primers (White et al. 1990) in 20 µL reactions containing 1 mM MgCl2, 0.1 mM each dNTP, 0.36 µM each primer and 0.5 units Taq DNA polymerase (Promega). Thermal cycling conditions were an initial denaturation step of 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, 52°C for 40 s and 72°C for 40s, and a final elongation step of 72°C for 10 min. We genotyped each individual at nine microsatellite loci (Sutton et al. 2015) in multiplex PCR reactions (Sutton et al. 2016). Genotyping was performed on a 3500 Genetic Analyzer (Applied Biosystems), and genotypes were visualised and scored using GeneMapper v4.1 (Applied Biosystems).

5.3.3 DNA sequence analyses

We aligned and edited DNA sequences by eye in Sequencher v4.10.1 (Gene Codes), and constructed a Bayesian haplotype phylogeny in MrBayes v3.2 (Ronquist et al. 2012) using all unique cytb haplotypes, which were extracted using the GetHaplo function in the SIDIER package (Muñoz-Pajares 2013) in R v3.2.2 (R Core Team 2015). We included reference sequences for both P. imperialis sp. 1 populations (Haine et al. 2006; Darwell et al. 2014), P. imperialis sp. 2 (Darwell et al. 2014), and outgroup sequences for Pleistodontes nigriventris (Lopez-Vaamonde et al. 2001) and Ceratosolen galili (Kerdelhue et al. 1999). We used jModelTest v2 (Guindon & Gascuel 2003; Darriba et al. 2012) to select the most appropriate model of nucleotide substitution for the cytb data. We ran two independent runs of four Metropolis coupled Monte Carlo Markov chains for 10 million generations, sampling every 1,000 generation and with a relative burn-in period of 25%. If the standard deviation of the split frequencies of the two runs was below 0.01 after 10 million generations, the two runs were assumed to have converged.

We tested for genetic signals of population expansion for the two P. imperialis sp. 1 populations using neutrality tests (Tajima's D and Fu's FS; Tajima 1989; Fu 1997) and mismatch distributions using all DNA sequences in ARLEQUIN v3.5.1.3

(Excoffier & Lischer 2010). Negative values for Tajima’s D and Fu’s FS indicate recent population expansion or purifying selection, while positive values indicate population bottlenecks or over-dominant selection. Mismatch distributions are expected to be smooth and unimodal for populations that have recently undergone expansion (Slatkin & Hudson 1991; Ray et al. 2003).

For ITS2 data, we calculated within- and between-population genetic distances in MEGA v6.0 (Tamura et al. 2013). We selected the most appropriate nucleotide substitution model using jModelTest.

5.3.4 Microsatellite analyses

We calculated observed (HO) and expected (HE) heterozygosity and tested for deviation from Hardy-Weinberg equilibrium (HWE) for each locus independently and across pooled loci in GENEPOP v4.2 (Raymond & Rousset 1995; Rousset

2008). Furthermore, we calculated allelic richness (NA) and private allelic richness

(NP) with rarefaction in ADZE (Szpiech et al. 2008) to account for differences in sample size for each population.

We estimated genetic differentiation by computing pairwise FST (Weir & Cockerham

1984) and RST (Slatkin 1995) within the AMOVA (Excoffier et al. 1992) framework in GENALEX v6.501 (Peakall & Smouse 2006, 2012). FST provides an estimate of genetic divergence based on allele identity, while RST takes into account allele size as expressed in number of repeat units. According to the Stepwise Mutation Model, microsatellites evolve by the addition or subtraction of repeat units. RST is therefore expected to be more efficient at identifying distinct gene pools. We also performed exact tests of genotypic differentiation (Goudet et al. 1996) and corrected for multiple testing using the False Discovery Rate (FDR) method (Benjamini & Hochberg 1995). We also conducted a Principal Components Analysis (PCA) in R using the dudi.pca function in the ade4 package (Dray & Dufour 2007) to visualise

microsatellite genotypes in multidimensional space. Prior to the analysis, allele frequencies were transformed using the scaleGen function (method = “mean”).

5.4 Results

5.4.1 DNA sequence analyses

We sequenced a 339 bp cytb fragment of 214 P. imperialis wasps, and a 284 bp fragment in one sample (AT28P), which was excluded from further mtDNA analyses. We identified 76 unique cytb haplotypes and revealed consistent levels of nucleotide diversity across the three populations (Table 5.1). We used the HKY + G (Hasegawa et al. 1985) model to construct the Bayesian haplotype phylogeny, which showed clear separation and reciprocal monophyly of the two P. imperialis sp. 1 populations with 100% posterior probability (Figure 5.2). Tajima’s D and Fu’s FS were significantly negative for both of the P. imperialis sp. 1 populations (Table 5.2), suggesting sudden population expansion. However, mismatch distributions supported population expansion in the southern species 1 population only (Figure 5.3).

We sequenced a fragment (311 to 315 bp) of the ITS2 region in 25 P. imperialis samples (see Table 5.3 for sample sizes per population). Within each population, all sequences were identical. Mean Jukes-Cantor (Jukes & Cantor 1969) distance was zero between the southern and northern species 1 populations, but there was a 4 bp indel in a dinucleotide repeat region that was diagnostic for each of the three populations (Figure 5.4). Genetic distance between species 2 and each of the species 1 populations was 0.01 (Table 5.3).

5.4.2 Microsatellite analyses

We genotyped 215 wasps at nine microsatellite loci. The northern species 1 population was monomorphic at three loci, while species 2 was monomorphic at one locus (Table 5.4). All loci deviated from HWE in the southern species 1 population, while the northern species 1 population (Pli3) and species 2 (Pli10) had only one locus each that did not deviate from HWE. Deviation from HWE is expected in fig

wasps as reproduction occurs within the natal fig, where sib-mating is very common and inbreeding is typically very highTable 5.. Genetic variation (NA, HE) was similar in the southern species 1 population and species 2, but severely reduced in the northern species 1 population (Table 5.5). The number of private alleles (NP) was highest in species 2 and lowest in the northern species 1 population. Across pooled loci, all populations deviated from HWE, and FIS ranged from 0.36 (southern species 1) to 0.53 (northern species 1).

All populations were significantly differentiated (Table 5.6). Pairwise FST was lowest for the southern species 1/species 2 comparison and highest for the northern species

1/species 2 comparison. The pairwise RST value for the two species 1 populations was lower than the corresponding FST value. Furthermore, pairwise RST values for population comparisons involving species 2 were considerably higher than the southern/northern species 1 comparison. The first two axes of the PCA explained a total of 20.16% of the variation (Figure 5.5). PCA scores were plotted on the first two axes and formed three clusters which corresponded to the three populations. Axis 1 explained 12.02% of the variation, which accounted for variation between species 2 and the species 1 populations. The second PCA axis (8.14%) was due to variation between the two species 1 populations.

Figure 5.2: Bayesian haplotype tree of three P. imperialis populations based on a 339 bp fragment of the mitochondrial cytb gene. Posterior probabilities are indicated. The scale bar represents the number of substitutions per site.

Table 5.1: mtDNA (cytb) diversity of P. imperialis sp. 1 and P. imperialis sp. 2 populations

Population N No. haplotypes Nucleotide diversity (± S.E.) Species 1 south 101 46 0.008 ± 0.005 Species 1 north 34 11 0.005 ± 0.004 Species 2 79 19 0.008 ± 0.005

Table 5.2: Tests of selective neutrality under a population expansion model for P. imperialis sp. 1 populations based on cytb sequences.

Population Tajima's D Fu's FS Species 1 south -2.24** -26.62** Species 1 north -2.35** -4.42* Asterisks denote statistical significance (*P < 0.01, **P < 0.001).

Table 5.3: Mean within- and between-population Jukes-Cantor distances based on a 311-315 bp fragment of the ITS2 region. Species 1 south Species 1 north Species 2 Species 1 south (7) 0.00 Species 1 north (8) 0.00 0.00 Species 2 (10) 0.01 0.01 0.00 Sample size of each population is given in parentheses. Mean within-population distance is given on the diagonal and mean between-population distance is given below the diagonal.

Figure 5.3: Mismatch distributions of mtDNA lineages for the (a) southern and (b) northern P. imperialis sp. 1 populations. Observed distributions are represented by open circles and solid lines. Simulated distributions are represented by black triangles and dotted lines. The southern population shows signals of sudden population expansion.

Table 5.4: Genetic characteristics of microsatellite loci. Species 1 south Species 1 north Species 2

Locus HO/HE FIS HO/HE FIS HO/HE FIS Pli2 0.27/0.49*** 0.43 Monomorphic - 0.28/0.75*** 0.63

Pli3 0.46/0.68*** 0.33 0.03/0.03 0.00 0.54/0.75*** 0.27

Pli4 0.43/0.60*** 0.29 0.40/0.67** 0.41 0.59/0.91*** 0.35

Pli5 0.59/0.88*** 0.33 0.31/0.67*** 0.53 0.33/0.56*** 0.41

Pli6 0.44/0.75*** 0.42 Monomorphic - 0.57/0.90*** 0.37

Pli7 0.22/0.55*** 0.60 0.26/0.47* 0.45 0.33/0.56*** 0.42

Pli8 0.66/0.91*** 0.28 0.43/0.88*** 0.52 0.67/0.91*** 0.26

Pli10 0.53/0.77*** 0.31 Monomorphic - 0.04/0.06 0.39

Pli11 0.34/0.49*** 0.32 0.11/0.52*** 0.78 Monomorphic -

Observed (HO) and expected (HE) heterozygosity with significance for deviation from HWE (*P < 0.05,**P < 0.01, ***P < 0.001) after correction for multiple testing, and inbreeding coefficient (FIS)

Table 5.5: Global (across pooled loci) genetic variability.

Population N NA NP HO/HE FIS Species 1 south 101 13.63 ± 2.65 10.97 ± 1.87 0.44/0.68* 0.36 Species 1 north 35 5.89 ± 2.14 3.44 ± 1.13 0.17/0.35* 0.53 Species 2 79 13.15 ± 3.46 11.61 ± 2.98 0.37/0.60* 0.38

Table 5.6: Pairwise population genetic differentiation based on microsatellite data. Species 1 south Species 1 north Species 2 Species 1 south - * * Species 1 north 0.398/0.170 - * Species 2 0.355/0.976 0.495/0.977 -

Population pairwise FST/RST are given in the lower diagonal, and significance for G-tests of genotypic differentiation are given in the upper diagonal (*P < 0.001).

5.5 Discussion

We successfully applied a comparative population genetic approach to scale genetic differentiation of allopatric fig wasp populations relative to a closely related species. As hypothesised, the observed high levels of genetic differentiation based on microsatellite data, coupled with moderate and low levels of differentiation based on mtDNA and nuclear DNA sequences respectively, were consistent with genetic signals of incipient speciation. Thus, the P. imperialis species complex comprises at least six species. Our results highlight the importance of multilocus approaches to

species delimitation, and suggest that species diversity in this complex has previously been underestimated (Darwell et al. 2014).

Fixed indels of dinucleotide repeats at the ITS2 region, reciprocal monophyly at the cytb gene, extremely high microsatellite differentiation, and microsatellite genotype clustering provide strong evidence that allopatric P. imperialis sp. 1 populations are incipient species. Darwell et al. (2014) independently applied three DNA barcoding loci (COI, cytb and ITS2) and two species delimitation methods to resolve P. imperialis into five distinct molecular operational taxonomic units (MOTUs; Blaxter 2004). Use of DNA barcoding tools for species delimitation, however, is known to yield inconsistent results depending on the methods used (Linares et al. 2009; Sauer & Hausdorf 2012; Fujisawa & Barraclough 2013; Hamilton et al. 2014). The task of delimiting recently diverged species is particularly problematic (Hickerson et al. 2006), even when multiple loci are employed (Satler et al. 2013). Satler et al. (2013), for example, applied four species ‘discovery’ methods (i.e. methods that do not require a priori expectations or data partitioning) to a multilocus dataset for the trapdoor spider genus Aliatypus and found support for three, four, five and eighteen species, respectively. Consequently, it may be difficult to have confidence in any one estimate of species diversity.

Figure 5.5: PCA plot of P. imperialis individuals based on multilocus microsatellite genotypes. (Black circles = southern P. imperialis sp. 1, open circles = northern P. imperialis sp. 1, black square = P. imperialis sp. 2). The delimitation of allopatric lineages is particularly difficult as it is not clear whether or not the observed genetic divergence is indicative of reproductive isolation. Not only are measures of behavioural and ecological traits much simpler for sympatric lineages; they also lend themselves to relatively straightforward tests of gene flow (e.g. Donnelly et al. 2013). Gómez et al. (2007) validated the species status of divergent allopatric lineages of the marine bryozoan Celleporella hyaline using laboratory mating trials. Similar tests, however, are problematic in fig wasps because (i) they are extremely difficult to keep in culture, (ii) mating takes place among offspring within the fig, making controlled crosses almost impossible, and (iii) cryptic taxa are often geographically separated (in the case of P. imperialis sp. 1, by approximately 1,500 km), which can pose logistical, and even biosecurity, problems.

Our molecular results show that P. imperialis sp. 1 populations are incipient species; however, it is clear that an exhaustive and integrative revision of the taxonomy of this species complex is required. Expert morphological taxonomic revision in 2002 concluded that P. imperialis comprised a single species across its wide geographic

range (Lopez-Vaamonde et al. 2002). It has been suggested since, however, that morphological differences (colour and head length-to-width ratio) might be useful in delimiting at least two P. imperialis species (J.-Y. Rasplus, personal communication). The geographic separation of the two P. imperialis sp. 1 populations could yield intriguing insights into ecological diversification in the early stages of speciation. The northern P. imperialis sp. 1 population coexists with two other pollinator species (Darwell et al. 2014), and therefore is subject to a selective pressure (interspecific competition for oviposition sites) that the southern population does not experience throughout most of its range. Furthermore, both P. imperialis sp. 1 populations interact with a single geographically widespread population of the parasitoid Sycoscapter sp. A (Sutton et al. 2016). It is possible that they will develop adaptive responses to a common antagonist (Kjellberg & Proffit 2016), but this requires further investigation.

Despite several molecular studies within the P. imperialis complex (Haine et al. 2006; Darwell et al. 2014; Sutton et al. 2016), a thorough understanding of the phylogenetic relationships within it are lacking. Given that standard phylogenetic markers lack the resolution required for this (Haine et al. 2006; Darwell et al. 2014), a phylogenomic approach may be required. A resolved phylogeny would provide a platform to form hypotheses regarding major factors influencing speciation in this complex and a better understanding of how diversity is generated in host-specific nursery pollination systems. Tests of population expansion in the present study support sudden expansion of the southern P. imperialis sp. 1 population. Mismatch distribution analysis of the northern P. imperialis sp. 1 population, however, indicates that it has not experienced population expansion. This suggests a northern origin of P. imperialis sp. 1 with rapid southward expansion (which was likely concurrent with the range expansion of the host plant) and subsequent splitting and reduction of the smaller northern population. This explanation could account for higher levels of genetic diversity in the southern population despite a northern origin of these species.

There are three other possibilities that could account for the extremely low genetic diversity in the northern P. imperialis sp. 1 population. Firstly, low sample size of the northern P. imperialis sp. 1 population could result in decreased estimates of

genetic diversity due to subsampling. This seems unlikely, though, given that genetic diversity within each of the four southern P. imperialis sp. 1 sites was higher than in northern P. imperialis sp. 1 despite the fact that sample size for the northern P. imperialis sp. 1 population (N = 35) was higher than the sample size for any of the southern P. imperialis sp. 1 sites (N ≤ 30; Chapter 3; Sutton et al. 2016). Another possibility for the low genetic diversity of the northern P. imperialis sp. 1 population, including the fact that three of the loci were monomorphic, is that these microsatellite markers were developed using individuals from the southern P. imperialis sp. 1 population (Sutton et al. 2015). However, this explanation seems less likely given the high level of genetic diversity in the more distantly related P. imperialis sp. 2. Finally, Wolbachia-induced selective sweeps have the potential to concurrently decrease intraspecific diversity and increase interspecific diversity (e.g. Shoemaker et al. 2004; Delgado & Cook 2009; Xiao et al. 2012). This is certainly a possibility given the extremely high frequency of Wolbachia infection in fig wasps (Shoemaker et al. 2002; Haine & Cook 2005), but remains to be investigated in P. imperialis.

5.5.1 Conclusions

We show that allopatric P. imperialis sp. 1 populations are incipient species. Our results support the use of multilocus approaches to species delimitation, and highlight major gaps in our understanding of evolution within this system. Further studies are required to establish morphological, behavioural and/or ecological characters that will help to refine the taxonomy of P. imperialis.

The allure of DNA taxonomy is strong, but while DNA barcoding methods offer a quick and convenient ‘first glance’ at species diversity, they should not be used as the sole tool for species delimitation (Will & Rubinoff 2004b; Ebach & Holdrege 2005). Where possible, future efforts to define species boundaries should integrate multiple sources of data (Will et al. 2005), whether they be molecular, morphological, behavioural or ecological.

Chapter 6

Vulnerability of temperate fig- pollinating wasps to heat and desiccation stress

6.1 Abstract

Climate change can threaten insect pollination systems in different ways. While temperate zone ectotherms typically experience mean temperatures below their thermal optima, high variation can still expose them to dangerous extremes. We investigated the effects of short bouts of realistic high temperatures and low humidity on the only fig wasp species that pollinates Port Jackson fig in temperate New South Wales (Australia). Wasp offspring develop only inside the fig fruits and female wasps rely on males to release them from figs to disperse. We measured wasp emergence from figs at a range of temperatures, and tested the effects of heat wave conditions (hot and dry) on adult female longevity. Emergence from figs was unaffected by temperatures up to 39°C, but declined drastically above 39°C. Meanwhile, adult longevity decreased gradually with temperatures above 30°C and median lifespan was just 3 hours at 39°C. Low humidity further reduced wasp longevity. Field data indicated that fig fruits exposed to solar radiation were significantly hotter than those in shade, suggesting that figs in shaded areas could provide protected microhabitats for fig wasps. Our data indicate that temperate pollinating insects may be vulnerable to high temperatures that already occur occasionally, and are predicted to increase in frequency and intensity, coupled with low levels of humidity. When considering the prospects for temperate insect pollinators under climate change, humidity effects should be considered alongside temperature. Temperature extremes may be at least as important as increasing means, and their impacts may be exacerbated by reduced levels of humidity.

6.2 Introduction

Climate change can influence biotic processes in many ways. With regard to plants and their insect pollinators, interest has focused primarily on predictions that global warming may uncouple plant and pollinator phenologies, reducing their temporal overlap and, thus, overall pollination success (Hegland et al. 2009). Phenological matching is only one aspect of insect pollination, however, and Forrest (2015) recently proposed a shift in focus towards (1) the effects of acute aspects of climate change (e.g. drought) on pollination mutualisms, and (2) the effects on pollinator populations themselves (rather than the plants they pollinate).

Direct effects of climate stresses on insects are likely to be of increasing importance, because current models predict an increase in the frequency of extreme weather (e.g. heat waves, drought) (Meehl & Tebaldi 2004; Collins et al. 2013). Tropical insects typically occupy relatively narrow latitudinal ranges compared to temperate species (Rapoport’s Rule; Stevens 1989) and are thought to have adapted to narrower thermal ranges as a result (Stevens 1989). Conversely, outside of tropical zones, many insects experience mean temperatures below their thermal optima (Addo- Bediako et al. 2000; Deutsch et al. 2008), and are also subjected to large variation in both temperature and relative humidity. Consequently, tropical insects are much more vulnerable to small increases in mean temperature than temperate species (Addo-Bediako et al. 2000). It may therefore be more ecologically relevant to explore the effects of thermal range extremes than increased average temperatures when investigating the climate change factors that are likely to threaten temperate insect species (Jentsch et al. 2007; Kingsolver et al. 2013). For example, both thermal tolerance and desiccation resistance in drosophilids are wider in species from higher latitudes than in those restricted to the tropics (Parkash & Munjal 1999; Sgrò et al. 2010; Mitchell et al. 2011).

There have been relatively few studies of the responses of pollinating insects to heat stress and drought. Butterflies are the exception, with field studies highlighting their vulnerability to drought (Pollard et al. 1997) and heat wave events (Piessens et al. 2009). These patterns are supported by laboratory experiments that highlight the severe effects of acute climatic events on butterfly fitness, influencing various

aspects of physiology and life history, such as reductions in body size and longevity (Bauerfeind & Fischer 2014; Fischer et al. 2014).

One of the largest evolutionary radiations of pollinating insects involves the tiny wasps (Hymenoptera: Agaonidae) that pollinate fig trees (Ficus, Moraceae). This is an obligate reproductive mutualism with very high partner specificity, which provides food (figs) for a diverse range of animals on all tropical continents (Shanahan et al. 2001). Fig trees can be keystone resources because their unusual phenology, with at least some trees in a population in flower at any given time, favours year-round fruit production. However, key details of fig wasp biology also make their populations, and as a result their interactions with figs, vulnerable to climate change. Importantly, fig wasps are the only vectors of fig pollen, while fig wasp offspring develop only in the figs of the trees they pollinate. The tree population effectively “rears” its own host-specific pollinator population inside its fruits. While the larvae of most species take a few weeks to mature inside figs, the free-living adults live for only a day or two. During this brief, non-feeding adult life, the females must disperse potentially long distances (Harrison 2003; Ahmed et al. 2009) to find receptive figs in which to lay their eggs. Abiotic factors reducing the presence or activity of adult wasps can therefore cause drastic reductions in pollination, or even local collapse of pollinator populations (e.g. Bronstein & Hossaert-Mckey 1995).

Harrison (2000, 2001) first demonstrated drought-induced mortality of tropical fig- pollinators in Borneo. Subsequent laboratory experiments have further highlighted the sensitivity of tropical fig wasps to small increases in temperature (Jevanandam et al. 2013) and decreases in humidity (Dunn et al. 2008b; Warren et al. 2010). These results imply threats to the persistence of tropical fig wasp populations and, consequently, the figs they pollinate, under climate change. Overgaard et al. (2011), however, showed that Drosophila species with broad latitudinal ranges have a higher tolerance to thermal shifts than do species with strictly tropical distributions. Consequently, temperate fig wasps and fig-pollination systems may be more stress- tolerant than are tropical ones.

We investigated the effects of heat stress on the emergence and longevity of a temperate fig wasp. Pleistodontes imperialis sp. 1 (sensu Haine et al. 2006) is the

pollinator of Port Jackson fig (Ficus rubiginosa) in southeastern Australia. In temperate zones, it experiences a typically warm, but variable climate (January mean maximum temperature for Sydney is 25.9°C; Bureau of Meteorology 2015), with mild winters. The frequency, intensity and duration of extreme heat events and heat waves in Australia has been increasing strongly in the last 60 years, and current climate models predict further increases (Hughes 2003; Steffen 2015). Furthermore, evidence suggests that relative humidity over land at low and mid latitudes has been decreasing (Simmons et al. 2010). If the maximum temperatures currently observed in Sydney cause high mortality in adult fig wasps, then such mortality events are likely to increase in frequency with climate change and the number of days per year above key thresholds could increase rapidly even with minor changes in mean conditions, due to the high daily variability at this latitude.

The importance of conditions inside the fig prior to adult female emergence raises the question of how such conditions vary under field conditions. A study in Panama showed that temperatures experienced by wasp offspring inside figs are higher for larger figs (across species) and for those exposed to direct sunlight (relative to shaded figs) (Patiño et al. 1994). While most figs in the canopy may be exposed to sun, shading within a tree may result in some protected microhabitats (figs) that could provide safeguards against the hottest ambient conditions.

We conducted a series of separate experiments to test the effects of climatic stresses on different key aspects of temperate fig wasp biology:

1. The effect of temperature on the emergence of wasps from figs, mediated by male activity inside figs; 2. The effect of temperature and humidity on the lifespan of adult female wasps, the key dispersing and pollinating stage of the life cycle; 3. Preliminary surveys to determine the core temperatures of figs on trees and ascertain if those in shade experience lower core temperatures than those in direct sunlight.

6.3 Materials and methods

6.3.1 Study system

We carried out three independent experiments to test the effects of temperature and humidity on: (1) the emergence of fig-pollinating wasps, (2) the longevity of adult fig-pollinating wasps and (3) the core temperature of figs.

Pollen-bearing female fig wasps enter receptive figs to lay eggs and pollinate fig ovules. Several weeks later, mating occurs between wasp offspring that have matured inside the fig, after which females collect pollen and exit the natal fig through holes created by the males. Development of pollinator wasps occurs over six-to-eight weeks in average field conditions (Harrison 2005; Xiao et al. 2008). As F. rubiginosa figs ripen, they turn bright yellow (they are green during early stages of development) and become soft, and are thus easily identified.

The fruiting of fig trees in a population is asynchronous, and monoecious figs generally have low population and fruiting densities, so, as in many studies of fig wasps, selection of sites was dictated by the availability of ripe figs when we performed the experiments (Table 6.1). Limitations on incubator space and availability of ripe figs meant that we were unable to perform all experiments at the same time. For the experiment on wasp longevity detailed below, ripe F. rubiginosa fruits were collected in autumn 2014 from the Sydney region in New South Wales, and we exposed the emergent wasps to the temperature treatments sequentially over a five week period with a two week interval between experiments at the Balmoral and Penrith sites. Importantly, for the experiment testing differences in the core temperature of figs in the sun and shade, we required ripe figs which were accessible in order to take temperature readings without removing the figs from the tree. As the only accessible figs were located in direct sunlight during the time this experiment was conducted, we had to provide artificial shading for some of the figs.

6.3.2 Experiment 1: Effects of increasing ambient temperature on wasp emergence

Ripe F. rubiginosa figs (N = 410) were collected from 15 trees in the Wollongong area (~70 km south of Sydney) in December 2014, and put into cold boxes with ice (6.02 ± 0.01°C) for approximately two hours to restrict wasp emergence prior to incubation. While it is possible that such low temperatures might affect responses to temperature treatments, it is not uncommon for these sites to experience temperatures of 6°C overnight during spring and autumn. Furthermore, cold transport was essential to avoid emergence and the confounding effects of the presence of exit holes on internal fig temperatures during temperature treatments. Prior to incubation, fig samples were removed from the ice box and transferred to individual clear plastic pots with mesh-covered lids for ventilation, allowing figs to acclimate to room temperature (~25°C) for approximately 45 – 60 minutes. Fig diameter was measured (ventrally, at the widest point) using Vernier callipers.

The pots were then placed in a TLM-590 incubator (Thermoline Scientific, NSW, Australia) at temperatures of 25°C (control), 33°C, 36°C, 39°C, 42°C, 45°C and 48°C. Temperatures in the range of 35-40°C are frequently observed during hot summers in Wollongong, and temperatures greater than 40°C have been recorded there a total of nine times in the past decade. We coupled temperature treatments with one of two time treatments (3 or 7 hours exposure) to compare the potential effects of short bursts and extended periods of heat stress. Sample size was N = 50 figs for the control group and N = 30 figs for all other temperature/time treatments. Figs from individual trees were randomly allocated across treatment groups.

Relative humidity (RH) of the climate chamber was maintained at 55 to 60% and monitored using Hygrochron temperature/humidity data loggers (Maxim Integrated, California, USA). Although it was expected that increased temperatures would be accompanied by an increase in humidity within the fig fruit, we did not measure internal fig humidity because it would compromise the objective of the experiment through the creation of artificial holes that would affect internal fig temperatures and encourage wasp emergence. Artificial light was provided for the duration of the initial heat treatment. After the specified exposure times, the pots were transferred to an incubator with a diurnal temperature cycle of 25/19°C and a 14:10 hour light

100

cycle. After 72 hours from the start of the incubation period, the pots were filled with 70% ethanol to prevent further wasp emergence and the number of emerged female pollinators were counted. From previous experience, we have found that 72 hours at ambient temperature is sufficient to allow almost all wasps to emerge from a ripe fig, while preventing the fig from becoming mouldy (Segar et al. 2014).

Data were analysed using zero-inflated regression models (Greene 1994). Zero- inflated regression models are employed for overdispersed count data with excess zeroes that result from processes other than experimental treatment(s). For example, in this experiment, zero counts resulted from either (1) high temperature treatments or (2) error in predicting the timing of fig wasp emergence (i.e. figs may look and feel ripe, but wasps may not be ready to emerge). The former are considered ‘true’ zeroes, while the latter are ‘excess’ zeroes. We used the pscl package (Jackman 2015) in R v3.2.0 to determine that a zero-inflated negative binomial model was a better fit for our data than a zero-inflated Poisson model. Subsequent analyses were performed in the glmmADMB package (Fournier et al. 2012; Skaug et al. 2013), which allows for the inclusion of random effects in the model. ‘Tree’ was included in the model as a random variable.

101

Table 6.1: Site locations and summary climate data.

Mean summer statistics Total no. days ≥ 39°C Site GPS coordinates Max. temperature (°C) Min. temperature (°C) Rainfall (mm) 3pm RH (%) in last decade Wollongonga 34°25'30"S, 150°53'35"E 24.0 18.1 121.3 74.3 12 Balmoralb 33°49'44"S, 151°15'18"E 25.6 18.4 98.9 61.7 8 Penrithc 33°45'18"S, 150°40'27"E 29.8 18.0 72.6 48.3 59 Putney Parkd 33°50'00"S, 151°06'34"E 28.0 18.9 84.1 52.3 21e

Max., maximum; Min., minimum; RH, relative humidity (Bureau of Meteorology 2015). ‘3pm RH’ is the mean relative humidity as measured at 3pm local time. Climate data obtained for aAlbion Park (Wollongong Airport), bSydney (Observatory Hill), cPenrith Lakes AWS, and dSydney Olympic Park weather stations. eData for 2005 – 2011 obtained from Sydney Olympic Park weather station, and data from 2011 – 2015 obtained from Sydney Olympic Park (Archery Centre) weather station.

102

6.3.3 Experiment 2: Effects of temperature and relative humidity on wasp lifespan

In order to test the hypothesis that provenance plays a role in the response of wasps to increasing temperatures and aridity, we compared the effects of temperature and humidity on adult wasp lifespan using samples collected in autumn 2014 from two sites in Sydney, Balmoral and Penrith, which differ in local climatic conditions. Balmoral is situated on the Sydney Harbour foreshore, and Penrith is located inland in western Sydney, ~60 km west of Balmoral. Both sites are at similar altitudes (5 – 30m above sea level) and represent two extremes along a sharp temperature gradient in the Sydney region (Table 6.1). Due to constraints described above, it was necessary to conduct this experiment as a sequence of trials over a five-week period. In the six weeks preceding sample collection at the respective sites, mean maximum temperature was 26.68 ± 0.04°C in Balmoral and 28.09 ± 0.04°C in Penrith, and mean minimum temperature was 20.20 ± 0.03°C in Balmoral and 17.80 ± 0.03°C in Penrith. We expected that wasps from the Penrith site would survive longer than those from Balmoral under hotter conditions, given the higher mean maximum temperature experienced during development.

Ripe F. rubiginosa figs were collected from 11 trees at Balmoral in March 2014, and placed in individual plastic pots with mesh-covered holes in the lids. Wasps were allowed to emerge at room temperature (22 – 25°C) for 1.5 hours, after which they were randomly assigned to a temperature and humidity treatment (N = 2 to 49 wasps per pot). Temperature treatments were 25°C, 30°C, 35°C, 38°C, 39°C and 40°C, and this was coupled with either a high (~70%) or low (~30%) RH treatment (Table 6.2). For the low RH treatment, a 0.2 mL tube containing a single bead of desiccated silica gel was placed in the pot, which was sealed with a solid lid. We pierced two holes in the tube so that the RH within the pot could be reduced but wasps could not enter the tube. All pots were placed in a CLIMATRON-520-SL-H incubator (Thermoline Scientific, NSW, Australia) with RH set to 70%. RH was recorded with Hygrochron temperature/humidity data loggers placed in pots with mesh lids (high RH treatment) and solid lids with a 0.2 mL tube containing a desiccated silica gel bead (low RH treatment). Three data loggers were used for each humidity treatment.

103

Table 6.2: Samples sizes (no. of wasps) used to test the effects of temperature and humidity on fig wasp lifespan.

Balmoral Penrith Temperature (°C) High RH Low RH High RH Low RH 25 292 308 - - 30 296 264 398 349 35 209 214 318 361 38 251 212 340 337 39 283 333 451 488 40 189 161 - - (RH, relative humidity).

The number of dead female pollinators in each individual pot was recorded every three hours until all had died. Silica beads in the low RH treatment pots (and data logger pots) were replaced with fresh ones at each three-hourly check. Mean RH was 31.89% ± 0.38% and 72.63% ± 0.23% for the ‘high’ and ‘low’ RH treatments, respectively. The experiment was repeated two weeks later using figs collected from seven trees (N = 1 to 47 wasps per pot) at the Penrith site. However, given the results from the Balmoral site, the experimental design was simplified and only temperature treatments of 30°C, 35°C, 38°C and 39°C were included, coupled with ‘high’ and ‘low’ humidity treatments. Mean RH was 27.77% ± 0.80% and 71.40% ± 0.21% for ‘high’ and ‘low’ RH treatments, respectively.

Median lifespan estimates were obtained using Kaplan-Meier analysis in SPSS v22.0 (IBM Corp 2013), and log rank tests were performed to determine the effect of temperature on wasp lifespan. Cox Proportional Hazards models were constructed using the coxme package (Therneau 2015) in R v3.2.0 to obtain mortality rates. Within- and between-site tests were performed to test the effects of temperature, RH and site on lifespan. ‘Tree’ was included in the model as a random variable for within-site tests. No data were censored.

6.3.4 Experiment 3: Core temperature of figs in sun versus shade

This experiment was performed using 141 ripe figs from a single tree in Putney Park, Sydney, on 2 April 2015. Data were collected between 1300 and 1500 h local time in

104

sunny conditions. Ambient air temperature ranged from 27.8°C to 29.7°C and 27.0°C to 30.5°C for samples in sun (N = 68) and shade (N = 73), respectively. Three measurements were recorded for each fig: (1) fig diameter, (2) ambient air temperature, and (3) fig core temperature. Vernier callipers were used to measure the diameter of the fig ventrally at its widest point. Ambient air temperature was measured using a Testo 925 1-channel thermometer with a Type K thermocouple air probe (Testo, Lenzkirch, Germany). We defined ambient air temperature as the air temperature directly adjacent to the fig, and so the air temperature probe was exposed to direct sunlight when measuring ambient air temperature for figs in direct sunlight, but was in similar shaded conditions for ambient air temperature measurements for shaded figs. Between measurements of ambient air temperature, the air temperature probe was placed in deep shade to protect it from excess solar radiation. Core temperature, defined as the temperature in the centre of the fig, was measured with a Digi-Sense long-stem digital thermometer (Cole-Parmer, Illinois, USA) by inserting the probe directly into the centre of the fig.

Samples were divided into two groups based on sun exposure: figs exposed to sunlight (N = 68) and figs in shade (N = 73). Although many fig fruits do grow in shade, the vast majority of ripe figs accessible at the time this experiment was conducted were naturally in direct sunlight, so artificial shading was provided for figs in the shaded group using an umbrella for 15 minutes prior to recording any temperatures. We tested the effects of ambient air temperature, fig diameter and sun exposure on core temperature using the lme4 package (Bates et al. 2013; Bates et al. 2014) in R v3.2.0 (R Core Team 2015).

6.4 Results

6.4.1 Experiment 1: Effects of increasing ambient temperature on wasp emergence

Wasp emergence gradually increased with temperature, before falling sharply at 39°C and 36°C for the ‘3 hours’ and ‘7 hours’ treatments, respectively (Figure 6.1). No statistical support could be given to these trends as post-hoc tests could not be performed on data analysed using zero-inflated regression models. Exposure time

105

was not significant and was thus excluded from the final model. Temperature was significant in the final model (χ2 = -4.03, DF = 1, P < 0.001), and the number of emerged wasps increased with fig diameter (χ2 = 5.50, DF = 1, P < 0.001). Fig size did not vary significantly across temperature/time treatments (F1,394 = 1371.82, P = 0.62).

Figure 6.1: Mean (± S.E.) number of female P. imperialis sp. 1 wasps emerged from F. rubiginosa figs subject to various temperature/time treatments. (Black circles = 3 hours exposure, open triangles = 7 hours exposure). Sample size for control group (25°C) was N = 50, and sample size for each other treatment was N = 30.

6.4.2 Experiment 2: Effects of temperature and relative humidity on wasp lifespan

Increasing temperatures and lower RH were associated with significantly higher mortality rates for the Balmoral site, and there was a significant temperature × humidity interaction (Table 6.3). At 25 and 30°C, median lifespan was 21 and 15 hours for wasps at high and low RH, respectively (Figure 6.2a). Lifespan steadily decreased at temperatures above 30°C for both RH treatments. At 39°C, less than 5% of wasps survived longer than 3 hours for each RH treatment, and the only wasps to survive past 3 hours at 40°C were three individuals in the low RH treatment. 106

Pollinator lifespan in the Penrith site was significantly affected by increasing temperature and low RH, and we observed a significant interaction between temperature and humidity (Table 6.3). Low RH treatments were characterised by shorter lifespan across most temperature treatments, and lifespan declined at temperatures of 35°C and above (Figure 6.2b). Median lifespan at 39°C was just 3 hours for both RH treatments. There was no difference in mortality between sites at high humidity. However, mortality was higher for wasps from the Penrith site at low humidity across the range of temperatures (P < 0.001) (Table 6.4).

Table 6.3: The effect of increasing temperature and low humidity on mortality rates.

Mortality rate N HR coeff. S.E. P 3012 Balmoral 1.37 0.31 0.01 < 0.001 Temperature 20.92 3.04 0.27 < 0.001 Humidity (low) 0.93 -0.07 0.01 < 0.001 Temperature:Humidity 3042 Penrith 1.48 0.39 0.01 < 0.001 Temperature 55.72 4.02 0.44 < 0.001 Humidity (low) Temperature:Humidity 0.91 -0.09 0.01 < 0.001 Mortality rates were obtained with Cox Proportional Hazards regression. Hazard ratios indicate risk of death relative to the baseline group (i.e. basal temperature and high humidity) for each site. (N, no. samples; HR, hazard ratio; coeff., regression coefficient; S.E., standard error).

Table 6.4: Comparison of mortality rate between sites for each humidity treatment.

Mortality rate Site Humidity N HR coeff. S.E. P Balmoral High 1520 - Low 1492 - Penrith High 1508 1.06 0.06 0.04 0.159 Low 1534 1.20 0.18 0.04 < 0.001 Mortality rates were obtained with Cox Proportional Hazards regression. (N, no. samples; HR, hazard ratio; coeff., regression coefficient; S.E., standard error). HR is given for the high risk site only (the low risk site is considered the baseline).

107

Figure 6.2: Median lifespan duration (± S.E.) of adult P. imperialis sp. 1 females across temperature treatments for (a) Balmoral site and (b) Penrith site. The Penrith site was analysed at a subset of temperatures based on results from the Balmoral site. (Black circles = 70%RH, open inverted triangles = 30%RH).

108

6.4.3 Experiment 3: Core temperature of figs in sun versus shade

Ambient air temperature ranged from 27.8°C to 29.7°C (mean = 28.72 ± 0.06°C) and 27.0°C to 30.5°C (mean = 28.61 ± 0.11°C) for samples in sun (N = 68) and shade (N = 73), respectively. Fig core temperature increased with ambient temperature only for shaded figs (F1,71 = 27.09, P < 0.001) (Figure 6.3), and was higher for figs in direct sunlight (30.03 ± 0.01°C) than for figs provided with artificial shade (29.09 ±

0.01°C) (F1,133 = 30.29, P < 0.001). We also observed a significant interaction between sun exposure treatment and fig diameter (P = 0.048). However, fig diameter was not significant when sun exposure treatments were analysed independently.

Figure 6.3: Fig core temperature as a function of ambient temperature in shaded figs. (R2 = 0.27, P < 0.001).

6.5 Discussion

We provide the first empirical data on the vulnerability of a temperate fig-pollinating wasp species to extreme levels of heat, which have increased in frequency with recent climate change (Steffen 2015). Our results support our expectation that these temperate insect populations are highly vulnerable to extreme heat conditions. Most notably, while most studies have focussed on the effects of temperature alone on

109

insect longevity, we have shown that the risks associated with increased temperature may be amplified by reductions in humidity, providing a more realistic scenario given future climate change predictions for temperate regions (Hughes 2003; Meehl & Tebaldi 2004; Simmons et al. 2010; Collins et al. 2013; Steffen 2015).

The frequency of extreme heat events in Australia has increased over the last 50 years (Hughes 2003; Steffen 2015), and this is forecast to further increase into and beyond the 21st century (Meehl & Tebaldi 2004; Collins et al. 2013). By 2090, the hottest day for each season is expected to increase by 2 to 4°C in eastern Australia (CSIRO and Bureau of Meteorology 2015). Alarmingly, temperatures already experienced during summers in southeastern Australia were observed to have severe negative effects on fig wasp emergence and longevity.

Emergence from the natal fig did not appear to be negatively affected by temperatures up to 39°C (13°C above the local summer mean maximum temperature), and may actually have increased with temperature up to this point in the ‘3 hours’ exposure treatment. However, there was almost no emergence of pollinators at temperatures of 42°C or above. This is consistent with typical thermal performance curves in ectotherms, which predict that performance will increase with temperature until optimal conditions are met, but will decrease rapidly at temperatures slightly exceeding optimal (Huey & Stevenson 1979; Gilchrist 1995; Martin & Huey 2008). Similarly, Adamo & Lovett (2011) showed that temperatures of 7°C above mean field temperature resulted in increased fecundity, faster development and greater body mass in the cricket Gryllus texensis (although temperature treatments used in their study apparently did not exceed optimal).

The transportation of fig samples in cold boxes was required to restrict wasp emergence until samples were incubated under experimental conditions, and it is unclear how this might have affected wasp emergence. However, given the short time of exposure, we suspect that the subsequent acclimation period was sufficient for wasps to make a full recovery from any effects of cold temperature. Optimal temperature for emergence for wasps collected for this experiment was 10 to 13°C above the local summer mean maximum temperature, however, this is likely to vary in a natural setting due to weather variability and potential active transpiration by fig fruits still attached to the tree (Patiño et al. 1994). The positive correlation between

110

fig diameter and wasp emergence is likely due to higher wasp/seed production in larger F. rubiginosa figs (Cook & Power 1996), rather than differential effects of temperature on figs of different sizes.

Median lifespan was not affected by temperatures of 30°C, but gradually declined at 35°C and above. Median lifespan was just 3 hours at 39°C, and this is likely to be an overestimate, as most individuals likely died prior to the first mortality check. In contrast, Jevanandam et al. (2013) reported a significant reduction in the lifespan of fig-pollinating wasps under tropical conditions in Singapore at 30°C, which is 3°C above mean daily temperature, but 1°C below mean maximum temperature. We observed a strong reduction in adult lifespan at sub-lethal temperatures (> 30°C), even under favourable humidity conditions. A decrease in the already short lifespan of fig wasps could have severe consequences for pollination of sparsely-distributed monoecious figs, and potentially the ecosystems they support. Firstly, it may limit the dispersal potential for female wasps to locate fruits in which to lay their eggs and consequently provide pollination services, thus reducing the reproductive output of the fig trees. Secondly, reduced pollination may lead to increased abortion of fig fruits, reducing food resources for local vertebrate frugivores.

Low RH treatments were associated with higher mortality rates. Dunn et al. (2008b) and Warren et al. (2010) showed that lifespan was significantly reduced at low ambient humidity in three tropical fig wasp species, however, in their studies they used a crude measure of humidity and did not quantify actual RH. Mean monthly RH for the populations sampled by Dunn et al. (2008b) is high, ranging seasonally from 61 to 81% (Bureau of Meteorology 2015). In contrast, mean RH for the sites that we sampled ranges seasonally from 40 to 64% (Bureau of Meteorology 2015). Humidity treatments applied during this experiment therefore reflect conditions that are commonly experienced by P. imperialis sp. 1 in its natural environment. Significant differences in mortality between sites at low RH was not unusual considering that Jevanandam et al. (2013) found differences in mortality between wasps on the same tree. Genetic factors have been shown to contribute to heat tolerance in Drosophila (Krebs & Feder 1997; Krebs et al. 2001). It is not clear how temperature variation during larval development may have influenced resilience under high temperature conditions at the two sites used in this experiment, and although (epi)genetic factors

111

were not explored in this study, they are likely to contribute to differences in the thermal tolerance of fig wasps, and deserve attention in future studies.

Patiño et al. (1994) showed that figs exposed to direct sunlight can experience internal temperatures of 2-to-3°C above ambient air temperature. Our data suggest that sun-exposed figs experience internal temperatures of almost 1°C hotter than shaded figs. Given the effect of small increases in temperature on fig wasp emergence, it could be considered that shaded figs provide protected microhabitats for fig wasps. Our observations of F. rubiginosa, however, indicate that figs are typically located on the outside of the tree canopy, suggesting that any shade protection may be more relevant to whole trees located in a shady environment (e.g. on the southern side of a steep gradient), rather than individual figs located in shaded parts of a tree. Patiño et al. (1994) showed that higher fig core temperatures were experienced in species with larger figs, but we found no evidence to suggest that size variation within a fig species affects fig core temperature. Future work should aim to characterise how the internal temperature profiles of figs change over a range of ambient (and more stressful) temperatures, and how this differs between figs that are naturally shaded and those which are sun-exposed.

Few studies of climate change and pollination have focussed on the effects of observed and predicted increases in the frequency and intensity of heat wave conditions. Piessens et al. (2009) found that small population size and habitat fragmentation played a major role in the extinction of local populations of the butterfly Cupido minimus following a heat wave in Europe in 2003. Similarly, Bauerfeind & Fischer (2014) demonstrated that a reduction in longevity and body size in the widespread butterfly Pieiris napi resulted from heat wave simulations. Interestingly, they also showed that increases in mean temperature did not have any significant fitness effects in this species. Our study bears similarities to studies on the responses of non-pollinating insects to heat waves. Using Drosophila as a model system, Feder et al. (Feder et al. 1997) reported (i) lethality of exposure to temperatures of 39°C for short periods, and (ii) fruit core temperatures exceeding ambient temperature. Assessment of the microclimatic conditions in citrus trees determined that citrus rust mites (Phyllocoptruta oleivora) actively avoid solar exposure, opting instead to inhabit the cool shady parts of the tree (Allen & McCoy 1979). Such behaviours are not possible for fig wasps which have developed within a

112

fig fruit over several weeks, but it is not known whether or not pollen-bearing females actively seek out shaded figs when locating receptive figs to pollinate. This certainly warrants further investigation.

Although our study highlights some difficulties associated with performing climate- controlled experiments in non-model systems that cannot be kept in culture, we have clearly demonstrated that temperate pollinating fig wasp species are vulnerable to extreme conditions that are currently experienced in parts of temperate Australia. Future studies should aim to characterise how conditions experienced during development affect the timing of fig wasp development and how they respond to different climate scenarios throughout different life history stages. The frequency of extreme conditions varies across the Sydney Basin, and some areas of Sydney receive, on average, as many as six days per year of 39°C or above (Table 6.1) (Bureau of Meteorology 2016). Climate models predict that the frequency and intensity of extreme heat events is likely to increase; indeed, this has already been observed in some temperate regions (Hughes 2003; Steffen 2015), and so we can expect more extreme conditions similar to those simulated in our experiments. Low pollinator emergence and reduced longevity under these conditions could have potentially cascading effects on fig trees and the ecosystems they support. Failure of fig-pollinating wasps to emerge or locate receptive figs could lead to the collapse of local pollinator populations, thereby reducing pollination services to figs and thus depleting an important food source for vertebrate frugivores.

Extreme heat is known to have negative consequences for the longevity and fitness of terrestrial ectotherms (Roux et al. 2010; Gillespie et al. 2012; Jeffs & Leather 2014), and our results further emphasise the need for a sharper focus on the impact of extreme conditions, such as heat waves, on temperate biodiversity. Furthermore, the combined impact of heat stress and humidity conditions must be considered in order to provide more realistic and ecologically-relevant data on the future of our ecosystems.

113

Chapter 7

General discussion

114

Plant-insect interactions are not only responsible for generating a significant proportion of extant biodiversity, but are also crucial for maintaining stability in terrestrial food webs. The research presented within my thesis addresses several outstanding questions about the evolutionary ecology of an important group of insects: fig wasps. Specifically, it improves our understanding of their diversity, biogeography and speciation through the use of molecular tools, and highlights the potential challenges they may be exposed to in the face of climate change. Below, I summarise my key findings, discuss the implications and limitations of my work and outline potential directions for future research arising from my thesis.

7.1 Key findings, implications and limitations

7.1.1 Host-parasitoid coevolution

In Chapter 2, I described the development and characterisation of microsatellite markers for P. imperialis sp. 1. I provided detailed methods for the amplification of nine microsatellite markers and demonstrated their utility in other P. imperialis species. In Chapter 3, I applied these markers to the first comparison of population genetic structure in a fig-pollinating wasp (P. imperialis sp. 1) and its parasitoid (Sycoscapter sp. A).

Consistent with my hypotheses, I found that (i) the pollinator comprises two genetically-distinct populations, while the parasitoid exists as a single population throughout eastern Australia, and (ii) the parasitoid disperses as far, if not farther than the pollinator. The first finding suggests that geographic mosaics (Thompson 2005) likely play an important role in the coevolution of host-parasitoid interactions within the fig – fig wasp mutualism. The significance of geographic mosaics for evolution in the fig – fig wasp system has previously been suggested (Thompson 2005; Kobmoo et al. 2010; Segar & Cook 2012; Darwell et al. 2014; Tian et al. 2015), but this is the first time they have been considered for an interaction involving a NPFW. Furthermore, this study is also (to the best of my knowledge) the first comparative population genetic study undertaken with strict co-sampling across the geographic range of the host fig species. Ecological divergence of the pollinator

115

populations seems extremely likely given not only their genetic divergence and geographic isolation, but also the added ecological and evolutionary pressure that the northern P. imperialis sp. 1 population faces in having to compete with heterospecific pollinators for habitat patches.

The second finding – that the parasitoid disperses as far, if not farther than the pollinator – challenges common views about the dispersal behaviour of NPFWs (Compton et al. 2000; Harrison 2003; Kjellberg & Proffit 2016). Nevertheless, these views are not necessarily incompatible. Early studies (Compton et al. 2000; Harrison 2003) suggested that NPFWs disperse at lower heights than do pollinators in Bornean rainforests, and are thus more likely to disperse actively and locally beneath the tree canopy, rather than passively on wind currents far above the canopy. Given the habitat differences between previous studies and the present study (Bornean rainforest versus open woodland, fragmented eucalypt forest and urbanised areas of eastern Australia), it is not difficult to envisage long-distance dispersal of NPFWs (on wind currents), even if absolute flight heights remained consistent between the two habitats.

There were two limitations of this study that pertained to the parasitoid data. Firstly, although parasitoids were initially genotyped at nine microsatellite loci, only six of these amplified consistently and were used in subsequent analyses. Although this situation was not ideal, power analyses suggested that these six loci had sufficient power to detect small levels of genetic differentiation in my data. Furthermore, I lacked parasitoid samples from a latitudinal stretch of almost 1,500 km (where the pollinator, P. imperialis sp. 1 does not occur). Inclusion of parasitoid samples from this stretch would probably not have altered the conclusions of this study, but they could have provided more robust estimates of parasitoid dispersal.

7.1.2 Assessment of DNA barcoding in fig wasps

In Chapters 4 and 5, I applied a population genetic approach to test the reliability of DNA barcoding in fig wasps under two geographic scenarios. Firstly, in Chapter 4, I assessed levels of gene flow and hybridisation among sympatric taxa that have been previously delimited by DNA barcoding (Haine et al. 2006; Darwell et al. 2014).

116

Secondly, in Chapter 5, I scaled genetic differentiation between allopatric fig wasp populations that were not split by previous DNA barcoding efforts (Haine et al. 2006; Darwell et al. 2014), with reference to a closely-related, but distinct, species.

I predicted that there would be minimal gene flow between sympatric P. imperialis species and that allopatric conspecific P. imperialis populations would exhibit genetic signals consistent with incipient speciation. My results confirmed my hypotheses by revealing that (i) sympatric P. imperialis sp. 2, 3 and 4 rarely exchange genes, despite considerable opportunity to interbreed at the patch level, and thus represent true (reproductively isolated) species, and (ii) allopatric P. imperialis sp. 1 populations are more than likely incipient species. The use of DNA barcoding to identify and define species limits can be a contentious endeavour as the discovery of distinct mtDNA groups is not necessarily evidence of reproductive isolation (Rubinoff et al. 2006; Giska et al. 2015; Martinsson et al. 2015). Furthermore, maternally inherited endosymbionts can confound DNA barcoding efforts via selective sweeps that simultaneously decrease mtDNA variation within infected populations and increase mtDNA variation between infected and uninfected populations (Shoemaker et al. 2004; Delgado & Cook 2009; Xiao et al. 2012). Interestingly, however, while Shoemaker et al. (2003) found a correlation between Wolbachia infection and host mtDNA lineages, this was not coupled with a reduction of mtDNA diversity within host lineages. In contrast, Wolbachia has also been known to facilitate hybrid introgression, resulting in shared mtDNA barcodes among genetically distinct species (Klopfstein et al. 2016).

The issues surrounding the effects of inherited endosymbionts on DNA barcoding efforts are especially pertinent to fig wasps as they have among the highest Wolbachia infection frequencies of all insects (Shoemaker et al. 2002; Haine & Cook 2005). Common Wolbachia infections are known to be shared across several P. imperialis species (Haine et al. 2006); however, Wolbachia infection does not appear to be a confounding factor in the molecular taxonomy of this species complex (Chapter 4; Haine et al. 2006; Darwell et al. 2014). The results presented in Chapter 4 thus represent an important step forward in our understanding of fig wasp diversity, ecology and evolution in that they validate this widely used method of DNA barcoding for the discovery of cryptic fig wasp species.

117

Nevertheless, Chapter 5 highlights the limitations of DNA barcoding using one or a few genetic markers for species delimitation of allopatric taxa. Particularly for recently diverged populations, it is important to consider applying multilocus approaches using a variety of genetic markers (Dupuis et al. 2012). The use of mtDNA and nuclear DNA sequences and microsatellites in Chapter 5 is a very good illustration of this. Although different markers may tell a different story when considered independently, taken together, they may provide a strong insight into ongoing processes of divergence and speciation.

Although I present strong support for species status among allopatric P. imperialis sp. 1 populations, in reality, it cannot be definitively shown that they are indeed reproductively isolated. It is likely, however, that over longer evolutionary timeframes these populations will continue to diverge genetically (and perhaps ecologically) to the point where their species status is irrefutable.

7.1.3 The resilience of temperate fig-pollinating wasp species to climate change

In Chapter 6, I assessed the resilience of temperate fig-pollinating wasps to extreme weather events. Specifically, I recorded wasp longevity under different temperature × humidity regimes, and quantified wasp emergence from the natal fig at increasing temperatures. I also took field measurements to compare the internal temperature of figs (still growing on trees) exposed to direct sunlight with those protected by shade.

My results support my hypotheses that (i) extreme temperatures can cause a severe reduction in pollinator emergence, (ii) longevity is reduced under similar extreme temperatures (but not at temperatures only slightly above summer mean maximum), and (iii) longevity decreases under reduced humidity. I showed that pollinator emergence was significantly reduced at 39-42°C, but relatively unaffected at temperatures up to this point. Furthermore, wasp longevity was severely reduced at extreme temperatures (>5°C above summer mean maximum) and under low relative humidity conditions (across all temperatures). These results suggest that, unlike for tropical fig wasps (Jevanandam et al. 2013), small increases in mean temperature alone are not likely to have severe impacts on temperate fig wasp populations. In fact, similar to some tropical fig wasp species (Dunn et al. 2008b; Wang et al. 2013),

118

variation in humidity had a greater impact on longevity than a 5°C increase in temperature. Temperate fig wasp species are, however, likely to suffer due to the predicted increase in frequency and intensity of extreme heat events (Hughes 2003; Steffen 2015). Alarmingly, I observed severe reductions in emergence and longevity at temperatures currently experienced during Australian summers (Bureau of Meteorology 2016), and found that critical temperatures for both wasp emergence and longevity were in the range of 39-42°C. The frequency of days in this temperature range in the Sydney region is forecast to increase; in fact, from December 2015 to February 2016, six days with temperature maximums of 39°C or higher were recorded in Penrith (one of the sites used in this study), including one instance of temperatures exceeding 42°C (Bureau of Meteorology 2016).

Consistent with my hypothesis, I present evidence that sun-exposed figs experience hotter internal temperatures (+0.94°C) than do shaded figs. While this difference does not seem great at face value, it becomes substantial when one considers that an increase of just 1°C could potentially push internal fig temperatures over the critical temperature for fig wasp emergence. Consequently, fig wasps contained inside figs that are protected by shade may be at an advantage during periods of intense heat. This could be a good example of the wider idea that spatial microhabitat variation can provide important refugia for some individuals in populations under stress.

Logistical constraints on laboratory space and the location of field sites, and the availability of ripe figs, dictated the timing of these experiments. Consequently, I was forced to conduct each experiment at a different time point and use figs sourced from different sites around the Sydney area. Furthermore, samples used to test the effect of temperature on wasp emergence were exposed to periods of cold temperatures to reduce emergence during transportation. The resilience of insects to climatic variation is known to be affected by genetic and environmental (i.e. during development) conditions (Krebs & Feder 1997; Krebs et al. 2001), and so it may be difficult to disentangle the relative influence of genetics, developmental conditions and experimental pre-treatment on my results. Nevertheless, these experiments provide important insights into the potential fate of temperate fig wasp species under predicted climate change and provide the foundation for future research that takes the

119

study of climate stress effects on fig wasps beyond the simple exposure of hatched adults to different conditions in controlled laboratory environments.

7.2 Future research

Interspecific interactions are known to evolve along different paths in isolated conspecific populations (Carroll et al. 1997; Benkman 1999; Brodie et al. 2002; Parchman & Benkman 2002). Carroll et al. (1997), for example, revealed fitness- associated morphological differences among populations of the soapberry bug on an introduced host. Observed discordance in the genetic structure of P. imperialis sp. 1 and Sycoscapter sp. A provides an excellent opportunity to compare the ecological and evolutionary responses of isolated conspecific populations to a common antagonist. This approach could also be expanded to the other P. imperialis species, and would improve our understanding of their diversification and the factors that facilitate their coexistence in many regions. Moreover, variation in the ecological response of pollinator populations (or even species) to parasitism pressure requires further attention. Parasitoids exert a strong selective pressure on host populations that manifest in their ecology. Dunn et al. (2008a), for example, revealed that female P. imperialis wasps typically favour oviposition in long-styled inner ovules that are less accessible to parasitoids. Despite including multiple P. imperialis complex species in their analyses, they did not test for differences in ovule preferences between species, due to difficulties in distinguishing between morphologically cryptic wasp species. Future studies should address the ecological impacts of parasitism and how it varies across populations.

However, in order to truly improve our view of diversification within the P. imperialis species complex, a thorough grasp of the phylogenetic relationships and distinguishing taxonomic characters is required. The former is absolutely critical for the formation and testing of hypotheses regarding mechanisms of speciation in this system. Genetic studies within this complex have so far been limited to DNA barcoding (Haine et al. 2006; Darwell et al. 2014) and population genetics of one or a few species in isolation (Chapters 3-5). It is likely that robust phylogenetic analyses will require genomic approaches, such as restriction site-associated DNA sequencing

120

(RAD-seq), due to the recent nature of diversification in P. imperialis (Darwell 2012). Indeed, evidence supports the reliability of RAD-seq in phylogenetic studies (Cariou et al. 2013; Cruaud et al. 2014; Gonen et al. 2015), even in problematic taxa (Hou et al. 2015) and recent radiations (Wagner et al. 2013). An improved taxonomy of P. imperialis species would no doubt improve the efficiency with which ecological studies can proceed. Subtle differences in head morphology and melanism between some P. imperialis species (J.-Y. Rasplus, personal communication) offer promise that a better taxonomic revision is also achievable.

Warren et al. (2010) demonstrated that geographic distribution in two African Ceratosolen fig wasps is dictated by physiological limits on thermal tolerance and desiccation resistance. Darwell (2012) used environmental modelling to predict the distribution of P. imperialis species, and subsequently suggested that tests of physiological limits may help to explain geographic distribution in P. imperialis. Interestingly, the P. imperialis species complex comprises some species that occupy purely temperate (P. imperialis sp. 1) and tropical (P. imperialis sp. 2) environments, as well as more widespread species found throughout tropical and subtropical climate zones (P. imperialis sp. 3 and 4). There is potential for reciprocal climate experiments to reveal evidence of adaptation among P. imperialis species by measuring performance of different species under various temperature × humidity conditions. Furthermore, similar experiments on conspecific individuals from different locations could elucidate the effects of local environmental conditions to thermal tolerance. The P. imperialis species complex could thus be an important model for understanding species distributions, and could potentially challenge Rapoport’s Rule and the Climate Variability Hypothesis (Stevens 1989).

Furthermore, an extension of studies presented in Chapter 6 to other P. imperialis species would provide invaluable information on the resilience of this pollination system to climate change. More realistic predictions could be made if (i) tests of wasp emergence at different temperatures were made in the field and (ii) fig core temperature could be monitored over a range of ambient temperatures, including those that could be considered ‘extreme’. Given the effect of extreme temperatures on fig-pollinating wasp emergence and longevity, it would be useful to understand how NPFWs might cope under similar conditions. Preliminary data suggest that they

121

live much longer than pollinators at both ambient and high temperatures (T.L. Sutton & J.L. DeGabriel, unpublished data). The widespread knockdown of pollinator populations may have dire consequences for their parasitoids (due to lack of hosts); however, our observations suggest that large gallers are among the most resilient to heat stress in the F. rubiginosa NPFW community. Not only can these species reproduce in figs that are not pollinated, but they can also release themselves from the natal fig without the help of male fig-pollinating wasps (Segar et al. 2014). This could be due to their larger size and smaller surface area to volume ratio, which may reduce water loss under low humidity conditions, giving them an advantage over pollinators. Incorporating NPFWs into these studies would provide a more complete overview on the fate of the fig – fig wasp system under predicted climate change.

7.3 Concluding remarks

The research presented in this thesis advances our understanding of the ecology and evolution of fig wasps. More broadly, it promotes population genetic approaches to genetic studies of coevolution and biodiversity by demonstrating their sensitivity at detecting important evolutionary processes at fine spatial and temporal scales. Importantly, these approaches are crucial to enhancing our knowledge of species diversity – a necessary first step in making predictions about the fate of biodiversity under climate change, and essential for the development of mitigation strategies.

122

References

Adamo SA, Lovett MME (2011) Some like it hot: the effects of climate change on reproduction, immune function and disease resistance in the cricket Gryllus texensis. Journal of Experimental Biology, 214, 1997-2004. Addo-Bediako A, Chown SL, Gaston KJ (2000) Thermal tolerance, climatic variability and latitude. Proceedings of the Royal Society B: Biological Sciences, 267, 739-745. Agrawal AA, Fishbein M (2006) Plant defense syndromes. Ecology, 87, S132-S149. Ahmed S, Compton S, Butlin R, Gilmartin P (2009) Wind-borne insects mediate directional pollen transfer between desert fig trees 160 kilometers apart. Proceedings of the National Academy of Sciences of the United States of America, 106, 20342-20347. Aigner PA (2005) Variation in pollination performance gradients in a Dudleya species complex: can generalization promote floral divergence? Functional Ecology, 19, 681-689. Allen JC, McCoy CW (1979) The thermal environment of the citrus rust mite. Agricultural Meteorology, 20, 411-425. Althoff DM, Thompson JN (1999) Comparative geographic structures of two parasitoid-host interactions. Evolution, 53, 818-825. Alvarez N, McKey D, Kjellberg F, Hossaert-McKey M (2010) Phylogeography and historical biogeography of obligate specific mutualisms. In: The biogeography of host – parasite interactions (eds. Morand S, Krasnov B), p. 31. Oxford University Press, Oxford, UK. Anderson B, Johnson SD (2009) Geographical covariation and local convergence of flower depth in a guild of fly-pollinated plants. New Phytologist, 182, 533- 540. Anton C, Zeisset I, Musche M, et al. (2007) Population structure of a large blue butterfly and its specialist parasitoid in a fragmented landscape. Molecular Ecology, 16, 3828-3838. Avise JC (2000) Phylogeography: the history and formation of species Harvard university press, Cambridge, MA. Avise JC, Arnold J, Ball R, et al. (1987) Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual review of ecology and systematics, 18, 489-522. Baker H (1961) Ficus and Blastophaga. Evolution, 15, 378-379. Barbour RC, Potts BM, Vaillancourt RE, Tibbits WN, Wiltshire RJE (2002) Gene flow between introduced and native Eucalyptus species. New Forests, 23, 177-191. Bastos-Silveira C, Santos SM, Monarca R, Mathias MDL, Heckel G (2012) Deep mitochondrial introgression and hybridization among ecologically divergent vole species. Molecular Ecology, 21, 5309-5323. Bates D, Mächler M, Bolker B, Walker S (2014) Fitting linear mixed-effects models using lme4. arXiv preprint,, arXiv:1406.5823. Bates D, Maechler M, Bolker B, Walker S (2013) lme4: linear mixed-effects models using Eigen and S4, R package version 1.1.

123

Bauerfeind SS, Fischer K (2014) Simulating climate change: temperature extremes but not means diminish performance in a widespread butterfly. Population Ecology, 56, 239-250. Baum DA, Donoghue MJ (1995) Choosing among alternative "phylogenetic" species concepts. Systematic botany, 20, 560-573. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 57, 289-300. Benkman CW (1999) The selection mosaic and diversifying coevolution between crossbills and lodgepole pine. American Naturalist, 153, S75-S91. Berg C (1989) Classification and distribution of Ficus. Experientia, 45, 605-611. Berlocher S, Feder J (2002) Sympatric speciation in phytophagous insects: moving beyond controversy? Annual Review of Entomology, 47, 773-815. Bernasconi C, Cherix D, Seifert B, Pamilo P (2010) Molecular taxonomy of the Formica rufa group (red wood ants) (Hymenoptera: Formicidae): a new cryptic species in the Swiss Alps? Myrmecological News, 14, 37-47. Bickford D, Lohman DJ, Sodhi NS, et al. (2007) Cryptic species as a window on diversity and conservation. Trends in Ecology and Evolution, 22, 148-155. Biedrzycka A, Solarz W, Okarma H (2012) Hybridization between native and introduced species of deer in Eastern Europe. Journal of Mammalogy, 93, 1331-1341. Blaxter ML (2004) The promise of a DNA taxonomy. Philosophical Transactions of the Royal Society B: Biological Sciences, 359, 669-679. Bouteiller-Reuter C, Hale M, Cook J (2009) Characterization of microsatellite markers for Sycoscapter nonpollinating fig wasps. Molecular Ecology Resources, 9, 832-835. Brittain C, Kremen C, Klein AM (2013) Biodiversity buffers pollination from changes in environmental conditions. Global Change Biology, 19, 540-547. Brodie EDJ, Ridenhour BJ, Brodie EDI (2002) The evolutionary response of predators to dangerous prey: hotspots and coldspots in the geographic mosaic of coevolution between garter snakes and newts. Evolution, 56, 2067-2082. Bronstein JL (1991) The nonpollinating wasp fauna of Ficus pertusa: exploitation of a mutualism? Oikos, 61, 175-186. Bronstein JL, Hossaert-Mckey M (1995) Hurricane Andrew and a Florida fig pollination mutualism: resilience of an obligate interaction. Biotropica, 27, 373-381. Broufas GD, Pappas ML, Koveos DS (2009) Effect of relative humidity on longevity, ovarian maturation, and egg production in the olive fruit fly (Diptera: Tephritidae). Annals of the Entomological Society of America, 102, 70-75. Brower AVZ (2006) Problems with DNA barcodes for species delimitation: 'ten species' of Astraptes fulgerator reassessed (Lepidoptera: Hesperiidae). Systematics and Biodiversity, 4, 127-132. Brown WM, George Jr M, Wilson AC (1979) Rapid evolution of animal mitochondrial DNA. Proceedings of the National Academy of Sciences of the United States of America, 76, 1967-1971. Bull JK, Sands CJ, Garrick RC, et al. (2013) Environmental complexity and biodiversity: the multi-layered evolutionary history of a log-dwelling velvet worm in montane temperate Australia. PLoS ONE, 8, (e84559).

124

Bull JK, Sunnucks P (2014) Strong genetic structuring without assortative mating or reduced hybrid survival in an onychophoran in the Tallaganda State Forest region, Australia. Biological Journal of the Linnean Society, 111, 589-602. Bureau of Meteorology (2016) Climate Data Online. Australian Government. http://www.bom.gov.au/climate/data/ Accessed March 1, 2016. Cariou M, Duret L, Charlat S (2013) Is RAD-seq suitable for phylogenetic inference? An in silico assessment and optimization. Ecology and Evolution, 3, 846-852. Carroll SP, Dingle H, Klassen SP (1997) Genetic differentiation of fitness-associated traits among rapidly evolving populations of the soapberry bug. Evolution, 51, 1182-1188. Chaves-Campos J, Johnson SG, Hulsey CD (2011) Spatial geographic mosaic in an aquatic Predator-Prey network. PLoS ONE, 6, (e22472). Chen C, Durand E, Forbes F, François O (2007) Bayesian clustering algorithms ascertaining spatial population structure: a new computer program and a comparison study. Molecular Ecology Notes, 7, 747-756. Chen Y, Compton S, Liu M, Chen X (2012) Fig trees at the northern limit of their range: the distributions of cryptic pollinators indicate multiple glacial refugia. Molecular Ecology, 21, 1687-1701. Chesson P (1991) A need for niches? Trends in Ecology and Evolution, 6, 26-28. Chiu MC, Kuo JJ, Kuo MH (2015) Life stage-dependent effects of experimental heat waves on an insect herbivore. Ecological Entomology, 40, 175-181. Clare EL (2011) Cryptic species? patterns of maternal and paternal gene flow in eight Neotropical bats. PLoS ONE, 6, (e21460). Collins M, Knutti R, Arblaster J, et al. (2013) Long-term climate change: projections, commitments and irreversibility. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the fifth assessment report of the Intergovernmental Panel on Climate Change (eds. Stocker T, Qin D, Plattner G-K, et al.), pp. 1029-1136. Cambridge University Press, Cambridge, UK and New York, USA. Compton S, Ellwood M, Davis A, Welch K (2000) The flight heights of chalcid wasps (Hymenoptera, Chalcidoidea) in a lowland Bornean rain forest: fig wasps are the high fliers. Biotropica, 32, 515-522. Cook J, Bean D, Power S, Dixon D (2004) Evolution of a complex coevolved trait: active pollination in a genus of fig wasps. Journal of Evolutionary Biology, 17, 238-246. Cook J, Power S (1996) Effects of within-tree flowering asynchrony on the dynamics of seed and wasp production in an Australian fig species. Journal of Biogeography, 23, 487-493. Cook J, Segar S (2010) Speciation in fig wasps. Ecological Entomology, 35, 54-66. Cook J, West S (2005) Figs and fig wasps. Current Biology, 15, R978-980. Cornille A, Underhill J, Cruaud A, et al. (2012) Floral volatiles, pollinator sharing and diversification in the fig-wasp mutualism: insights from Ficus natalensis, and its two wasp pollinators (South Africa). Proceedings of the Royal Society B: Biological Sciences, 279, 1731-1739. Cosacov A, Nattero J, Cocucci AA (2008) Variation of pollinator assemblages and pollen limitation in a locally specialized system: the oil-producing Nierembergia linariifolia (Solanaceae). Annals of Botany, 102, 723-734. Coyne J, Orr H (2004) Speciation Sinauer Associates, Inc., Sunderland, MA.

125

Cracraft J (1989) Speciation and its ontology: the empirial consequences of alternative species concepts for understanding patterns and processes of differentiation. In: Speciation and its consequences (eds. Otte D, Endler J). Sinauer Associates, Sunderland. Cruaud A, Gautier M, Galan M, et al. (2014) Empirical assessment of rad sequencing for interspecific phylogeny. Molecular Biology and Evolution, 31, 1272- 1274. Cruaud A, Ronsted N, Chantarasuwan B, et al. (2012) An extreme case of plant- insect codiversification: figs and fig-pollinating wasps. Systematic Biology, 61, 1029-1047. CSIRO and Bureau of Meteorology (2015) Climate change in Australia website. http://www.climatechangeinaustralia.gov.au/en/climate-projections/explore- data/extremes-data-explorer/ Accessed May 25, 2015. Darriba D, Taboada GL, Doallo R, Posada D (2012) jModelTest 2: more models, new heuristics and parallel computing. Nature Methods, 9, 772. Darwell C (2012) Species composition and patterns of diversity in fig wasp communities. PhD thesis, University of Reading. Darwell C, Al-Beidh S, Cook J (2014) Molecular species delimitation of a symbiotic fig-pollinating wasp species complex reveals extreme deviation from reciprocal partner specificity. BMC Evolutionary Biology, 14, (189). Darwin C (1859) On the origin of species by means of natural selection, or the preservation of favored races in the struggle for life John Murray, London. de Queiroz K (2005) Ernst Mayr and the modern concept of species. Proceedings of the National Academy of Sciences of the United States of America, 102, 6600- 6607. de Queiroz K (2007) Species concepts and species delimitation. Systematic Biology, 56, 879-886. de Queiroz K, Donoghue MJ (1988) Phylogenetic systematics and the species problem. Cladistics, 4, 317-338. Delgado AM, Cook JM (2009) Effects of a sex-ratio distorting endosymbiont on mtDNA variation in a global insect pest. BMC Evolutionary Biology, 9, (49). Dépraz A, Hausser J, Pfenninger M (2009) A species delimitation approach in the Trochulus sericeus/hispidus complex reveals two cryptic species within a sharp contact zone. BMC Evolutionary Biology, 9, (171). Deutsch CA, Tewksbury JJ, Huey RB, et al. (2008) Impacts of climate warming on terrestrial ectotherms across latitude. Proceedings of the National Academy of Sciences of the United States of America, 105, 6668-6672. Dieckmann U, Doebeli M (1999) On the origin of species by sympatric speciation. Nature, 400, 354-357. Dixon D, Jackes B, Bielig L (2001) Figuring out the figs: The Ficus obliqua - Ficus rubiginosa complex (Moraceae: Urostigma sect. Malvanthera). Australian Systematic Botany, 14, 133-154. Dobzhansky T (1935) A critique of the species concept in biology. Philosophy of Science, 2, 344-355. Doi H, Gordo O, Katano I (2008) Heterogeneous intra-annual climatic changes drive different phenological responses at two trophic levels. Climate Research, 36, 181-190. Donnelly RK, Harper GL, Morgan AJ, et al. (2013) Nuclear DNA recapitulates the cryptic mitochondrial lineages of Lumbricus rubellus and suggests the

126

existence of cryptic species in an ecotoxological soil sentinel. Biological Journal of the Linnean Society, 110, 780-795. Downey MH, Nice CC (2011) Experimental evidence of host race formation in Mitoura butterflies (Lepidoptera: Lycaenidae). Oikos, 120, 1165-1174. Dray S, Dufour AB (2007) The ade4 package: implementing the duality diagram for ecologists. Journal of Statistical Software, 22, 1-20. Duminil J, Kenfack D, Viscosi V, Grumiau L, Hardy OJ (2012) Testing species delimitation in sympatric species complexes: the case of an African tropical tree, Carapa spp. (Meliaceae). Molecular Phylogenetics and Evolution, 62, 275-285. Dunn D, Segar S, Ridley J, et al. (2008a) A role for parasites in stabilising the fig- pollinator mutualism. PLoS biology, 6, 490-496. Dunn D, Yu D, Ridley J, Cook J (2008b) Longevity, early emergence and body size in a pollinating fig wasp - implications for stability in a fig-pollinator mutualism. Journal of Animal Ecology, 77, 927-935. Dupuis JR, Roe AD, Sperling FAH (2012) Multi-locus species delimitation in closely related animals and fungi: one marker is not enough. Molecular Ecology, 21, 4422-4436. Durand E, Jay F, Gaggiotti OE, François O (2009) Spatial inference of admixture proportions and secondary contact zones. Molecular Biology and Evolution, 26, 1963-1973. Ebach MC, Holdrege C (2005) DNA barcoding is no substitute for taxonomy. Nature, 434, 697. Engeler B, Reyer HU (2001) Choosy females and indiscriminate males: mate choice in mixed populations of sexual and hybridogenetic water frogs (Rana lessonae, Rana esculenta). Behavioral Ecology, 12, 600-606. Ereshefsky M (1992) Eliminative pluralism. Philosophy of Science, 671-690. Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources, 10, 564-567. Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics, 131, 479-491. Feder ME, Blair N, Figueras H (1997) Natural thermal stress and heat-shock protein expression in Drosophila larvae and pupae. Functional Ecology, 11, 90-100. Fenster C, Armbruster W, Wilson P, Dudash M, Thomson J (2004) Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution and Systematics, 35, 375-403. Ferguson JWH (2002) On the use of genetic divergence for identifying species. Biological Journal of the Linnean Society, 75, 509-516. Fischer K, Klockmann M, Reim E (2014) Strong negative effects of simulated heat waves in a tropical butterfly. Journal of Experimental Biology, 217, 2892- 2898. Forrest JRK (2015) Plant–pollinator interactions and phenological change: what can we learn about climate impacts from experiments and observations? Oikos, 124, 4-13. Fournier DA, Skaug HJ, Ancheta J, et al. (2012) AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optimization Methods and Software, 27, 233-249.

127

François O, Durand E (2010) Spatially explicit Bayesian clustering models in population genetics. Molecular Ecology Resources, 10, 773-784. Fu YX (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics, 147, 915-925. Fujisawa T, Barraclough TG (2013) Delimiting species using single-locus data and the generalized mixed yule coalescent approach: a revised method and evaluation on simulated data sets. Systematic Biology, 62, 707-724. Galil J, Eisikowitch D (1968) On the pollination ecology of Ficus sycomorus in East Africa. Ecology, 49, 259-269. Galil J, Neeman G (1977) Pollen transfer and pollination in the common fig (Ficus carica L.). New Phytologist, 79, 163-171. Gaston KJ (2009) Geographic range limits of species. Proceedings of the Royal Society B: Biological Sciences, 276, 1391-1393. Gebiola M, Lopez-Vaamonde C, Nappo AG, Bernardo U (2014) Did the parasitoid Pnigalio mediterraneus (Hymenoptera: Eulophidae) track the invasion of the horse chestnut leafminer? Biological Invasions, 16, 843-857. Ghara M, Kundanati L, Borges R (2011) Nature's swiss army knives: ovipositor structure mirrors ecology in a multitrophic fig wasp community. PLoS ONE, 6, (e23642). Gibson L, McNeill A, Tores PD, Wayne A, Yates C (2010) Will future climate change threaten a range restricted endemic species, the quokka (Setonix brachyurus), in south west Australia? Biological Conservation, 143, 2453- 2461. Gilchrist GW (1995) Specialists and generalists in changing environments. I. Fitness landscapes of thermal sensitivity. American Naturalist, 146, 252-270. Gillespie DR, Nasreen A, Moffat CE, Clarke P, Roitberg BD (2012) Effects of simulated heat waves on an experimental community of pepper plants, green peach aphids and two parasitoid species. Oikos, 121, 149-159. Giska I, Sechi P, Babik W (2015) Deeply divergent sympatric mitochondrial lineages of the earthworm Lumbricus rubellus are not reproductively isolated. BMC Evolutionary Biology, 15, (488). Godfray H (1994) Parasitoids: behavioral and evolutionary ecology Princeton University Press, Princeton. Gómez A, Wright PJ, Lunt DH, et al. (2007) Mating trials validate the use of DNA barcoding to reveal cryptic speciation of a marine bryozoan taxon. Proceedings of the Royal Society B: Biological Sciences, 274, 199-207. Gómez JM, Abdelaziz M, Camacho JPM, Muñoz-Pajares AJ, Perfectti F (2009a) Local adaptation and maladaptation to pollinators in a generalist geographic mosaic. Ecology Letters, 12, 672-682. Gómez JM, Perfectti F, Bosch J, Camacho JPM (2009b) A geographic selection mosaic in a generalized plant-pollinator-herbivore system. Ecological Monographs, 79, 245-263. Gonen S, Bishop SC, Houston RD (2015) Exploring the utility of cross-laboratory RAD-sequencing datasets for phylogenetic analysis. BMC Research Notes, 8, (299). Gordo O, Sanz JJ (2005) Phenology and climate change: a long-term study in a Mediterranean locality. Oecologia, 146, 484-495. Goudet J (1995) FSTAT (version 1.2): a computer program to calculate F-statistics. Journal of Heredity, 86, 485-486.

128

Goudet J, Raymond M, De Meeüs T, Rousset F (1996) Testing differentiation in diploid populations. Genetics, 144, 1933-1940. Grant WS (2015) Problems and cautions with sequence mismatch analysis and Bayesian skyline plots to infer historical demography. Journal of Heredity, 106, 333-346. Greeff J, Jansen Van Vuuren G, Kryger P, Moore J (2009) Outbreeding and possibly inbreeding depression in a pollinating fig wasp with a mixed mating system. Heredity, 102, 349-356. Greene W (1994) Accounting for excess zeros and sample selection in Poisson and negative binomial regression (Working paper EC-94-10). New York University, New York. Griffiths KE, Trueman JWH, Brown GR, Peakall R (2011) Molecular genetic analysis and ecological evidence reveals multiple cryptic species among thynnine wasp pollinators of sexually deceptive orchids. Molecular Phylogenetics and Evolution, 59, 195-205. Grillenberger BK, Gadau J, Bijlsma R, Van De Zande L, Beukeboom LW (2009) Female dispersal and isolation-by-distance of Nasonia vitripennis populations in a local mate competition context. Entomologia Experimentalis et Applicata, 132, 147-154. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology, 52, 696-704. Haine E, Cook J (2005) Convergent incidences of Wolbachia infection in fig wasp communities from two continents. Proceedings of the Royal Society B: Biological Sciences, 272, 421-429. Haine E, Martin J, Cook J (2006) Deep mtDNA divergences indicate cryptic species in a fig-pollinating wasp. BMC Evolutionary Biology, 6, (83). Hale ML, Squirrell J, Borland AM, Wolff K (2002) Isolation of polymorphic microsatellite loci in the genus Clusia (Clusiaceae). Molecular Ecology Notes, 2, 506-508. Hall AAG, Gherlenda AN, Hasegawa S, et al. (2015) Anatomy of an outbreak: the biology and population dynamics of a Cardiaspina psyllid species in an endangered woodland ecosystem. Agricultural and Forest Entomology. Hamilton CA, Hendrixson BE, Brewer MS, Bond JE (2014) An evaluation of sampling effects on multiple DNA barcoding methods leads to an integrative approach for delimiting species: a case study of the North American tarantula genus Aphonopelma (Araneae, Mygalomorphae, Theraphosidae). Molecular Phylogenetics and Evolution, 71, 79-93. Hamilton W (1979) Wingless and fighting males in fig wasps and other insects. In: Reproductive competition and sexual selection in insects (eds. Blum M, Blum N), pp. 167-220. Academic Press, New York. Hao Y, Tong-Xin Z, Hao-Yuan H, et al. (2008) Isolation of microsatellite loci in the pollinating fig wasp of Ficus hispida, Ceratosolen solmsi. Insect Science, 15, 303-306. Hardy OJ, Charbonnel N, Fréville H, Heuertz M (2003) Microsatellite allele sizes: a simple test to assess their significance on genetic differentiation. Genetics, 163, 1467-1482. Hardy OJ, Vekemans X (1999) Isolation by distance in a continuous population: reconciliation between spatial autocorrelation analysis and population genetics models. Heredity, 83, 145-154.

129

Hardy OJ, Vekemans X (2002) SPAGeDI: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes, 2, 618-620. Harrison R (2000) Repercussions of El Nino: drought causes extinction and the breakdown of mutualism in Borneo. Proceedings of the Royal Society B: Biological Sciences, 267, 911-915. Harrison R (2001) Drought and the consequences of El Niño in Borneo: a case study of figs. Population Ecology, 43, 63-75. Harrison R (2003) Fig wasp dispersal and the stability of a keystone plant resource in Borneo. Proceedings of the Royal Society B: Biological Sciences, 270, S76- S79. Harrison R, Rasplus J (2006) Dispersal of fig pollinators in Asian tropical rain forests. Journal of Tropical Ecology, 22, 631-639. Harrison RD (2005) Figs and the diversity of tropical rainforests. BioScience, 55, 1053-1064. Hasegawa M, Kishino H, Yano Ta (1985) Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution, 22, 160-174. Hausdorf B, Wilkens H, Strecker U (2011) Population genetic patterns revealed by microsatellite data challenge the mitochondrial DNA based taxonomy of Astyanax in Mexico (Characidae, Teleostei). Molecular Phylogenetics and Evolution, 60, 89-97. Hayward A, Stone GN (2006) Comparative phylogeography across two trophic levels: The oak gall wasp Andricus kollari and its chalcid parasitoid Megastigmus stigmatizans. Molecular Ecology, 15, 479-489. Hebert PDN, Cywinska A, Ball SL, DeWaard JR (2003a) Biological identifications through DNA barcodes. Proceedings of the Royal Society B: Biological Sciences, 270, 313-321. Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W (2004) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proceedings of the National Academy of Sciences of the United States of America, 101, 14812-14817. Hebert PDN, Ratnasingham S, DeWaard JR (2003b) Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society B: Biological Sciences, 270, S96-S99. Hegland SJ, Nielsen A, Lázaro A, Bjerknes AL, Totland Ø (2009) How does climate warming affect plant-pollinator interactions? Ecology Letters, 12, 184-195. Heller R, Chikhi L, Siegismund HR (2013) The confounding effect of population structure on Bayesian skyline plot inferences of demographic history. PLoS ONE, 8, (e62992). Herre E, Machado C, Bermingham E, et al. (1996) Molecular phylogenies of figs and their pollinator wasps. Journal of Biogeography, 23, 521-530. Herre E, West S, Cook J, Compton S, Kjellberg F (1997) Fig wasp mating systems: pollinators and parasites, sex ratio adjustment and male polymorphism, population structure and its consequences. In: The evolution of mating systems in insects and arachnids (eds. Choe J, Crespi B), pp. 226-239. Cambridge University Press, Cambridge. Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society B: Biological Sciences, 359, 183-195.

130

Heywood JS (1991) Spatial analysis of genetic variation in plant populations. Annual Review of Ecology and Systematics, 22, 335-355. Hickerson MJ, Meyer CP, Moritz C (2006) DNA barcoding will often fail to discover new animal species over broad parameter space. Systematic Biology, 55, 729-739. Hossaert-McKey M, Gibernau M, Frey J (1994) Chemosensory attraction of fig wasps to substances produced by receptive figs. Entomologia Experimentalis et Applicata, 70, 185-191. Hou Y, Nowak MD, Mirré V, et al. (2015) Thousands of RAD-seq loci fully resolve the phylogeny of the highly disjunct arctic-alpine genus Diapensia (Diapensiaceae). PLoS ONE, 10, (e0140175). Hubby JL, Lewontin RC (1966) A molecular approach to the study of genic heterozygosity in natural populations. I. The number of alleles at different loci in Drosophila pseudoobscura. Genetics, 54, 577-594. Hubert N, Meyer CP, Bruggemann HJ, et al. (2012) Cryptic diversity in Indo-Pacific coral-reef fishes revealed by DNA-barcoding provides new support to the centre-of-overlap hypothesis. PLoS ONE, 7, (e28987). Huey RB, Stevenson RD (1979) Integrating thermal physiology and ecology of ectotherms: A discussion of approaches. Integrative and Comparative Biology, 19, 357-366. Hughes L (2003) Climate change and Australia: trends, projections and impacts. Austral Ecology, 28, 423-443. IBM Corp (2013) IBM SPSS Statistics for Windows, Jackman S (2015) pscl: classes and methods for R developed in the Political Science Computational Laboratory, Jackson A (2004) Cophylogeny of the Ficus microcosm. Biological Reviews of the Cambridge Philosophical Society, 79, 751-768. Jackson A, Machado C, Robbins N, Herre E (2008) Multi-locus phylogenetic analysis of neotropical figs does not support co-speciation with the pollinators: the importance of systematic scale in fig/wasp cophylogenetic studies. Symbiosis, 45, 57-72. Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics, 23, 1801-1806. Jandér K, Herre E (2010) Host sanctions and pollinator cheating in the fig tree-fig wasp mutualism. Proceedings of the Royal Society B: Biological Sciences, 277, 1481-1488. Janzen DH (1979) How to be a fig. Annual Review of Ecology and Systematics, 10, 13-51. Jeffs CT, Leather SR (2014) Effects of extreme, fluctuating temperature events on life history traits of the grain aphid, Sitobion avenae. Entomologia Experimentalis et Applicata, 150, 240-249. Jentsch A, Kreyling J, Beierkuhnlein C (2007) A new generation of climate-change experiments: events, not trends. Frontiers in Ecology and the Environment, 5, 365-374. Jermiin L, Crozier R (1994) The cytochrome b region in the mitochondrial DNA of the ant Tetraponera rufoniger: sequence divergence in Hymenoptera may be associated with nucleotide content. Journal of Molecular Evolution, 38, 282- 294.

131

Jevanandam N, Goh AGR, Corlett RT (2013) Climate warming and the potential extinction of fig wasps, the obligate pollinators of figs. Biology Letters, 9, (20130041). Johannesen J, Seitz A (2003) Comparative population genetic structures of the fruit fly Urophora cardui and its primary parasitoid Eurytoma robusta. Entomologia Experimentalis et Applicata, 108, 149-157. Joly S, Davies TJ, Archambault A, et al. (2014) Ecology in the age of DNA barcoding: the resource, the promise and the challenges ahead. Molecular Ecology Resources, 14, 221-232. Jones M, Ghoorah A, Blaxter M (2011) jMOTU and taxonerator: turning DNA barcode sequences into annotated operational taxonomic units. PLoS ONE, 6, Jourdie V, Alvarez N, Molina-Ochoa J, et al. (2010) Population genetic structure of two primary parasitoids of Spodoptera frugiperda (Lepidoptera), Chelonus insularis and Campoletis sonorensis (Hymenoptera): to what extent is the host plant important? Molecular Ecology, 19, 2168-2179. Jousselin E, Kjellberg F (2001) The functional implications of active and passive pollination in dioecious figs. Ecology Letters, 4, 151-158. Jousselin E, Rasplus JY, Kjellberg F (2003) Convergence and coevolution in a mutualism: evidence from a molecular phylogeny of Ficus. Evolution, 57, 1255-1269. Jukes T, Cantor C (1969) Evolution in protein molecules. In: Mammalian protein metabolism (ed. Munro H). Academic Press, New York. Kahilainen KK, Østbye K, Harrod C, et al. (2011) Species introduction promotes hybridization and introgression in Coregonus: is there sign of selection against hybrids? Molecular Ecology, 20, 3838-3855. Kalinowski ST (2011) The computer program STRUCTURE does not reliably identify the main genetic clusters within species: simulations and implications for human population structure. Heredity, 106, 625-632. Kankare M, van Nouhuys S, Gaggiotti O, Hanski I (2005) Metapopulation genetic structure of two coexisting parasitoids of the Glanville fritillary butterfly. Oecologia, 143, 77-84. Katabuchi M, Isagi Y, Nakashizuka T (2008) Development of 17 microsatellite markers for Ceratosolen constrictus, the pollinating fig wasp of Ficus fistulosa. Molecular Ecology Resources, 8, 840-842. Kerdelhue C, Le Clainche I, Rasplus J-Y (1999) Molecular phylogeny of the Ceratosolen species pollinating Ficus of the subgenus Sycomorus sensu stricto: biogeographical history and origins of the species-specificity breakdown cases. Molecular Phylogenetics and Evolution, 11, 401-414. Kerr KCR, Stoeckle MY, Dove CJ, et al. (2007) Comprehensive DNA barcode coverage of North American birds. Molecular Ecology Notes, 7, 535-543. Kingsolver JG, Diamond SE, Buckley LB (2013) Heat stress and the fitness consequences of climate change for terrestrial ectotherms. Functional Ecology, 27, 1415-1423. Kjellberg F, Jousselin E, Bronstein J, et al. (2001) Pollination mode in fig wasps: the predictive power of correlated traits. Proceedings of the Royal Society B: Biological Sciences, 268, 1113-1121. Kjellberg F, Proffit M (2016) Tracking the elusive history of diversification in plant- herbivorous insect-parasitoid food webs: insights from figs and fig wasps. Molecular Ecology, 25, 843-845.

132

Klopfstein S, Kropf C, Baur H (2016) Wolbachia endosymbionts distort DNA barcoding in the parasitoid wasp genus Diplazon (Hymenoptera: Ichneumonidae). Zoological Journal of the Linnean Society. Knowles LL, Carstens BC (2007) Delimiting species without monophyletic gene trees. Systematic Biology, 56, 887-895. Kobmoo N, Hossaert-Mckey M, Rasplus J, Kjellberg F (2010) Ficus racemosa is pollinated by a single population of a single agaonid wasp species in continental South-East Asia. Molecular Ecology, 19, 2700-2712. Kohnen A, Wissemann V, Brandl R (2011) No host-associated differentiation in the gall wasp Diplolepis rosae (Hymenoptera: Cynipidae) on three dog rose species. Biological Journal of the Linnean Society, 102, 369-377. Koressaar T, Remm M (2007) Enhancements and modifications of primer design program Primer3. Bioinformatics, 23, 1289-1291. Kraaijeveld AR, van Alphen JJM (1995) Geographic variation in capsulation ability of Drosophila melanogaster larvae and evidence for parasitoid-specific components. Evolutionary Ecology, 9, 10-17. Krebs RA, Feder ME (1997) Natural variation in the expression of the heat-shock protein hsp70 in a population of Drosophila melanogaster and its correlation with tolerance of ecologically relevant thermal stress. Evolution, 51, 173-179. Krebs RA, Roberts SP, Bettencourt BR, Feder ME (2001) Changes in thermotolerance and Hsp70 expression with domestication in Drosophila melanogaster. Journal of Evolutionary Biology, 14, 75-82. Kudo G (2014) Vulnerability of phenological synchrony between plants and pollinators in an alpine ecosystem. Ecological Research, 29, 571-581. Kudo G, Ida TY (2013) Early onset of spring increases the phenological mismatch between plants and pollinators. Ecology, 94, 2311-2320. Kusumi J, Su ZH (2014) Isolation and characterization of 15 polymorphic microsatellite markers for the fig-pollinating wasp, Blastophaga nipponica (Hymenoptera: Agaonidae). Applied Entomology and Zoology, 49, 487-491. Kvist S (2014) Does a global DNA barcoding gap exist in Annelida? Mitochondrial DNA, In press. Labandeira CC, Dilcher DL, Davis DR, Wagner DL (1994) Ninety-seven million years of angiosperm-insect association: paleobiological insights into the meaning of coevolution. Proceedings of the National Academy of Sciences of the United States of America, 91, 12278-12282. Lara A, Ponce de León JL, Rodríguez R, et al. (2010) DNA barcoding of Cuban freshwater fishes: evidence for cryptic species and taxonomic conflicts. Molecular Ecology Resources, 10, 421-430. Larkin MA, Blackshields G, Brown NP, et al. (2007) Clustal W and Clustal X version 2.0. Bioinformatics, 23, 2947-2948. Lin R, Huang C, Li S (2007) Development and characterization of microsatellite markers for a fig-pollinating wasp (Ceratosolen sp. 1) of Ficus septica. Molecular Ecology Notes, 7, 1360-1362. Lin R, Yeung C, Fong J, Tzeng H, Li S (2011) The lack of pollinator specificity in a dioecious fig tree: sympatric fig-pollinating wasps of Ficus septica in southern Taiwan. Biotropica, 43, 200-207. Lin R, Yeung C, Li S (2008) Drastic post-LGM expansion and lack of historical genetic structure of a subtropical fig-pollinating wasp (Ceratosolen sp. 1) of Ficus septica in Taiwan. Molecular Ecology, 17, 5008-5022.

133

Linares MC, Soto-Calderón ID, Lees DC, Anthony NM (2009) High mitochondrial diversity in geographically widespread butterflies of Madagascar: a test of the DNA barcoding approach. Molecular Phylogenetics and Evolution, 50, 485- 495. Liu GX, Yang DR, Peng YQ, Compton SG (2015a) Complementary fruiting phenologies facilitate sharing of one pollinator fig wasp by two fig trees. Journal of Plant Ecology, 8, 197-206. Liu M, Compton S, Peng F, Zhang J, Chen X (2015b) Movements of genes between populations: are pollinators more effective at transferring their own or plant genetic markers? Proceedings of the Royal Society B: Biological Sciences, 282, 1-9. Liu M, Wang X, Cui M, Chen Y (2009) Fifteen polymorphic microsatellite loci in Wiebesia pumilae (Hill) (Agaonidae). Conservation Genetics Resources, 1, 189-191. Liu M, Zhang J, Chen Y, Compton S, Chen X (2013) Contrasting genetic responses to population fragmentation in a coevolving fig and fig wasp across a mainland-island archipelago. Molecular Ecology, 22, 4384-4396. Loiselle BA, Sork VL, Nason J, Graham C (1995) Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany, 82, 1420-1425. Lopez-Vaamonde C, Dixon D, Cook J, Rasplus J (2002) Revision of the Australian species of Pleistodontes (Hymenoptera: Agaonidae) fig-pollinating wasps and their host-plant associations. Zoological Journal of the Linnean Society, 136, 637-683. Lopez-Vaamonde C, Rasplus J, Weiblen G, Cook J (2001) Molecular phylogenies of fig wasps: partial cocladogenesis of pollinators and parasites. Molecular Phylogenetics and Evolution, 21, 55-71. Lu Y, Wu K (2011) Effect of relative humidity on population growth of Apolygus lucorum (Heteroptera: Miridae). Applied Entomology and Zoology, 46, 421- 427. Lukhtanov VA, Dantchenko AV, Vishnevskaya MS, Saifitdinova AF (2015) Detecting cryptic species in sympatry and allopatry: analysis of hidden diversity in Polyommatus (Agrodiaetus) butterflies (Lepidoptera: Lycaenidae). Biological Journal of the Linnean Society, 116, 468-485. Lynch HJ, Rhainds M, Calabrese JM, et al. (2014) How climate extremes - not means - define a species' geographic range boundary via a demographic tipping point. Ecological Monographs, 84, 131-149. Lysyk TJ (1998) Relationships between temperature and life history parameters of Trichomalopsis sarcophagae (Hymenoptera: Pteromalidae). Environmental Entomology, 27, 488-498. Lysyk TJ (2001) Relationships between temperature and life history parameters of Muscidifurax zaraptor (Hymenoptera: Pteromalidae). Environmental Entomology, 30, 147-156. Ma G, Rudolf VHW, Ma CS (2015) Extreme temperature events alter demographic rates, relative fitness, and community structure. Global Change Biology, 21, 1794-1808. Machado C, Jousselin E, Kjellberg F, Compton S, Herre E (2001) Phylogenetic relationships, historical biogeography and character evolution of fig- pollinating wasps. Proceedings of the Royal Society B: Biological Sciences, 268, 685-694.

134

Machado C, Robbins N, Gilbert M, Herre E (2005) Critical review of host specificity and its coevolutionary implications in the fig/fig-wasp mutualism. Proceedings of the National Academy of Sciences of the United States of America, 102, 6558-6565. Maddison WP (1997) Gene trees in species trees. Systematic Biology, 46, 523-536. Mallet J (1995) A species definition for the modern synthesis. Trends in Ecology & Evolution, 10, 294-299. Mallet J, Willmott K (2003) Taxonomy: renaissance or Tower of Babel? Trends in Ecology and Evolution, 18, 57-59. Martin TL, Huey RB (2008) Why "suboptimal" is optimal: Jensen's inequality and ectotherm thermal preferences. American Naturalist, 171, E102-E118. Martinsson S, Rhodén C, Erséus C (2015) Barcoding gap, but no support for cryptic speciation in the earthworm Aporrectodea longa (Clitellata: Lumbricidae). Mitochondrial DNA, 1-9. Marussich W, Machado C (2007) Host-specificity and coevolution among pollinating and nonpollinating New World fig wasps. Molecular Ecology, 16, 1925-1946. Mayr E (1942) Systematics and the origin of species Columbia University Press, New York. Mayr E (1963) Animal species and evolution Belknap Press of Harvard University Press, Cambridge, Massachusetts. Mayr E (1995) Species, classification, and evolution. In: Biodiversity and evolution (eds. Arai R, Kato M, Doi Y), pp. 3-12. National Science Museum Foundation, Tokyo. McBride CS, Van Velzen R, Larsen TB (2009) Allopatric origin of cryptic butterfly species that were discovered feeding on distinct host plants in sympatry. Molecular Ecology, 18, 3639-3651. McLeish M, van Noort S (2012) Codivergence and multiple host species use by fig wasp populations of the Ficus pollination mutualism. BMC Evolutionary Biology, 12, McLeish MJ, Beukman G, Van Noort S, Wossler TC (2012) Host-plant species conservatism and ecology of a parasitoid fig wasp genus (Chalcidoidea; Sycoryctinae; Arachonia). PLoS ONE, 7, (e44804). McPherson J (2005) Phenology of six Ficus L., Moraceae, Species and its effects on pollinator survival, in brisbane, Queensland, Australia. Geographical Research, 43, 297-305. Meehl GA, Tebaldi C (2004) More intense, more frequent, and longer lasting heat waves in the 21st century. Science, 305, 994-997. Meyer C, Paulay G (2005) DNA barcoding: error rates based on comprehensive sampling. PLoS biology, 3, Miller-Struttmann NE, Geib JC, Franklin JD, et al. (2015) Functional mismatch in a bumble bee pollination mutualism under climate change. Science, 349, 1541- 1544. Mironidis GK, Savopoulou-Soultani M (2010) Effects of heat shock on survival and reproduction of Helicoverpa armigera (Lepidoptera: Noctuidae) adults. Journal of Thermal Biology, 35, 59-69. Mishler BD, Donoghue MJ (1982) Species concepts: a case for pluralism. Systematic zoology, 31, 491-503.

135

Mitchell KA, Sgrò CM, Hoffmann AA (2011) Phenotypic plasticity in upper thermal limits is weakly related to Drosophila species distributions. Functional Ecology, 25, 661-670. Moe A, Weiblen G (2010) Molecular divergence in allopatric Ceratosolen (Agaonidae) pollinators of geographically widespread Ficus (Moraceae) species. Annals of the Entomological Society of America, 103, 1025-1037. Molbo D, Machado C, Herre E, Keller L (2004) Inbreeding and population structure in two pairs of cryptic fig wasp species. Molecular Ecology, 13, 1613-1623. Molbo D, Machado C, Sevenster J, Keller L, Herre E (2003) Cryptic species of fig- pollinating wasps: implications for the evolution of the fig-wasp mutualism, sex allocation, and precision of adaptation. Proceedings of the National Academy of Sciences of the United States of America, 100, 5867-5872. Moore J, Compton S, Hatcher M, Dunn A (2002) Quantitative tests of sex ratio models in a pollinating fig wasp. Animal Behaviour, 64, 23-32. Moore J, Obbard D, Reuter C, West S, Cook J (2008) Fighting strategies in two species of fig wasp. Animal Behaviour, 76, 315-322. Mora C, Tittensor DP, Adl S, Simpson AGB, Worm B (2011) How many species are there on earth and in the ocean? PLoS Biology, 9, (e1001127). Mulcahy DL (1979) The rise of the angiosperms: a genecological factor. Science, 206, 20-23. Muñoz-Pajares AJ (2013) SIDIER: substitution and indel distances to infer evolutionary relationships. Methods in Ecology and Evolution, 4, 1195-1200. Nason JD, Heard SB, Williams FR (2002) Host-associated genetic differentiation in the goldenrod elliptical-gall moth, Gnorimoschema gallaesolidaginis (Lepidoptera: Gelechiidae). Evolution, 56, 1475-1488. Nicholls JA, Fuentes-Utrilla P, Hayward A, et al. (2010a) Community impacts of anthropogenic disturbance: natural enemies exploit multiple routes in pursuit of invading herbivore hosts. BMC Evolutionary Biology, 10, (322). Nicholls JA, Preuss S, Hayward A, et al. (2010b) Concordant phylogeography and cryptic speciation in two Western Palaearctic oak gall parasitoid species complexes. Molecular Ecology, 19, 592-609. Nyabuga FN, Loxdale HD, Heckel DG, Weisser WW (2012) Coevolutionary fine- tuning: evidence for genetic tracking between a specialist wasp parasitoid and its aphid host in a dual metapopulation interaction. Bulletin of Entomological Research, 102, 149-155. Nyamukondiwa C, Weldon CW, Chown SL, le Roux PC, Terblanche JS (2013) Thermal biology, population fluctuations and implications of temperature extremes for the management of two globally significant insect pests. Journal of Insect Physiology, 59, 1199-1211. Overgaard J, Kristensen TN, Mitchell KA, Hoffmann AA (2011) Thermal tolerance in widespread and tropical Drosophila species: Does phenotypic plasticity increase with latitude? American Naturalist, 178, S80-S96. Palmé A, Laikre L, Ryman N (2013) Monitoring reveals two genetically distinct brown trout populations remaining in stable sympatry over 20 years in tiny mountain lakes. Conservation Genetics, 14, 795-808. Parchman TL, Benkman CW (2002) Diversifying coevolution between crossbills and black spruce on Newfoundland. Evolution, 56, 1663-1672. Parkash R, Munjal AK (1999) Climatic selection of starvation and desiccation resistance in populations of some tropical drosophilids. Journal of Zoological Systematics and Evolutionary Research, 37, 195-202.

136

Patel A (1996) Variation in a mutualism: phenology and the maintenance of gynodioecy in two Indian fig species. Journal of Ecology, 84, 667-680. Patiño S, Herre EA, Tyree MT (1994) Physiological determinants of Ficus fruit temperature and implications for survival of pollinator wasp species: comparative physiology through an energy budget approach. Oecologia, 100, 13-20. Peakall R, Smouse PE (2006) GENALEX 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes, 6, 288- 295. Peakall R, Smouse PE (2012) GenALEx 6.5: Genetic analysis in Excel. Population genetic software for teaching and research-an update. Bioinformatics, 28, 2537-2539. Perry WL, Feder JL, Lodge DM (2001) Implications of hybridization between introduced and resident Orconectes crayfishes. Conservation Biology, 15, 1656-1666. Piessens K, Adriaens D, Jacquemyn H, Honnay O (2009) Synergistic effects of an extreme weather event and habitat fragmentation on a specialised insect herbivore. Oecologia, 159, 117-126. Pietri C, Alves PC, Melo-Ferreira J (2011) Hares in Corsica: high prevalence of Lepus corsicanus and hybridization with introduced L. europaeus and L. granatensis. European Journal of Wildlife Research, 57, 313-321. Polce C, Garratt MP, Termansen M, et al. (2014) Climate-driven spatial mismatches between British orchards and their pollinators: increased risks of pollination deficits. Global Change Biology, 20, 2815-2828. Pollard E, Greatorex-Davies JN, Thomas† JA (1997) Drought reduces breeding success of the butterfly Aglais urticae. Ecological Entomology, 22, 315-318. Pons J, Barraclough TG, Gomez-Zurita J, et al. (2006) Sequence-based species delimitation for the DNA taxonomy of undescribed insects. Systematic Biology, 55, 595-609. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics, 155, 945-959. R Core Team (2015) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Rader R, Reilly J, Bartomeus I, Winfree R (2013) Native bees buffer the negative impact of climate warming on honey bee pollination of watermelon crops. Global Change Biology, 19, 3103-3110. Raja S, Suleman N, Compton S (2008) Why do fig wasps pollinate female figs? Symbiosis, 45, 25-28. Ramírez W (1970) Host specificity of fig wasps (Agaonidae). Evolution, 24, 680- 691. Ramírez W (1974) Coevolution of Ficus and Agaonidae. Annals of the Missouri Botanical Gardens, 61, 770-780. Randi E (2008) Detecting hybridization between wild species and their domesticated relatives. Molecular Ecology, 17, 285-293. Rasplus J, Kerdelhué C, Le Clainche I, Mondor G (1998) Molecular phylogeny of fig wasps Agaonidae are not monophyletic. Comptes Rendus de l'Academie des Sciences - Serie III, 321, 517-527. Ray N, Currat M, Excoffier L (2003) Intra-deme molecular diversity in spatially expanding populations. Molecular Biology and Evolution, 20, 76-86.

137

Raymond M, Rousset F (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of heredity, 86, 248-249. Regal PJ (1977) Ecology and evolution of flowering plant dominance. Science, 196, 622-629. Riley SPD, Shaffer HB, Voss SR, Fitzpatrick BM (2003) Hybridization between a rare, native tiger salamander (Ambystoma californiense) and its introduced congener. Ecological Applications, 13, 1263-1275. Roe AD, Sperling FAH (2007) Patterns of evolution of mitochondrial cytochrome c oxidase I and II DNA and implications for DNA barcoding. Molecular Phylogenetics and Evolution, 44, 325-345. Ronquist F, Teslenko M, Van Der Mark P, et al. (2012) MrBayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Systematic Biology, 61, 539-542. Rosenberg NA (2004) DISTRUCT: a program for the graphical display of population structure. Molecular Ecology Notes, 4, 137-138. Rosselló-Mora R (2003) Opinion: the species problem, can we achieve a universal concept? Systematic and Applied Microbiology, 26, 323-326. Rousset F (2008) GENEPOP'007: a complete re-implementation of the GENEPOP software for Windows and Linux. Molecular Ecology Resources, 8, 103-106. Roux O, Le Lann C, Van Alphen JJM, Van Baaren J (2010) How does heat shock affect the life history traits of adults and progeny of the aphid parasitoid Aphidius avenae (Hymenoptera: Aphidiidae)? Bulletin of Entomological Research, 100, 543-549. Rubinoff D, Cameron S, Will K (2006) A genomic perspective on the shortcomings of mitochondrial DNA for "barcoding" identification. Journal of Heredity, 97, 581-594. Ruiz-Montoya L, Núñez-Farfán J, Vargas J (2003) Host-associated genetic structure of Mexican populations of the cabbage aphid Brevicoryne brassicae L. (Homoptera: Aphididae). Heredity, 91, 415-421. Ryman N, Palm S (2006) POWSIM: A computer program for assessing statistical power when testing for genetic differentiation. Molecular Ecology Notes, 6, 600-602. Saitoh T, Sugita N, Someya S, et al. (2015) DNA barcoding reveals 24 distinct lineages as cryptic bird species candidates in and around the Japanese Archipelago. Molecular Ecology Resources, 15, 177-186. Saltonstall K, Castillo HE, Blossey B (2014) Confirmed field hybridization of native and introduced Phragmites australis (Poaceae) in North America. American Journal of Botany, 101, 211-215. Samways MJ, Osborn R, Hastings H, Hattingh V (1999) Global climate change and accuracy of prediction of species' geographical ranges: establishment success of introduced ladybirds (Coccinellidae, Chilocorus spp.) worldwide. Journal of Biogeography, 26, 795-812. Satler JD, Carstens BC, Hedin M (2013) Multilocus species delimitation in a complex of morphologically conserved trapdoor spiders (Mygalomorphae, Antrodiaetidae, Aliatypus). Systematic Biology, 62, 805-823. Sauer J, Hausdorf B (2012) A comparison of DNA-based methods for delimiting species in a Cretan land snail radiation reveals shortcomings of exclusively molecular taxonomy. Cladistics, 28, 300-316. Schluter D (2001) Ecology and the origin of species. Trends in Ecology and Evolution, 16, 372-380.

138

Schneider C, Moritz C (1999) Rainforest refugia and evolution in Australia's Wet Tropics. Proceedings of the Royal Society B: Biological Sciences, 266, 191- 196. Segar S (2011) The ecology and evolution of fig wasp communities. PhD thesis, University of Reading. Segar S, Pereira R, Compton S, Cook J (2013) Convergent structure of multitrophic communities over three continents. Ecology Letters, 16, 1436-1445. Segar ST, Cook JM (2012) The dominant exploiters of the fig/pollinator mutualism vary across continents, but their costs fall consistently on the male reproductive function of figs. Ecological Entomology, 37, 342-349. Segar ST, Dunn DW, Darwell C, Cook JM (2014) How to be a fig wasp down under: the diversity and structure of an Australian fig wasp community. Acta Oecologica, 57, 17-27. Senn HV, Barton NH, Goodman SJ, et al. (2010) Investigating temporal changes in hybridization and introgression in a predominantly bimodal hybridizing population of invasive sika (Cervus nippon) and native red deer (C. elaphus) on the Kintyre Peninsula, Scotland. Molecular Ecology, 19, 910-924. Sgrò CM, Overgaard J, Kristensen TN, et al. (2010) A comprehensive assessment of geographic variation in heat tolerance and hardening capacity in populations of Drosophila melanogaster from Eastern Australia. Journal of Evolutionary Biology, 23, 2484-2493. Shanahan M, Samson S, Compton S, Corlett R (2001) Fig-eating by vertebrate frugivores: a global review. Biological Reviews of the Cambridge Philosophical Society, 76, 529-572. Shoemaker D, Machado C, Molbo D, et al. (2002) The distribution of Wolbachia in fig wasps: Correlations with host phylogeny, ecology and population structure. Proceedings of the Royal Society B: Biological Sciences, 269, 2257-2267. Shoemaker DD, Dyer KA, Ahrens M, McAbee K, Jaenike J (2004) Decreased diversity but increased substitution rate in host mtDNA as a consequence of Wolbachia endosymbiont infection. Genetics, 168, 2049-2058. Shoemaker DD, Keller G, Ross KG (2003) Effects of Wolbachia on mtDNA variation in two fire ant species. Molecular Ecology, 12, 1757-1771. Silvieus S, Clement W, Weiblen G (2008) Cophylogeny of figs, pollinators, gallers, and parasitoids. In: Specialization, speciation and radiation: The evolutionary biology of herbivorous insects (ed. Tilmon K), pp. 225-239. University of California Press, Berkeley & Los Angeles. Simmons AJ, Willett KM, Jones PD, Thorne PW, Dee DP (2010) Low-frequency variations in surface atmospheric humidity, temperature, and precipitation: inferences from reanalyses and monthly gridded observational data sets. Journal of Geophysical Research Atmospheres, 115, (D01110). Simonsen KL, Churchill GA, Aquadro CF (1995) Properties of statistical tests of neutrality for DNA polymorphism data. Genetics, 141, 413-429. Skaug HJ, Fournier DA, Nielsen A, Magnusson A, Bolker B (2013) Generalized linear mixed models using AD Model Builder, R package version 0.7.5. Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139, 457-462. Slatkin M, Hudson RR (1991) Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics, 129, 555-562.

139

Steffen W (2015) Quantifying the impact of climate change on extreme heat in Australia. Climate Council of Australia. Stevens GC (1989) The latitudinal gradient in geographical range: how so many species coexist in the tropics. American Naturalist, 133, 240-256. Stoeckle MY, Thaler DS (2014) DNA barcoding works in practice but not in (neutral) theory. PLoS ONE, 9, (e100755). Stone GN, Lohse K, Nicholls JA, et al. (2012) Reconstructing community assembly in time and space reveals enemy escape in a western palearctic insect community. Current Biology, 22, 532-537. Sun X, Xiao J, Cook J, Feng G, Huang D (2011) Comparisons of host mitochondrial, nuclear and endosymbiont bacterial genes reveal cryptic fig wasp species and the effects of Wolbachia on host mtDNA evolution and diversity. BMC Evolutionary Biology, 11, (86). Sunnucks P (2000) Efficient genetic markers for population biology. Trends in Ecology and Evolution, 15, 199-203. Sutton T, Reuter C, Riegler M, Cook J (2015) Characterisation of microsatellite markers for fig-pollinating wasps in the Pleistodontes imperialis species complex. Australian Journal of Zoology, 63, 122-126. Sutton T, Riegler M, Cook J (2016) One step ahead: a parasitoid disperses farther and forms a wider geographic population than its fig wasp host. Molecular Ecology, 25, 882-894. Szpiech ZA, Jakobsson M, Rosenberg NA (2008) ADZE: A rarefaction approach for counting alleles private to combinations of populations. Bioinformatics, 24, 2498-2504. Tajima F (1989) The effect of change in population size on DNA polymorphism. Genetics, 123, 597-601. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6: Molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30, 2725-2729. Tang W, Chen Y (2012) Hybridization between native barbless carp (Cyprinus pellegrini) and introduced common carp (C. carpio) in Xingyun Lake, China. Zoological Science, 29, 311-318. Therneau T (2015) coxme: mixed effects Cox models, R package version 2.2-4. Thompson JN (1994) The coevolutionary process University of Chicago Press, Chicago. Thompson JN (2005) The geographic mosaic of coevolution University of Chicago Press, Chicago. Thompson JN, Cunningham BM (2002) Geographic structure and dynamics of coevolutionary selection. Nature, 417, 735-738. Thompson JN, Fernandez CC (2006) Temporal dynamics of antagonism and mutualism in a geographically variable plant-insect interaction. Ecology, 87, 103-112. Tian E, Nason JD, MacHado CA, et al. (2015) Lack of genetic isolation by distance, similar genetic structuring but different demographic histories in a fig- pollinating wasp mutualism. Molecular Ecology, 24, 5976-5991. Tian E, Yu H, Zhang D, Nason J (2011) Development of microsatellite loci for Blastophaga javana (Agaonidae), the pollinating wasp of Ficus hirta (Moraceae). American Journal of Botany, 98, e41-e43.

140

Tochen S, Woltz JM, Dalton DT, et al. (2016) Humidity affects populations of Drosophila suzukii (Diptera: Drosophilidae) in blueberry. Journal of Applied Entomology, 140, 47-57. Untergasser A, Cutcutache I, Koressaar T, et al. (2012) Primer3-new capabilities and interfaces. Nucleic Acids Research, 40. van Noort S, Compton S (1996) Convergent evolution of agaonine and sycoecine (Agaonidae, Chalcidoidea) head shape in response to the constraints of host fig morphology. Journal of Biogeography, 23, 415-424. van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P (2004) MICRO- CHECKER: Software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes, 4, 535-538. Van Valen L (1973) A new evolutionary law. Evolutionary Theory, 1, 1-30. Vaughn TT, Antolin MF (1998) Population genetics of an opportunistic parasitoid in an agricultural landscape. Heredity, 80, 152-162. Wagner CE, Keller I, Wittwer S, et al. (2013) Genome-wide RAD sequence data provide unprecedented resolution of species boundaries and relationships in the Lake Victoria cichlid adaptive radiation. Molecular Ecology, 22, 787-798. Wallace A (1858) On the tendency of species to form varieties to depart indefinitely from the original type. Journal of the proceedings of the Linnean Society of London: Zoology, 3, 53-62. Wang HY, Hsieh CH, Huang CG, et al. (2013) Genetic and physiological data suggest demographic and adaptive responses in complex interactions between populations of figs (Ficus pumila) and their pollinating wasps (Wiebesia pumilae). Molecular Ecology, 22, 3814-3832. Wang YT, Gong J, Zhang LY (2014) Development of polymorphic microsatellite loci for Eupristina altissima, the pollinating wasp of Ficus altissima (Moraceae). Conservation Genetics Resources, 6, 625-627. Ware AB, Kaye PT, Compton SG, van Noort S (1993) Fig volatiles: Their role in attracting pollinators and maintaining pollinator specificity. Plant Systematics and Evolution, 186, 147-156. Warren DL, Cardillo M, Rosauer DF, Bolnick DI (2014) Mistaking geography for biology: inferring processes from species distributions. Trends in Ecology and Evolution, 29, 572-580. Warren M, Robertson M, Greeff J (2010) A comparative approach to understanding factors limiting abundance patterns and distributions in a fig tree-fig wasp mutualism. Ecography, 33, 148-158. Wei ZD, Kobmoo N, Cruaud A, Kjellberg F (2014) Genetic structure and hybridization in the species group of Ficus auriculata: can closely related sympatric Ficus species retain their genetic identity while sharing pollinators? Molecular Ecology, 23, 3538-3550. Weiblen G (2004) Correlated evolution in fig pollination. Systematic Biology, 53, 128-139. Weiblen G, Bush G (2002) Speciation in fig pollinators and parasites. Molecular Ecology, 11, 1573-1578. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution, 38, 1358-1370. West S, Cook J, Werren J, Godfray H (1998) Wolbachia in two insect host-parasitoid communities. Molecular Ecology, 7, 1457-1465.

141

West S, Herre E (1994) The ecology of the New World fig-parasitizing wasps Idarnes and implications for the evolution of the fog-pollinator mutualism. Proceedings of the Royal Society B: Biological Sciences, 258, 67-72. West S, Herre E, Windsor D, Green P (1996) The ecology and evolution of the New World non-pollinating fig wasp communities. Journal of Biogeography, 23, 447-458. White T, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal genes for phylogenetics. In: PCR protocols: A guide to methods and applications (eds. Glefand D, White T, Innis M, Sninsky J), pp. 315-322. Academic Press, San Diego. Wiebes J (1961) On the variability of Agaon paradoxum (Dalman) Grandi and Seres armipes Waterston, with remarks on other African Agaonidae (Hymanoptera, Chalcidoidea). Zoologische Mededelingen, 37, 231-240. Wiebes J (1979) Co-evolution of figs and their insect pollinators. Annual Review of Ecology and Systematics, 10, 5143-4164. Wiemers M, Fiedler K (2007) Does the DNA barcoding gap exist? - a case study in blue butterflies (Lepidoptera: Lycaenidae). Frontiers in Zoology, 4, (8). Will K, Rubinoff D (2004a) Myth of the molecule: DNA barcodes for species cannot replace morphology for identification and classification. Cladistics, 20, 47- 55. Will KW, Mishler BD, Wheeler QD (2005) The perils of DNA barcoding and the need for integrative taxonomy. Systematic Biology, 54, 844-851. Will KW, Rubinoff D (2004b) Myth of the molecule: DNA barcodes for species cannot replace morphology for identification and classification. Cladistics, 20, 47-55. Williams PH, An J, Brown MJF, et al. (2012) Cryptic bumblebee species: consequences for conservation and the trade in greenhouse pollinators. PLoS ONE, 7, (e32992). Witt JDS, Threloff DL, Hebert PDN (2006) DNA barcoding reveals extraordinary cryptic diversity in an amphipod genus: implications for desert spring conservation. Molecular Ecology, 15, 3073-3082. Xiao C, Jin Y, Yi Z, Cook J, Crozier R (2008) The phenology and potential for self- pollination of two Australian monoecious fig species. Symbiosis, 45, 91-96. Xiao J, Wang N, Murphy R, et al. (2012) Wolbachia infection and dramatic intraspecific mitochondrial DNA divergence in a fig wasp. Evolution, 66, 1907-1916. Yang LY, Machado CA, Dang XD, et al. (2015) The incidence and pattern of copollinator diversification in dioecious and monoecious figs. Evolution, 69, 294-304. Yeo CK, Tan HTW (2009) Variation in reproductive output of Ficus superba despite aseasonal reproduction. Plant Ecology, 205, 235-248. Yu D (2001) Parasites of mutualisms. Biological Journal of the Linnean Society, 72, 529-546. Yu H, Nason JD (2013) Nuclear and chloroplast DNA phylogeography of Ficus hirta: obligate pollination mutualism and constraints on range expansion in response to climate change. New Phytologist, 197, 276-289. Zammit J, Schwarz M (2000) Intersexual sibling interactions and male benevolence in a fig wasp. Animal Behaviour, 60, 695-701. Zavodna M, Arens P, Vosman B, Van Dijk P, Van Damme J (2002) Development and characterization of microsatellite markers for Liporrhopalum tentacularis

142

Grandi, the pollinator fig wasp of Ficus montana Blume. Molecular Ecology Notes, 2, 569-571. Zhang D, Lin K, Hanski I (2004) Coexistence of cryptic species. Ecology Letters, 7, 165-169.

143

Appendices

144

Appendix A

Figure A1: TESS barplots for K values ranging from 2 to 8. According to the TESS manual, the optimal Kmax can be determined by viewing barplots for increasing Kmax from 2 to n. The optimal Kmax becomes clear when an increase in Kmax does not produce an additional unambiguous cluster. In this instance, Kmax = 4 provides the same information as Kmax = 3 (it merely splits the yellow cluster into two). The optimal Kmax was therefore determined to be 3.

145

Appendix B

Table B1: Within-site genetic structure (Kinship coefficient)

Pollinator Parasitoid (a) Logarithmic distance b P b P Sydney -0.002 ± 0.005 0.506 -0.004 ± 0.005 0.247 Newcastle -0.008 ± 0.004 0.140 0.002 ± 0.000 0.348 Port Macquarie 0.000 ± 0.001 0.909 0.003 ± 0.002 0.277 Byron Bay -0.009 ± 0.007 0.551 0.002 ± 0.007 0.540 Atherton 0.002 ± 0.003 0.662 -0.009 ± 0.004 0.409 (b) Linear distance b P b P Sydney -0.000 ± 0.000 0.379 -0.000 ± 0.000 0.175 Newcastle -0.004 ± 0.004 0.395 0.000 ± 0.000 0.030 Port Macquarie 0.000 ± 0.000 0.511 0.000 ± 0.000 0.230 Byron Bay 0.000 ± 0.001 0.551 0.000 ± 0.000 0.476 Atherton 0.000 ± 0.000 0.902 0.000 ± 0.000 0.605

b : slope of regression against (a) logarithmic and (b) linear geographic distance for kinship coefficient (F ) within sites. P: Probability following Mantel test (10,000 permutations)

146

Appendix C

Table C1. Microsatellite power analysis.

Pollinator Parasitoid Fisher' Exact Chi-Squared Fisher' Exact Chi-Squared FST Test Test Test Test 0.001 0.1315 0.1385 0.1010 0.0880 0.0025 0.3615 0.3810 0.1955 0.1960 0.005 0.7840 0.8180 0.4665 0.5105 0.01 0.9980 0.9995 0.8695 0.8870 0.013 - - 0.9695 0.9570 0.015 1.0000 1.0000 0.9825 0.9905 0.02 1.0000 1.0000 0.9995 1.0000 0.025 1.0000 1.0000 1.0000 1.0000 0.05 1.0000 1.0000 1.0000 1.0000

Probability of pollinator and parasitoid microsatellite loci detecting true FST values according to Fisher’s exact test and Chi-Squared tests. Analyses were conducted in Powsim with an effective population size (Ne) of 4000. Simulations for FST = 0.013 were conducted for parasitoid only, as this is the global parasitoid FST obtained from our real data.

147